[Jose M. Caballero, Francisco Hens, Andreu GuimerÃ(BookFi.org)

Installation and Maintenance of SDH/SONET, ATM, xDSL and Synchronization Networks

Jose Caballero

Francisco Hens

Roger Segura

Andreu Guimerá

To Benny Moré and his big band

—because life is not only made of broadband

SDH/SONET, ATM, xDSL, and Synchronization Networksvi

vii

Contents

Preface ....................................................................................................xv

Chapter 1 PDH and T-Carrier: The Plesiochronous Hierarchies ...... 11.1 An Introduction to Communications Systems.................................1

1.1.1 Signals and Information ...................................................... 11.1.2 Transmission Medium ......................................................... 21.1.3 Channel Coding ................................................................... 71.1.4 Multiplexing and Multiple Access ...................................... 10

1.2 Pulse Code Modulation ................................................................... 111.3 PDH and T-Carrier ..........................................................................13

1.3.1 Basic Rates: T1 and E1 ....................................................... 141.4 The E1 Frame .................................................................................. 15

1.4.1 Frame Alignment .................................................................151.4.2 Frame Alignment Signal ...................................................... 161.4.3 Multiframe CRC-4 ............................................................... 161.4.4 Supervision Bits ................................................................... 181.4.5 NFASs - Spare Bits ............................................................. 191.4.6 NFAS - Alarm Bit ............................................................... 191.4.7 Signaling Channel ............................................................... 201.4.8 CAS Signaling Multiframe .................................................. 20

1.5 The Plesiochronous Digital Hierarchy ............................................ 221.5.1 Higher Hierarchical Levels .................................................. 231.5.2 Multiplexing Level 2: 8 Mbps ............................................. 231.5.3 Multiplexing Level 3: 34 Mbps ........................................... 241.5.4 Multiplexing Level 4: 140 Mbps ......................................... 241.5.5 Service Bits in Higher Level Frames ................................... 241.5.6 Plesiochronous Synchronization ......................................... 261.5.7 Positive Justification ............................................................ 27

1.6 Managing Alarms in Higher Level Hierarchies .............................. 281.7 The T-Carrier Hierarchy.................................................................. 29

1.7.1 The DS1 Frame .................................................................... 291.7.2 The DS2 Frame .................................................................... 321.7.3 The DS3 Frame .................................................................... 32

Selected Bibliography............................................................................. 33

SDH/SONET, ATM, xDSL, and Synchronization Networksviii

Chapter 2 SDH/SONET: The Synchronous Hierarchies ..................... 352.1 The Emergence of SDH/SONET Networks .................................... 35

2.1.1 Limitations of Plesiochronous Networks ............................ 362.1.2 The SDH/SONET Challenge ............................................... 37

2.2 Comparison of SDH and SONET.................................................... 392.3 Functional Architecture ................................................................... 40

2.3.1 Network Elements ............................................................... 402.3.2 Network Topology ............................................................... 412.3.3 Topology Partitioning .......................................................... 422.3.4 SDH/SONET Layers ........................................................... 42

2.4 SDH/SONET Formats and Procedures ........................................... 452.4.1 SDH/SONET Frame Structure ............................................ 462.4.2 Multiplexing Map ................................................................ 49

2.5 SDH Transport Services .................................................................. 492.6 Transporting PDH/T-Carrier Tributaries ......................................... 53

2.6.1 Transport on VC-4 or STS-3c SPE ..................................... 542.6.2 Transport on VC-3 ............................................................... 552.6.3 Transport of 2-Mbps Circuits .............................................. 57

2.7 Pointers and Timing Compensation................................................. 602.7.1 Payload Synchronization ..................................................... 602.7.2 Pointer Formats and Procedures ..........................................61

2.8 Overheads ........................................................................................ 642.8.1 Path Overhead ..................................................................... 662.8.2 Section Overhead .................................................................672.8.3 The SDH/SONET Hierarchy ............................................... 69

2.9 Concatenation .................................................................................. 712.9.1 Contiguous Concatenation of VC-4 .................................... 722.9.2 Virtual Concatenation .......................................................... 722.9.3 Link Capacity Adjustment Scheme ..................................... 74

2.10Maintenance .................................................................................... 742.10.1 SDH/SONET Events ......................................................... 762.10.2 Monitoring Events ............................................................. 782.10.3 Event Tables ...................................................................... 78

2.11 Performance Monitoring .................................................................782.11.1 Bit Error Checking ............................................................ 802.11.2 Tandem Connection Monitoring ....................................... 812.11.3 Forward Error Correction .................................................. 82

2.12Defects ............................................................................................. 832.13SDH Resilience ............................................................................... 85

2.13.1 Protection Basics ............................................................... 872.13.2 Multiplex Section or Line Protection ................................ 89

2.14Operation, Administration, and Management .................................932.14.1 The TMN Standard ............................................................ 932.14.2 TMN Benefits .................................................................... 94

Contents ix

2.15Next Generation SDH...................................................................... 95Selected Bibliography............................................................................. 97

Chapter 3 ATM Architectures................................................................ 993.1 Introduction ..................................................................................... 993.2 Basic Principles of ATM .................................................................101

3.2.1 ATM Cell Format ................................................................ 1013.2.2 Virtual Channels and Virtual Paths ..................................... 1033.2.3 Basic Principles of ATM Switching .................................... 104

3.3 ATM Network Architecture............................................................. 1063.3.1 Introduction ......................................................................... 1063.3.2 AAL Layer ........................................................................... 1073.3.3 ATM Layer ..........................................................................1093.3.4 Physical Layer ..................................................................... 113

3.4 ATM Adaptation Level Structures................................................... 1153.4.1 AAL1 Format ...................................................................... 1153.4.2 AAL2 Format ...................................................................... 1183.4.3 AAL3/4 Format ................................................................... 1203.4.4 AAL5 Format ...................................................................... 121

3.5 Quality of Service............................................................................ 1223.5.1 Traffic Characterization Parameters .................................... 1223.5.2 Negotiated QoS Parameters ................................................. 1243.5.3 Service Categories ............................................................... 1253.5.4 Traffic Contract ................................................................... 127

3.6 Resource Management .................................................................... 1293.6.1 Connection Admission Control ........................................... 1303.6.2 UPC and NPC Policing Functions ....................................... 1323.6.3 Other Control Functions ...................................................... 138

3.7 ATM in Access Networks................................................................ 1393.7.1 ATM as Transport in ADSL ................................................ 1393.7.2 Wireless Local Loop ............................................................ 144

3.8 Conclusions ..................................................................................... 149Selected Bibliography............................................................................. 149

Chapter 4 ADSL Technology.................................................................. 1514.1 The Origin of DSL Technologies ................................................... 151

4.1.1 The Birth of DSL Technologies: HDSL .............................. 1524.1.2 New Modulation Technologies ........................................... 1544.1.3 Asymmetry ..........................................................................155

4.2 Reference Models ............................................................................ 1574.2.1 ADSL System ...................................................................... 1574.2.2 ADSL Transceivers ............................................................. 158

4.3 Framing............................................................................................ 1614.3.1 Data and Overhead Buffers ................................................. 161

SDH/SONET, ATM, xDSL, and Synchronization Networksx

4.3.2 Superframes ......................................................................... 1634.4 Coding ............................................................................................. 165

4.4.1 Error Protection ................................................................... 1664.4.2 Scrambling ........................................................................... 1684.4.3 Interleaving ..........................................................................168

4.5 Modulation ...................................................................................... 1694.5.1 Organizing the Tones .......................................................... 1704.5.2 Constellation Coders ........................................................... 1704.5.3 DMT Modulation .................................................................1714.5.4 Cyclic Prefix and Synchronization Symbol ........................ 174

4.6 Operation and Maintenance Channel (EOC)................................... 1744.6.1 EOC Message Format .......................................................... 1744.6.2 EOC Commands .................................................................. 175

4.7 Initialization..................................................................................... 1764.7.1 Handshake ........................................................................... 1774.7.2 Training ............................................................................... 1794.7.3 Analyzing the Channel ........................................................ 1804.7.4 Exchanging Information ...................................................... 182


Chapter 5 Network Synchronization ..................................................... 1855.1 Architecture of Synchronization Networks ..................................... 185

5.1.1 Synchronization Network Topologies .................................1875.2 Interconnection of Nodes ................................................................ 189

5.2.1 Synchronization Signals ...................................................... 1895.2.2 Holdover Mode .................................................................... 1915.2.3 Global Positioning System .................................................. 191

5.3 Disturbances in Synchronization Signals ........................................ 1925.3.1 Frequency Offset .................................................................1925.3.2 Phase Fluctuation .................................................................194

5.4 Synchronization of Transmission Networks.................................... 1975.4.1 Synchronization in SONET and SDH .................................1985.4.2 Synchronization Models ...................................................... 1995.4.3 Timing Loops ...................................................................... 201

5.5 Digital Synchronization and Switching........................................... 2015.6 SSU in a Synchronization Network................................................. 203

5.6.1 Functions of SSU .................................................................204Selected Bibliography............................................................................. 205

Chapter 6 Test and Measurement .......................................................... 2076.1 Areas of Application for Test and Measurement............................. 2076.2 Tests in the Interfaces ...................................................................... 208

6.2.1 Line Interfaces ..................................................................... 2096.2.2 Connection Modes in Electrical Interfaces .......................... 209

Contents xi

6.2.3 Measurements in Electrical Interfaces .................................2116.2.4 Measurements in Optical Interfaces .................................... 2136.2.5 Measuring Frequency .......................................................... 215

6.3 In-Service and Out-of-Service Measurements ................................ 2166.3.1 Bit Error Rate ...................................................................... 2176.3.2 Out-of-service Measurements .............................................. 2176.3.3 In-Service Measurements .................................................... 2256.3.4 Connecting a Measurement Device for ISM ....................... 228

6.4 Synchronization of NE-Test Set in SDH ......................................... 230Selected Bibliography............................................................................. 231

Chapter 7 SDH/SONET and PDH Roll-Out ......................................... 2337.1 Bit Error Rate Test ........................................................................... 233

7.1.1 BERT of Virtual Container ................................................. 2337.1.2 Overhead Transparency Test ............................................... 234

7.2 Stimulus-Response Tests .................................................................2357.2.1 Preliminary Definitions ....................................................... 2367.2.2 Line and Test Sequence Events ........................................... 2367.2.3 PDH Events ......................................................................... 2367.2.4 SDH/SONET Events ........................................................... 2367.2.5 Interaction of Maintenance Signals ..................................... 236

7.3 Stress Tests ...................................................................................... 2387.3.1 Introducing Frequency Offset .............................................. 2397.3.2 Generating Pointer Movements ........................................... 239

7.4 Mux/Demux Tests............................................................................ 2407.4.1 PDH Mux/Demux Test ........................................................ 2417.4.2 SDH/SONET Mux/Demux Test ..........................................242

7.5 Measuring Round Trip Delay .......................................................... 2427.6 APS Measurements ......................................................................... 244

7.6.1 Network Security: Concept and Classification .................... 2447.6.2 Characterizing the Measurement ......................................... 2467.6.3 Measurement Procedure ...................................................... 247

7.7 Performance Measurements ............................................................ 2517.7.1 Introduction to G.821, G.826, and M.2100 ......................... 2517.7.2 Measurements in Line with G.821 ...................................... 2527.7.3 Measurements in Line with G.826 ...................................... 2567.7.4 Measurements in Line with M.2100 .................................... 2617.7.5 Recommendations M.2110 and M.2120 ............................. 2687.7.6 Open Network Provision for Leased Lines at 2,048 Kbps .. 268

7.8 Tests on ADMs and DXC................................................................ 2687.8.1 Tributary Continuity Test .................................................... 269

7.9 Tests on SDH/SONET Rings........................................................... 2707.9.1 Transparency Tests .............................................................. 2707.9.2 Multiplexers ......................................................................... 2717.9.3 Synchronization Measurements ........................................... 272

SDH/SONET, ATM, xDSL, and Synchronization Networksxii

7.9.4 Protection Switching Tests .................................................. 2737.9.5 Defect Indicators in the Network Management System ...... 2737.9.6 Path Trace Tests .................................................................. 274


Chapter 8 ATM Performance................................................................. 2778.1 Introduction ..................................................................................... 2778.2 Performance Parameters in ATM Networks.................................... 277

8.2.1 Cell-Based ATM Reference Events .................................... 2788.2.2 ATM Cell Transfer Outcomes ............................................. 2788.2.3 ATM Cell Transfer Performance Parameters ...................... 2798.2.4 Performance of Permanent Connections ............................. 282

8.3 OAM Functions: In-Service Measurements .................................... 2838.3.1 Presentation of OAM Functions ..........................................2848.3.2 Physical Layer OAM Procedures ........................................ 2858.3.3 ATM Layer OAM Procedures ............................................. 2878.3.4 ATM Layer OAM Cells ...................................................... 2918.3.5 Fault Management Functions .............................................. 2928.3.6 Performance Management: Performance ISM .................... 2968.3.7 Activation/Deactivation Functions ...................................... 299

8.4 Test Traffic for Out-of-Service Measurements................................ 3008.4.1 Generating Test Traffic ....................................................... 3008.4.2 Estimating Performance Parameters in OOS Mode ............ 302

8.5 Measurement Cycle in ATM Networks ........................................... 3058.5.1 Properties of ATM Switches ............................................... 3068.5.2 Installing the Network ......................................................... 3148.5.3 Network Commissioning ..................................................... 3158.5.4 Bringing-Into-Service in ADSL Environments ................... 3188.5.5 Commissioning in Wireless Local Loop Environments ...... 3228.5.6 In-Service Measurements .................................................... 326


Chapter 9 xDSL Qualification................................................................ 3319.1 Qualification Strategies and Protocols ............................................ 331

9.1.1 Prequalification .................................................................... 3329.1.2 Qualification During Commissioning .................................3339.1.3 Commissioning Without Qualification ................................ 334

9.2 Copper Pair ...................................................................................... 3349.2.1 Attenuation and Distortion .................................................. 3349.2.2 Return Losses ...................................................................... 3379.2.3 Noise .................................................................................... 3409.2.4 Longitudinal Conversion Loss ............................................ 3439.2.5 Crosstalk .............................................................................. 3489.2.6 Other Reasons for Defects ................................................... 353

9.3 Analog Measurements ..................................................................... 3589.3.1 One-End Measurements ...................................................... 359

Contents xiii

9.3.2 Two-End Measurements ...................................................... 3649.3.3 Bridged Measurements ........................................................ 3659.3.4 Digital Measurements .......................................................... 366


Chapter 10 Jitter and Wander Control ................................................. 36910.1Dealing with Jitter ........................................................................... 369

10.1.1 Phase Fluctuation ............................................................... 36910.1.2 Jitter Metrics and Measurement ........................................ 37010.1.3 Measuring Jitter in Output Interfaces ................................ 37310.1.4 Measuring Jitter Tolerance ................................................ 37410.1.5 Measuring Jitter Transfer .................................................. 37610.1.6 Mapping Jitter and Combined Jitter .................................. 37910.1.7 Jitter in Leased Lines ......................................................... 382

10.2Dealing with Wander ....................................................................... 38210.2.1 Synchronization of SDH/SONET Networks ..................... 38310.2.2 Measuring Relative and Absolute Wander ........................ 38310.2.3 The Metrics of Wander: TIE, MTIE, and TDEV .............. 38410.2.4 Measuring Output Wander ................................................ 38610.2.5 Measuring Tolerance to Input Wander .............................. 38710.2.6 Measuring Wander Transfer .............................................. 38910.2.7 Response to Phase Transients ............................................ 38910.2.8 Operating in Holdover Mode ............................................. 391

10.3Tests on ADMs and DXC................................................................ 39310.3.1 Measuring Jitter .................................................................39310.3.2 Synchronization Tests ....................................................... 395


Appendix A Error Detection and Correction Techniques .................. 403 A.1 Cyclic Redundancy Check ............................................................. 403 A.2 RS and BCH Codes........................................................................ 404 A.3 Bit Interleaved Parity ..................................................................... 406

Appendix B Masks for Copper Qualification ...................................... 409 B.1 ANSI Masks ................................................................................... 409 B.2 ETSI Masks .................................................................................... 412

Appendix C Two-Wire Transmission Line Model .............................. 415 C.1 Characteristic Parameters of the Line ............................................ 416

About the Authors...................................................................................421

Index 423

xv

Preface

For about 200,000 years, Neanderthal man inhabited and ruled the chilly forestsand steppes of Eurasia. The Neanderthal society was gregarious and hierarchical,and it was formed by scattered tribes that brought up their children and took care ofthe wounded, the sick, and the elderly. The world view of these early human beingsalready showed signs of capacity for abstract and synthetic thinking, as they prac-ticed rites and decorated their bodies with necklaces, paintings, and earrings.Thanks to more than 2 million years of human evolution, they had their tools andtechniques to make fire and procure and store food. They were also able to tanleather to make clothes and to protect their feet. But their weapons were what con-verted them into the most extraordinary predators in the food chain.

However, some 40,000 years ago, a new hominid species of African origin start-ed to compete for space with homoneanderthalensis. The newcomers were slightlydifferent physically; their skin was darker and they were taller, although less mus-cular. This made them physically less prepared for the cold climate. They did notseem to be more intelligent either; at least if we look at the size of their skull, whichwas about 10% smaller than that of Neanderthal man. And if this is not enough, thechildren of this new species took twice as long to grow up; in this way forcing theirparents to have fewer descendants. Evolution had made their reproductive periodshorter, so that they could feed and take care of their descendants. In spite of all this,after a relatively short period of coexistence, homoneanderthalensis mysteriouslydisappeared. Perhaps they were just simply wiped away by their competitors, orkilled by the new viruses coming from the south, or maybe they just disappeared be-cause they were unable to adapt themselves to the rapid changes between the Ice Ag-es.

So, did mankind benefit in any way from the extinction of Neanderthal man?

Yes, communication. The unusual form of the larynx and the gullet of the newspecies, known as homosapiens-sapiens, enabled them to generate and modulate so-phisticated sounds. Neanderthal men did not have this capacity, without which it isimpossible to create a human language. This theory explains how the hominidsmoved from waiting for genetic changes to using communication as the vital surviv-al tool. This proved to be more useful than the slow biological evolution in adaptingthe hominids to their environment. However, acquiring language skills and prepar-

SDH/SONET, ATM, xDSL, and Synchronization Networksxvi

ing the brain for learning is a slow process that in this case made new generationsmature more slowly, and parents had to spend more years taking care of their imma-ture descendants. Despite this, and other physiological difficulties, the new humanbeings who, as you probably already know, are nothing less than us, took over in arelatively short period of time, and ended up populating most of the planet.

Another significant milestone in the history of communications was the discov-ery of writing, probably the most important intellectual tool ever discovered by man.Writing enables us to store information and transmit it between two distant pointsand even between generations, without distorting or losing the message. There is ev-idence of earlier attempts, although the first effective form of writing was developedby the Sumerians about 5,000 years ago. The Sumerians lived in city-states on thebanks of the Tigris and Euphrates rivers, where such activities as agriculture, cattleraising, craft work, metallurgy, and construction flourished in an extraordinary way.Writing was born in the heart of these urban societies as a means to increase com-merce, and solve both legal and social problems. Originally, the Sumerian codeswere iconographic, whereby each sign was an icon resembling the object it repre-sented. This way, it was possible to sell or buy a herd of 53 lambs, for instance, orlegally divide a property of 180 ikus of surface between heirs. When numbers werelater developed, this was a huge step forward, as it was no longer necessary to repeatthe same icon a number of times. But what really made a change was the inventionof the phonetic writing system. Now it was possible to describe battles or the posi-tion of stars, or write down laws, such as Hammurabi’s Code of Law.

This was the start of our civilization.

For thousands of years, writing was done by hand on clay, stones, papyrus, orleather, until in 1450 the workshop of Gutenberg started to mechanically producewhat became the first books. Fifty years later, the few books kept in monasteries andpalaces were transformed into more than 10 million volumes. It was finally possibleto store and produce a large amount of information at lower cost than before, andwithout changing the original contents. Knowledge, literature, and science were nolonger tools of power for a small elite of scribes, priests, and courtiers. This way, byhaving the medium to broadcast information to thousands of recipients, the printingindustry had an important role in marking the end of the dark Middle Ages.

Some centuries had to pass before electricity was managed in such a way thatthe first telephone patented by Alexander Graham Bell in 1876 could be developed.A few years later, around 1900, the first radio transmissions took place, and televi-sion appeared in the 1930s. Without underestimating telephone and television, thereis something that makes the telephone special, in that it enables direct interpersonalcommunication at a distance. We can even see the telephone as an extension of ourlarynx and ears, while radio and television are one-way media, where the receiver

Preface xvii

can only connect and disconnect, the same way as you can close this book, but notmodify its contents.

This difference is notable, and it explains why both radio and TV tend to be de-sired, controlled and even manipulated by political, economical or religious power,while private telephone conversations offer more liberty and independence. Thetelephone is by definition a tool where the contents, the language, and the recipientcan be decided by the users themselves.

Finally, we arrive at the mid-1990s, when the Internet became an important me-dium for mass communication. It combines two fundamental inventions: writing andtelecommunications. Writing can be very precise and it enables us to store informa-tion, crossing the time barrier; while telecommunications overcomes space barriers.It is so efficient that many times we prefer to send electronic messages even withinthe same office, although it would be easier just to have a short conversation. But theInternet is a lot more than an efficient two-way communication medium. It is also away to access the immense “universal library,” with an impact that can only be com-pared to the Alexandria Library 2,000 years ago. In the Internet, we have millions ofdocuments with information that can be reached from any part of the world in just afew seconds.

The third generation wireless networks will also bring some changes in the nearfuture, by improving the human-machine relationship. Our mobile telephones willbecome terminals with Internet access or radio and video broadcast, accessible fromanywhere in the world with reasonable costs. During the next years, our writtenworks, both in the office and at home, will depend less and less on paper. There willbe a need for new devices to substitute for paper. These devices should be connect-able, autonomous, light, easy to handle, and shock-proof. With a resolution of about300 dots per inch we could read our newspaper in the train or read a book in the gar-den as comfortably as before, but saving the cost of cutting tons of wood and usingchemical substances to make paper.

We can conclude that globalization is, basically, a communication matter as oldas mankind. Therefore human interaction and multiculturalism are acceleratedwhenever a new communication milestone is reached, such as the larynx, the art ofwriting, the printing press, the telephone, radio, television, or the Internet.

As authors of this book, we wanted to make our modest contribution to theworld of telecommunications, with this work oriented toward the installation, com-missioning, and maintenance of telecommunications networks.

This book is divided into two main parts. The first part covers Chapters 1 to 5,which describe PDH, SDH/SONET, ATM, ADSL, and synchronization architec-

SDH/SONET, ATM, xDSL, and Synchronization Networksxviii

tures. We have chosen these technologies because they are the base on which access,switching; and the majority of voice, data, and video networks, such as POTS, IS-DN, FRL, TCP/IP, GSM, UMTS-3G, and DVB are constructed. The second part,formed by Chapters 6 to 10, is the where installation, commissioning, maintenace,and monitoring technics are explained.

This book can be read in many ways. On the one hand, university students mayfind valuable information in the first five chapters, without needing to read the sec-ond part. On the other hand, companies and engineers that already know the theoret-ical basis of these technologies might only be interested in Chapters 6 to 10.

It is also possible to read this book selectively, since there may be readers whoare only interested in the DSL access network. For them, Chapters 4 and 9 are themost useful, although we would like to recommend Chapters 3 to 8 as well, whichare dedicated to ATM. For those who are interested in the transmission network, themost interesting chapters are 1, 2, 6, and 7. Finally, synchronization, due to its rele-vancy in minimizing transmission level errors, is dealt with independently in Chap-ters 5 to 10. This could naturally also be seen as a continuation for those chaptersthat deal with transmission.

Most of this book was originally written in Spanish. Mia Kosma translated theSpanish text into English and created the English version. We would also like to ex-press our gratitude to Salvador Borràs, who helped us a lot with his jitter and wanderexpertise.

We want to dedicate this book to Benny Moré, the great musician who reachedhis height in the 1950s and who still continues to be a symbol of communication andconcord among human beings.

José M. Caballero

International Marketing Manager - Trend Communications

High Wycombe, England

July 2003

1

Chapter 1

PDH and T-Carrier: The Plesiochronous Hierarchies

1.1 AN INTRODUCTION TO COMMUNICATIONS SYSTEMS

One of the first communications networks known was built by Mediterranean cul-tures more than 1,000 years ago and consisted of a series of successive towers witha distance of about 5 to 12 km between them. A message could be coded and trans-mitted from the first tower to the second one by using optical signals, and then bepassed on along the line until it reached its final destination.

In this primitive system we can already identify all the elements of a genuinecommunications network (see Figure 1.1):

• Information consists of the messages interchanged between final users. In or-der to be introduced into the network, information needs to be coded into sig-nals.

• Signals are a physical magnitude, specific for each transmission medium, thatchange with respect to time.

• The transmission medium consists of the links that connect distant nodes.

• Nodes are those network elements that receive the signals and retransmit themfurther along until reaching the final users.

In other words, in a telecommunications network, user information is distributed assignals from one point to another through the transmission medium that connectsthe nodes in the system.

1.1.1 Signals and Information

The messages to be transmitted are meaningful for the users and are structured hier-archically in lexical, syntactic, and semantic layers, in line with the grammar of the

SDH/SONET, ATM, xDSL, and Synchronization Networks2

natural language used, whereas signals, by comparison, are only meaningful insidethe telecommunications network. The signals used in telecommunications systemscan be of two types (see Table 1.1):

1. Analog or continuous: They can take any of an unlimited number of valueswithin a given range.

2. Digital or discrete: They can only take a limited number of values. In a binarysystem, the only valid values are 0 and 1.

1.1.2 Transmission Medium

The transmission medium can be defined as the environment where a signal istransmitted, be it material (electrical wires, optical fiber, open air, etc.) and nonma-terial, or vacuum, through which only electromagnetic waves are propagated.

The material transmission medium can be divided into two main groups:

1. A conductive medium, in which the information is transmitted in the form ofelectrical impulses. Typical examples of this medium are twisted-pair andcoaxial cables.

Table 1.1 Combinations of signals and information.

Signal Analog Information Digital Information

Analog Modulation (e.g., AM/FM radio and TV) Digital modulation (e.g., ADSL)

Digital Digitalization (e.g., audio CD, GSM) Coding (e.g., frame relay)

Figure 1.1 Elements of a telecommunications network.

Source SinkCoder DecoderTransmitter Receiver

line coding

Transmissionmedium

line decoding data decodingdata codingmessage message

Terminal Node

Terminal Node

Network Nodes

Access Network Access NetworkTransmission Network

InformationSignalsInformation

UserUser

PDH and T-Carrier: The Plesiochronous Hierarchies 3

2. A dielectric medium, in which the information is transmitted in the form ofradioelectrical or optical signals; for example, the atmosphere and optical fiber.

The propagation of signals over one of these media is what we call transmission.The success of transmission of information in telecommunications networks de-pends basically on two factors: the quality of the signal transmitted, and the qualityof the transmission medium used. In addition, there are natural forces that can resisttransmission and modify the original characteristics of the signals, which may endup being degraded by the time they reach their destination.

The most significant impairments are attenuation, noise, and distortion. Welook at these below in respect to a communications channel, which is defined as ameans of unidirectional transmission of signals between two points.

1.1.2.1 Attenuation

Attenuation weakens the power of the signal proportionally to the transmission me-dium length. It is expressed in decibels (AdB) through the logarithmic ratio of thetransmitted power (PTx) and received power (PRx), measured at both ends of the

Transmitted signal

Attenuation

Distortion

Noise

Received signal

SourceTransmitter

Tran

smis

sion

med

ium

Receiver Sink

dist

ance

( d)

PTx

PRx

Figure 1.2 Effects of attenuation, distortion, and noise on transmission.

Sampling timesData receivedOriginal data

1 0 0 01 0 1 0


distance (d) being examined (see Figure 1.2). Transmission media can usually becharacterized by their attenuation per unit of length (AdB / Km):

Example: Thus for a transmission medium with A=0.2 dB/Km, after 15 Km, the at-tenuation is AdB=3 dB. If the transmitted power is PTx=1W. After 10 Km receivedpower is PRx= 0.5W, because 10 log (1/PRx) = 3 dB (see Figure 1.3).

At the far end the received signal must have enough power (PRx) to be interpret-ed, otherwise amplifiers (also known as repeaters or regenerators in digital transmis-sion) must be inserted along the transmission medium to improve the power of thereceived signal.

1.1.2.2 Distortion

Distortion produces a change in the original shape of the signal at the receiver end.There are two types: amplitude distortion and delay distortion.

• When the impairments affect the amplitudes of the frequency components ofthe signal differently, this is said to produce amplitude distortion (sometimescalled absorption). Amplitude distortion is caused because the transmissionchannel is limited to certain frequencies (see Figure 1.4). To overcome thisproblem amplifiers must equalize the signal, separately amplifying each bandof frequencies.1

1. Note that attenuation is a specific case of amplitude distortion that equally affects allfrequencies of the signal.

10 PTx PRx⁄( )log d AdB Km⁄⋅=

AdB d A⋅ dB Km⁄=

0.0

0.5

1.0

2.5

3.0

101000 1310 1400 1550Wavelength (nm)

Atte

nuat

ion

(dBm

/Km

)

2.5

1.5

Figure 1.3 Typical attenuation values for single mode optical fiber and coaxial cable.

1

100

Atte

nuat

ion

(dBm

/Km

)

10

1200 1KHz 100 1MHz 10Frequency

800

Optical fiber Coaxial cable


• When the velocity of propagation of a signal varies with the frequency, there issaid to be delay distortion (sometimes called dispersion). Delay distortion isparticularly disturbing in the digital transmission producing intersymbol inter-ference (ISI), where a component of the signal of one bit is misplaced in thetime slot reserved for another bit. ISI limits the capacity to extract digital infor-mation from the received signal.

Harry Nyquist showed that the maximum transmission capacity (C) is limited byISI and depends on the channel bandwidth (B) and the number of signal elements(M) coding the information.

Example: For a modem using 16 signal elements and a channel bandwidth (B) of4,000 hertz (Hz), the maximum data transfer rate (C) is 32,000 bits per second(bps).

1.1.2.3 Noise

Noise refers to any undesired and spurious signal that is added to an informationsignal. It is usually divided into five categories:

1. Thermal noise: This is caused by the agitation of electrons in any conductor ina temperature different than absolute zero. The noise (N) is independent of thefrequency and proportional to the bandwidth (B) and the temperature (T) indegrees Kelvin:

(k is the Boltzmann’s constant in joules/kelvin, k = 1.3803 x 10-23)

Figure 1.4 The two basic transmission channels. In the frequency domain the channel transfer function H(f) determines the attenuation of each frequency and consequently the amplitude distortion.

fffo f2f1

Bandwidth Bandwidth

H(f)H(f)

Lowpass Bandpass

Cbps 2Blog2M=

N k T B⋅ ⋅=


2. Intermodulation noise: This is caused when two or more signals of frequenciesf1 and f2, transmitted in the same medium, produce a spurious signal at fre-quencies that are a linear combination of the previous ones.

3. Atmospheric noise: This is caused by the static discharge of clouds, or ionizedgas from the sun, or high frequency signals radiated by the stars.

4. Impulse noise: Of short duration but high amplitude, these energy bursts arecaused by sources such as electrical machinery, a drop in voltage, atmosphericinterference, and so on. These do not tend to be a problem for analog signals,but are a prime cause of errors in digital transmission.

5. Crosstalk: Whenever a current flows through a conductor a magnetic field isset up around it that can induct a current into a second conductor collocated ina short distance.

Noise is always present in transmission channels, even when no signal is beingtransmitted. A key parameter at the receiver end to distinguish between informationand spurious power is the signal-to-noise ratio (S/N):

Claude Shannon proved that the signal-to-noise ratio (S/N) determines the the-oretical maximum transmission capacity (C) in bits per second of channel with a lim-ited bandwidth (B):

Example: A typical value of S/N for a voice grade line is 30 dB (equivalent to apower ratio of 1,000:1). Thus for a bandwidth of 3,100 Hz the maximum data trans-fer rate (C) should be 30,894 bps.

If we pay attention only to the Nyquist formula (see Section 1.1.2.2) we couldinaccurately conclude that for a given bandwidth (B) the data rate can be increasedendlessly, by increasing the number of signal elements. However in reality, the sig-nal-to-noise ratio sets up the theoretical limit of the channel capacity.

The Shannon theorem makes no statement as to how the channel capacity isachieved. In fact, channels only approach this limit. The task of providing high chan-nel efficiency is the goal of coding techniques.

S N⁄( )dB 10 PowerSignal PowerNoise⁄( )log=

Cbps Blog2 1 S N⁄+( )=


1.1.2.4 The transmission channel

A digital channel is a communication subsystem with capacity to send and receiveinformation between two points: a source and a sink. Related concepts are:

• Bandwidth, expressed in hertz (Hz). This is the difference between the highestand the lowest frequency that can be transmitted across a line or a network.

• Data rate, expressed in bits per second (bps). This is a measure of the speedwith which information is transferred. It depends on the bandwidth, transmis-sion medium impairments, and the technological capacity to efficiently use theavailable bandwidth.

• Performance, expressed in bit error rate (BER). This is the probability of a sin-gle bit being corrupted in a defined interval. Performance is on indication ofthe quality of the channel.

Channel capacity is the data rate that can be transmitted over a communication pathunder specific conditions.When two channels define a two-way communication, itis more usual to talk about a circuit.

1.1.3 Channel Coding

Channel coding is the process that transforms binary data bits into signal elementsthat can cross the transmission medium. In the simplest case, in a metallic wire a bi-nary 0 is represented by a lower voltage, and a binary 1 by a higher voltage. How-ever, before selecting a coding scheme it is necessary to identify some of thestrengths and weaknesses of line codes:

• High-frequency components are not desirable because they require more chan-nel bandwidth, suffer more attenuation, and generate crosstalk in electricallinks.

• Direct current (dc) components should be avoided because they require physi-cal coupling of transmission elements. Since the earth/ground potential usuallyvaries between remote communication ends, dc provokes unwanted earth-re-turn loops.

• The use of alternating current (ac) signals permits a desirable physical isola-tion using condensers and transformers.

• Timing control permits the receiver to correctly identify each bit in the trans-mitted message. In synchronous transmission, the timing is referenced to thetransmitter clock, which can be sent as a separate clock signal, or embeddedinto the line code. If the second option is used, then the receiver can extract itsclock from the incoming data stream thereby avoiding the installation of an ad-ditional line.


In order to meet these requirements, line coding is needed before the signal is trans-mitted, along with the corresponding decoding process at the receiving end. Thereare a number of different line codes that apply to digital transmission, the mostwidely used ones are alternate mark inversion (AMI), high-density bipolar three ze-ros (HDB3), and coded mark inverted (CMI).

1.1.3.1 Nonreturn to zero

Nonreturn to zero (NRZ) is a simple method consisting of assigning the bit “1”to the positive value of the signal amplitude (voltage), and the bit “0” to the nega-tive value (see Figure 1.5). There are two serious disadvantages to this:

1. No timing information is included in the signal, which means that synchronismcan easily be lost if, for instance, a long sequence of zeros is being received.

2. The spectrum of the signal includes a dc component.

1.1.3.2 Alternate mark inversion

Alternate mark inversion (AMI) is a transmission code, also known as pseudo-ternary, in which a “0” bit is transmitted as a null voltage and the “1” bits are repre-

Figure 1.5 Line encoding technologies. AMI and HDB3 are usual in electrical signals, while CMI is often used in optical signals.

0

B8ZS

HDB3

CMI

0

+V

-V

0

+V

-V

0

+V

-V

0

+V

-V

0 0 0 V

B 0 0 V

B 0 0 V

BipolarEight-ZeroSuppression

HighDensityBipolarThreeZeros

CodedMarkInverted

B: balancingV: violation

NRZ 0

+V

-VNon- Return toZero

AMIAlternateMarkInversion

- 0 0 0 V + 0

0 1 0 0 01 1 1 0 0 0 00 0 0 0 1 0

V -


sented alternately as positive and negative voltage. The digital signal coded in AMIis characterized as follows (see Figure 1.5):

• The dc component of its spectrum is null.

• It does not solve the problem of loss of synchronization with long sequences ofzeros.

1.1.3.3 Bit eight-zero suppression

Bit eight-zero suppression (B8ZS) is a line code in which bipolar violations are de-liberately inserted if the user data contains a string of eight or more consecutive ze-ros. The objective is to ensure a sufficient number of transitions to maintain thesynchronization when the user data stream contains a large number of consecutivezeros (see Figure 1.5 and Figure 1.6).

The coding has the following characteristics:

• The timing information is preserved by embedding it in the line signal, evenwhen long sequences of zeros are transmitted, which allows the clock to be re-covered properly on reception

• The dc component of a signal that is coded in B8Z3 is null.

1.1.3.4 High-density bipolar three zeroes

High-density bipolar three zeroes (HDB3) is similar to B8ZS, but limits the maxi-mum number of transmitted consecutive zeros to three (see Figure 1.5). The basicidea consists of replacing a series of four bits that are equal to “0” with a code word“000V” or “B00V,” where “V” is a pulse that violates the AMI law of alternate po-larity, and B it is for balancing the polarity.

Figure 1.6 B8ZS and HDB3 coding. Bipolar violations are: V+ a positive level and V- negative.

+

–

Last pulsepolarity

B8ZS Number of ones

B-00V-+

–

Last ‘1’ polarity

HDB3

000V- B+00V+

000V+

Odd EvenSubstitution

000V+–0V-+

000V-+0V+–


• “B00V” is used when, until the previous pulse, the coded signal presents a dccomponent that is not null (the number of positive pulses is not compensatedby the number of negative pulses).

• “000V” is used under the same conditions as above, when, until the previouspulse, the dc component is null (see Figure 1.6).

• The pulse “B” (for balancing), which respects the AMI alternation rule and haspositive or negative polarity, ensuring that two consecutive “V” pulses willhave different polarity.

1.1.3.5 Coded mark inverted

The coded mark inverted (CMI) code, also based on AMI, is used instead of HDB3at high transmission rates, because of the greater simplicity of CMI coding and de-coding circuits compared to the HDB3 for these rates. In this case, a “1” is transmit-ted according to the AMI rule of alternate polarity, with a negative level of voltageduring the first half of the period of the pulse, and a positive level in the secondhalf. The CMI code has the following characteristics (see Figure 1.5):

• The spectrum of a CMI signal cancels out the components at very low frequen-cies.

• It allows for the clock to be recovered properly, like the HDB3 code.

• The bandwidth is greater than that of the spectrum of the same signal coded inAMI.

1.1.4 Multiplexing and Multiple Access

Multiplexing is defined as the process by which several signals from different chan-nels share a channel with greater capacity (see Figure 1.7). Basically, a number ofchannels share a common transmission medium with the aim of reducing costs andcomplexity in the network. When the sharing is carried out with respect to a remoteresource, such as a satellite, this is referred to as multiple access rather than multi-plexing.

Some of the most common multiplexing technologies are:

1. Frequency division multiplexing/frequency division multiple access (FDM/FDMA): Assigns a portion of the total bandwidth to each of the channels.

2. Time-division multiplexing/time division multiple access (TDM/TDMA):Assigns all the transport capacity sequentially to each of the channels.

3. Code-division multiplexing access (CDMA): In certain circumstances, it ispossible to transmit multiple signals in the same frequency, with the receiverbeing responsible for separating them. This technique has been used for years


in military technology, and is based on artificially increasing the bandwidth ofthe signal according to a predefined pattern.

4. Polarization division multiple access (PDMA): Given that polarization can bemaintained, the polarization direction can be used as a multiple access tech-nique, although when there are many obstacles, noise can make it unsuitable,which is why it is not generally used in indoor installations. Outside, however,it is widely exploited to increase transmission rates in installations that usemicrowaves.

5. Space division multiple access (SDMA): With directional antennas, the samefrequency can be reused, provided the antennas are correctly adjusted. There isa great deal of interference, but this system lets frequencies obtain a highdegree of reusability.

1.2 PULSE CODE MODULATION

The pulse code modulation (PCM) technology (see Figure 1.8) was patented anddeveloped in France in 1938, but could not be used because suitable technologywas not available until World War II. This came about with the arrival of digital

DTE-AB1

DTE-BB2

DTE-nBn

.

.

.

Figure 1.7 Multiplexing consolidates lower capacity channels into a higher capacity channel. Frequency division multiplexing access (FMDA) is used by radio, TV, and global system mobile (GSM). Time division multiplexing access (TDMA) is used by the integrated services digital network (ISDN), frame relay (FRL), and GSM. Code divi-sion multiplexing access (CDMA) is used by the third generation networks (3G).

AAB

CDE

FBCDEFAB

TDMAFDMA

time

0 0 1 0 1 1 1 0 1 1 1 0 1 1 1 0 0 1

1 1 0 1 0 0 0 1 0 1 1 0 1 1 1 0 0 1

code bit

CDMA

frequency

DTE-A B1

DTE-B B2

DTE-nBn

.

.

. Transmission media

ΣBi

n

m

n

m

Multiplexer Demultiplexer

Multiplexing

Bi = bandwidth

.

.

.

.

.

.

Multiplexing technologies

pattern

data

signal

0 1


systems in the 1960s, when improving the performance of communications net-works became a real possibility. However, this technology was not completelyadopted until the mid-1970s, due to the large amount of analog systems already inplace and the high cost of digital systems, as semiconductors were very expensive.PCM’s initial goal was that of converting an analog voice telephone channel into adigital one based on the sampling theorem (see Figure 1.9):

The sampling theorem states that for digitalization without information loss, thesampling frequency (fs) should be at least twice the maximum frequency component(fmax) of the analog information:

The frequency 2·fmax is called the Nyquist sampling rate. The sampling theoremis considered to have been articulated by Nyquist in 1928, and mathematically prov-en by Shannon in 1949. Some books use the term Nyquist sampling theorem, andothers use Shannon sampling theorem. They are in fact the same theorem.

PCM involves three phases: sampling, encoding, and quantization:

1. In sampling, values are taken from the analog signal every 1/fs seconds (thesampling period).

Figure 1.8 Pulse code modulation (PCM) was the technology selected to digitalize the voice in telephone networks. Other pulse techniques are pulse amplitude modulation (PAM), pulse duration modulation (PDM), and pulse position modulation (PPM).

PAM3

7

3

-3-1to

PDM1 3 1 5 4

t

tot

PPMto

t

tot

PCM

1 3 1 5 4

0 1 10 0 1 0 0 1 1 0 1 1 0 0

Amplitude

Sampling

to t

7

3

-1-3

V

Puls

e m

odul

atio

n te

chni

ques

fs 2 f⋅ max>


2. Quantization assigns these samples a value by approximation, and in accor-dance with a quantization curve (i.e., A-law of ITU-T 2).

3. Encoding provides the binary value of each quantified sample.

A telephone channel admits frequencies of between 300 Hz and 3,400 Hz. Becausemargins must be established in the channel, the bandwidth is set at 4 kHz. Then thesampling frequency must be ; equivalent to a sampleperiod of .

In order to codify 256 levels, 8 bits are needed, where the PCM bit rate (v) is:

This bit rate is the subprimary level of transmission networks.

1.3 PDH AND T-CARRIER

At the beginning of the 1960s, the proliferation of analog telephone lines, based oncopper wires, together with the lack of space for new installations, led the transmis-

2. This is a International Telecommunication Union (ITU-T) ratified audio encoding andcompression technique (Rec. G.711). Among other implementations, A-law was orig-inally intended as a phone-communications standard.

Figure 1.9 The three steps of digitalization of a signal: sampling of the signal, quantization of the amplitude, and binary encoding.

Ampl

itude

(vol

ts)

0

Sampling timeT

n

Sampling

2T

Cod

e Am

plitu

de

1

Quantization

1

23

56

4

Analog Signal

000001

011

101

111

010

100

110

Coding

t

t

t

1 1 1 1 1 1 1 1 1 10 0 0 0 0 0 0 0 0 0 0

0 1

2

3

3T 4T 5T

V

T 2T t3T 4T 5T

fs 2 4 000, 8 000,= Hz⋅≥T 1 8 000, 125µs=⁄=

v 8 000, samples s⁄ 8bits sample⁄× 64Kbps= =


sion experts to look at the real application of PCM digitalization techniques andTDM multiplexing. The first digital communications system was set up by BellLabs in 1962, and consisted of 24 digital channels running at what is known as T1.

1.3.1 Basic Rates: T1 and E1

In 1965, a standard appeared in the U.S. that permitted the TDM multiplexing of 24digital telephone channels of 64 Kbps into a 1.544-Mbps signal with a formatcalled T1 (see Figure 1.10). For the T1 signal, a synchronization bit is added to the24 TDM time slots, in such a way that the aggregate transmission rate is:

125 µs is the sampling period

Europe developed its own TDM multiplexing scheme a little later (1968), al-though it had a different capacity: 32 digital channels of 64 Kbps (see Figure 1.10).The resulting signal was transmitted at 2.048 Mbps, and its format was called E1which was standardized by the ITU-T and adopted worldwide except in the U.S.,

139264 Kbps

34368 Kbps

8448 Kbps

97728 Kbps

6312 Kbps

Figure 1.10 The PDH and T-carrier hierarchies, starting at the common 64-Kbps channel and the multiplexing levels. Most of the narrowband networks are built on these stan-dards: POTS, FRL, GSM, ISDN, ATM (asynchronous transfer mode), and leased lines to transmit voice, data, and video.

139264 Kbps

x4

34368 Kbps

x4

8448 Kbps

x4

2048 Kbps

64 Kbps

1544 Kbps

x2

44736 Kbps 32064 Kbps

97728 Kbps

x3

x30 x24

x3

x7 x5

x3

4th Level

3rd Level

2nd Level

PDH T-carrierJapan

Single Channel

1st Level

E4

E3

E2

E1 T1 J1

T2 J2

J3

J4

T3

6312 Kbps

U.S. and Canada

3152 Kbps

x2

T1c J1c

worldwide

PDH

24channels 8bit channel⁄ 1bit+×( ) 125µs⁄ 1.544Mbps=


Canada, and Japan. For an E1 signal, the aggregate transmission rate can be obtainedfrom the following equation:

1.4 THE E1 FRAME

The E1 frame defines a cyclical set of 32 time slots of 8 bits. The time slot 0 is de-voted to transmission management and time slot 16 for signaling; the rest were as-signed originally for voice/data transport (see Figure 1.11).

The main characteristics of the 2-Mbps frame are described in the following.

1.4.1 Frame Alignment

In an E1 channel, communication consists of sending consecutive frames from thetransmitter to the receiver. The receiver must receive an indication showing whenthe first interval of each frame begins, so that, since it knows to which channel theinformation in each time slot corresponds, it can demultiplex correctly. This way,

30channels 8bit channel⁄×( ) 125µs⁄ 2.048Mbps=

0 1C1 0 1 0 1 1 0 00 0 S A S SA S0 1 S S S S c 1 d 1a 1 b 1 a16 b16 c16 d16

c 2 d 2a 2 b 2 a17 b17 c17 d170 1C2 0 1 0 1 1A S0 1 S S S S c 3 d 3a 3 b 3 a18 b18 c18 d18

c 4 d 4a 4 b 4 a19 b19 c19 d190 1C3 0 1 0 1 1

Frame 01234

A S0 1 S S S S c 5 d 5a 5 b 5 a20 b20 c20 d20


567

A S0 1 S S S S c 9 d 9a 9 b 9 a24 b24 c24 d24

c10 d10a10 b10 a25 b25 c25 d250 1C2 0 1 0 1 1A S0 1 S S S S c11 d11a11 b11 a26 b26 c26 d26

c12 d12a12 b12 a27 b27 c27 d270 1C3 0 1 0 1 1

9101112

A SE 1 S S S S c13 d13a13 b13 a28 b28 c28 d28

c15 d15a15 b15 a29 b29 c29 d290 1C4 0 1 0 1 1A SE 1 S S S S c16 d16a16 b16 a30 b30 c30 d30

131415

c 8 d 8a 8 b 8 a23 b23 c23 d230 1C1 0 1 0 1 18

Time Slot 0 1 15 Time Slot 16. . . 17 31. . .

125 µs

Submultiframe I

Submultiframe II

2 msChannel 1 15 16 30. . .. . .

Remote Alarm IndicatorChannel CAS BitsAlignment Bits

CRC-4 Bits

CRC-4 Error Signaling Bits

C1

A

SE

a17 b17 c17 d17

C2 C3 C4

Spare Bits

1 0...

Channel Bytes

Figure 1.11 The E1 frame is the first hierarchy level, and all the channels are fully synchronous.


the bytes received in each slot are assigned to the correct channel. A synchroniza-tion process is then established, and it is known as frame alignment.

1.4.2 Frame Alignment Signal

In order to implement the frame alignment system so that the receiver of the framecan tell where it begins, there is what is called a frame alignment signal (FAS) (seeFigure 1.12). In the 2Mbps frames, the FAS is a combination of seven fixed bits(“0011011”) transmitted in the first time slot in the frame (time slot zero or TS0).For the alignment mechanism to be maintained, the FAS does not need to be trans-mitted in every frame. Instead, this signal can be sent in alternate frames (in thefirst, in the third, in the fifth, and so on). In this case, TS0 is used as the synchroni-zation slot. The TS0 of the rest of the frames is therefore available for other func-tions, such as the transmission of the alarms.

1.4.3 Multiframe CRC-4

In the TS0 of frames with FAS, the first bit is dedicated to carrying the cyclic re-dundancy checksum (CRC). It tells us whether there are one or more bit errors in aspecific group of data received in the previous the previous block of eight framesknown as submultiframe (see Figure 1.13).

1.4.3.1 The CRC-4 procedure

The aim of this system is to avoid loss of synchronization due to the coincidentalappearance of the sequence “0011011” in a time slot other than the TS0 of a framewith FAS. To implement the CRC code in the transmission of 2-Mbps frames, aCRC-4 multiframe is built, made up of 16 frames. These are then grouped in twoblocks of eight frames called submultiframes, over which a CRC checksum or wordof four bits (CRC-4) is put in the positions Ci (bits #1, frames with FAS) of the nextsubmultiframe.

At the receiving end, the CRC of each submultiframe is calculated locally andcompared to the CRC value received in the next submultiframe. If these do not co-incide, one or more bit errors is determined to have been found in the block, and an

0 1C1 0 1 0 1 1 0 00 0 S A S SA S0 1 S S S S c 1 d 1a 1 b 1 a16 b16 c16 d16


c 4 d 4a 4 b 4 a19 b19 c19 d190 1C3 0 1 0 1 1

01234

A S0 1 S S S S c 5 d 5a 5 b 5 a20 b20 c20 d20


567

A S0 1 S S S S c 9 d 9a 9 b 9 a24 b24 c24 d24

c10 d10a10 b10 a25 b25 c25 d250 1C2 0 1 0 1 1A S0 1 S S S S c da b a b c d

91011

c 8 d 8a 8 b 8 a23 b23 c23 d230 1C1 0 1 0 1 18

Time Slot 0 1 15 Time Slot 16. . . 17 31. . .

125 µs

Submultiframe I

Sub-multiframe II

c 6 d 6a 6 b 6 a21 b21 c21 d21C4

A S0 S S S S c 7 d 7a 7 b 7 a22 b22 c22 d22

67

A S0 S S S S c 9 d 9a 9 b 9 a24 b24 c24 d24

c10 d10a10 b10 a25 b25 c25 d25C2

A S0 S S S S c da b a b c d

91011

c 8 d 8a 8 b 8 a23 b23 c23 d23C18

Submultiframe II

0 1 1 0 1 1

0 1 1 0 1 1

0 1 1 0 1 11

1

1

0

0

0

FAS NFAS

0 1 1 0 1 1

0 1 1 0 1 11

1

0

0

0 1 1 0 1 1

0 1 1 0 1 1

0 1 1 0 1 1

10

0

0

Figure 1.12 The E1 multiframe uses the FAS code only transmitted in even frames. The NFAS frames are the odd ones, using a bit equal to “1” to avoid coincidences.


alarm is sent back to the transmitter, indicating that the block received at the far endcontains errors (see Table 1.2).

1.4.3.2 CRC-4 multiframe alignment

The receiving end has to know which is the first bit of the CRC-4 word (C1). Forthis reason, a CRC-4 multiframe alignment word is needed. Obviously, the receiverhas to be told where the multiframe begins (synchronization).

The CRC-4 multiframe alignment word is the set combination “001011,” whichis introduced in the first bits of the frames that do not contain the FAS signal.

1.4.3.3 Advantages of the CRC-4 method

The CRC-4 method is mainly used to protect the communication against a wrongframe alignment word, and also to provide a certain degree of monitoring of the biterror rate (BER), when this has low values (around 10-6). This method is not suit-able for cases in which the BER is around 10-3 (where each block contains at leastone errored bit).

Another advantage in using the CRC is that all the bits transmitted are checked,unlike those systems that only check seven bits (those of the FAS, which are the onlyones known in advance) out of every 512 bits (those between one FAS and the next).However, the CRC-4 code is not completely infallible, since there is a probability ofaround 1/16 that an error may occur and not be detected; that is, that 6.25% of theblocks may contain errors that are not detected by the code.

Figure 1.13 The CRC-4 provides error monitoring by means of four Ci bits that correspond to the previous submultiframe. If the receiver detects errors, it sets the E-bit to indi-cate the error. The “001011”sequence is used to synchronize the submultiframe.

0 10 1 0 1 1A S1 S S S S0 10 1 0 1 1A S1 S S S S0 10 1 0 1 1

1

3

A S1 S S S S0 10 1 0 1 1A S1 S S S S

5

7

A S1 S S S S0 10 1 0 1 1A S1 S S S S0 10 1 0 1 1

9

11

A S1 S S S S0 10 1 0 1 1A S1 S S S S

13

15

0 10 1 0 1 18

0

0

0

0

0

0

E

E

C1

0 C1

2

4

6

10

C2

C3

C4

C2

14 C4

12 C3

Submultiframe I

Submultiframe II

0 10 1 0 1 1A S1 S S S S0 10 1 0 1 1A S1 S S S S0 10 1 0 1 1

0

A S1 S S S S0 10 1 0 1 1A S1 S S S S

A S1 S S S S0 10 1 0 1 1A S1 S S S S0 10 1 0 1 1

10

12A S1 S S S S0 10 1 0 1 1A S1 S S S S

14

0 10 1 0 1 1

C1

C2

C3

2

4

C46

C2

C3

C4

C18

01

11

13

15

03

1

0

5

7

1

1

9

E

E

Submultiframe I

Submultiframe II


1.4.3.4 Monitoring errors

The aim of monitoring errors is to continuously check transmission quality withoutdisturbing the information traffic and, when this quality is not of the required stan-dard, taking the necessary steps to improve it. Telephone traffic is two way, whichmeans that information is transmitted in both directions between the ends of thecommunication. This in its turn means that two 2-Mbps channels and two direc-tions for transmission must be considered.

The CRC-4 multiframe alignment word only takes up six of the first eight bitsof the TS0 without FAS. There are two bits in every second block or submultiframe,whose task is to indicate block errors in the far end of the communication. The mech-anism is as follows: Both bits (called E-bits) have “1” as their default value. Whenthe far end of the communication receives a 2Mbps frame and detects an erroneousblock, it puts a “0” in the E-bit that corresponds to the block in the frame being sentalong the return path to the transmitter (see Figure 1.14). This way, the near end ofthe communication is informed that an erroneous block has been detected, and bothends have the same information: one from the CRC-4 procedure and the other fromthe E bits. If we number the frames in the multiframe from 0 to 15, the E-bit offrame 13 refers to the submultiframe I (block I) received at the far end, and the E-bitof frame 15 refers to the submultiframe II (block II).

1.4.4 Supervision Bits

The bits that are in position 2 of the TS0 in the frame that does not contain the FASare called supervision bits and are set to “1,” to avoid simulations of the FAS sig-nal.

Figure 1.14 The A multiplexer calculates and writes the CRC code, and the multiplexer B reads and checks the code. When errors affect the 2-Mbps frame, the multiplexer B indi-cates the problem by means of the E-bit of the frame which travels toward the mul-tiplexer B.

REBE (bit E=1)

CRC-4 Writer CRC-4 Reader

2

2 Errors

Multiplexer A Multiplexer B

Error Indication Reader Error Indication Writer

.

.

.

.

.

.

64

64

2 Mbps

64 Kbps 64 Kbps


1.4.5 NFASs - Spare Bits

The bits of the TS0 that do not contain the FAS in positions 3 to 8 make up what isknown as the nonframe alignment signal or NFAS. This signal is sent in alternateframes (frame 1, frame 3, frame 5, etc.). The first bit of the NFAS (bit 3 of the TS0)is used to indicate that an alarm has occurred at the far end of the communication.When operating normally, it is set to “0,” while a value of “1” indicates an alarm.

The bits in positions 4 to 8 are spare bits (see Figure 1.15), and they do not haveone single application, but can be used in a number of ways, as decided by the tele-communications carrier. In accordance with the ITU-T Rec. G.704, these bits can beused in specific point-to-point applications, or to establish a data link based on mes-sages for operations management, maintenance or monitoring of the transmissionquality, and so on. If these spare bits in the NFAS are not used, they must be set to“1” in international links.

1.4.6 NFAS - Alarm Bit

The method used to transmit the alarm makes use of the fact that in telephone sys-tems, transmission is always two way (see Figure 1.16). Multiplexing/demultiplex-ing devices (known generically as multiplex devices) are installed at both ends ofthe communication for the transmission and reception of frames. An alarm must besent to the transmitter when a device detects either a power failure or a failure of thecoder/decoder, in its multiplexer; or any of the following in its demultiplexer: lossof the signal (LOS), loss of frame alignment (LOF), or a BER greater than 10-3.

0C3 04 1 1 0 1 1

0C2 02 1 1 0 1 1

0C1 0 0 00 00 1 c 1 d 1a 1 b 1 a16 b16 c16 d16

c 2 d 2a 2 b 2 a17 b17 c17 d17

0 1 c 3 d 3a 3 b 3 a18 b18 c18 d18

c 4 d 4a 4 b 4 a19 b19 c19 d19

Frame 01

3

0 1 c 5 d 5a 5 b 5 a20 b20 c20 d20

c 6 d 6a 6 b 6 a21 b21 c21 d210 1C4 0 1 0 1 156

A

A

A

AS S S1 1 0 1 1S S S S S

S S S S S

S S S S S

Figure 1.15 Spare bits in the E1 frame.

0 1C1 0 1 0 1 1 0 00 0 S S SS0 1 S S S S c 1 d 1a 1 b 1 a16 b16 c16 d16

c 2 d 2a 2 b 2 a17 b17 c17 d170 1C2 0 1 0 1 1S0 1 S S S S c 3 d 3a 3 b 3 a18 b18 c18 d18

c 4 d 4a 4 b 4 a19 b19 c19 d190 1C3 0 1 0 1 1

Frame 01234

S0 1 S S S S c 5 d 5a 5 b 5 a20 b20 c20 d20

c 6 d 6a 6 b 6 a21 b21 c21 d210 1C4 0 1 0 1 156

Submultiframe II

A

A

A

A

Figure 1.16 The alarm indication signal is used to send alarms to the remote end to indicate a power fault, loss of incoming signal, loss of frame, coder/decoder fault or a high bit error rate, among others.


The remote alarm indication (RAI) is sent in the NFAS of the return frames,with bit 3 being set to “1.” The transmitter then considers how serious the alarm is,and goes on generating a series of operations, depending on the type of alarm con-dition detected (see Table 1.2).

1.4.7 Signaling Channel

As well as transmitting information generated by the users of a telephone network,it is also necessary to transmit signaling information. Signaling refers to the proto-cols that must be established between exchanges so that the users can exchange in-formation between them.

There are signals that indicate when a subscriber has picked up the telephone,when he or she can start to dial a number, and when another subscriber calls, as wellas signals that let the communication link be maintained, and so on.

In the E1 PCM system, signaling information can be transmitted by two differ-ent methods: the common channel signaling (CCS) method and the channel associ-ated signaling (CAS) method. In both cases, the time slot TS16 of the basic 2-Mbpsframe is used to transmit the signaling information (see Figure 1.17).

For CCS signaling, messages of several bytes are transmitted through the64-Kbps channel provided by the TS16 of the frame, with these messages providingthe signaling for all the channels in the frame. Each message contains informationthat determines the channel that is signaling. The signaling circuits access the64-Kbps channel of the TS16, and they are also common to all the channels signaled.There are different CCS systems that constitute complex protocols. In the followingsection and by way of example, channel associated signaling will be looked at in de-tail. CAS is defined in the ITU-T Rec. G.704, which defines the structure of the E1frame.

In CAS signaling, a signaling channel is associated with each information chan-nel (there is no common signaling channel), meaning that the signaling circuits arepersonalized for each channel.

1.4.8 CAS Signaling Multiframe

In the case of channel associated signaling, each 64Kbps telephone channel is as-signed 2 Kbps for signaling. This signaling is formed by a word of 4 bits (generical-ly known as a, b, c, and d) that is situated in the TS16 of all the frames sent. EachTS16 therefore carries the signaling for two telephone channels.

Given that there are only four signaling bits available for each channel, to trans-mit all the signaling words from the 30 PCM channels that make up a 2-Mbps frame


(120 bits), it is necessary to wait until the TS16 of 15 consecutive frames have beenreceived. The grouping of frames defines a CAS signaling multiframe, which con-sists of a set of the TS16 of 16 consecutive E1 frames.

1.4.8.1 CAS multiframe alignment signal

In order to synchronize the CAS multiframe, that is to identify where it begins, amultiframe alignment signal (MFAS) is defined, made up of the sequence of bits“0000” located in the first four bits of the TS16 of the first frame of the CAS multi-frame.

1.4.8.2 CAS nonmultiframe alignment signal

The remaining four bits of the interval are divided between one alarm bit and threespare bits, making up the nonmultiframe alignment signal (NMFAS). In short, thesignaling information for the 30 channels is transmitted in 2 ms, which is fastenough if we consider that the shortest signaling pulse (the one that corresponds todialing the number) lasts for 100 ms.

The alarm bit in the NMFAS is dealt with in a similar way to the NFAS. In thiscase, the alarms are transmitted between multiplex signaling devices connected tothe 64-Kbps circuits that correspond to signaling (TS16). The alarm is sent when theCAS multiplexer detects:

• A power failure;

• Loss of incoming signaling;

• Loss of CAS multiframe alignment.

An indication must be sent to the multiplex signaling device at the far end (see Ta-ble 1.2), setting bit 6 of the TS16 in the return frame 0 to “1.” Additionally, the val-ue “1” is applied to all the signaling channels (see Figure 1.21).



c 4 d 4a 4 b 4 a19 b19 c19 d190 1C3 0 1 0 1 1

Frame 01234

S0 1 S S S S c 5 d 5a 5 b 5 a20 b20 c20 d20

c 6 d 6a 6 b 6 a21 b21 c21 d210 1C4 0 1 0 1 156

Submultiframe II

A

A

A

ATime Slot 0 1 15 Time Slot 16. . . 17 31. . .

Figure 1.17 When the CAS method is used, each of the channels has an associated 2-Kbps channel (ai bi ci di) in the time slot 16. This multiframe also has an alignment signal “0000”; spare and alarm bit to be used specifically by the signaling multiframe.


Example: A remote multiplexer is considered to have lost multiframe alignmentwhen it receives two consecutive MFAS words with error, that is, with a value otherthan “0000.” In this case, bit 6 of the TI16 of the frame 0 that this multiplexer trans-mits to the near-end multiplexer is set to “1.” When it receives this indication of lossof multiframe alignment at the far end, the near end multiplexer sends a signal madeup entirely of bits at “1,” known as AIS64 (alarm indication signal - 64 Kbps) in theTS16 (64-Kbps channel).

1.5 THE PLESIOCHRONOUS DIGITAL HIERARCHY

Based on the E1 signal, the ITU-T defined a hierarchy of plesiochronous signalsthat enables signals to be transported at rates of up to 140 Mbps (see Figure 1.18).This section describes the characteristics of this hierarchy and the mechanism fordealing with fluctuations in respect to the nominal values of these rates, which areproduced as a consequence of the tolerances of the system.

Table 1.2 2-Mbps events: Alarms, errors, and event indications.

ID Explanation

AIS Alarm indication signal. It is detected if there are two or less zeros (ITU-T G.775).LOF Loss of frame alarm. It is raised after three consecutive frames with FAS error or

three consecutive signalling bits (ITU-T G.706).LOS Loss of frame signal alarm.RAI Remote alarm indication. It is detected after three consecutive frames with the A

bit equals to 1 (ITU-T G.732).FAS error Frame alignment signal error, indicationg an incorrect bit in the alignment word.Bit error Bit sequence mismatch (when the transmitted pattern is known).Code error Violation on coding sequence.CRC-LOM Cyclic redundancy checksum - loss of multiframe. It is activated if there is LOF

and deactivated after one correct FAS and two correct CRC-MFAS (ITU-T G.706).CAS-LOM Channel associated signaling-loss of multiframe. It is raised after two consecutive

MFAS errors or two multiframes with time-slot 16 bits equal to 0 (ITU-T G.732).CAS-MRAI Channel associated signaling-multiframe remote alarm indication. Detected after

two consecutive frames with the MRAI bit equal to 1 (ITU-T G.732).CAS-MAIS Channel associated signaling-multiframe alarm indication signal. It is detected if

there are less than three zeros in the time slot 16 during two consecutive multiframes.CRC error Cyclic redundancy check error. It is raised whenever one or more bits are errone-

ous, whenever CRC-LOM is off (ITU-T G.706). REBE Remote end block error. It is erased if the first bit of the frames 14 and 16 is 0

(ITU-T G.706).


1.5.1 Higher Hierarchical Levels

As is the case with level 1 of the plesiochronous digital hierarchy (2 Mbps), thehigher levels of multiplexing are carried out bit by bit (unlike the multiplexing of64-Kbps channels in a 2-Mbps signal, which is byte by byte), thus making it impos-sible to identify the lower level frames inside a higher level frame. Recovering thetributary frames requires the signal to be fully demultiplexed.

The higher hierarchical levels (8,448, 34,368, and 139,264 Mbps, etc.; referredto as 8, 34, and 140 Mbps for simplicity) are obtained by multiplexing four lowerlevel frames within a frame whose nominal transmission rate is more than four timesthat of the lower level (see Table 1.3), in order to leave room for the permitted vari-ations in rate (justification bits), as well as the corresponding FAS, alarm, and sparebits (see Figure 1.18).

1.5.2 Multiplexing Level 2: 8 Mbps

The 8-Mbps frame structure is defined in the ITU-T Rec. G.742 (see Figure 1.19).The frame is divided into four groups:

• Group I contains the FAS, with sequence “1111010000”; the A-bit (remotealarm); the S-bit (spare); and 200 T-bits (tributary) transporting data.

• Groups II and III contain a block of four J-bits (justification control) and 208T-bits transporting data.

Figure 1.18 The PDH hierarchy, whith four levels from 2 to 140 Mbps. A bit-oriented justifica-tion process is used to fit tributaries created with different clocks in the second, third, and fourth hierarchy. The first hierarchy, 2 Mbps, is the only fully synchro-nous frame.

.64

2

34

140

8

34x4 x4x30

34Mbps

2

8x4

34Mbps8Mbps2Mbps64Kbps .64

2

2

140

8

34x4 x4x30

2Mbps

x16

140Mbps2

140

2Mbps

x64

34

140

140Mbps

140Mbps

E1 E2 E3 E4

E1 E3

E1

E4

E4


• Group IV contains a block of four J-bits, a block of R-bits (justification oppor-tunity), one per tributary, and 204 T-bits. To check whether R-bits have beenused, the J-bits are analyzed in each of the groups II, III, and IV (there are threeper tributary). Ideally the R-bit does not carry useful information on 42.4% ofthe occasions. In other words, this percentage is the probability of justificationor the insertion of stuffing bits.


The structure of this frame is described in the ITU-T Rec. G.751 (see Figure 1.19).As in the previous case, the frame is divided into four groups:

• Group I contains the FAS, with sequence “1111010000”; the A-bit (remotealarm); the S-bit (spare); and 372 T-bits (tributary) transporting data.

• Groups II and III contain a block of four J-bits (justification control) and 380T-bits transporting data.

• Group IV contains a block of four J-bits, a block of R-bits (justification oppor-tunity) one per tributary, and 376 T-bits. To check whether R-bits have beenused, the J-bits are analyzed in each of the groups II, III, and IV (there are threeper tributary). Ideally the R-bit does not carry useful information on 43.6% ofthe occasions.


The structure of this frame is described in the ITU-T Rec. G.751 (see Figure 1.19).In this case, the frame is divided into six groups:

• Group I contains the FAS, with sequence “111110100000;” the A-bit (remotealarm); the S-bit (spare); and 472 T-bits (tributary) transporting data.

• Groups II, III, IV, and V contain a block of four J-bits (justification control)and 484 T-bits transporting data.

• Group VI contains a block of four J-bits, a block of R-bits (justification oppor-tunity), one per tributary, and 376 T-bits. To check whether R-bits have beenused, the J-bits are analyzed in each of the groups II, III, IV, V, and VI (thereare five per tributary). Ideally the R-bit does not carry useful information on41.9% of the occasions.

1.5.5 Service Bits in Higher Level Frames

In any of the groups containing the FAS in the 8-, 34-, and 140-Mbps frames, alarmbits and spare bits are also to be found. These are known as service bits. The A-bits(alarm) carry an alarm indication to the remote multiplexing device, when certain


J1 J2 J3 J4 R1 R2 R3 R4 T1 T2 T3 T4

29282441

T1 T2 T3 T4

11 1 1 0 1 0 00 0 A S T1 T2 T3 T1 T2 T3 T4

384134 Mbps

11

Remote Alarm Indicator Justification Opportunity Bits

Frame Alignment Signal (Fas)

AS Spare Bits

1 0...Justification Control Bits

T1 T2 T3 T4 Tributary Bits

Figure 1.19 The PDH higher hierarchies. A bit-oriented justification process is used to fit tribu-taries created with clock impairments.

T4

1 11 1 1 0 1 0 00 0 0 A S S S T1 T2 T3 T4T1 T2 T3 T4

17 4881140 Mbps

13

J1 J2 J3 J4 T1 T2 T3 T4 T1 T2 T3 T4

976489

J1 J2 J3 J4 T1 T2 T3 T4 T1 T2 T3 T4

1464977

J1 J2 J3 J4 T1 T2 T3 T4 T1 T2 T3 T4

19521465

J1 J2 J3 J4 T1 T2 T3 T4 T1 T2 T3 T4

24401953

T1 T2 T3 T4

T1 T2 T3 T4

T1 T2 T3 T4

T1 T2 T3 T4

J1 J2 J3 J4 T1 T2 T3 T4 T1 T2 T3 T4T1 T2 T3 T4

385 768


769 1152

J1 J2 J3 J4 R1 R2 R3 R4 T1 T2 T3 T4T1 T2 T3 T4

1153 1536

J1 J2 J3 J4

11 1 1 0 1 0 00 0 A S T1 T2 T3 T1 T2 T3 T4

21218 Mbps

11

T4


213 424


425 636

J1 J2 J3 J4 R1 R2 R3 R4 T1 T2 T3 T4T1 T2 T3 T4

637 848

: Group I

: Group II

: Group III

: Group IV

: Group V

: Group VI

R1 R2 R3 R4

44,7 µs

21,02 µs

42,4 µs

: Group I

: Group II

: Group III

: Group IV

: Group I

: Group II

: Group III

: Group IV

Bits

per

trib

utar

y(Σ

Ti+R

i): 2

05+1

bits

Bits

per

trib

utar

y(Σ

Ti+R

i): 3

77+1

bits

Bits

per

trib

utar

y(Σ

Ti+R

i): 7

22+1

bits


breakdown conditions are detected in the near-end device. The spare bits are de-signed for national use, and must be set to “1” in digital paths that cross internation-al boundaries.

1.5.6 Plesiochronous Synchronization

As far as synchronization is concerned, the multiplexing of plesiochronous signalsis not completely trouble free, especially when it comes to demultiplexing the cir-cuits. In a PCM multiplexer of 30 + 2 channels, a sample of the output signal clock(1/32) is sent to the coders, so that the input channels are synchronized with the out-put frame. However, higher level multiplexers receive frames from lower levelmultiplexers with clocks whose value fluctuates around a nominal frequency valuewithin certain margins of tolerance. The margins are set by the ITU-T recommen-

Table 1.3 The PDH hierarchy, with four levels from 2 to 140 Mbps. A bit-oriented justification process is

used to fit tributaries created with different clocks in the second, third, and fourth hierarchy.

Standard Binary Rate Size Frame/s Code Amplitude Attenuation

G.704/732 2,048 Kbps±50 ppm 256 bits 8,000 HDB3 2.37-3.00V 6 dBG.742 8,448 Kbps±30 ppm 848 bits 9,962.2 HDB3 2.37V 6 dBG.751 34,368 Kbps±20 ppm 1536 bits 22,375.0 HDB3 1.00V 12 dBG.751 139,264 Kbps±15 ppm 2928 bits 47,562.8 CMI 1.00V 12 dB

Figure 1.20 The PDH and the T-carrier hierarchies are not synchronous and variations can be expected in the bit rate clock, shown in this figure as parts per million (ppm). The justification mechanism is implemented in the E2, E3, and E4 frames. If all Ji=1, then Ri is a justification bit that does not contain information. If all Ji=0, then Ri contains information. If all are not 0 or 1, the decision is based on the majority.

Ji Justification control bit of the ith tributary

R1 R2 R4R3 4 justification bits, one per tributary

8

34

8448 Kbps (+5 ppm)

8448 Kbps (+7 ppm)

8448 Kbps (+2 ppm)

8448 Kbps (-10 ppm)

34368 Kbps (-10 ppm)

A S T1 T2 T3 J1 J2 J3 J4T4 J1 J2 J3 J4 R1 R2 R4

1

T4T1T4T1 T4T1R3FASn

Multiplexer


dations for each hierarchical level. The signals thus formed are almost synchro-nous, except for differences within the permitted margins of tolerance, and for thisreason they are called plesiochronous (see Figure 1.20).

1.5.7 Positive Justification

In order to perform bit-by-bit TDM, each higher-order PDH multiplexer has elasticmemories in each of its inputs in which the incoming bits from each lower level sig-nal line or tributary are written. Since the tributary signals have different rates, theyare asynchronous with respect to each other. To prevent the capacity of the elasticmemories from overflowing, the multiplexer reads the incoming bits at the maxi-mum rate permitted within the range of tolerances. When the rate of the incomingflow in any of the tributary lines is below this reading rate, the multiplexer cannotread any bits from the elastic memory, and so it uses a stuffing bit or justificationbit (called justification opportunity) in the output aggregate signal. Its task is that ofadapting the signal that enters the multiplexer to the rate at which this signal istransmitted within the output frame (its highest clock value). This type of justifica-tion is called positive justification.

Justification bits, together with other overhead bits, make the output rate higherthan the total of the input signals.

1.5.7.1 Justification opportunity bits

The task of the justification opportunity bits (R-bits) is to be available as extra bitsthat can be used when the rate of the incoming tributaries is higher than its nominalvalue (within the margin specified by ITU-T) by an amount that makes this neces-sary. In this case, the opportunity bit is no longer mere stuffing, but becomes an in-formation bit instead.

In order for the device that receives the multiplexed signal to be able to deter-mine whether a justification opportunity bit contains useful information (i.e., infor-mation from a tributary), justification control bits (J-bits) are included in the frame.Each group of control bits refers to one of the tributaries of the frame. All of themwill be set to “0” if the associated opportunity bit is carrying useful information; oth-erwise they will be set to “1.” Several bits are used instead of just one, to provideprotection against possible errors in transmission. On examining the control bits re-ceived, if they do not all have the same value, it is decided that they were sent withthe majority value (a “1” if there are more 1s than 0s, for instance; it is assumed thatthere has been an error in the bits that are at 0).

It can be seen that there is a dispersion of the control bits referring to a tributarythat causes them to be located in separate groups. Spreading out the J-bits (control


bits), reduces the probability of errors occurring in them, and a wrong decision beingmade as to whether or not they have been used as a useful data bit. If the wrong de-cision is made, there is not only an error in the output data, but also a slip of one bit;that is, the loss or repetition of one bit of information.

1.6 MANAGING ALARMS IN HIGHER LEVEL HIERARCHIES

The A-bit of the FAS in 8-, 34-, and 140-Mbps frames enables the multiplexers thatcorrespond to these hierarchies to transmit alarm indications to the far ends (seeFigure 1.21) when a multiplexer detects an alarm condition (see Table 1.4).

In addition, 140-Mbps multiplexers also transmit an alarm indication when facedwith the loss of frame alignment of the 34-Mbps signals received inside the140-Mbps signals, as well as in the NFAS of the 34-Mbps signal that has lost itsalignment (bit 11 of group I changes from “0” to “1”) in the return channel (see Fig-ure 1.20).



Frame 0123

A

A

ATime Slot 0 Time Slot 16

A S T1 T2 T3 T4

1

T1FASn

2 Mbps 8, 34, 140 Mbps 2Mbps CAS

RAI (bit A=1)

LOF AIS

: X: 1

Figure 1.21 When a multiplexer detects an LOS or LOF, it sends a remote alarm indication (RAI) to its partner multiplexer and forwards an AIS to the next network element, because it has not been possible to recover any information.

Multiplexer B Network ElementMultiplexer A

AIS Formats:

Alarm Management:

1 n

RAI Formats:

NE

Network Element

NE


1.7 THE T-CARRIER HIERARCHY

As is the case of the PDH, the T-carrier higher levels multiplexing is carried out bitby bit (unlike the multiplexing of 64-Kbps channels in a DS1 frame, which is byteby byte), thus making it impossible to identify the lower level frames inside a high-er level frame. Recovering the tributary frames requires the signal to be fully de-multiplexed (see Figure 1.22).

1.7.1 The DS1 Frame

The DS1 frame is made up of 24 byte-interleaved DS0s, the 64-Kbps channels ofeight bits, plus one framing bit that indicates the beginning of the DS1 frame. The

Table 1.4 PDH events: Alarms, errors, and event indications.

ID Explanation

AIS Alarm indication signal. This is detected if less than six zeros in a frame in the case of 140 Mbps, or less than three zeros in 34 Mbps, and 8 Mbps (ITU-T G.751 and ITU-T G.742).

LOF Loss of frame alarm. It is raised after four consecutive frames with FAS error (ITU-T G.751 and ITU-T G.742).

LOS Loss of frame signal alarm.RAI Remote alarm indication. It is detected after two consecutive frames with the A

bit equal to 1 (ITU-T G.751 and ITU-T G.742).FAS error Frame alignment signal error. One or more incorrect bits in the alignment word.

Figure 1.22 T-carrier multiplexing hierarchy. M1 is a TDM; the rest (M12, M23, M11c, and M13) are bit interleaving multiplexers.

DS2

x7x24

DS1

x444.736 Mbps

6.312 Mbps1.544 Mbps64 Kbps DS0

DS0 T1 T2T3

DS1 DS3

M1 M12 M23

x24

DS1

x2

64 Kbps DS0

DS0 T1T1c

DS1

M1 M11c

DS2

DS1c

3.152 Mbps1.544 Mbps

x24

DS1

x28

64 Kbps DS0

DS0 T1T3

DS1

M1 M13

DS3

44.736 Mbps1.544 Mbps


DS0 channels are synchronous with each other, and are then time division multi-plexed in the DS1 frame. Depending on the application, the DS1 frames aregrouped in superframe (SF), 12 consecutive DS1 frames, and extended superframe(ESF), 24 consecutive frames (see Figure 1.23). Depending on the application, theDS1 signal is coded in AMI or in B8ZS.

Frame bit

The F-bit delimits the beginning of the frame and has different meanings. UsingESF, the F-bit sequence has a pattern for synchronization, but if ESF is used, thenthere is a synchronization pattern, CRC control, and a data link control channel of4 Kbps.

DCDFDCDFDCDF

1

x xx x x x x S

125 µs

Superframe (SF)

Data Link, Remote Configuration, and Monitoring (4 Kbps)

CRC Bits

Framing Bits

C1

S

F

D

C2 C3 C4

x

Figure 1.23 The T1 frame and superframe. Depending on the application, the frame bit has dif-ferent interpretations.

00011011100

234567

9101112

8

Frame 1. . . 2423

Frame bit

1.5 µs

D 125 µsCDFDCDFDCDF

24

1.5 µs

3 µs

Extended Superframe (ESF)

Signaling Bits

User Information Bits

User Channels

Frame bit

234567

9101112

8

Frame 1

1314151617

192021

18

222324

Chan. 1 2 Chan.1 2 3 . . . 233


Signaling

When inband signaling is used, the signaling goes in the least significant channelbit of every sixth DS1 frame (SF and ESF framing), leaving the effective through-put of those channels at 56 Kbps, to keep distortion minimal. Although inband is

Figure 1.24 (a) The DS2 frame or M12 multiplexing.(b) The DS3 frame. If C-bit parity framing is used, the Ci bits are not necessary for stuffing, and they are used for end-to-end signaling instead.

T1 T2 T3 T4T1 T2 T3 T4

1 48

M1M-Subframe 1 C1 F1 C2 C3 F2

M2 C1 F1 C2 C3 F2

M3 C1 F1 C2 C3 F2

Mx C1 F1 C2 C3 F2

Si

1 48C3

F2

Si

1174 bits

Control Stuffing Bits

Subframe Alignment “01”Tributary BitsStuffing Bits

1 2 .... 294

M1 M2 M3 Mx Multiframe Alignment “011X”

F1

C2C1

X1 M-Subframe 1

Si

1 48

C3

F4

Si

Control Stuffing BitsSub-frame Alignment “1001”Tributary BitsStuffing Bits (i=1..7)

1 2 ....

M1 M2 M3 Multiframe Alignment “010”

F3

C2C1

F1 C1 F2 C2 F3 C3 F4

680

X2 F1 C1 F2 C2 F3 C3 F4

P1 F1 C1 F2 C2 F3 C3 F4

P2 F1 C1 F2 C2 F3 C3 F4

M1 F1 C1 F2 C2 F3 C3 F4

M2 F1 C1 F2 C2 F3 C3 F4

M3 F1 C1 F2 C2 F3 C3 F4

234567

F2F1

P1 P2 Parity BitsX1 X2 Monitoring and Maintenance

DS3 frame

DS2 frame

8218 bits

234

(b)

(a)


still in use, signaling (call setup, teardown, routing, and status) is generally carriedout of band over a separate network using a protocol called signaling system 7(SS7).

1.7.2 The DS2 Frame

As long as the DS1 frames are asynchronous from each other, each DS1 line istreated as a bit stream rather than individual frames, and the DS2 signal is formedby bit interleaving of four DS1 signals. Stuffing bits are added to the DS2 to com-pensate for the slightly different rates. Framing, signaling, alarm, frame alignment,and parity are also added to the frame (see Figure 1.24).

1.7.3 The DS3 Frame

A DS3 frame is formed by bit-interleaving 28 DS1 frames or seven DS2 frames.There are two framing formats, called M13 or C-bit parity (see Figure 1.24), de-scribed in the following:

1. M-13, the multiplexion DS1 to DS3, is done in two steps: (a) the first four DS1frames form a DS2 frame using M12 multiplexing; (b) the second seven DS2frames form a DS3 frame using M23 multiplexing. M13 multiplexing uses bitstuffing to bring each asynchronous DS1/DS2 line up to a common data ratefor transmission.

2. C-12, the multiplexion DS1 to DS3, is done in one step. The stuffing controlbits (C-bits) at the M23 multiplexing level are not required and can be used fora maintenance link between the end points, applications like far-end alarm andcontrol (FEAC) and far-end block error (FEBE).

If C-bit parity framing is used, the Ci-bits in the DS3 frame take on signaling defi-nitions as described in the ANSI T1.107 (see Table 1.5).

Table 1.5 C-bit parity: alarm and status signal codes.

Code Event Type Explanation

00110010 1111111100011110 1111111100000000 1111111100001010 1111111100011100 1111111100101010 1111111100111100 1111111100000000 1111111100101100 11111111

DS3 equipment failure Service affecting (SA), requires quick attentionDS3 equipment failure Nonservice affecting (NSA)DS1 equipment failure Service affecting (SA), requires quick attentionDS1 equipment failure Nonservice affecting (NSA)DS3 LOS Loss of signalMultiple DS1 LOS Multiple loss of frame in the DS1 tributariesSingle DS1 LOS Alarm indication signal received from a DS1DS3 OOF Out of frame in the DS3 signalDS2 AIS Alarm indication signal received from a DS2


Selected Bibliography

• William Stallings, Data and Computer Communications, Prentice Hall, 1997.

• Fred Halsall, Data Communications, Computer Networks and OSI, Addison-Wesley, 1998.

• Llorens Cerdá, Transmissió de Dades, Universitat Oberta de Catalunya, 1999.

• Stanatios V. Kartalopoulos, Understanding SDH/SONET and ATM, IEEE Press, 1999.

• ANSI T1.403, DS1 metallic interface.

• ANSI T1.107, Digital hierarchy formats.

• ITU-T Rec. G.703 (10/98), Physical/electrical characteristics of hierarchical digital interfaces.

• ITU-T Rec. G.704 (10/98), Synchronous frame structures used at 1,544, 6,312, 2,048, 8,448 and44,736 Kbit/s hierarchical levels.

• ITU-T Rec. G.742 (11/88), Second order digital multiplex equipment operating at 8,448 Kbit/s andusing positive justification.

• ITU-T Rec. G.751 (11/88), Digital multiplex equipment operating at the third order bit rate of 34,368Kbit/s and the fourth order bit rate of 139,264 Kbit/s and using positive justification.

• José M. Caballero, Redes de Banda Ancha, Marcombo, 1997.

• José M. Caballero and Andreu Guimera, Jerarquías Digitales de Multiplexión, PDH y SDH, Sincroni-zación de Redes, L&M Data Communications, 2001.

35

Chapter 2

SDH/SONET: The Synchronous Hierarchies

Synchronous digital hierarchy (SDH) is an ITU-T universal standard that defines acommon and reliable architecture for transporting telecommunications services ona worldwide scale. Synchronous optical network (SONET) is today a subset ofSDH, promoted by American National Standards Institute (ANSI) and used in theU.S., Canada, Taiwan, and Korea.

From now on we will use the acronym SDH to refer to the generic ITU-T stan-dard that includes also SONET.

2.1 THE EMERGENCE OF SDH/SONET NETWORKS

During the 1980s, progress in optical technologies and microprocessors offerednew challenges to telecommunications in terms of bandwidth and data processing.At that time, plesiochronous hierarchies (T-carrier and PDH) dominated transportsystems, but a series of limitations and the necessity to introduce new transmissiontechnologies moved to develop a new architecture.

Antitrust legislation was the final factor that hastened the development of SO-NET. It was applied to the telecommunications business and forced the giant, Bell,to be split up into small companies, the regional Bell operating companies(RBOCs). SONET, developed at Bellcore labs in 1984, grew out of the need to inter-connect RBOCs using standardized optical interfaces. Telecom liberalization wasconfirmed around the world during the ‘90s, and this has inevitably led to globalcompetition and interoperation. In 1988, the Comité Consultatif InternationalTélégraphique Et Téléphonique CCITT (now ITU) proposed creating broadband-ISDN (B-ISDN) to simultaneously transport data, voice, video, and multimedia overcommon transmission infrastructures. Asynchronous transfer mode (ATM) was se-lected for the switching layer, and SDH for transport at the physical layer.


2.1.1 Limitations of Plesiochronous Networks

Plesiochronous networks have the following limitations:

• Their management, supervision, and maintenance capabilities are limited, asthere are no overhead bytes to support these functions. One example of this isthat if a resource fails, there is no standard function whereby the network canbe reconfigured.

• Access to 64-Kbps digital channels from higher PDH hierarchical signals re-quires full demultiplexing, because the use of bit-oriented procedures removesany trace of the channels.

• In PDH it was not possible to create higher bit rates directly; one could do soonly after following all the steps and hierarchies (see Figure 2.1).

• Plesiochronous ANSI and European Telecommunications Standard Institute(ETSI) hierarchies were not compatible.

• There were no standards defined for rates over 45 Mbps in T-carrier, and over140 Mbps in PDH.

Figure 2.1 SDH and SONET allow for direct multiplexing and demultiplexing.

MUX155 Mbps

Multiplexer

155 Mbps

1

n

n x 155 MbpsDS2

x7x24

DS1

x4 44.736 Mbps

64 Kbps DS0

DS0

T1 T2 T3

DS1 DS3

M1 M12 M23

DS2 (n=1, 4, 64, 256)

PDH/T-carrier SDH/SONET Step-by-step Multiplexing Direct Multiplexing

Multiplexer

Figure 2.2 SONET and SDH assumed legacy T-carrier and PDH as native transport inter-faces. New networks became hybrid, as the interfaces remained plesiochronous while the long-haul transport network was synchronous.

T-carrier SONET

DS1

M12

DS2

DS1

M12

DS2

DS1

M12

DS2

DS1

M12

DS2SONET

T-carrierT-carrier

SDH/SONET: The Synchronous Hierarchies 37

• Different manufacturers of plesiochronous equipment could not always be in-terconnected, because they implemented additional management channels orproprietary bit rates.

These limitations meant that it was necessary to design a new transmission archi-tecture to increase the flexibility, functionality, reliability, and interoperability ofnetworks.

2.1.2 The SDH/SONET Challenge

What had to be decided first was how to provide smooth migration from legacy in-stallations. Then a basic frame period of 125 µs was selected, the same of E1 andT1 frames, in order to guarantee compatibility with existing services such as plainold telephone service (POTS), integrated services digital network (ISDN), framerelay (FRL) or any n x 64 Kbps (see Figure 2.2). Note that a byte constantly carriedon a 125-µs frame period defines a 64-Kbps channel. (see Figure 2.3).

Some of the remarkable features of SDH compared with its predecessors are:

Synchronous versus plesiochronous

Plesiochronous means “almost synchronous.” This in its turn means that nodes tryto do work in the same frequency, but in fact they do not, because each PDH islanduse its own clock. In synchronous networks, all digital transitions should occur si-multaneously, and all the nodes must be fed with the same master clock (see Chap-ter 5). There may, however, be a phase difference between the transitions of the twosignals but this must lie within standardized limits.

Bytes versus bits

In SDH and SONET, such basic operations as multiplexing, mapping, or alignmentare byte oriented, to keep transported elements identified throughout the wholetransmission path (see Figure 2.4).

Frame 3Frame 2Frame 1

Figure 2.3 SDH frames have a period of 125 µs (8,000 frames per second) and a byte always defines a 64-Kbps channel, independent of the bit rate.

125 µs

8 bits125 µs

= 64 Kbps

125 µs 125 µs1 byte


Direct access

The main difference between SDH and its predecessors is in synchronization andbyte-oriented operations. Synchronization enables us to insert and extract tributar-ies directly at any point and at any bit rate, without delay or extra hardware. For thisreason, PDH/T-carrier must completely demultiplex signals of various megabitsper second, to access any embedded channel of n x 64 Kbps.

2.1.2.1 Full management

In SDH and SONET, payload and overheads are always accessible, and there is noneed to demultiplex the signal. This drastically improves operation, administration,and maintenance (OA&M) functions, which are essential to enable centralizedmanagement independently of the bit rate.

SDH and SONET also provide embedded mechanisms to protect the networkagainst link or node failures, to monitor network performance, and to manage net-work events.

2.1.2.2 Providing circuits for public networks

The basic function of SDH, just like any transmission network, is that of providingmetropolitan or long-haul transport to networks such as POTS, ISDN, FRL, GigabitEthernet (10GbE), Universal Mobile Telecommunications System (UMTS) or Inter-net (see Figure 2.5). Signaling, switching, routing, and billing do not depend onSDH, as it is only in charge of providing bandwidth between two points. (see Fig-ure 2.6).

Figure 2.4 SDH multiplexing is based on byte interleaving.

DS2

x7x24

DS1

x444.736 Mbps

6.312 Mbps1.544 Mbps64 Kbps DS0

DS0 T1 T2T3

DS1 DS3

M1 M12 M23

DS2

VC11T1

VC12E1

VC4E4

STM-1

Bit interleaving

mapping

MUX

Byte interleaving

STM-n


2.1.2.3 Universal standard

SDH and SONET standards enable transmission over multiple media, including fi-ber optics, radio, satellite, and electrical interfaces. They allow internetworking be-tween equipment from different manufacturers by means of a set of genericstandards and open interfaces. Scalability is also an important point, as transmis-sion rates of up to 40 Gbps have been defined, making SDH a suitable technologyfor high-speed trunk networks.

2.2 COMPARISON OF SDH AND SONET

SDH and SONET are compatible but not identical. SDH is used worldwide exceptin the U.S., Canada, Japan, and partially in South Korea, and Taiwan. Both define asimilar set of structures and functions; however, there are differences in usage.

The 51.84-Mbps STS-1 is the basic building block of SONET (OC-1 if this signalis transmitted over fiber optics). STS-1 was enough to transport all T-carrier tribu-

Table 2.1Terminology comparison.

SDH SONET SDH SONET SDH SONET

STM-0 OC-1 RSOH SOH Regenerator section SectionSTM-X OC-3X MSOH LOH Multiplex section LineAU-3 STS-1 HO POH STS POH Higher-order path VT pathAU-4 STS-3c F3 Z3 Lower-order path STS pathAU-4-Xc STS-3Xc Z4 K3AU-4-Yv STS-3Yv LO POH VT POHTU VT Z6 N2VC SPE Z7 K4

Figure 2.5 SDH and SONET networks offer reliable, efficient, and flexible transport.

Access

Transport

Transmission

DVB

TCP/IP

ATM10GE

UMTS

ISDN

POTSGSM

FRLClientsInternet

Cable SatelliteFiber Radio

PDH/T-carrier

DVB

TCP/IP

ATM10GE

UMTS

ISDN

POTSGSM

FRL

SDH/SONET

Internet

CableSatellite FiberRadio

PDH/T-carrier

SDH/SONET


taries but not to carry the 140Mbps PDH tributary. Then SDH chose a basic trans-port frame three times faster at 155.53 Mbps, in order to allocate the full Europeanhierarchy. SDH and SONET terminology have some differences, in referring to thesame concepts, bytes, and structures. Nevertheless, beyond the names, the function-ality is equivalent (see Table 2.1).

Internetworking is always possible because the evolution of both technologieshas been the same with the new hierarchies up to 40 Gbps and to the last standardslike link capacity adjustment scheme (LCAS) (see Section 2.9.3). The objective is toguarantee universal connectivity.

2.3 FUNCTIONAL ARCHITECTURE

Traditionally, telecommunications networks have been described using a layeredmodel to facilitate their design, implementation, and management. Standardizedformats and protocols describe peer interchanges between separate nodes. Interfac-es and services define client/server relationships inside each node.

2.3.1 Network Elements

SDH systems make use of a limited number of network elements (NEs) withinwhich all the installations are fitted (see Figure 2.7):

• Regenerators (REGs) or section terminating equipments (STEs): Every signalsent through any transmission medium (optical, electrical or radio-electrical)experiences attenuation, distortion, and noise. Regenerators supervise the re-

Figure 2.6 The aim of the SDH network is to provide transmission services to other networks.

Internet

ADM

ADM

ADM

ADM

STM-64OC-192

2 Mbps45 Mbps

140 Mbps DXC

ADM

ADM

ADM

ADM

STM-16OC-48

POTS

ATM

UMTS

Leased line

ISDN

Ethernet1,5 Mbps140 Mbps

2 Mbps

STM-1

OC-12

OC-3

45 Mbps

155 Mbps

DWDM


ceived data and restore the signal’s physical characteristics, including shapeand synchronization. They also manage the monitoring and maintenance func-tions of the regenerator section (RS) or section in SONET (see Figure 2.10).

• Line terminal multiplexers (LTMUX) or path terminal equipment (PTE): theyare common in line and access topologies. Their function is to insert and ex-tract data in synchronous frames (see Figure 2.10).

• Add and drop multiplexers (ADMs) can insert or extract data directly into orfrom the traffic that is passing across them, without demultiplexing/multiplex-ing the frame. Direct access to the contents of the frame is a key feature ofSDH, as it enables us to turn any point of the network into a service node, justby installing an ADM.

• Digital cross-connects (DXCs) configure semipermanent connections toswitch traffic between separate networks. The switched traffic can be eitherSDH streams or selected tributaries. Although it is not common, DXCs canalso insert and drop tributaries in transport frames.

2.3.2 Network Topology

Synchronous multiplexers provide great flexibility for building topologies, which iswhy point-to-point, linear, ring, hub, meshed, and mixed topologies are all possible(see Figure 2.8):

• Linear point-to-multipoint: this topology follows the basic point-to-point struc-ture, but now includes ADM multiplexers performing add and drop functionsat intermediate points.

• Ring: this topology closes itself to cover a specific area, with ADM multiplex-ers installed at any point. It is flexible and scalable, which makes it very suit-

Figure 2.7 SDH network elements (NE).

Line Terminal MultiplexerMUX

Add and Drop Multiplexer

REG

STM-nOC-m West

STM-nOC-m

STM-nOC-m

STM-nOC-m

STM-YOC-Z

Regenerator Multiplexer

Digital Cross-Connect

STM-nOC-m

STM-nOC-m

DXC

STM-nOC-m

East

West East

Tributaries

MUX

STM-nOC-m

Tributaries Tributaries

ADM


able for wide area and metropolitan networks. Rings are frequently used tobuild fault-tolerant architectures.

• Hub or star: this topology concentrates traffic at a central point, to make topol-ogy changes easier. A hub can join several networks with different topologies.

2.3.3 Topology Partitioning

Topology describes the potential connections in a network. Moving from top down,a network can be split repeatedly in interconnected subnetworks. We can describe asubnetwork by means of linked nodes. Nodes are network elements, such as switch-es, multiplexers, and regenerators. Links can be optical, electrical, and radioelectri-cal (see Figure 2.10).

2.3.4 SDH/SONET Layers

In plesiochronous networks, interactions are simple and direct. In synchronous net-works they are more sophisticated, so responsibilities have been divided among

PTE PTE

Figure 2.8 Network topologies.

Tributaries

ADM

ADM

ADM

ADM

STM-nOC-m

STM-nOC-m

Linear

Ring

ADM ADM

Tributaries

Meshed

ADM

ADM

ADM

ADM

STM-nOC-m

Hub

Core

Metro

Access

10 Gbps

2.5 Gbps

155 Mbps

Mixed

MUX MUX

DXC

DXC

DXC

DXCDXC

DXC

PTE PTE

Tributaries Tributaries

STM-nOC-m

Point to point

MUX MUX


several layers that communicate with their counterparts by making use of specificoverheads, formats, and protocols. This architecture is equivalent to the layeredopen system interconnection (OSI) model to define and design communication net-works (see Figure 2.9).

2.3.4.1 Path layers

Path layers are the route to transport clients’ information across the synchronousnetwork from its source to its destination, where the multiplexers interface with cli-ent equipment (see Figure 2.9). At this layer clients’ information is mapped/demapped into a frame and path overhead is added. There are two specialized pathlayers (see Figure 2.10):

1. Lower-order path (LP), or virtual tributary path (VT Path) in SONET, to trans-port lower-rate services. Associated overhead is lower-order path overhead(LO-POH) or virtual tributary path overhead (VT-POH) in SONET.

2. Higher-order path (HP), synchronous transport signal path (STS Path) inSONET, to transport higher-rate services or a combination of lower-rate ser-vices. Associated overhead is higher-order path overhead HO-POH or syn-chronous transport signal path overhead (STS-POH) in SONET.

Some of the path layer functions are routing, performance monitoring, anomaliesand defect management, security and protection, as well as specific path OAMfunctions support.

Figure 2.9 SDH and SONET standards define a layered client/server model that can be divided into up to four layers in order to manage transmission services.

SectionMux

HO Path

SectionLine/Mux

VC-4 (VC-3)MSOH MSOH

RSOH

LO Path

SectionMux

HO PathLO Path

VC-11, VC-12, VC-2, VC-3

RSOH RSOH

Path

sSe

ctio

ns

Formats and ProtocolsLa

yers

RSOHSection Section

SectionLine

STS Path

SectionLine/Mux

STS SPELOH LOH

SOH

VT Path

SectionLine

STS PathVT Path

VTn SPE

SOH SOH

Path

sSe

ctio

ns

Laye

rs

SOHSection Section

SDH

SONET

Terminal Multiplexer Multiplexer Terminal MultiplexerRegenerator Regenerator

Physical media


2.3.4.2 Multiplex section or line layer

Multiplexer section (MS), or line section in SONET, is a route between two adja-cent multiplexers. This layer has several capabilities such as bit error detection, andcircuit protection when an intermediate link or node collapses. It also carries syn-chronization reference information and OAM information between nodes. Associ-ated overhead is multiplex section overhead (MSOH) or line overhead (LOH)in SONET (see Figure 2.9).

2.3.4.3 Regeneration section or section layer

The regeneration section (RS), (the section layer in SONET) is the link betweentwo successive NEs. It reads and writes specific overheads and management func-tions for each type of transmission media. Its most typical functions are framing, biterror detection, and regenerator OAM functions support. Associated overhead is re-generation section overhead (RSOH) or <Default Para Font>section over-head (SOH) in SONET (see Figure 2.9).

2.3.4.4 Physical layers

Fiber optics and metallic cable, together with terrestrial radio and satellite links canbe used as the physical layer (PL). Fiber optics is the most common medium be-cause of its capacity and reliability. Radio is a cost-effective medium when dis-tance, geographical difficulties, or low-density areas make the optical alternativeless practical. Nevertheless, radio has some important weaknesses; for example,noise and frequency allocation, that limit bit rates to 622 Mbps. Electrical cablesare also used in some legacy installations (see Figure 2.10).

Figure 2.10 SDH and SONET line topology and network layers.

Lower-Order Path (LP)

Higher-Order Path (HP)

Regenerator Section Reg. Sect.

Multiplex Section

Path MUX

PTE

MUX

STM-nOC-m

MUX

STS Path

VT Path

Line

Section

LineLine

Section SectionSection

SON

ET

REG

Reg. Sect. Regenerator Section

Multiplex SectionMultiplex Section

STM-nOC-m

Path

SDH

CPE LTE STE LTE

PathMUX

PTE

Path

CPE


2.4 SDH/SONET FORMATS AND PROCEDURES

SDH defines a set of structures to transport adapted payloads over physical trans-mission networks (ITU-T Rec. G.707). Five basic procedures are involved here(see Figure 2.11):

• Mapping: A procedure by which tributaries are adapted into virtual containersat the boundary of an SDH network.

• Stuffing: This is a mapping procedure to adapt the bit rate of client data streamsinto standardized, fixed-size containers.

• Multiplexing: A procedure by which multiple lower-order signals are adaptedinto a higher-order path signal, or when the higher-order path layer signals areadapted into a multiplex section.

• Overhead addition: This procedure is to attach information bytes to a data sig-nal for internal routing and management.

• Aligning: A procedure by which a pointer is incorporated into a tributary unit(TU) or an administrative unit (AU). TU and AU pointers are used to find unitsanywhere in the transmission network.

Figure 2.11 SDH and SONET digital hierarchies multiplexing map.

STM-1

STM-64

STM-16

STM-4

STM-256OC-768

OC-192

OC-48

OC-12

OC-3/STS-3

STM-0OC-1/STS-1

TUG-3

DS1

E1

DS2

DS3 E3

E4

x3

x4

x7

x7

TUG-2

x3

x3

MappingMultiplexingAligning Pointer processing

SDH ContainerGroup

VC-4

VC-3

VC4-4c

VC4-16c

AU-4

AU4-4c

AU4-16c

STS-3c SPE

STS-1 SPE

VT-Group

VC4-64c

VC4-256c

AU4-64c

AU4-256c

ATM

ATM

ATM

ATM

ATM

ATM

ATM

AU-3

GbE

AUG-256

STS-12c SPE

STS-48c SPE

STS-192c SPE

STS-768c SPE

STS-3c

STS-1

STS-12c

STS-48c

STS-192c

STS-768c STS-768

AUG-64STS-192

AUG-16STS-48

AUG-4STS-12

AUG-1STS-3

x4

x4

x4

x4

40Gbps

10Gbps

2.5Gbps

622Mbps

155Mbps

52Mbps

Frame transmission

VC-3

VC-2

VC-12

VC-11

VT-6 SPE

VT-2 SPE

VT-1.5 SPEPOH addition

x1

x1

x1

x1

x1

x1

x1

x1

x1

x1

x1

x1

C-3

C-2

C-12

C-11

C-4-256c

C-4-64c

C-4-16c

C-4-4c

C-4

TU-3

TU-2

TU-12

TU-11

VT-6

VT-2

VT-1.5

x1

GbE

MbE


When the tributary reaches the end of the transport network, demapping, demulti-plexing, and overhead removal procedures are performed to extract and deliver thetributary (see Figure 2.12).

2.4.1 SDH/SONET Frame Structure

The basic transport frame in SONET is synchronous transport signal (STS-1),while in SDH it is synchronous transmission module (STM-1) (see Figure 2.13):

• STS-1 is a 3 x 9 byte structure transmitted at 52 Mbps, which is equivalent toSTM-0.

• STM-1 is a 9 x 9 byte structure transmitted at 155 Mbps, which is equivalent tooptical carrier 3 (OC-3) and electrical STS-3.

Both have the same structure that is based on three types of information blocks:

1. Overhead blocks: These blocks contain information that is used to managequality, anomalies, defects, data communication channels, service channels,and so on. There are two types of overhead blocks, RSOH (managed by theregenerator section layer) and MSOH (managed by the multiplex sectionlayer).

2. Payload blocks or virtual containers (VCs): They contain a combination of cli-ent signals and overhead blocks. VC does not have a fixed position in theframe, but it floats in the frame to accommodate clock mismatches.

Figure 2.12 Two examples: E4 (140 Mbps) and E1 (2 Mbps) transport. Transmission and reception operations are symmetrical, and any bit, pointer, or multiplexing per-formed by the transmission multiplexer is reversed by the receiver multiplexer.

C-4STM-1 VC-4AU-4

C-2VC-12TUG-12 TU-12

AUG-1

- Stuffing bits- Justification bits- Overhead bits

+LPOH

+AU pointer

+TU pointer

+HP-OH

E4

E1

C-4STM-1 VC-4AU-4

C-2VC-12TUG-12 TU-12

AUG-1

-LPOH

-AU pointer

-TU pointer

-HP-OH

E4

E1

Add

+ Stuffing bits+ Justification bits+ Overhead bits

Drop

Higher Order Path

SDHnetwork

TUG-3

TUG-3

x3

Higher Order Path

x7

Lower Order Path

Lower Order Path

x3

div3div7

div3


3. Pointers: They track the VC position, pointing to its first byte, while movinginside the frame (see Figure 2.13).

2.4.1.1 Containers as transport interfaces

Containers (C-n) are used to map client bit streams. Adaptation procedures havebeen defined to suit most telecom transport requirements. These include PDH, met-ropolitan area network (MAN), asynchronous transfer mode (ATM), high-leveldata link control (HDLC), internet protocol (IP), and Ethernet streams.

Placing signals inside a container requires a stuffing function to match the clientstream with the container capacity. The justification function is necessary for asyn-chronous mappings, to adapt clock differences and fluctuations.

2.4.1.2 Virtual containers or virtual tributaries

Virtual containers (VC-n) or virtual tributaries (VTs) in SONET (see Figure 2.14),support end-to-end path layer connections; that is, between the point where the cli-ent stream is inserted into the network and the point where it is delivered. Nobodyis allowed to modify the VC contents across the entire path.

VCs consist of a C-n payload and a path overhead (POH). Fields are organizedinto a block structure that repeats every 125 or 500 µs. Containers hold client data,and the POH provides information to guarantee end-to-end data integrity.

There are two types of VCs:

• The lower-order VC, such as VC-11, VC-12, VC-2, and VC-31. These consistof a small container (C-11, C-12, C-2, and C-3), plus a 4-byte POH attached tothe container (9 bytes for VC-3).

• The higher-order VC, such as VC-3 or VC-4. These consist of either a big con-tainer (C-3, C-4) or an assembly of tributary unit groups (TUG-2, TUG-3). Inboth cases, a 9-byte POH is attached.

2.4.1.3 Tributary units and tributary unit groups

A tributary unit is a structure for adaptation between the lower-order and higher-or-der path layer.

1. VC-3 can be transported through a lower-order path or a higher-order path, dependingon the multiplexing map used (see Figure 2.11).


A TUG is an SDH signal made up of byte-interleaved multiplexing of one ormore TUs. In other cases, lower-order TUGs are multiplexed to form a higher-orderTUG (for instance, seven multiplexed TUG-2s form one TUG-3), and in other cases,a TUG is formed by a single TU (for instance, a single TU-3 is enough to form aTUG-3). TUGs occupy fixed positions in higher-order VCs.

2.4.1.4 Administrative unit

An administrative unit (AU-n) provides adaptation between the higher-order pathlayer and the multiplex section layer. It consists of an HO-VC payload and an AUpointer indicating the payload offset.

STS-1 SPE

Figure 2.13 SONET and SDH differences are minimal, both are highly compatible, and most differences fall into certain names and acronyms. SDH-to-SONET gateways need to adapt just a few bytes. In the figure, the STS-1 and STM-1 represent the basic frames of SONET and SDH.

904

SOH

LOH

SPE ptr

1 3STS-1

STS POH

90B1 E1

*A1 *A2 *

F1

D1 D2 D3

B2 K1 K2

D4 D5 D6

D7 D8 D9

S1 E2

D10 D11 D12

J0

M19

1

Section Overhead)

Line Overhead

9

1

H1 H2 H3 pointer

Transport Overhead

4

9

Synchronous Transport Signal

Z7

VT POHPath OverheadVirtual Tributary

G1F2H4

J1B3C2

Z3Z4N1

RSOH

MSOH

AU4 ptr

1

VC4 27010

9

G1F2H4

J1B3C2

F3K3N1

270101 9STM-1

HP POH

TU pointersptr ptr

V5

V5

Z6J2

V5

VT pointersptr ptr

V5

V5

VT SPE VT SPE

VCn VCn K4

LP POHLower Order

N2J2

V5

500 µs

Path Overhead

375 µs250 µs

125 µs

500 µs375 µs

250 µs125 µs

x3

MUX

STM-1

3:1

STS-1STS-1STS-1

9

1


2.4.2 Multiplexing Map

A multiplexing map is a road map that shows how to transport and multiplex anumber of services in STM/OC frames (see Figure 2.11).

• The client tributary (PDH, T-carrier, ATM, IP, Ethernet, etc.) needs to bemapped into a C-n container, and a POH added to form a VC-n, or a VT forSONET.

• The VC/VT is aligned with a pointer to match the transport signal rate. Pointerstogether with VCs form TUs or AUs.

• A multiplexing process is the next step, whereby TUG-n and AUG-n groupsare created.

• When it comes to TUGs, they are multiplexed again to fill up a VC, synchro-nous payload envelope (SPE) in SONET, and a new alignment operation is per-formed.

• Finally, an administrative unit group (AUG) is placed into the STM/OC trans-port frame.

2.5 SDH TRANSPORT SERVICES

Today’s telecommunications services (voice, data, TV, Internet) are heterogeneous,based on a diverse combination of technologies. Most of them are clients of SDHwhen they need to extend their service range to wider areas.

Figure 2.14 SONET virtual tributary superframes.

V5

25 bytes

J2

25 bytes

Z6

25 bytes

Z7

25 bytes

500µs

125µs

V5

34 bytes

J2

34 bytes

Z6

34 bytes

Z7

34 bytes

V5

52 bytes

J2

52 bytes

Z6

52 bytes

Z7

52 bytes

V5

106 bytes

J2

106 bytes

Z6

106 bytes

Z7

106 bytes

VT1.5 VT2 VT6VT3

V1

26, 35, 53 or

V2

V3

V4

107 bytes

26, 35, 53 or107 bytes

26, 35, 53 or107 bytes

26, 35, 53 or107 bytes

VT Superframe


Channelized networks in 64-Kbps circuits (POTS, ISDN, FRL, GSM, FRL) aremapped transparently in SDH synchronous containers designed to transport PDH orT-carrier tributaries. Packet technologies, such as IP, Ethernet or ATM, also havespecial mapping procedures (see Figure 2.6).

PDH/T-carrier over SDH

To guarantee smooth migration from legacy installations, SDH standards have de-fined procedures to transport all legacy circuits (E1, E2, E3, E4, T1, T2, and T3).This way, all former PDH/T-carrier services (ISDN or FRL) are today transportedby hybrid PDH/SONET or T-carrier/SONET networks.

• Synchronous mapping maintains the original 64-Kbps channel structure duringthe whole transmission, making it possible to access these channels directly, asbit justification is not needed. This is common in such services as ISDN orFRL, where there are nodes synchronized with the SDH reference clock. Smallclock differences are adjusted with pointer movements.

• Asynchronous mapping is used when PDH and SDH do not share the sameclock. Here we need bit-oriented justification to adjust any clock differencesbetween the PDH signal and the SDH container. Due to this, the existing bytestructure is lost. This scheme is rather common in POTS and in old plesiochro-nous circuits.

ATM over SDH

ATM cells are mapped into containers at different bit rates. The range goes from afew Mbps up to several Gbps, using any concatenation technique (see Section 2.9).

Table 2.2VC types and capacity.

SDH SONET Bandwidth Payload

VC-11 VT 1.5 SPE 1,664 Kbps 1,600 KbpsVC-12 VT 2 SPE 2,240 Kbps 2,176 KbpsVC-2 VT 6 SPE 6,848 Kbps 6,784 KbpsVC-3 STS-1 SPE 48,960 Kbps 48,384 KbpsVC-4 STS-3c SPE 150,336 Kbps 149,760 KbpsVC-4-4c STS-12c SPE 601,344 Kbps 599,040 KbpsVC-4-16c STS-48c SPE 2,405,376 Kbps 2,396,160 KbpsVC-4-64c STS-192c SPE 9,621,504 Kbps 9,584,640 KbpsVC-4-256c STS-768c SPE 38,486,016 Kbps 38,338,560 Kbps


ATM cells are mapped by aligning every cell with the structure of virtual or con-catenated containers. Since capacity may not be an integer multiple of the ATM celllength (53 bytes), a cell is allowed to cross the container frame boundary (see Figure2.15). The ATM cell information field (48 bytes) is scrambled before mapping, toguarantee delineation. An ATM cell stream with a data rate that can be mapped isequal to the VC payload capacity (see Table 2.2).

Mapping HDLC-framed signals

HDLC-framed signals are mapped by aligning the byte structure of every framewith the byte structure of the VC. The range also goes from 1.5 Mbps up to severalGbps using concatenation techniques (see Section 2.9). 7Ex HDLC flags are usedbetween frames to fill the buffer, due to the discontinuous arrival of HDLC-framedsignals. Since HDLC frames are of variable length, a frame may cross the containerboundary.

Figure 2.15 Mapping ATM cells in VC-n and concatenated VC-4-Xc.

G1F2H4

J1B3C2

F3K3N1

..... R

VC-4-Xc

No. of columns = X-1

260X+101

9

53 bytes

VC-12

ATM cell

V1

V2

V3

V4

VC-n

R = fixed stuff

Figure 2.16 Mapping HDLC frames enables IP transport.

PPP

G1F2H4

J1B3C2

F3K3N1

..... R

VC-4-Xc

Columns = X-1

260X+101

9

R = fixed stuff

IP

IPV1

V2

V3

V4

VC-n

PPP

PPP IP

PPPIP

IPPPP IP

‘7Ex‘framer


Packet over SDH

Packet over SDH (PoS) enables core routers to send native IP packets directly overSDH container frames, using point to point protocol (PPP) for framing and bit errordetection. The request for comments 2615 (RFC 2615) defines the use of PPP en-capsulation over SDH circuits.

IP traffic is treated as a serial data stream that travels hop by hop through thenetwork. At each node, IP packets are unwrapped from the PPP frame, destinationaddresses are examined, routing paths are determined, and, finally, packets are re-wrapped in a new PPP frame and sent on their way (see Figure 2.16).

PoS is more reliable and has lower overhead than its alternatives, such as ATMor frame relay encapsulation.

Ethernet over SDH

Ethernet has become the standard technology for local area networks (LANs). It ischeap, easy to use, well-known, and always in constant evolution toward higherrates. Now it is also being considered as a good technology for access and metronetworks, but carriers still need SDH to route high volumes of Ethernet traffic toget long haul. There are several schemes:

• Ethernet over LAPS: defined in ITU-T X.86. This is an HDLC family protocol,including performance monitoring, remote fault indication, and flow control.However, it calls for contiguous concatenated bandwidth techniques (see Sec-tion 2.9) that do not match the burst nature of Ethernet.

• Generic framing procedure (GFP): defined in ITU-T Rec. G.7041. This is aprotocol for mapping any type of data link service, including Ethernet, resilientpacket ring (RPR), and digital video broadcasting (DVB).

• Virtual concatenation: defined in ITU-T Rec. G.707, creates right-sized pipesfor the traffic, providing quite a lot of flexibility and high compatibility withlegacy SDH installation techniques (see Section 2.9.2).

• Link capacity adjustment scheme (LCAS): defined in ITU-T Rec. G.7042. Thisdynamically allocates/deallocates new bandwidth to match Ethernet require-ments in a flexible and efficient way. It calls for virtual concatenation.


2.6 TRANSPORTING PDH/T-CARRIER TRIBUTARIES

Transporting tributaries always calls for a set of operations (mapping, aligning,multiplexing, etc.) before inserting data into STM-n/OC-m frames. The number andtype of operations may vary, depending on the tributary rate (see Figure 2.11). Forhigher bit rates (45 or 140 Mbps), the operation is straightforward; basically, amapping process followed by aligning (see Figure 2.17). For lower rate tributaries,

Figure 2.17 Asynchronous mapping of 139,264 Kbps into a VC-4 and STM-1 frame.

27010

RSOH

MSOH

AU4 ptr

11 9

VC4 = C4 + POH

STM1

VC4

STM-1 = AUG + RSOH + MSOH

RSOH: Regenerator section overhead

AUG-1 = AU4 ptr + VC4

POH

MSOH: Multiplex section overheadAU ptr: Administration unit pointerPOH: Path overhead VC4: Virtual container 4STM: Synchronous transport module

D D X D Y D Y D Y D

X D Y D Z D

X D Y D Y D Y DX D Y D Y D Y DX D Y D Y D Y D

1 121 121 121 12260 bytes: 1 12

Column 11 Column 270

:DDDDDDDDD:CRRRRROOX:RRRRRRRRY

R: Fixed stuff bitsO: Overhead bitsC: Justification control bitsS: Justification opportunity bit

C-411

9

1270

One row

27010

9

1

G1F2H4

J1B3C2

F3K3N1

C4

C-4STM-1 VC-4AU-4AUG-1

+AU Pointer +HPOH

E4

+ StuffingHigher Order Path

+ Justification + Overhead

1234

5

1

2

34

if CCCCC = 00000, then S is a data bitif CCCCC = 11111, then S is a justification bit

5

:DDDDDDSRZ

D: Data bits

in other cases, majority vote counts

125 µs


several multiplexing operations are also needed, to fill up the whole STM-n/OC-mframe capacity (see Figure 2.19).

2.6.1 Transport on VC-4 or STS-3c SPE

In a VC-4 container, several client signals can be mapped, including PDH/T-carriercircuits, HDLC-like protocols, and ATM cells (see Figure 2.19). In our example wewill look at 140Mbps mapping, to describe this in detail:

1. The mapping operation in C-4, whereby a 140 Mbps bit stream is fitted into anine row container. Each row has a justification bit opportunity, so mapping isasynchronous, as PDH circuit itself is asynchronous and does not have a byte-oriented structure.

2. Creation of VC-4 when higher-order path overhead (HO-POH) is added. HO-POH provides end-to-end management and performance monitoring.

3. Alignment or an AU-4 pointer addition, which enables locating the VC-4 float-ing in the STM-1 frame. AU-4 always points to the first VC-4 byte.

4. A unitary multiplexing operation to create an AUG-1.5. Adding MSOH and RSOH overheads to build the STM-1 frame.

Now the STM-1 frame is ready to be sent (see Figure 2.18). On reception, in orderto deliver the 140-Mbps tributary, we must perform these same operations —from5 to 1.

125 µsPayload

B1 ^ ^ E1 ^

A1A1 A1 A2 A2 A2

F1

D1 ^ ^ D2 ^ D3

Pointer (s)B2 B2 B2 K1 K2

D4 D5 D6

D7 D8 D9

S1 E2

D10 D11 D12

M1

STM-1/STS-3/OC-3X

X

X

X

X

J0

X9

1

2

A1A1A1A2A2A2J0

1

Pointer (s) Payload

Overhead

Sequence to transmit

125 µsFigure 2.18 STM and STS frames are prepared for serial transmission every 125 µs.


2.6.2 Transport on VC-3

In a VC-3 container several client signals can be mapped, including PDH/T-carriercircuits, HDLC-like protocols, and ATM cells. Here, we will look at the 45-Mbpsand 34-Mbps transport; in both cases all procedures except mapping are identical.

VC-3 has two transport schemes: (a) Lower-order (LO) transport, where VC-3sare allocated directly into the STM-1 frame; (b) Higher-order (HO) transport, whereVC-3s are multiplexed into a VC-4 which is finally placed into the STM-1.

Higher-order transport

The operations and steps to follow are depicted in Figure 2.19:

1. The mapping operation in C-3. The 45 Mbps transport uses a one-row struc-ture, while for 34 Mbps, three rows are repeated three times.

2. A higher-order path overhead (HO-POH) addition to create a VC-3.3. An alignment or an AU-3 pointer addition to locate the VC-3.4. A unitary multiplexing operation to create an AUG-1.5. Adding MSOH and RSOH overheads to build an STM-1 frame.

Now the STM-1 frame is ready to be sent. On reception, to deliver the tributary, wemust perform the same operations vice versa; that is, from 5 to 1.

Lower-order transport

The operations and steps to follow are depicted in Figure 2.19:

a. Mapping is identical to that of higher-order transport.b. A lower-order path overhead (LO-POH) addition to form a VC-3.c. An alignment operation or AU-3 pointer addition to locate the VC-3.

The new structure is called TU-3.d. Multiplexing of three different TU-3s to create a TU-3 Group (TUG-3) e. Adding an HO-POH to produce a VC-4.f. A new alignment operation to find the VC-4.g. A unitary multiplexing operation to create an AUG-1.h. Adding MSOH and RSOH overheads to build an STM-1 frame.

The STM-1 frame is now ready to be sent. Again, to deliver the tributary, the sameoperations must be performed vice versa, from 8 to 1, on reception.


27010

RSOH

MSOH

AU4 ptr

9

11 9

STM-1

270

+HO-POH

9

110

G1F2H4

J1B3C2

F3K3N1

R

13

byte interleaving3 TUG-3

Figure 2.19 Asynchronous mapping of 44,736 Kbps and 34,368 Kbps into VC-3 via AU-4 and also via AU-3.

STM-1 VC-4AU-4 C-3VC-3TUG-3 TU-3AUG-1

+LPOH+AU-4 pointer

+TU pointer

+HPOH

E3

+ Stuffing+ Justification

+ Overhead +RSOH+MSOH

: Data bitsD: RRCDDDDDK: CCRRRRRRL

R: Stuffing bitsO: Overhead bits (future)C: Justification control bitsS: Justification bit

1 84

R DR DR D

RRR

X DX DS YZD

R D

Column 86

R K DD R L D R P D1 2 3 84

three

44 736 Kbps (DS3)

34 368 Kbps (E3)

: CCRROORSP

: RRRRRRC1C2X: RRRRRRRS1A: S2DDDDDDDB

C-3

9

2

84

rows

onerow

C-3

21TUG-3

VC3

G1F2H4

J1B3C2

F3K3N1

9

186

H1H2H3

R

x3

+LO-POH

C-3

ab

a

c

dx 3

Bytes bytes

b

R DR DR D

R DR DR D

R DR DR D

R DR DR D

R DR DR D

RRR

R DR DR D

R DR DR D

R DR DR D

X DX DX D

5 9

R DR DR D

RRR

R DR DR D

R DR DR D

R DR DR D

R DR DR D

R DR DR D

RRR

R DR DR D

R DR DR D

R DR DR D

: SSSSSSSSS

if CiCiCiCiCi=11111, then Si is a justification bitif CiCiCiCiCi=00000, then Si is a data bit

Byte

DS3

10

RSOH

MSOH

9

11 9

STM-1

87+HO-POH

9

11

G1F2H4

J1B3C2

F3K3N1

R

4

C-3

VC-4

3 pointers

H1a H1b H1c H2a H2b H2c H3a H3b H3c

VC3

x 3

3xAU3 ptr

ptr a ptr b ptr c

fg

e

cdefg

VC-3

+HO-POH

2AU-3

+AU-3 pointer

+TU pointer

3

2

x3

345

1

h

5 4

Byte-interleaving3 VC-3

28 56

otherwise majority vote

1

h


Note that transporting VC-3 via AU-3 calls for three AU-3 pointers in theSTM-1 frame. Pointer bytes appear interleaved in the space assigned for them in thestructure of the synchronous transport module. In the case of VC-3 transport viaAU-4, the first pointer indicates the beginning of the VC-4, and, inside the VC-4payload, a second level of pointers is needed to locate VC-3 containers.

2.6.3 Transport of 2-Mbps Circuits

Transporting 2 Mbps can be synchronous or asynchronous. Synchronous mappingis possible only if PDH and SDH networks use the same reference clock. In thiscase, mapping has certain advantages because it is byte oriented, which means thatthe 64-Kbps frame structure of the tributary will be maintained throughout thewhole transport, allowing direct access from SDH premises to the voice or datachannel.

This does not happen in asynchronous mapping, as it has to use bit-oriented jus-tification mechanisms that break the E1 frame structure. Except for mapping, the op-eration sequence for both cases is similar (see Figures 2.20 and 2.21):

1. The mapping operation in container C-12. If asynchronous mapping is used, C-12 adopts a 500-µs multiframe format.

2. Adding lower-order path overhead to create VC-12.3. Alignment or a TU pointer addition to indicate VC-12 offset.

TU-2 is created this way.4. Multiplexing of three TU-12s, which creates a TUG-2. 5. Multiplexing of seven TUG-2s, which creates a TUG-3. 6. A new multiplexing operation of three TUG-3s plus a HO-POH, together form

a VC-4.7. Alignment or an AU-4 pointer addition enables us to locate the VC-4.8. A unitary multiplexing operation to create an AUG-1.9. Adding MSOH and RSOH overheads to build an STM-1 frame.10. Since VC-12 is a multiframe, then transmission is a four STM-1 multiframe

operation (see Figure 2.22).

The STM-1 frame can now be sent. On reception, to deliver the 2-Mbps circuit, wemust again perform these operations vice versa, from 9 to 1. Note the VC-4 capaci-ty up to 63 x 2 Mbps circuits can be transported simultaneously.

At the reception end, to locate each circuit, we must first find the VC-4 usingthe AU-4 pointer and then, by reading the TU-12 pointer, it is possible to find theVC-12 offset.


V5

34 bytes

J2

34 bytes

N2

34 bytes

K4

34 bytes

Data32 bytes

Figure 2.20 Synchronous and asynchronous transport of a 2-Mbps circuit (I).

Synchronous

125 µs

500 µs

R: Fixed stuff bitO: Overhead bitsCi: Justification control bitsSi: Justification

D: Data bit

STM-1 VC-4AU-4 C-12VC-12TUG-2 TU-12AUG

+LO-POH+AU pointer +HO-POH

E2

+ Stuffing+ Justification+ Overhead

+RSOH+MSOH

x3

12346789

TUG-3

5

x7x3

S2 DDDDDDDC1C2RRRRRS1

TS0Channels

1 to 15TS16

Channels16 to 30

CAS RRRRRRRR

RRRRRRRR

32 bytes

RRRRRRRR

RRRRRRRR

Data32 bytes

C1C2OOOORR

RRRRRRRR

Data32 bytes

C1C2OOOORR

RRRRRRRR

Data32 bytes

RRRRRRRR

Data31 bytes

Asynchronous

TS0

Time Slot1 to 31

CCS 1

byte 34

VC-12

22

11 C-12C-12

V1

35 bytes

V2

35 bytes

V3

35 bytes

V4

35 bytes

500 ms

3

TU-12

H4=00

H4=01

H4=10

H4=11

V311 4

9

125 µs

9

12

V3 V3

4

TUG-2

1

140

V3 V3V31

Other TU-12

LO-POH = V5 J2 N2 K4TU-12 pointer = V1 V2 V3 V4

125 µs

if CiCiCi=111, then Si is a justification bitif CiCiCi=000, then Si is a data bit


RSOH

MSOH

AU4 ptr

R

NPI

Figure 2.21 Synchronous and asynchronous transport of a 2-Mbps circuit (II).

STM-1 VC-4AU-4 C-12VC-12TUG-2 TU-12AUG

+LO-POH+AU pointer +HO-POH

E2

+ Stuffing+ Justification+ Overhead

+RSOH+MSOH

x3

12346789

TUG-3

5

x7x3

9

1 TUG-2 12

9

86TUG-3

1

9

1 TUG-2 12

9

1 TUG-2 12

#1 #2 #7

9

1 TUG-2 12

#3

R1 2 3 4 5 6 7 1 2.......

5

TUG-3x 7

TUG-3

R

NPI

RR

NPI

R

10

9

11 9

VC-4

270+HO-POH

10

G1F2H4

J1B3C2

F3K3N1

R

13

6

Byte-interleaving3 TUG-3

7

8

STM-19

TU pointer

x 3

8

#1 #2 #3

11 12

. . . #7


2.7 POINTERS AND TIMING COMPENSATION

SDH supports two types of timing mismatches: asynchronous tributaries, and timevariations of NE clocks. Justification bits are used to compensate differences withtributaries during the mapping operation (see Section 2.6). Pointer adjustments arenecessary to compensate slight clock differences of the synchronous equipment(basically ADM and DXC).

2.7.1 Payload Synchronization

Pointers allow for dynamic alignment of the payload within transmission frames.These are necessary, as payloads are floating within the frame to compensate forclock phase fluctuations between NEs (see Figure 2.23).

At this point, we come across a paradox: If SDH is based on node and signalsynchronization, why do fluctuations occur? The answer lies in the practical limita-

Figure 2.22 VC-12 needs 4 x 125 µs intervals for full mapping. This means that a full VC-12 extends to cover four STM-1 frames.

VC-4

RSOH

MSOH

125 µs

500 µs

V1

V2

V3

V4

H4=xx00

H4=xx01

H4=xx10

H4=xx11

VC-12

125 µsTUG-2 TUG-2 TUG-2

#1 #2 #7

500 µs H4=xx11

STM-1

TUG-3#1

TUG-3#2

TUG-3#3

VC-4

RSOH

MSOH

H4=xx10

STM-1

VC-4

RSOH

MSOH

H4=xx01

STM-1

VC-4

RSOH

MSOH

H4=xx00

STM-1

x7

x3

x3 x7 x3

x3 x7 x3

x3x7x3


tions of synchronization. SDH networks use high-quality clocks feeding network el-ements. However, we must consider the following:

• A number of SDH islands use their own reference clocks, which may be nomi-nally identical, but never exactly the same.

• Cross services carried by two or more operators always generate offset andclock fluctuations whenever a common reference clock is not used.

• Inside an SDH network, different types of breakdown may occur and cause atemporary loss of synchronization. When a node switches over to a secondaryclock reference, it may be different from the original, and it could even be theinternal clock of the node.

• Jitter and wander effects (see Chapter 5).

2.7.2 Pointer Formats and Procedures

Although pointers have different names (AU-4, AU-3, TU-3, TU-2, TU-1, STS ptror VT ptr), they all share the same format and procedures (see Figure 2.24):

• Two bytes allocate the pointer (H1-H2 or V1-V2) that indicates the first byte ofthe payload (see Table 2.3).

• The pointer value 0 indicates that the payload starts after the last H3 or V3byte.

• Each pointer has its valid range of values.

• The offset is calculated by multiplying n times the pointer value, and n dependson the payload size.

2.7.2.1 Pointer Generation

In normal operation, pointers are located at fixed positions, and the new data flag(NDF) is 0110. However, sometimes it is necessary to change the pointer value, inwhich case the following rules apply:

Table 2.3SDH and SONET pointers.

SDH Payload SONET Payload Allocation Range Hops Justification

AU-4 VC-4 STS-3 ptr STS-3c H1, H2 0 - 782 3 bytes 3 bytesAU-3 VC-3 STS-1 ptr STS-1 H1, H2 0 - 782 3 bytes 1 bytesTU-3 VC-3 — — H1, H2 0 - 764 1 byte 1 byteTU-2 VC-2 VT-6 ptr VT-6 V1, V2 0 - 427 1 byte 1 byteTU-12 VC-12 VT-2 ptr VT-2 V1, V2 0 - 139 1 byte 1 byteTU-11 VC-2 VT-15 ptr VT-15 V1, V2 0 - 103 1 byte 1 byte


• Minimum time period: The minimum time period between two consecutivepointer changes is 500 µs.

• Pointer increment: If a positive justification is required, the pointer value issent with the I-bits inverted. The new pointer value is the previous value, incre-mented by one. If the pointer is H1-H2, the position of the payload is shiftedthree bytes forward, and void bytes are left after H3. If it is V1-V2, the payloadis shifted one byte forward, and a void byte is left after V3.

• Pointer decrement: If a negative justification is required, the pointer value issent with the D-bits inverted. In this case, the new pointer value is the previousvalue decremented by one. If the pointer is H1-H2, the position of the payload

Figure 2.23 The pointer adjustment operation used by LTE to compensate for clock mismatch. Consecutive pointer operations must be separated by at least 500 µs.

J1

J1

J1

payload

payload

payload

No Justification

J1

payload

payload

VC-4

J1

J1

payload

payload

Positive Justification Negative Justification

J1

payload

J1

00 0 0

J1

payload

125µ

sH2H1 H3

0H2H1 H3

H2H1 H3

H2H1 H3

H2H1 H3

H2H1 H3

H2H1 H3

H2H1 H3

H2H1 H3

0 0

ADM

MUX MUXCPE CPE

fo f1

Clock-2

f2

Clock-1

Tributaryfo

Tributary

Stuffing PointerMovement Destuffing

f1 = f2 (no change) f1 > f2 (ptr decrement) f1 < f2 (ptr increment)

PTE PTE

n

n+1

n+2


is shifted three bytes backwards, and H3 provides spare bytes. If the pointer isV1-V2, the payload is shifted one byte backwards, and either V3 provides thespare byte.

• New pointer: If the VC-n alignment changes for any reason, and it cannot betracked by pointer increments or decrements, then a new pointer value is sent,and the NDF is set to 1001 to reflect the new value.

N N N N S S I D I D I D I D I DPointer Value

H1 or V1 H2 or V2

SS: Unspecified I: Invert if increment

Concatenation

NDF

N N N N S S I D I D I D I D I D0 1 1 0 S S 0 1 0 1 1 0 1 1 0 1

0 1 1 0 S S 1 1 1 1 0 0 0 1 1 10 1 1 0 S S 0 1 0 1 1 0 1 1 1 0

0 1 1 0 S S 0 0 0 0 1 1 1 0 0 00 1 1 0 S S 0 1 0 1 1 0 1 1 0 0

0 1 1 0 S S 0 1 0 1 1 0 1 1 0 1

0 1 1 0 S S 0 1 0 1 1 0 1 1 0 1

ii+1i+2

jj+1j+2

Frame

Incr

emen

tD

ecre

men

t

Pointer MovementN N N N S S I D I D I D I D I D1 0 0 1 S S x x x x x x x x x x

New Pointer

NDF enabled: NNNN=1001

N: New Data Flag

NDF disabled: NNNN=0110(0001, 1101, 1011, 1000 also accepted)

(1110, 1010, 0100, 0111 also accepted)

New pointer value

H3 or V3

Negativejustification

payload

Positivejustification

Figure 2.24 Pointer formats, codification, and procedures.

D: Invert if decrement

1 0 0 1 S S 1 1 1 1 1 1 1 1 1 1

C-3G1F2H4

J1B3C2

F3K3N1

H1H2H3

AU-4 ptr / STS ptr

G1F2H4

J1B3C2

F3K3N1

YH1 Y 1H2 1 H3H3 H3

AU-4 / STS3c

TU-3

Y: 1001SS111: all 1s byte

V1

V2

V3

V4

TU-n / VT-m SPE

x00

x01

x10

x11

payload

N N N N S S I D I D I D I D I D

H1H1 H1 H2H2 H2 H3H3 H3

AU-3 / STS-1AU-3 ptr / STS ptr

C-3G1F2H4

J1B3C2

F3K3N1

H1H2H3

payload

TU3 ptr / STS ptr

V1

V2

V3

TU-n ptr

VC-n

VT ptr

payload


2.8 OVERHEADS

The key difference between SDH and its plesiochronous predecessors is in themanagement and monitoring capabilities SDH provides at the transmission layer.These features are based on peer protocols, standardized formats, and overheadfields. Network elements themselves generate a suitable response to managementactions, reconfigurations, performance monitoring, failures, or any type of events.Overheads are also a key difference between SDH and its potential successors,based on any combination of gigabit-Ethernet (GbE), dense wavelength divisionmultiplexed (DWDM), and IP protocols. These networks are always said to be moreefficient, because they do not support most of these management facilities and,eventually, will not need overheads or protocols to support them.

Table 2.4Nine-byte path overhead for VC-3, VC4, VC-4-Xc, STS-1 SPE, and STS-Xc SPE.

Byte Description

J1 HP trace: Its position is indicated by the AU-n or the TU-3 pointer. It carries a config-urable sequence identifier of 16 or 64 bytes (including a CRC-7 byte), so that the receiving path terminal can continuously verify its connection with the transmitter.

B3 HP error monitoring: This is a bit interleaved parity 8 (BIP-8) code using even parity, computed over all bits of the previous VC-3, VC-4, or VC-4-Xc before scrambling.

C2 Path signal label: This indicates the composition or mapping of the VC-n.0x: Unequipped, 01x: Reserved, 02x: TUG structure, 03x: Locked TU-n, 04x: 34-Mbps or 45-Mbps mapping, 12x: 140-Mbps mapping, 13x: ATM mapping, 14x: distributed queue dial bus (DQDB) mapping, 15x: fiber distributed data interface (FDDI) mapping, 16x: HDLC/PPP mapping, 17x: simple data link (SDL) mapping, 18x: Mapping of HDLC/LAPS, 19x: SDL mapping, 1Ax: 10 GbE, 1Bx: GFP map-ping, CFx: Obsolete mapping of HDLC/PPP, from E1x to FC: reserved for national use, FEx: test signal O.181.

G1 HP status and performance: This byte enables continuous monitoring of anomalies and defects either at path end or at any point along the trail. Bits 1-4: remote error indi-cation (HP-REI) conveys the number of bit errors detected by B3. Bit 5: remote defect indication (HP-RDI), is sent back if a signal failure is detected. Bits 6-7 can be used to provide enhanced RDI information to differentiate between payload defects (HP-PLM), server defects (HP-AIS, LOP), and connectivity defects (HP-TIM, HP-UNEQ).

F2, F3 HP user channel: User communication purposes between path terminations.H4 Sequence indication for virtual VC-3/4 concatenation: If the payload is VC-2, VC-12,

or VC-11, it is used as a multiframe indicator.K3(bit1-4) APS signaling: Allocated for the VC-3/4 protection protocol in case of a failure

K3(bit7-8) HP data communication channel of 16 Kbps.

N1 HP tandem connection monitoring function (HP-TCM): Two options are described in the G.707 (Appendix C and D). Bits 1-4: incoming error count (IEC), bit 5: TC remote error indication (TC-REI), bit 6: outgoing error indication (OEI), bits 7-8: operate in a 76-byte multiframed string including access point identifier (TC-APId), a generic 16-byte identifier, and a remote defect indication (TC-RDI).


REI (Remote Error Indication) BIP-8 violation count

REI RDIG1:

RDI (Remote Defect Indication) is sent back

00: Unequipped01: Reserved02: TUG03: Locked TU04: E3, T3

14: DQDB15: FDDI xx11xx00: pointer to V1

xx11xx01: pointer to V2xx11xx10: pointer to V3xx11xx11: pointer to V4

Nine-byte Path Overhead (POH)

Path trace, message with CRC

BIP-8 parity control

Signal label (mapping)

Path status

Path user channel (voice or data)

Position and sequence indicator

Path user channel (voice or data)

Automatic Protection Switch

Tandem Connection Monitoring

G1F2H4

J1B3C2

F3Z3Z4

16: HDLC/PPP

18: HDLS/LAPS1A: 10GbEthernetFE: Test Signal

SpareE-RDI

IEC Incoming Error Count, BIP-8 errors in Tandem Conn.

IEC TCN1:REI OEI TC-API, TC-RDI

ODI, reserved

E-RDI (Enhanced RDI information)

TC-REI: Remote Error Indication in a TC subnetworkOEI: Outgoing Error IndicationMultiframe: TC-API (Access Point Identifier)

multiframe

17: SDL

(RDI=0) 10: Payload defect (PLM)(RDI=1) 01: Server defect (AIS, LOP), (RDI=1) 10: Connectivity defect (TIM, UNEQ)

LO Multiframe Sequence

BIP-2V5:

ESL (Extended Signal Label) 32 bits multiframe

REI RFI RDI

BIP-2 bit 1: Odd bit parity of the previous VCbit 2: Even bit parity

Signal Label

RDI (Remote Defect Indication)

VC signal label (mapping)000 - Unequipped001 - Reserved010 - Asynchronous floating011 - Bit synchronous 100 - Byte synchronous

Path Overhead

Reserved or TCM

APS: Automatic Protection Switching channel

Path Trace

V5

J2

N2

K4

V5

J2

N2

K4

500µ

s

101 - Extended signal label110 - Test Signal O.181

Additional Path Overhead

REI (Receive Error Indication)RFI (Remote Failure Indication)

bits 1-11 Multiframe Alignmentbits12-19:

0A: HDLC/PPP0B: HDLC/LAPS0C: Concatenated test signal

DL: Lower Order Data Link

BIP2 for Tandem Connection calculated over the VC

BIP-2 I-AISN2:

OEI TC-API, TC-RDIODI, reserved

TC-REI: Remote Indication Error in a TC subnetworkOEI (Outgoing Error Indication)

multiframe

1 TCREI

I-AIS: Incoming AIS

111 - VC- AIS

K4:APS DLE-RDIESL VC

Figure 2.25 Nine-byte path overhead is attached to VC3, VC4, and VC4-Xc. Four-byte path overhead is attached to VC11, VC12, and VC2.

Four-byte Path Overhead (POH)

bits 20-32: 0 (reserved for future)VC (Lower Order Virtual Concatenation)

ODI (Outgoing Defect Indication)

VC-3/4-Xv sequencebit 5-8: MFI1 multiframe indicator (0 to 15)frame 0 bit 1-4 MFI2 MSB Multiframe Indicator 2frame 1 bit 1-4 MFI2 LSBframe 14 bit 1-4 SQ MSB sequence indicatorframe 15 bit 1-4 SQ LSB sequence indicator

12: E413: ATM

H4: LO Seqx

C2:

APS: Automatic Protection

K3:

HODL: Higher Order Data Link

xx x 1 1

H4: Multiframe Indicator 1MFI2 (frames 0 and 1)SQ (frames 14 and 15)

09: ATM

E-RDI (Enhanced RDI information)(RDI=0) 010: Payload defect (PLM)(RDI=1) 101: Server defect (AIS, LOP), (RDI=1) 110: Connectivity defect (TIM, UNEQ)

TC-RDI (RDI in Tandem Connection)

Multiframe: TC-API (Access Point Identifier)

ODI (Outgoing Defect Indication)TC-RDI (RDI in Tandem Connection)

G1F2H4

J1B3C2

F3K3N1

SDH SONET

V5

J2

Z6

Z7

SDH SONET

APS SpareHODL

(76 frames)

(76 frames)

Z3:

Z6:

Z7:

Z4:


2.8.1 Path Overhead

The POH provides a communication protocol between the two ends of a VC path.Among these protocols are path performance monitoring, error and alarm indica-tions, path protection, signals for maintenance purposes, and multiplex structure in-dications. There are two categories of virtual container POH (see Figure 2.25):

Table 2.5Four-byte path overhead for VC-11, VC-12, VC-2, VC-2-Xc, VT-11, VT-12, and VT-6.

Byte Description

V5 LP general overhead: Its position is indicated by the TU-n pointer, and it provides path status, performance monitoring, and signal label functions for VC-2, VC-12, and VC-11 paths. This byte enables continuous monitoring of anomalies and defects, and payload composition either at path end or at any point along the trail.

V5(bit1-2) LP bit error monitoring: A BIP-2 is calculated by the transmitter over all the bits of the previous VC-n. The calculation includes POH bytes, but excludes V1, V2, V3 (except when used for negative justification), and V4.

V5(bit3) LP remote error indication (LP-REI): This is set to 1 and sent back toward an LP orig-inator, if one or more bit errors is detected by the BIP-2.

V5(bit4) LP remote failure indication (LP-RFI), only VC-11: This is set to 1 and sent back if a failure is declared. Otherwise it is cleared (i.e., set to 0).

V5(bit5-7) LP signal label: This indicates the payload composition.0x: Unequipped, 1x: Reserved, 2x: Asynchronous, 3x: Bit-synchronous, 4x: Byte-syn-chronous, 5x: Extended signal label, see K4 bit 1, 6x: Test signal, O.181, 7x: VC-AIS.

V5(bit8) LP remote defect indication (LP-RDI): This is set to 1 and sent back towards the trail termination source if a failure condition is detected.

J2 LP trace: It carries on a configurable 16 sequence identifier (including a CRC-7 byte) so that the receiving path terminal could continuously verify its connection with the transmitter.

N2 LP tandem connection monitoring function (LP-TCM): Bits 1-2: BIP-2 for TC bit error checking; bit 3: fixed to 1, bit 4: incoming AIS indicator (I-AIS), bit 5: indicates errored blocks (TC-REI), bit 6: OEI to indicate errored blocks, bits 7-8: operate as a 76-multiframe string, including access point identifier (TC-APId), TC-RDI, and ODI.

K4(bit1) Extended signal label (if V5(bit5-7) are 5x): This is a 32-bit multiframed string. Bits 12 to 19 contain the label. 09x: ATM mapping, 0Ax: HDLC/PPP mapping, 0Bx: HDLC/LAPS mapping, 0Cx: test signal O.181 mapping, 0Dx: flexible topology data link mapping.

K4(bit2) LP virtual concatenation: A 32-bit multiframed string.

K4(bit3-4) LP automatic protection switching channel (APS).

K4(bit5-7) LP enhanced remote defect indication: Provides enhanced RDI information. 1x: no defect, 2x: payload defect (LP-PLM; loss of cell delineation or LCD), 5x server defects (LP-AIS, TU-LP), 6x: connectivity defects (LP-TIM, LP-UNEQ).

K4(bit8) LP data link.


1. Nine-byte path overhead, or STS POH in SONET: When this structure isattached to C-4 or TUG-3, it creates a VC-4. It can also be attached to C-3 orTUG-2, in which case it creates a VC-3 (see Figure 2.25).

2. Four-byte path overhead, or VT POH in SONET: This structure is added toC-2, C-12, and C-11 to form a VC-2, VC-12, and VC-11 respectively. The fourbytes are not just contiguous, but also part of a multiframe (see Figure 2.25).

The functionality of the nine-byte POH (see Table 2.4) and the four-byte POH(see Table 2.5) are very similar.

2.8.2 Section Overhead

Section overhead (SOH) information is attached to the information payload to cre-ate an STM-n/OC-m frames (see Figure 2.26). This includes block framing infor-mation for maintenance, performance monitoring, and other operational functions.SOH information is classified into:

• Regenerator section overhead (RSOH): which is the interchange data unit be-tween regenerator section layers (see Table 2.6).

• Multiplex section overhead (MSOH): which, passing transparently through re-generators, is the interchange data unit between multiplex section layers (seeTable 2.7).

The SOH needs plenty of bytes to manage a wide range of functions. Amongother things, it is responsible for the frame alignment, performance monitoring man-agement channels, voice channels for communication between nodes, data channelsused for synchronization, and protection services in case of physical layer failures.

Table 2.6Regenerator section overhead (RSOH) or section overhead (SOH).

Byte Description

A1, A2

Framing pattern A1=F6x, A2=28x: Indicates the beginning of the STM frame. For STM-n, there are 3 x n A1 bytes followed by 3 x n A2 bytes.

J0 Regenerator section trace: This is used to transmit a 16- or 64-byte identifier (including a CRC-7 byte) repeatedly, so that every regenerator can verify its connection.

Z0 Spare: Reserved for future international standards.B1 RS bit error monitoring: BIP-8 code using even parity, computed over all bits of the pre-

vious STM-n frame after scrambling. The value is placed into B1 before scrambling. E1 RS orderwire: Provides a 64-Kbps voice channel between regenerators. F1 User channel: This can be used to provide data/voice channel for maintenance purposes.D1-D3 192-Kbps data communication channel (DCCR): between regenerators providing OAM

functions.


Figure 2.26 STM-n/OC-m section and line overheads.

* Nonscrambled bytesX bytes reserved for national use^ Media-dependent bytes

B1 ^ ^ E1 ^

*A1 *A1 *A1 *A2 *A2 *A2 * *

F1

D1 ^ ^ D2 ^ D3

Pointer (s)

B2 B2 B2 K1 K2

D4 D5 D6

D7 D8 D9

S1 E2

D10 D11 D12

M1

STM-1/STS-3/OC-3

X

XXX

X

J0 *

X

B1 E1

*A1 *A2 *F1

D1 D2 D3

B2 K1 K2

D4 D5 D6

D7 D8 D9

S1 E2

D10 D11 D12

STS-1/STM-0J0

M1

B1

*A1 *A1 *

D1

B2 B2 B2

D4

D7

S1

D10

A1 *A1

B2

^

*A1

^

B2

^

*A1

^

B2

E1

*A2

D2

K1

D5

D8

D11

*A2 *A2

M1

*A2^

*A2

^

^

*A2

^

*A2 *A2F1

*J0

D3

K2

D6

D9

E2

D12

X

*Z0

X

X

*Z0

X

X

*Z0

X

X

*X

X

X

*X

X

^

*A1

^

B2

^

*A2

^

*A2X

*X

X

Pointer (s)

1 13 25 36 STM-4/OC-4

B1 +*A1 *A1 *

+

D1 + +

B2 B2 B2

D4

D7

S1 +D10

A1

+

^

*A1

^

B2

^

*A1

^

B2

+

E1

*A2

D2

K1

D5

D8

D11

+*A2

++*A2

+

M1

+*A2

+^

*A2

^

^

*A2

^

+

*A2+*A2

+

+

F1

*J0

D3

K2

D6

D9

E2

D12

X

*Z0

+

+++

X

+

X

*Z0

+

+++

X

+

X

*X

X

X

*X

+

+++

X

+

X

*X

+

+++

X

+

^

*A1

^

B2

+

^

*A2

^

+

+*A2

+

+

X

*X

$

+++

X

+

Pointer (s)

1 49 97 144 STM-16/OC-48

D1-D3: 192-Kbps OA&M dataD4-D12: 576-Kbps OA&M dataE1, E2: 64-Kbps orderwire channels

M0, M1: Resending of B2 errors

F1: 64-Kbps user channelH1, H2, H3: pointer bytes

Z0: reserved

K1, K2: Request /answer APS channelsF1: 64-Kbps user channel

A1= 11110110, A2= 00101000: Frame AlignmentB1: Section Parity Code BIP-8B2: Bit interleaved parity,

J0: Section Trace

17

^

*A1

^

B2

^

*A1

^

B2

+

^

*A1

^

B2

+

33 65 81 113

X

*X

X

129

X

*X

$

+++

X

+

B1 +*A1 *A1 *

+

D1 + +

B2 B2 B2

D4

D7

S1 +D10

A1

+

^

*A1

^

B2

^

*A1

^

B2

+

E1

*A2

D2

K1

D5

D8

D11

+*A2

+

M0

+*A2

+

M1

+*A2

+^

*A2

^

^

*A2

^

+

*A2+*A2

+

+

F1

*J0

D3

K2

D6

D9

E2

D12

X

*Z0

+

+++

X

+

X

*Z0

+

+++

X

+

X

*X

X

X

*X

+

+++

X

+

X

*X

+

+++

X

+

^

*A1

^

B2

+

^

*A2

^

+

+*A2

+

+

X

*X

$

+++

X

+

Pointer (s)

1 193 385 576 STM-64/OC-192

^

*A1

^

B2

^

*A1

^

B2

+

^

*A1

^

B2

+

X

*X

X

X

*X

$

+++

X

+

+ + P1151

FEC sequence:

$ $ Q1P1P1P1151

FEC sequence:

P1P1 P1161

P1P1 P16320

P1P1

Q1P1P1 P1161

P1P1 P16320

STM-16 STM-64

1 9

99

11

9

1

9

1

9

1

+, $: FEC sequencesQ1, P1: Optional Forward

Error Correction

H1 H2 H3


2.8.3 The SDH/SONET Hierarchy

We have already seen the STM-1 frame, equivalent to STS-3 and OC-1 in SONET,made up of a 9 x 270 byte matrix and transmitted at 155 Mbps. The hierarchy de-fines higher-order frames whose bit rates are obtained by multiplying successivelyby four. For simplicity, each bit rate is usually referred to by its rounded-off value(see Table 2.9).

The STM-n structure frame (n = 4, 16, 64, 256) consists of two section over-heads (RSOH and MSOH) plus an AUG-n (see Table 2.8). Four AUG-n are block-interleaved, to create a superior structure referred to as AUG-4n. For instance, fourAUG-4s are needed to create an AUG-16 (see Figure 2.27).

Table 2.7Multiplex section overhead (MSOH) or line overhead (LOH).

Byte Description

B2 MS bit error monitoring: This is an IP-n x 24 code is calculated over all bits of the pre-vious STM-n frame, except for the three rows of SOH, and it is placed into the B2 bytes of the current frame.

K1, K2 APS bytes: They carry on the APS protocol (see Chapter 7).K2(bit6-8)

MS-RDI: This is used to return an indication (110) to the transmitting end that a defect has been detected or that MS-AIS (111) is received.

D4, D12 MS data communication channel (DCCM): This is a 576-Kbps channel intended for OAM information for central management of multiplexer functions. The STM-256 frame has an additional 9,216-Kbps channel which is carried in bytes from D13 to D156.

S1(bit5-8)

Synchronization status messages: This is used to inform the remote multiplexer on the quality of the clock used to generate the signals. 0x: Unknown, 2x: G811, 4x: G.812 transit, 8x: G.812 local, Bx: G.813, Fx: Not used for synchronization.

M1 MS-REI: The M1 returns the number of the detected BIP-24 x n errors to the remote multiplexer.

M0 MS-REI: This byte is concatenated with M1 to indicate a number of BIP violations greater than 256, (only STM-64 and STM-256).

E2 MS orderwire: Provides a 64-Kbps voice channel for express orderwire between multi-plex section terminations.

P1, Q1 Optional forward error correction: Defined only for STM-16, STM-64, and STM-256.

Table 2.8AUG-n composition.

AUG-1 AUG-4 AUG-16 AUG-64 AUG-256

AU-4 AU4-4c AU4-16c AU4-64c AU4-256c3 x AU-3 4 x AUG-1 4 x AUG-4 4 x AUG-16 4 x AUG-64


Like the STM-1 frame, STM-n frames are represented as a rectangular structureof 270 x n columns and 9 rows, which gives a total of 270 x n x 9 = 2,430 x n bytes.Nonetheless, the frame period remains the same as that of the STM-1 frame: 125 µs.The new SOH is therefore 3 x 9n bytes, the multiplex section 3 x 9n bytes, and 9nbytes for AU-n pointers.

When looking at the STM-n/OC-m frames, the concept of indirect multiplexingcannot be ignored. To explain this, let us look at the example of the STM-16 frame.Forming one of these frames by direct multiplexing means that it has been formedby interleaving bytes from 16 STM-1 frames. An STM-16 structure can also be ob-tained from four STM-4 frames, by indirect multiplexing. In this case, interleavingis carried out in blocks of 4 bytes. The resulting structure is therefore the same forboth cases.

Table 2.9Signals and information combinations.

SDH SONETFrame

SONETOptical Size (Bytes) Rate (Mbps) Acronym Capacity Samples

STM-0 STS-1 OC-1 9x90 51.840 52M 28DS-1, DS-3, E3, 21E1STM-1 STS-3 OC-3 9x270 155.520 155M 84DS-1, 3DS-3, E4, 3E3,

2E3+21E2, E3+42E2, 63E2STM-4 STS-12 OC-12 9x1080 622.080 622M 4OC-3, 4 STM-1STM-16 STS-48 OC-48 9x4320 2488.320 2.5G 16OC-3, 16 STM-1STM-64 STS-192 OC-192 9x17280 9953.280 10G 64OC-3, 64 STM-1STM-256 STS-768 OC-768 9x69120 39814.120 40G 256OC-3, 256 STM-1

Figure 2.27 Multiplexing 4 AUG-n into an AUG-4n is block-interleaving, and the block size is exactly n bytes. 1, 4, and 16 are valid values for n.

#11 36

1

9

AUG-16

4 bytes STM-64 / OC-192

#21 36

AUG-4

1

9

#31 36

AUG-4

1

9

#41 36

AUG-4

1

9

AUG-4

4320

108037 108037 108037 108037

1 144

Blocks of 4consecutive bytes

4 bytes 4 bytes4 bytes


In addition to its functions outlined in previous sections, the pointer mechanismalso facilitates the construction of STM-n. In effect, due to the imperfection that isinherent to synchronization, STM frames reach multiplexers with random relativealignments; that is, some of them are out of phase in respect to others. Nonetheless,the STM-n higher-order frame that leaves the multiplexer keeps its bytes grouped to-gether in a single block. The payloads of the frames can be interleaved without prioralignment, since virtual containers will have a pointer value that has been recalcu-lated in the overhead of the outgoing frame.

2.9 CONCATENATION

Concatenation is the process of summing the bandwidth of X containers (C-i) into alarger container. This provides a bandwidth X times bigger than C-i. It is well indi-cated for the transport of big payloads requiring a container greater than VC-4, butit is also possible to concatenate low-capacity containers, such as VC-11, VC-12, orVC-2.

G1F2H4

J1B3C2

F3K3N1

G1F2H4

J1B3C2

F3K3N1

Figure 2.28 An example of contiguous concatenation and virtual concatenation. Contiguous concatenation requires support by all the nodes. Virtual concatenation allocates bandwidth more efficiently, and can be supported by legacy installations.

Bandwidth requirement

G1F2H4

J1B3C2

F3K3N1

SDH

G1F2H4

J1B3C2

F3K3N1

One concatenated payload

G1F2H4

J1B3C2

F3K3N1

G1F2H4

J1B3C2

F3K3N1

G1F2H4

J1B3C2

F3K3N1

G1F2H4

J1B3C2

F3K3N1

VC4-4c VC4-3v

Bandwidth delivery

One Path Several paths

(430 Mbps)

2

1

3 III

4

II

IV

622 Mbps

3 x155 Mbps

Contiguous concatenation Virtual concatenation

STS-12c SPE or STS-9v

(3 in SDH, or9 in SONET)

One concatenated payload


There are two concatenation methods (see Figure 2.28):

1. Contiguous concatenation: which creates big containers that cannot split intosmaller pieces during transmission. For this, each NE must have a concatena-tion functionality.

2. Virtual concatenation: which transports the individual VCs and aggregatesthem at the end point of the transmission path. For this, concatenation function-ality is only needed at the path termination equipment.

2.9.1 Contiguous Concatenation of VC-4

A VC-4-Xc provides a payload area of X containers of C-4 type. It uses the sameHO-POH used in VC-4, and with identical functionality. This structure can betransported in an STM-n frame (where n = X). However, other combinations arealso possible; for instance, VC-4-4c can be transported in STM-16 and STM-64frames. Concatenation guarantees the integrity of a bit sequence, because the wholecontainer is transported as a unit across the whole network (see Table 2.10).

Obviously, an AU-4-Xc pointer, just like any other AU pointer, indicates the posi-tion of J1, which is the first byte of the VC-4-Xc container. The pointer takes thesame value as the AU-4 pointer, while the remaining bytes take fixed values equalto Y=1001SS11 to indicate concatenation. Pointer justification is carried out thesame way for all the X concatenated AU-4s and X x 3 stuffing bytes (see Figure2.29).

However, contiguous concatenation, today, is more theory than practice, sinceother alternatives more bandwidth-efficient, such as virtual concatenation, are gain-ing more importance.

2.9.2 Virtual Concatenation

Connectionless and packet-oriented technologies, such as IP or Ethernet, do notmatch well the bandwidth granularity provided by contiguous concatenation. For

Table 2.10Contiguous concatenation of VC-4-Xc. X indicates the number of VC-n.

SDH SONET X Capacity Justification Unit Transport

VC-4 STS3c-SPE 1 149,760 Kbps 3 bytes STM-1/OC-3VC-4-4c STS12c-SPE 4 599,040 Kbps 12 bytes STM-4/OC-12VC-4-16c STS48c-SPE 16 2,396,160 Kbps 48 bytes STM-16/OC-48VC-4-64c STS192c-SPE 64 9,584,640 Kbps 192 bytes STM-64/OC-192VC-4-256c STS768c-SPE 256 38,338,560 Kbps 768 bytes STM-256/OC-768


instance, to implement a transport requirement of 1 Gbps, it would be necessary toallocate a VC4-16c container, which has a 2.4-Gbps capacity. More than double thebandwidth is needed.

Virtual concatenation is a solution that allows granular increments of bandwidthin single virtual container (VC-n) units, providing a payload capacity of X times theVC used. The payload is separated at the point where the VC-n-Xv is generated, tobe transmitted across separate paths. At the destination point, the X VC-n virtualcontainers are organized, according to the H4 or the V5 byte, and delivered to theclient (see Table 2.11).

Virtual concatenation needs to be supported only at end nodes, and it is compat-ible with legacy SDH networks that do not support concatenation. To get the fullbenefit out of this, individual containers should be routed to be as diverted as possi-

Table 2.11Capacity of virtually concatenated SDH VC-n-Xv or SONET STS-3Xv SPE.

SDH SONET Individual Capacity Number (X) Virtual Capacity

VC-11 VT.15 SPE 1,600 Kbps 1 to 64 1,600 to 102,400 KbpsVC-12 VT2 SPE 2,176 Kbps 1 to 64 2,176 to 139,264 KbpsVC-2 VT6 SPE 6,784 Kbps 1 to 64 6,784 to 434,176 KbpsVC-3 STS-1 SPE 48,384 Kbps 1 to 256 48,384 to 12,386 KbpsVC-4 STS-3c SPE 149,760 Kbps 1 to 256 149,760 to 38,338,560 Kbps

Figure 2.29 Contiguous concatenation: Pointers and containers. A VC-4-Xc (X = 1, 4, 16, 64, 256) structure, where X represents the level. The increment/decrement unit (justifi-cation) is 3 X, as it depends on the level: AU-4=3 bytes, AU-4-256c=768 bytes.

G1F2H4

J1B3C2

F3K3N1

R

VC-4-Xc / STS-3Xc SPE

Fixed columns

261X

C-4-Xc

260X1H1 Y Y H2 1 1 H3 H3 H3

H3

H1 Y Y....... H2 1 1....... H3 H3 H3.......

H2H1 1 3

1 9

1 36

3

12 24

H1 Y Y....... H2 1 1....... H3 H3 H3.......1 14448 96

AU-3 ptr

AU-4 ptr

AU-4-4c ptr

AU-4-16c ptr

H1 Y Y....... H2 1 1....... H3 H3 H3.......1 9X3X 6X

AU-4-Xc ptr

6

Y: 1001SS111: All 1s byte Justification unit = 3 X bytes size = X-1

(STM-0/STS-1)

(STM-1/OC-3)

(STM-4/OC-12)

(STM-16/OC-48)

(STM-n/OC-m)

STM / STS pointers

or SPE


ble across the network, so, if a link or a node goes down, the connection is affectedonly partially. This is a way of providing a resiliency service (see Figure 2.30).

2.9.3 Link Capacity Adjustment Scheme

Virtual concatenation provides granularity, but continues using predefined band-width allocation, which does not match the bit rate patterns of most data networks.

The link capacity adjustment scheme (LCAS) was designed to obtain dynamicbandwidth allocation. LCAS is a protocol transported in H4 bytes of a 16 frame se-quence, to allow for channel capacity resizing at any time, without disrupting thetraffic. The LCAS optimizes the efficiency of virtual concatenation by enabling dy-namic changes on the containers virtually concatenated, depending on traffic de-mand.

2.10 MAINTENANCE

SDH and SONET transmission systems are robust and reliable; however they arevulnerable to several effects that may cause malfunction. These effects can be clas-sified as follows:

• Natural causes: This include thermal noise, always present in regeneration sys-tems; solar radiation; humidity and Raleigh fading2 in radio systems; hardware

2. Raleigh fading is the phenomenon in which the field detected at the receiver is thesum of many random contributions of different phases and directions, due to multi-path effects.

Figure 2.30 Virtual concatenation uses bandwidth more efficiently. Individual VC-4s can be routed across different paths on the network. If a resource fails, only a part of the bandwidth is affected.

A B

FE

X YG

60%

40%

25%

35%VC4-7v

(1.05 Gbps)

Reassembly

VC4-7v(1.05 Gbps)

VC4-7v(1.05 Gbps)

VC4-7v(1.05 Gbps)

VC4-7v(1.05 Gbps)

VC4-7v(1.05 Gbps)

MFI=k+2

MFI=k+1MFI=k

MFI=kSQ=0..6

MFI=i

MFI=i-1

MFI=i-2

MFI=iSQ=0..6 t1

t1-125µs

t0+250µs

t0+125µs

t0

Segmentation

t1-250µs

Transmission

3

5

6

555

5554


Figure 2.31 Anomalies and defects management. (In regular characters for SDH; in italic for SONET.)

Reg. Section Multiplexer Section Higher-Order Path Lower-Order Path

LOF

LOSA1/A2RS-TIM / TIM-SJ0RS-BIP / CV- SB1

MS-BIP / CV- LB2

(M0),M1MS-REI / REI- L

MS-AIS / AIS-PK26-8

MS-REI / REI- L(M0),M1

AU-AIS / AIS- PAU ptr

MS-RDI / AIS-PK26-8

AU-LOP / LOP-PAU ptr

HP-UNEQ / UNEQ- PC2HP-PLM / PLM-PC2

HP-BIP / CV- PB3

G11-4HP-REI / REI-P

HP-RDI / RDI-PG15-7

HP-REI / REI-PG11-4

HP-LOM / LOMH47-8TU-AIS / AIS-VTU ptr

TU ptr TU-LOP / LOP-V

J2 LP-TIM / TIM-V

V55-7LP-PLM / PLM-V

LP-BIP/ CV-VV51-2

V53LP-REI / REI-V

LP-REI / REI-VV53

V58LP-RDI / RDI-V

Transmission Line

HP-TIM / TIM-PJ1

V55-7LP-UNEQ / UNEQ-V

Event detection

Signal flow

Error indication received

Error indication sent

Alarm indication sent

OR (logical function)

Event registration

SDH

SONET

Section Line STS Path VT PathTransmission

AIS

AIS

AIS

AIS

AIS

AIS

Signal

LP-RFI / RFI-VV55

Internal signal flow


aging; degraded lasers; degradation of electric connections; and electrostaticdischarge.

• A network design pitfall: Bit errors due to bad synchronization in SDH. Timingloops may collapse a transmission network partially, or even completely.

• Human intervention: This includes fiber cuts, electrostatic discharges, powerfailure, and topology modifications.

All these may produce changes in performance, and eventually collapse transmis-sion services.

2.10.1 SDH/SONET Events

SDH/SONET events are classified as anomalies, defects, damage, failures, andalarms depending on how they affect the service:

• Anomaly: This is the smallest disagreement that can be observed between mea-sured and expected characteristics. It could for instance be a bit error. If a sin-gle anomaly occurs, the service will not be interrupted. Anomalies are used tomonitor performance and detect defects (see Section 2.11).

Defect: A defect level is reached when the density of anomalies is high enough tointerrupt a function. Defects are used as input for performance monitoring, to con-trol consequent actions, and to determine fault causes.

• Damage or fault: This is produced when a function cannot finish a requestedaction. This situation does not comprise incapacities caused by preventivemaintenance.

• Failure: Here, the fault cause has persisted long enough so that the ability of anitem to perform a required function may be terminated. Protection mechanismscan now be activated (see Section 2.13).

• Alarm: This is a human-observable indication that draws attention to a failure(detected fault), usually giving an indication of the depth of the damage. Forexample, a light emitting diode (LED), a siren, or an e-mail.

• Indication: Here events are notified upstream to the peer layer for performancemonitoring and eventually to request an action or a human intervention that canfix the situation (see Figure 2.31).

Errors reflect anomalies, and alarms show defects. Terminology here is often usedin a confusing way, in the sense that people may talk about errors but actually referto anomalies, or use the word, “alarm” to refer to a defect.


Figure 2.32 OAM management. Signals are sent downstream and upstream when events are detected at the LP edge (1, 2); HP edge (3, 4); MS edge (5, 6); and RS edge (7, 8).

PDHAU-LOPAU-AIS

LP-RDIV5=xxxxxxx1

LPHPMUXHPLP MUX

PTE PTE

MUXSTM-nOC-m MUX R

LP or VT path

HP or STS path

MS or Line

RS or Section

BIP-2 (V5) with errors

LP-REI

V5=xx1xxxxx

AU-LOPAU-AIS

HP-RDIG1=xxxx1nnx

B3 with errors

HP-REIn=number of errors

G1=nnnnxxxx

AIS

TUAIS

PDHAIS

LP-RDIV5=xxxxxxx1

AU-AIS

HP-RDIG1=xxxx1xxx

B2 with errors

MS-RDIK2=xxxxx110

TUAIS

LP-RDIV5=xxxxxxx1

PDHAISLOS

LOF

MS-REIn=number of errors

M1=nnnnnnnn

HP-RDIG1=xxxx1xxx)

B2 with errors

MS-RDIK2=xxxxx110

LP-RDIV5=xxxxxxx1

LOSLOF

AU-AISTUAIS

PDHAISMS-AIS

(All 1sexcept A1,A2)

1

2

4

3

5

6

7

8

MS

LTELTE STE


In order to support a single-end operation the defect status and the number ofdetected bit errors are sent back to the far-end termination by means of indicationssuch an RDI, REI, or RFI (see Figures 2.31 and 2.32).

2.10.2 Monitoring Events

SDH frames contain a lot of overhead information to monitor and manage events(see Table 2.18). When events are detected, overhead channels are used to notifypeer layers to run network protection procedures or evaluate performance. Messag-es are also sent to higher layers to indicate the local detection of a service affectingfault to the far-end terminations.

Defects trigger a sequence of upstream messages using G1 and V2 bytes. Down-stream AIS signals are sent to indicate service unavailability. When defects are de-tected, upstream indications are sent to register and troubleshoot causes.

2.10.3 Event Tables

Tables 2.12-2.16 summarize events and indications associated with each SDH lay-er. Testing events have also been included. .

2.11 PERFORMANCE MONITORING

SDH has performance monitoring capabilities based on bit error monitoring. A bitparity is calculated for all bits of the previous frame, and the result is sent as over-head. The far-end element repeats the calculation and compares it with the received

Table 2.12Regenerator and multiplex section events and indications.

SDH SONET Type How Comments

LOS LOS Defect BER>limit Loss of signal detectionECOD ECOD Anomaly Code error Line code violationOOF OOF Anomaly A1-A2 Out of frame detectionLOF LOF Defect A1-A2 Loss of frame detectionRS-TIM TIM-S Defect J0 Trace identifier mismatchB1 error B1 error Anomaly B1 Bit error detected by BIP-8 verificationB2 error B2 error Anomaly B2 Bit error detected by BIP-24 verificationMS-REI REI-L Indication M1=xxx

M0= xxxNumber of errors detected using B2M0 is used by STM-64 and STM-256 only

MS-AIS AIS-L Alarm K2(6-8)=111 Mux/line alarm indication signal detection

MS-RDI RDI-L Indication K2(6-8)=110 Mux/line remote defect indication


Table 2.13Path events and indications.


HP-TIM TIM-P Defect J1 Trace identifier mismatch in pathB3 error B3 error Anomaly B3 Bit error detected by BIP-24 verificationHP-REI REI-P Indication G1(1-4)=xxxx Number of errors detected using B3

HP-UNEQ UNEQ-P Defect C2(1-8)=0 Unequipped or supervisory unequipped

HP-PLM PLM-P Defect C2(1-8)=x Payload label mismatch

HP-RDI RDI-P Indication G1(5-7)=010 Payload defect PLM

HP-RDI RDI-P Indication G1(5-7)=101 Server defect. AIS or loss of pointer (LOP)

HP-RDI RDI-P Indication G1(5-7)=110 Connectivity defect. TIM or UNEQ

HP-LOM LOM-V Defect H4 Loss of multiframe (H4 is in POH)LP-TIM TIM-V Defect J2 Trace identifier mismatch in pathBIP-2 error BIP-2 error Anomaly V5(1-2) Error detected by BIP-2 verification

LP-REI REI-V Indication V5(3)=1 One or more errors detected by BIP-2 in V5

LP-RFI RFI-V Indication V5(4)=1 Remote failure indication lower-order path

LP-UNEQ UNEQ-V Defect V5(5-7)=0 Unequipped or supervisory unequipped

LP-PLM PLM-V Defect V5(5-7)=x Payload label mismatch

LP-RDI RDI-V Indication V5(8)=1 Remote defect indication lower-order path

Table 2.14Pointer events.


AU-NDF NDF-P Ptr event H1, H2 New AU Pointer (STS pointer in SONET)AU-PJE PJE-P Ptr event H1, H2 Pointer justificationAU-Inv Inv-P Ptr event H1, H2 Pointer inversionAU-Inc Inc-P Ptr event H1, H2 Pointer incrementAU-Dec Dec-P Ptr event H1, H2 Pointer decrementAU-LOP LOP-P Defect H1, H2 Loss of pointerAU-AIS AIS-P Alarm H1, H2 Alarm indication signal detectionTU-NDF NDF-V Ptr event V1, V2 New TU pointerTU-PJE PJE-V Ptr event V1, V2 Pointer justificationTU-Inv Inv-V Ptr event V1, V2 Pointer inversionTU-Inc Inc-V Ptr event V1, V2 Pointer incrementTU-Dec Dec-V Ptr event V1, V2 Pointer decrementTU-LOP LOP-V Defect V1, V2 Loss of pointerTU-AIS AIS-V Alarm V1, V2 Alarm indication signal detection


overhead. If the result is equal, there is considered to be no bit error; otherwise, abit error indication is sent to the peer end.

2.11.1 Bit Error Checking

Bit error monitoring is based on checking the value of certain groups of bits thatmake up bit interleaved parity (BIP) words as a checksum. This way, there is paritychecking between regenerators, multiplex sections, and paths (see Figure 2.33). If

Table 2.15Tandem connection monitoring events.

SDH/SONET Type How Comments

HPTC-LTCLPTC-LTC

Defect N1(7-8) (frames 1-8)N2(7-8) (frames 1-8)

Higher/lower-order path tandem connection loss of tandem connection monitoring.

HPTC-TIMLPTC-TIM

Defect N1(7-8) (frames 9-72)N2(7-8) (frames 9-72)

Higher/lower-order path tandem connection trace identifier mismatch.

HPTC-UNEQLPTC-UNEQ

Defect N1=0N2=0

Higher/lower-order path tandem connection unequipped

HPTC-RDILPTC-RDI

Indication N1(8) (frame 73)N2(8) (frame 73)

Higher/lower-order path tandem connection remote defect indication

HPTC-AISLPTC-AIS

Indication N1(1-4)=1110N2(4)=1

Higher/lower-order path tandem connection alarms indication signal

HPTC-IECLPTC-IEC

Anomaly N1(1-4)=xxxxN2(1-2)=xx

Higher/lower-order path tandem connection incoming error count

HPTC-ODILPTC-ODI

Indication N1(7) (frame 74)N2(7) (frame 74)

Higher/lower-order path tandem connection outgoing defect indication

HPTC-REILPTC-REI

Indication N1(5)=1 N2(5)=1

Higher/lower-order path tandem connection remote error indication

HPTC-OEILPTC-OEI

Indication N1(6)=1 N2(6)=1

Higher/lower-order path tandem connection outgoing error indication

Table 2.16Line and test sequence events. These events do not depend on the digital hierarchy, but are only

related to the test sequences used, or to the characteristics of the signal.

SDH SONET Type Detection Cause

TSE TSE Anomaly Test pattern Test sequence error (O.181)LSS LSS Anomaly Test pattern Loss of sequence synchronizationSlip Slip Fault Test pattern PLL buffer overflow or underflowLTI LTI Fault Test equipment Loss of timing inputOptical saturation Fault Optical receiver Optical power mismatch


the received signal contains bit errors, a BIP indication is generated that is treatedas an anomaly, and an REI indication is sent to the far end (see Table 2.17).

Network elements themselves make certain decisions after evaluating the frame re-ceived and generating a new one to the next NE. BIP is a continuous monitoringprocess that provides the operator with a powerful error detection tool, and acti-vates the corresponding mechanisms to deal with detected anomalies and also topermit evaluation of the performance of the network (see Chapter 7).

2.11.2 Tandem Connection Monitoring

Tandem connection monitoring (TCM) is a sublayer between the multiplex sectionand path layers. A tandem connection transports the virtual container in a reliableway when it is routed via networks of different operators. When a bit error occurs,the TCM protocol informs the network about its location (see Figure 2.34).

Table 2.17Bit interleaved parity bytes

Byte BIP SDH BIP area REI

B1 BIP-8 Reg section STM/STS frame —B2 BIP-8/BIP-24 Multiplex section Frame excluding RSOH MS-REIB3 BIP-8 Higher-order path HO virtual container LO-REIV5(1-2) BIP-2 Lower-order path LO virtual container HO-REI

Figure 2.33 Bit interleaved parity (BIP-n) enables error monitoring. A transmitter performs the exclusive or (XOR) function (even parity) over the previous block. The value com-puted is placed in the n bits before the block is scrambled.

1 0 0 1 1 1 0 10 0 0 1 1 0 1 11 0 1 1 0 1 1 11 0 0 1 0 1 0 0

10100101

XOR

B1 =

1 01 11 0

10

XOR

V5 (bits 1-2) =

BIP-2

Multiplex Section

VC-11, VC-12

1 0 0 1 1 1 0 10 0 0 1 1 0 1 11 0 1 1 0 1 1 11 0 0 1 0 1 0 1

00100101

XOR

B2 =

1 0 0 1 1 0 0 00 0 0 1 1 0 1 00 0 1 1 0 0 1 10 0 0 1 0 1 0 0

10100101

1 0 0 1 1 1 0 10 0 0 1 1 0 1 11 0 1 1 0 1 1 11 0 0 1 0 1 1 0

11100101

1 0 0 1 1 1 0 10 0 0 1 1 0 1 11 0 1 1 0 1 1 11 0 0 1 0 1 0 0

10100101

VC4

= B3

Regenerator Section

BIP-8 VC3

BIP-24

110010100110

XOR table

++

+

+


This mechanism calculates the number of bit errors that occurs when the VC en-ters the subnetwork. When the VC arrives at the subnetwork end, the number of biterrors is computed again, and the two results are compared. The input point is calledthe TCM source and the output point the TCM sink:

1. The number of bit errors detected (by means of the BIP-8) in the incoming VC-n at the TCM source is written to the incoming error count (IEC) (see Table2.15).

2. When the VC-n arrives at the TCM sink, the number of bit errors is calculatedagain, and, if the figure is different from the IEC, this tells us that new biterrors have occurred.

2.11.3 Forward Error Correction

Forward error correction (FEC) can reduce bit error rate (BER) in optical trans-mission, providing correction capabilities at the receiving end. The mathematicalalgorithm used to implement FEC in SDH is Bose-Chaudhuri-Hocquenghem(BCH). BCH is performed in the data being transported, and the results are storedin the P1 and Q1 bytes of the RS and MS sections. At the receiving end, we check ifbit errors have occurred, and, if so, we correct them. FEC is defined for STM-16,

Figure 2.34 Tandem connection monitoring sample: Three operators transporting a higher-order payload.

OEI

MUX

VC Path

Tandem Connection

HP-RDI(G1=xxxx1xxx)

MUX

Operator X Operator Z

Source

Sync

B3:=BIP-8 errors

B3=BIP-8

IEC=no. of errors

(N1=xxxxx1xx)

N1=0B3<>BIP-8

MUX

Sync

SourceMUX

B3<>BIP-8

Operator Y

TC-REI, OEI

HP-RDI

(N1=xxxx11xx)

N1=0B3<>BIP-8

B3<>BIP-8errorsIEC<>BIP-8B3:=BIP-8

(G1=xxxx1xxx)

B3<>BIP-8

N1=0


STM-64, and STM-256, and it uses MSOH and RSOH overhead bytes, providingcorrection for the AUG-n area.

2.12 DEFECTS

A defect is understood as any serious or persistent event that holds up the transmis-sion service. SDH defect processing reports and locates failures in either the com-plete end-to-end circuit (HP-RDI, LP-RDI) or on a specific multiplex sectionbetween adjacent SDH nodes (MS-RDI) (see Figure 2.31).

Table 2.18Analysis criteria of SDH/SONET events.

Event Criterion

LOS Loss of signal: This parameter should be raised when incoming power at the receiver has dropped to a level that produces a high BER. LOS indicates either a transmitter failure or an optical path break. Timing requirements for detection and reset fall within regional standards.

OOF Out of frame: This is raised when five frames are received with error in the FAS (incor-rect patterns in A1 and A2). The maximum OOF detection is 625 µs and it should be cleared after receiving one correct frame (ITU-T G.783, ANSI T1.231).

LOF Loss of frame: OOF events are collectively referred to as LOF. If the OOF state persists for 2.5 ms ± 0.5 ms, an LOF should be declared. LOF should be left after 2.5 ms with-out OOF (ITU-T G.783, ANSI T1.231).

LOP Loss of pointer: The LOP state is entered in the case of n consecutive invalid point-ers or n consecutive new data flag (NDF) enable flags . The LOP state should be cleared after three consecutive valid pointers or three consec-utive AIS indications (G.783, ANSI T1.231).In SDH: AU-LOP, TU-LOP.In SONET: LOP-P, LOP-V.

LOM Loss of multiframe: H4 byte does not track the multiframe sequence during eight frames. (ITU-T G.783).In SDH: HP-LOM.In SONET: LOM.

UNEQ Unequipped connectivity defect: C2 or V5 is equal to “0” during five consecutive frames (ITU-T G.783, ANSI T1.231).In SDH: HP-UNEQ, LP-UNEQ.In SONET: UNEQ-P, UNEQ-V.

TIM Trace identifier mismatch connectivity defect: The CRC of the J1 or J2 identifier does not match during n consecutive frames (ITU-T G.783, ANSI T1.231).In SDH: HP-TIM, LP-TIM.In SONET: TIM-P, TIM-V.

8 n 10≤ ≤( ) 8 n 10≤ ≤( )


Alarm indication signal

An alarm indication signal (AIS) is activated under standardized criteria (see Table2.18), and sent downstream in a path in the client layer to the next NE to inform aboutthe event (see Figure 2.31). The AIS will arrive finally at the NE at which that pathterminates, where the client layer interfaces with the SDH network (see Figure 2.32).

As an answer to a received AIS, a remote defect indication is sent backwards.An RDI is indicated in a specific byte, while an AIS is a sequence of “1s” in the pay-load space. The permanent sequence of “1s” tells the receiver that a defect affectsthe service, and no information can be provided.

PLM Payload label mismatch payload defect: The C2 or V5 contents are not consistent with the specified label during five consecutive frames (ITU-T G.783, ANSI T1.231).In SDH: HP-PLM, LP-PLM.In SONET: PLM-P, PLM-V.

REI Remote error indication: This indication contains the number of bit errors detected at the receiving node. REI is sent back to the far end to allow bit error monitoring and sin-gle-end control (ITU-T G.707, ANSI T1.231).In SDH: MS-REI, HP-REI, LP-REI, TC-REI.In SONET: REI-L, REI-P, REI-V, TC-REI.

RDI Remote defect indication: This indication is sent to the transmission end upon detecting LOS, LOF, or AIS defect. This indication was known previously as FERF. RDI should be detected before five consecutive frames with G1 or V5 arisen. (ITU-T G.783, ANSI T1.231).In SDH: MS-REI, HP-RDI, LP-RDI, TC-REI.In SONET: RDI-L, RDI-P, REI-V, TC-REI.

AIS Alarm indication signal: This indication is an all-ones signal. It is generated to replace the normal traffic signal when it contains a defect. The receiver has to detect it after three consecutive frames with K2=xxxxx111 or H1, H2 = 11111111 (ITU-T G.783, ANSI T1.231). In SDH: MS-AIS or SONET: AIS-L. In SDH: AU-AIS, TU-AIS or SONET: AIS-P, AIS-L.

RFI Remote failure indication: This indication is sent when a defect persists for a period of time. RFI is returned to the transmission end when a LOS, LOF, or AIS surpasses a predetermined period of time to activate the protection switch protocol to provide an alternative path.In SDH: LP-RFI.In SONET: RFI-L, RFI-P, RFI-V.

LSS Loss of sequence synchronization: This signal is activated during a test when a pseudo-random pattern is generated in one extreme, and on the receiver side the BER > 0.20 dur-ing 1 second; or long duration AIS; or uncontrolled bit slip; or loss of signal (M.2100).

Table 2.18Analysis criteria of SDH/SONET events.


Depending on which service is affected, the AIS signal adopts several forms (seeFigure 2.35):

• MS-AIS: All bits except for the RSOH are set to the binary value “1.”

• AU-AIS: All bits of the administrative unit are set to “1” but the RSOH andMSOH maintain their codification.

• TU-AIS: All bits in the tributary unit are set to “1,” but the unaffected tributar-ies and the RSOH and MSOH maintain their codification.

• PDH-AIS: All the bits in the tributary are “1.”

Enhanced remote defect indication

Enhanced remote defect indication (E-RDI) provides the SDH network with addi-tional information about the defect cause by means of differentiating:

• Server defects: like AIS and LOP;

• Connectivity defects: like TIM and UNEQ;

• Payload defects: like PLM.

Enhanced RDI information is codified in G1 (bits 5-7) or in k4 (bits 5-7), depend-ing on the path.

2.13 SDH RESILIENCE

Security consists of a series of contingency procedures to recover the voice/dataservice through a new path when the previous becomes unavailable because a net-work resource, such as a link or a node, fails. Fault detection, excessive bit errorrate, AIS detection or a network management request are common reasons for secu-rity procedures to run. Security strategies in transmission networks can be groupedas follows:

Figure 2.35 AIS formats.

RSOH

MSOH

RSOH

MSOH

K2PTR PTR

RSOH

MSOH

PTR

: X= 1

AU-AIS TU-AIS PDH-AIS MS-AIS


Diversity

This strategy consists of dividing paths between two points into different routes(see Figure 2.36). A breakdown on one of the routes will affect only a portion of thetotal traffic. This method, which has largely been applied to legacy networks, canalso be applied to SDH by using virtual concatenation and sending the trafficthrough several paths (see Section 2.9.2). Service is restored only when the re-source is repaired.

Restoration

This scheme calls for special nodes and external control software permanently ana-lyzing service failures (see Figure 2.37). When the process is triggered, an alterna-tive route is selected from spare resources that are used on demand instead of beingpreassigned. Everything can be replaced, including terminal nodes. Recovery timeis within the range of several minutes.

Figure 2.36 Diversification strategy between points X and Y.

Path1: 60%

Path2: 40%

A B

FE

X YG

A B

FE

X YG

Path1: 0%

Path2: 40%

100% 40%

Normal operation After breakdown

Figure 2.37 Restoration sample, whereby every connection is defined by a couple (x, y), where “x” is the number of active circuits, and “y” is the number of protection circuits.

(5,2)

(4,2)(3,4)

(7,7)

(4,5)

(7,0)

(4,2)(5,2)

(11,3)

(0,0)

A

D

B

C

A

D

B

C

Normal operation Restoration


Protection

This mechanism preassigns spare resources when the working ones are faced withfailure. This type of recovery is controlled from the network elements themselves,using internal information rather than control software like in restoration. Recoverytime is within the range of milliseconds.

2.13.1 Protection Basics

There is set of strategies that can be called protection. All of them use internal in-formation of the SDH network to configure fault-tolerant networks. In the follow-ing, we shall briefly look at some protection-related concepts:

• Working or protection resources: Working resources (lines, nodes, paths, sec-tions) transport traffic during normal operation. Protection resources are thereto replace them after a network failure (see Figure 2.39).

• Active and passive protection: Active protection uses a protocol to specify theprotection action, and recovery time depends on the number of nodes whichneed to be controlled to be below the limits. Passive protection is less sophisti-cated and independent of the number of nodes.

• Automatic protection switching: APS is the standard capacity of automatic re-covery after a failure at the multiplex section layer. K1 and K2 bytes are usedto manage the protection protocol (see Figures 2.38 and 2.39).

• Dedicated protection and shared protection: In dedicated protection eachworking channel has a protection channel. In shared protection n protectionchannels are used by m channels to be protected (see Figure 2.38). See also 1+1 or 1:n configurations.

protection ringservice ring

Figure 2.38 Shared and dedicated protection architectures.

service/protection ringservice/protection ring

Dedicated Shared

Working lineProtection line


• 1+1 or 1:n configurations: 1+1 means dedicated protection, where the protect-ed signal is sent to the destination on two separate channels. No special proto-col is needed, since the best signal is selected at the reception end. The switchthreshold is programmable; usually based on BIP error rate. This is a simpleand fast configuration that performs a 100% restoration, but it is also expen-sive. 1:n configuration means shared protection, where the n number of work-ing channels are protected by one protection channel using the APS protocol.The protection channel can transmit an idle signal or extra traffic. It is cheaper,but its downside is that it does not perform 100% restoration (see Figure 2.39).

• Unidirectional or bidirectional protection: In a normal situation, a unidirec-tional ring routes traffic only in one direction (i.e., clockwise). A bidirectionalring routes traffic in both directions. After a failure has occurred in one direc-tion, a bidirectional strategy switches both directions, affected or unaffected.When it comes to the unidirectional strategy, only the affected direction isswitched (see Figure 2.40).

• Ring-switching or span-switching: During a ring switch, the traffic is carriedover the protection channels on the long path. During a span switch, the trafficis carried over the protection channels on the same span as the failure. Spanswitch is similar to 1:1 linear protection, but applies only to four fiber rings(see Figure 2.42).

Figure 2.39 MSP protection. In 1+1 dedicated protection incoming signals are permanently bridged at far end. In 1:n shared protection working facilities use a common facility and APS protocol is required to manage the procedure. To be more effective, work-ing and protection resources should take a different geographical path.

1+1

Near End Far End

1+1 Dedicated Protection 1:n Shared Protection

Near End Far End

1:2

LOS LOS MS-AIS

K1=SF, Ch2

K2=Ch2

1:2

Bridge toProtection

Switch toProtection

1+1

Working lineProtection line

Rx

Tx 12

12

12

12

Rx

Tx

12

12

12

12

PermanentBridge

Switch toProtection

APS Protocol


• Dual-ended or single-ended protection: See unidirectional/bidirectional pro-tection.

• Trail or subnetwork protection: Trail protection is used when the protected re-source is a path or a multiplex section. If this is not the case, the scheme is clas-sified as subnetwork protection; for instance, in the case of the route betweentwo DXCs (see Figure 2.43).

• Revertive or nonrevertive protection: If revertive protection is used, the normaldata flow reverts to the original working resources once a failure has been re-paired. This scheme is used in 1:n configuration. Protection is nonrevertive ifthe protection channel, is treated as a working channel and the flow does notreturn to the original resources. This is used in 1+1 configurations.

• Bridge/switch signals: A bridged signal means that it is sent over two fibers.On the reception side the best signal is switched or selected.

Protection is a wide concept that can be implemented using a number of differentstrategies. Some of the most common are presented below. For more information,refer to ITU-T Rec. G.707 and G.841.

2.13.2 Multiplex Section or Line Protection

Multiplex section protection (MSP) schemes protect all the traffic flowing througha multiplex section without any discrimination. Switching actions are generallymanaged by the APS protocol using the K1 and K2 bytes. The protection switchingactions must be initiated within 10 ms after detecting signal fails, and traffic mustbe restored within 50 ms. This means that transport service should be restored with-in 60 ms after the fault.

Figure 2.40 Unidirectional and bidirectional protection strategies.

Unidirectional Bidirectional


Multiplex section linear protection

In multiplex section linear protection (MSLP), a working line is protected by a ded-icated protection facility. The simplest implementation uses a 1+1 configuration,and traffic is transmitted simultaneously along working and protection lines, withthe better of the two signals selected at the receiving end. 1:n configuration is alsopossible (see Figure 2.39). In that case “n” working lines share a unique protectionline.

Multiplex section dedicated protection ring

A multiplex section dedicated protection ring (MSDPRING) is a unidirectional ringusing a 1:1 dedicated protection scheme. The ring has two fibers: one working fiberand one protection fiber. Since traffic only travels in one direction unless a fail oc-curs, affected traffic is bridged at the entry node (see Figure 2.41).

Multiplex section shared protection

Multiplex section shared protection (MSSPRING) is a bidirectional ring using a 1:nshared protection scheme. The principle of sharing is based on the idea that work-ing channels and protection channels share the same multiplex section. Any sectioncan have access to the protection channels when a failure occurs. MSSPRING canbe categorized into two types: two fiber and four fiber rings.

1. Two fiber ring: Each fiber carries both working and protection channelspermanently. Working channels in one fiber are protected by the protectionchannels traveling in the opposite direction around the ring. Only ringswitching is possible (see Figure 2.42).

Figure 2.41 Multiplex section dedicated protection ring (MSDPRING) also known as unidirec-tional self-healing ring (USHR). The counterrotating ring provides the protection.

One Fiber ringsA A

BBAB

AB

Protection ringWorking ring

Normal operation Under protection


2. Four fiber ring: Working and protection channels are carried over differentfibers. Two multiplex sections transmitting in opposite directions carry theworking channels, while two multiplex sections, also transmitting in oppositedirections, carry the protection channels. This scheme allows both span switch-ing and ring switching (see Figure 2.42).

Span switching is a simple scheme equivalent to 1:1 protection between two adja-cent nodes. Ring switching is more complex, but prevents node faults and multiplefiber failures when routing the traffic away from the problem.

As in the previous case, the protection channels can transport low-priority traf-fic when they are not carrying out their protection function.

2.13.2.1 VC path protection

Virtual container path protection (VC-P): This scheme allows the protection of in-dividual virtual containers across the whole path where physically separated routesexist. The protection can be across different sections and different operators. The

Figure 2.42 Multiplex section shared protection ring (MSSPRING) also known as bidirectional self healing ring (BSHR).Two fiber rings only allow for ring switching, while four fiber rings enable both span and ring switching.

2 fiber 2 fiber

4 fiber 4 fiber 4 fiber

Span-switching Ring-switchingNormal operationWorking and protection

ring

Span

MSS

PRIN

G T

wo

Fibe

rsM

SSPR

ING

Fou

r Fib

ers


switching actions are managed at a higher level using the K3 (to protect VC3, VC4)or K4 (to protect VC11, VC12, VC2).

VC-P is a dedicated end-to-end protection that can be used in meshed, linear,and rings topologies. The protection switching may be either unidirectional or bidi-rectional (see Figure 2.43).

2.13.2.2 Subnetwork connection protection

Subnetwork connection protection (SNC-P), equivalent to SONET unidirectionalpath switched rings, is a 1+1 linear protection scheme. If VC-P provides surveil-lance to the whole path, SNC-P offers protection between two points on a path. Theprotection can switch on server failures using either inherent monitoring, such asAIS and LOP, nonintrusive information obtained from the POH, or client-layer in-formation.

The SNC scheme can be used on any network topology. An example of a sub-network is a link between two DXCs with no path defined between them. The SNCcan, in some cases, be the simplest method of protecting services across an intercon-nection (see Figure 2.44).

SNC-P was the first ring protection scheme to be deployed and is still useful inaccess topologies, but it is probably not the best option for complex core networks.

Figure 2.43 VC-P protection provides transport resilience across a tandem connection service.

VC

HO/LO Path

Operator A Operator B

Operator A Operator B

VC

VC VC

Normal Operation

Under Protection


2.14 OPERATION, ADMINISTRATION, AND MANAGEMENT

Operation, administration, and management (OAM) functions have been standard-ized by telecommunications management networking (TMN) in the ITU-T M.3000recommendation series. TMN provides a framework for achieving a set of OAMservices across heterogeneous networks.

The TMN defines a way of carrying out operation and maintenance tasks. It en-ables the center, often called operation support system (OSS), to communicate withthe network elements of the installation.

There is a trend among operators to buy and install SDH from different vendors,because interoperation is guaranteed by transmission standards. This does not, how-ever, mean that the management programs are compatible.

2.14.1 The TMN Standard

The TMN standard includes processes called management entities to manage theinformation. A management entity may take on one of two possible roles:

1. Manager: where it is the application that controls the network. It sends direc-tives, and processes and stores the information received.

2. Agent: which is a process installed in the NE. Agents send responses thatinclude information on performance, anomalies, and defects. They control bothphysical (switches, multiplexers, registers) and logical resources (multiplexsection, paths, etc.) (see Figure 2.45).

Protection

Service

ADM

ADM

ADM

Two rings A B

A

B A

B

A

B

Figure 2.44 SNC-P is a dedicated protection mechanism. Traffic is sent simultaneously over both working and service lines. When a failure occurs, the far end switches to the alternative channel. Equivalent to SONET unidirectional path switched ring (UPSR).

Switch to protection

Bridging

AIS


2.14.1.1 Central management

The TMN describes the information exchange between management entities usingthe open systems interconnection (OSI) seven-layer model. The standard manage-ment includes:

• Common management information protocol (CMIP), which defines how themanager and the agent send and receive requests and responses. It is more ro-bust, scalable, and secure than the protocol used in data communications thesimple network management protocol (SNMP). However, its implementation ismore complex.

• The Q interface, which is the point of reference of CMIP where data is ex-changed between the TMN operating system and the network elements. Q3 isdefined for the operation system functions (OSF) and Qx for NE. A mediationdevice (MD) is usually in the middle. The data communication protocol at thispoint can be any of the following: Ethernet, X.25, ISDN, or TCP/IP.

• The X interface, which is used to communicate between two separate TMNsystems.

• Embedded channels, which are used by the TMN to exchange information be-tween entities. These channels are formed by STM/STS-frame Di-bytes (seeSection 2.8.2). The TMN sends and receives management information acrossthe network using the section overhead D-channels (see Figure 2.45).

NEs are manufactured with Qx interfaces to facilitate their integration into TMNarchitectures. The managed information itself, together with the rules by which it ispresented and managed, is called the management information base (MIB).

2.14.2 TMN Benefits

The TMN has various benefits, including scalability, object-oriented management,and the fact that it does not require proprietary solutions. Operators can managecomplex and dynamic SDH networks easily, while maintaining quality and protect-ing legacy investments. NEs have information exchange capabilities, using embed-ded Di channels that enable the OSF to reach to all the points of the network just byusing a gateway. This means that the TMN does not need to link all the nodes, justone or a couple of gateways is enough to reach any agent installed in any NE.

The information gathered is the base on which to elaborate performance analy-sis that determines the service level agreement (SLA) control. The TMN providescomprehensive event information, making it easier to diagnose, troubleshoot, andrepair the service. All of these are essential tasks for maintenance and operation.


2.15 NEXT GENERATION SDH

The technologies that today challenge the supremacy of SDH are a combination ofIP with GbEthernet, RPR, and multiprotocol label switching (MPLS). Apparently,they can replace SDH without being compatible with the installed base. There is avery good reason behind this strategy: If 95% of the traffic is generated and termi-nated in IP/Ethernet, why do we need intermediate protocols?

These challenger technologies are being installed and commissioned in newmetropolitan environments that provide data services. Voice and access are also onthe road map. IP and Ethernet completely dominate datacommunications networks,the Internet, virtual private networks (VPNs), and LANs. They have the followingkey features:

• Low cost, easy installation, simple maintenance;

• Direct bandwidth provisioning;

• High flexibility in topologies;

• Scalability in terms of speed and distance.

Unfortunately, some weaknesses (poor management, lack of quality of service, jit-ter) still limit their application. However, they are moving in two directions:

STM-nOC-m

Figure 2.45 The transmission network is managed and monitored in a centralized way. The embedded DCC channels of section overheads (D1-D3 and D4-D12) allow the interchange of management information through the NE until the gateway which delivers it to the OSF. Agents, installed in the NE, provide information to the man-agement system.

OSFOSFX

MD

Q3Q3

Qx

MUX

Agent

CMIP ManagementCommunication MD

Qx

GatewayManager

CMIP

STM-nOC-m

ControlCenter

DCC channels


1. Optical integration to improve performance and resilience with several solu-tions have already developed: EoFiber, EoRPR, EoDWDM, and EoSDH.

2. IP delayering process to provide quality voice and video as well.

Originally designed for telephony services, SDH has dominated the transmissionnetworks of the world since the early 90s, providing high-quality connections.Their key features and benefits are:

• Comprehensive OAM functions;

• Resilience mechanisms to configure fault-tolerant architectures;

• Performance monitoring and hierarchical event control;

• Synchronization reduces jitter and wander below limits set in standards.

Unfortunately, static point-to-point bandwidth provisioning is not the most efficientsolution for burst data. But SDH continues improving its granularity and flexibilityfor data transport by means of such new standards as GFP, virtual concatenation,and LCAS. In addition a new higher hierarchy, STM-256/OC-768, is ready to beinstalled. In spite of all this evolution, IP/Ethernet will probably dominate metro-politan networks for data and voice applications. Nevertheless, for services with

Figure 2.46 Next generation networks.

ISDN/POTS IP/MPLS

Data

(Gb)Ethernet

RPRHDLC/PPP

GFP

SDH/SONET GbEthernet

WDM

Fiber

FRL

PDH

PoS

Voice Video 3G Internet

xDSL

application

access

optic

core

media

metro

info

ATM


high-quality requirements, such as DVB, long-distance voice, E1/T1, and QoS pro-tocols, the circuit approach provided by SDH will still be preferred.

The answer to the dilemma is a new generation of multiservice nodes providingthe best of both worlds by a process of convergence (see Figure 2.46). They integrateall the remote access, routing, and switching capabilities into the same pipe. In ad-dition, there are:

• Network interfaces: GbEthernet, STM-n/OC-m, and DWDM;

• Protocols: TDM, ATM, IP, and virtual LAN;

• Client interfaces: VC, VT, Ethernet, ATM and IP, and DVB.

They maintain resilience, and any topology is possible be it linear, UPSR, BLSR, ormesh. The definite proof of convergence will be if the next rate defined for Ethernetis 40 Gbps rather than 100 Gbps.


• William Stallings, Data and Computer Communications, Prentice-Hall, 1997.

• Mike Sexton and Andy Reid, Broadband Networking, Norwood, MA: Artech House, 1997.

• Stamatios V. Kartalopoulos, Understanding SONET/SDH and ATM, IEEE Press, 1999.

• Dirceu Cavendish, “New Transport Services for Next Generation SONET/SDH System,” IEEE Mag-azine, May 2002.

• Enrique Hernández Valencia, “The Generic Framing Procedure (GFP),” IEEE Communications Mag-azine, May 2002.

• ISO/IEC 3309 Information Technology - Telecommunications and Information eXchange BetweenSystems - High-level Data Link Control (HDLC) Procedures - Frame Structure, 1993.

• ANSI T1, A Comparison of SONET and SDH, May 1994.

• ANSI T1.105.07, Synchronous optical network (SONET) - Sub-STS-1 Interface rates and formatsspecification, 1996.

• ANSI T1.105.05, Synchronous optical network (SONET) - Tandem connection maintenance, 1994.

• ANSI T1.106, Hierarchy optical interface specifications: Single-mode, 1988.

• ANSI T1.107, Digital hierarchy - Formats specifications, 1995.

• ANSI T1.105.02, Synchronous optical network (SONET) - Payload mappings, 1995.

• ANSI T1.119, Information systems - Synchronous optical network (SONET) - Operations, adminis-tration, maintenance, and provisioning (OAM&P), 1994.

• ITU-T Rec. X.85/Y.1321, IP over SDH using LAPS, Oct. 2000.

• ITU-T Rec. G.701, Vocabulary of digital transmission and multiplexing and PCM terms, March 1993.


• ITU-T Rec. G.702, Digital Hierarchy bit rates, Nov. 1988.

• ITU-T Rec. G.703, Physical/electrical characteristics of hierarchical digital interfaces, March 93.

• ITU-T Rec. G.704, Synchronous frame structures used at 1544, 6312, 2048, 8448, and 44736 Kbpshierarchical levels, Oct. 1988.

• ITU-T Rec. G.7041/Y.1303, Generic framing procedure (GPF), Dec. 2001.

• ITU-T Rec. G.707, Link capacity adjustment scheme (LCAS) for virtual concatenated signals, Oct.2001, Nov. 2001.

• ITU-T Rec. G.707, Network node interface for the SDH, Oct. 2000.

• ITU-T Rec. G.708, Sub STM-0 node interface for SDH, June 1999.

• ITU-T Rec. G.709, Interfaces for the optical transport network (OTN), Nov. 2001.

• ITU-T Rec. G.772, Protected monitoring points provided on digital transmission systems, March1993.

• ITU-T Rec. G.780, Vocabulary of terms for SDH networks and equipment, June 1999.

• ITU-T Rec. G.783, Characteristics of SDH equipment functional blocks, Oct. 2000.

• ITU-T Rec. G.784, SDH management, June 1999.

• ITU-T Rec. G.803, Architecture of transport networks based on the SDH, March 2000.

• ITU-T Rec. G.806, Characteristics of transport equipment, Oct. 2000.

• ITU-T Rec. G.831, Management capabilities of transport network based on SDH, March 2000.

• ITU-T Rec. G.841, Types and characteristics of SDH network protection architectures. Oct. 1998.

• ITU-T Rec. G.957, Optical interfaces for equipment and systems relating to the SDH, June 1999.

• ITU-T Rec. G.958, Digital line systems based on SDH for use on optical fibre cables, Nov. 1994.

• ITU-T Rec. I.432, B-ISDN user-network interface – Physical layer specification, 1993.

• José M. Caballero and Andreu Guimerá, Jerarquías Digitales de Multiplexión, PDH y SDH, Sincroni-zación de Redes, L&M Data Communications 2001.

99

Chapter 3

ATM Architectures

Asynchronous transfer mode (ATM) has strengthened its position as a solid and re-liable technology among the currently existing telecommunications networks.ATM is used in multiple environments, from access networks to backbones, to of-fer a variety of services, such as data, voice, and video transmission. Although it isfar from being the predominant technology on which all the end-to-end networksshould be based, we must know how it works in order to understand and be able tocarry out adequate measurements in many of the current networks.

3.1 INTRODUCTION

ATM is a connection-oriented packet-switching technology:

• It uses time division multiplexing to transmit multiple-user information acrossthe same link. However, unlike in SDH and PDH technologies, user-specificinformation does not have a fixed position inside a repetitive framed structure(synchronous TDM), but is transferred in blocks that do not have the same po-sition inside a repetitive structure (asynchronous TDM) (see Figure 3.1).

• The blocks into which the data of each user is grouped are called packets, andthey have a defined length and structure. Besides user information, packetscontain an overhead, which is a set of fields for control functions. One of thesecontrol fields is a label that enables the network to route packets through nodesfrom their source to their destination. In a packet-switching technology, usersjust transfer blocks when they wish to send data.

• Packet-switching networks may be connectionless or connection oriented, de-pending on the way in which the network routes packets in order to deliverthem to their destination.

In connection-oriented networks, a route is established between a source and a des-tination before any packets are transferred. In other words, a sequence of links andnodes is chosen to transmit packets between two users. This route is what we call a


connection. To establish a connection, the originating user sends a special requestpacket to the node to which the connection is to be made. Network nodes decide thebest possible path to transfer the data in question and send a connection request tothe destination. If the destination accepts this request, the connection is establishedand data transfer may begin.

As packets from different connections are multiplexed in the same link, eachpacket must contain an identification label. This label enables the network node toidentify the connection to which each packet belongs, and in this way to address thepacket to a certain output link. The receiving node may also change the value of thelabel that the following node uses to route the packet. For this reason the value of alabel only has a local significance in each link. A connection may be identified ineach link by a different label value. The label value identifying a connection in eachlink is reserved, even if packets were not transported through this connection, as longas none of the stations ends the connection by sending a connection release controlpacket. This is why these connections are called virtual circuits: Certain label valuesare reserved and there is an established route, but if data is not really transmitted,bandwidth resources are not taken up.

ATM technology simultaneously offers the advantages of both a circuit switch-ing and a packet switching network, and carries a wide range of services with certainrate, synchronization, and latency requirements. The idea is to use the best of bothworlds: the predictability of a circuit network, and the flexibility of a packet network(see Figure 3.2).

Synchronous:

Asynchronous:

Figure 3.1 In synchronous time division multiplexing each user is identified according to the position occupied inside a periodic framed structure. In asynchronous time divi-sion multiplexing user data is grouped in blocks including a header and a trailer (the latter might be absent), without a fixed position inside a repetitive structure.

Sync

1 2 n. . .

Sync

1 2 n. . .

Sync

1 2 n. . .channel

Sync

1 2 n. . .

Sync

2 n2 2 1 1channel

ATM Architectures 101

3.2 BASIC PRINCIPLES OF ATM

3.2.1 ATM Cell Format

ATM packets are called cells and they are of fixed size, 53 bytes. Their small sizeguarantees a minimum delay in filling them, therby minimizing one of the factorsthat causes transmission delays. These delays must be limited to transport such iso-chronous information as voice and video. For example, for 128 byte cells, the fill-ing time may be up to 85 ms, an unacceptable value.1 However, small packet size isa disadvantage in data transmission, as it also decreases the relation between theuser data transmitted and the control or overhead bytes introduced. That is, it low-ers the efficiency of transmission, as a larger proportion of the available bandwidthis used to transfer control information.

The 53 bytes of a cell are divided into two main fields (see Figure 3.3):

1. The header, formed by five bytes. The structure of a header depends onwhether it is transmitted by an interface between a user and a network, user-to-network interface (UNI), or by an interface between two network nodes, net-work-to-network interface (NNI). A header has three basic functions. It trans-ports the label that identifies the cell (channel), the information concerning thetype of data transmitted, and a code to detect errors in the header (error-detec-

1. Voice can be transmitted at a variable rate, between 12 and 24 Kbps, if it is coded andcompressed. At worst, 12 Kbps, filling a 128-byte packet means a maximum delay of128*8 / 12,000 = 85.3 ms.

Figure 3.2 ATM offers the advantages of both circuit switching and packet switching.

CircuitSwitchingReal Time

TransparentLow Latency

Simple, Predictable

ATM

Real Time, TransparentStatistical Multiplexing

Shared ResourcesFlexible, Efficient

PacketSwitching

Statistical MultiplexingShared Resources

Variable ThroughputEfficient, Flexible


tion code). A header is formed by the following fields:

• Generic flow control (GFC): This is only found in a UNI cell structure, and itwas originally intended for implementing a flow control algorithm between theuser and the network. Nowadays, it is not used, and therefore it is coded to 0.

• Virtual path identifier and virtual channel identifier (VPI/VCI): These areidentifying labels of the virtual channel across which cells are transmitted.They are the two parts into which an identifying label is divided in ATM tech-nology. A VPI varies in size, depending on whether the cell is of UNI or NNIformat.

• Payload type identifier (PTI): This field indicates the type of information trans-mitted in the cell payload. It is formed by three bits. If the first bit has 0 as itsvalue, this means that the cell in question is transporting information generatedby a user, or by a layer that is higher than the ATM layer in a piece of terminalequipment. Conversely, if the value of the first bit is 1, this means that the celltransports control information, be it related to network management and main-tenance or to resource reserving functions. If the first bit indicates that the cellis transporting user information, the second one tells us if the cell has passed bynodes with congestion; in other words, telling us if the cell has experiencedcongestion. The third bit in its turn allows for transporting indications betweentwo users, and it is called the ATM-user-to-ATM-user (AAU) bit.

• Cell loss priority (CLP): This bit indicates if the cell is of high or low priority.It is processed through ATM switches, and, in case these suffer from conges-

Figure 3.3 Cell format according to the ATM Forum: (a) in a UNI; (b) in an NNI.

123456···

53

GFC VPIVPI

VCI

PT CLHEC

Payload

VPIVPI

VCI

PT CLHEC

Payload

Header(5 bytes)

Data(48 bytes)

UNI Format NNI Format

byte

GFC: Generic Flow ControlVPI: Virtual Path IdentifierVCI: Virtual Channel Identifier

PT: Payload TypeCL: Cell Loss PriorityHEC: Header Error Control


tion and cells have to be discarded, the low-priority cells will be discarded first(CLP=1).

• Header error control (HEC): the header of the cell is protected by a sequencecode capable of single-bit error correction and multiple-bit error detection.

2. The payload with 48 bytes. This transports user data and control informationon protocols operating in an ATM network.

3.2.2 Virtual Channels and Virtual Paths

In an ATM link, there is always a continuous cell flow. Cells are transported emptyby default. An empty or idle cell can be distinguished by the following values in itsheader fields: VPI/VCI=0 and PTI=001. When a cell containing information isavailable, it replaces the next idle cell to be transmitted (see Figure 3.4). The cellthat contains information will have a VPI/VCI label value defining the connectionthrough which it will be transported.

Since a label can be divided into two fields, we can look at a link as a set of hi-erarchical connections (see Figure 3.5):

Cell generation is asynchronous

Cell tansmission is synchronous

Figure 3.4 Asynchronous generation of cells containing information. Substitution of idle user cells in a continuous physical-level cell flow.

Tcell = (53 bytes x 8 bits) / vt

Tcell

VPs VCs

32

446 33

0 516

13251

115

VPs VCs

Cell Stream

32

44633

0516

13251115

Figure 3.5 Hierarchical representation of VP and VC, although the link actually transports a cell stream.

Optical fiber,Coaxial Cable,or Twisted pair


• Virtual channel (VC) is the capacity to transfer user information between thetwo ends of a link, and distinguish this information from that of other users.Transmission here is one way. An ATM virtual channel is a pair of definedVPI/VCI values.

• Virtual path (VP) is a set of virtual channels of a link sharing the same VPIidentifier.

The connection in a link is one way. To carry out two-way transmission, or full-du-plex transmission between two points, two links are needed, one for each direction.

3.2.3 Basic Principles of ATM Switching

Operating mode

A network node that routes information, or an ATM switch, has several inputs andoutputs. When a switch receives cells through a certain input port, K, it processesthe header and reads the VPI/VCI label value; for example, i, j. Knowing thesethree values (K, i, j), it consults its switching table and chooses an output port, L,through which it will retransmit the cell. It also gives new values for the VPI/VCIlabel; for example, m, n. This means that an ATM switch routes the cells receivedthrough a virtual channel of a certain input link toward another virtual channel of acertain physical output (see Figure 3.6).

EhDg

Cfe

ac

Ab

Bd

n-1

n

Port 1In

Figure 3.6 Operation of a VC-type ATM switch.

input

output

Port VPI VCIK i jL m n

Out

In

Out

In

Out

h

g

f

e

a

dc

b


Switching table

A set of mapping entities forms the switching table of an ATM switch. These tablesare modified when establishing and ending a connection between two ends. Whenestablishing a new connection, each switch along the path where cells are transmit-ted will add a new mapping to its table. When the connection is released, each ofthese switches will delete the mapping entry added.

Usually, when a connection between two ends of an ATM network is estab-lished, it allows for bidirectional (two-way) communication. In an ATM network,bidirectional communication means that the path chosen for both directions is pro-cessed by the same switches and links. In other words, the mapping entry in theswitching table is valid for both directions. So, in the example given above, if aswitch receives a cell at its input port L with VPI/VCI = m, n, this will be addressedtoward the output port K with VPI/VCI = i, j.

VC and VP switches

There are two types of ATM switches, VC and VP:

• VC switches take the complete VCI/VPI value into account. This value is foundin cell headers, and is there to make a routing decision. So, VCs switch eachvirtual channel individually.

• VP switches only take the VPI field value into account (besides the obvious,the input port) to be able to perform switching. Consequently, all the cells re-ceived via an input port with a certain VPI value will be routed toward a de-fined output port and output VP, independently of the VCI value. In this case,the switch only modifies the VPI value and will not change the VCI of any cell(see Figure 3.7).

For this reason, in a VP switch, the whole set of channels belonging to the same VPis switched in the same way. We therefore see that the switching tables here aresmaller, since there is no need for a different input for each VC. Switching becomesmore rapid when table size is reduced. Furthermore, grouping the VCs also makesit easier to manage all the connections, resources, and switches.


3.3 ATM NETWORK ARCHITECTURE

3.3.1 Introduction

A communications network has the goal of transporting information between theend users connected to it. The series of functions carried out by an ATM network toreach this goal can be grouped, like in any communications network, into differentlayers. These layers form the ATM network architecture, and they are the follow-ing:

• ATM adaptation layer (AAL);

• ATM layer;

• Physical layer.

We can also distinguish between different planes in a network architecture. In theuser plane, layers group the functions necessary to transfer information betweenthe two ends of a network. In the control plane, layers group the functions to makesure the network is working correctly. For example, all the protocols and functionsneeded to establish and clear connections form a part of the control plane. There isalso what is called the management plane that enables us to access configurationparameters of different functions and obtain information on their state.

The different NEs that can be found in an ATM network may partially imple-ment all the different layers and planes, depending on their goal inside the transmis-sion chain. The picture (see Figure 3.8) shows all the different layers and planesimplemented in terminal equipment and ATM switches.

Figure 3.7 Operation of a VP-type ATM switch.

Cd

ac

AbBd

Port 1

input

output

Port VPI VCIK y -L z -

Habc

n-1

In

Out

nIn

Out

In

Out

A

B

C

H


3.3.2 AAL Layer

In the user plane, the AAL level in the source node receives user information to re-transmit it across the network. At the opposite end, it is responsible for deliveringthe data received to the end-user or application. This means that this layer is incharge of the ATM network’s relations with the outside world. It accepts all typesof information with different transmission needs, and allows for its transmissionacross the network. So, in the user plane, the AAL layer is only implemented in theend equipment providing access to the ATM network.

AAL layer structure

The AAL layer can be divided into two (see Figure 3.9):

Management

User

ATM Layer

Physical Layer

Control

SAALLayer

SignalingProtocol

ManagementUser

AAL

ATM Layer

Physical Layer

Control

SAALLayer Layer

HigherLayers

(Q.2931)

SignalingProtocol

VC SwitchTE VC Switch TE

Management

User

ATM Layer

Physical Layer

Control

SAALLayer

SignalingProtocol

ManagementUser

AAL

ATM Layer

Physical Layer

Control

SAALLayer Layer

HigherLayers

(Q.2931)

SignalingProtocol

Figure 3.8 Layers and planes implemented in terminal equipment (TE) and switches.

Management Plane

SAR Sublayer

ATM Layer

Physical Layer

User Plane

Convergence Sublayer

Figure 3.9 AAL sublayers.


• Convergence sublayer (CS): This is the outermost layer, and it groups multi-plexion functions of multiple users or applications, so that they can all use onesingle ATM connection, with error detection and correction functions, flowcontrol functions, as well as functions for synchronization information betweenterminal equipment.

• Segmentation and reassemble sublayer (SAR): At the transmission end, thislayer is in charge of segmenting the blocks generated by the CS layer intofixed-size blocks of 48 bytes, which will be delivered to the ATM layer wherecell headers will be added. At the reception end, this layer carries out the oppo-site process: It receives the 48 bytes delivered by the ATM layer, and reassem-bles them to obtain the original block and deliver it to the convergence layer.At reception, this sublayer is also in charge of detecting errored and lost cells.

AAL layer types

There are different types of AAL layers. Each type implements specific functionsand is intended to satisfy the transmission needs of certain types of information ortraffic. These types are the following (see Figure 3.10):

• AAL Type 1: This type was intended to meet the transmission requirements of aconstant data flow, with synchronization between the two ends. It provides fora circuit-emulation service.

• AAL Type 2: This type has been redefined. Nowadays, it provides for a multi-plexion service for low-rate connections in a single ATM connection, with avariable rate.

Figure 3.10 The adaptation level adapts all traffic to the rate of the source, segments/reassem-bles information into pieces of 48 bytes, detects errored and lost cells, and main-tains synchronization between users during the connection.

AAL 1

AAL 2

AAL 3/4

AAL 5

Circuit Emulation

Multiplexing of

Data

Low Overhead

48 bytes

To ATM Layer

Sources

Low-Rate Channels

Voice and Video


• AAL Type 3/4: This type is not used in practice. It was designed to transmit in-formation at a variable rate and without any synchronization needs. It wasmeant to support interconnection between some data transmission networks,and it is no longer used because it inserts a high level of overhead and is there-fore not very efficient in transmission.

• AAL Type 5: This type is a further development of AAL3/4, intended to supportpacket transmission services at variable rates. It is simpler than AAL3/4 and in-serts less overhead, which makes it a lot more efficient in data transmission.

The different types of AAL layers may be implemented in a single piece of terminalequipment to offer multiple transmission services. The type of AAL layer to use de-pends on the transmission needs of the information in question. Each AAL layerhas a defined access point to the service called service access point (SAP). Thispoint is used by the application that generates the information, to deliver it to anAAL layer, and, at reception by the AAL layer, to deliver the transmitted informa-tion to the end application.

3.3.3 ATM Layer

This layer is the core of transmission and therefore gives its name to the wholetechnology. It is in charge of manipulating cells. It is implemented both in terminalnetwork access equipment and in the intermediate nodes. However, the functions itcarries out in these elements are different, and they will be described as follows.

3.3.3.1 ATM layer in terminal equipment

In the source terminal equipment, ATM is in charge of forming ATM cell headers,whereas in the destination terminal equipment, it extracts and processes ATM cellheaders. The ATM layer in terminal nodes also maintains the identifiers of the es-tablished connections.

48-byte Slots

Transmission

Header Addition AAL Clients

Reception

Header Drop

Figure 3.11 ATM layer in originating and destination terminal equipment.

Physical Layer

and Delively48-byte Slots

ATM Source ATM Sink


Procedures in the originating terminal node

At the originating point, there may be multiple applications or users transmitting in-formation to multiple terminal destination nodes. This means that multiple ATMconnections must be used; each of them associated with a single AAL layer entitythat works as a bridge between the user and the ATM layer (see Figure 3.11).

Therefore, at the originating point, the ATM layer receives the information inthe form of 48-byte blocks from multiple AAL entities. Every AAL entity exchangesthe 48-byte blocks with the ATM layer through a specific service access point chosenwhen the ATM connection was established. The ATM layer adds a 5-byte header toall received 48-byte blocks and obtains the ATM cells. The use of specific SAPs forevery AAL entity enables the ATM layer to set the appropriate VPI/VCI value iden-tifying the virtual channel used to send the cells for all blocks generated by the sameAAL entity. Finally, the ATM layer multiplexes all the cells originating from differ-ent AAL entities in a single flow that is delivered to the physical layer (see Figure3.12).

CS-PDU

SAR-PDU

ATM-PDUATM-PDU

ATM-PDUATM-PDU

AAL-SAP

PHY-SAP

AAL

ATM

Physical Layer

ATM-SAP

AAL-SAP

SAR-PDUSAR-PDU

SAR-PDU

ATM-SAP

PacketsContinuous

48 bytes

53 bytes

Convergence

Segmentation& reassembly

ATM cells

n bytes

ATM-PDU

ATM-PDU

ATM-PDU

sublayer

Data

Figure 3.12 Protocol data unit (PDU) exchange between layers.

StreamData


Procedures in the destination terminal node

At the destination point, the ATM layer receives a cell flow from the physical layerand extracts the header and processes it. If header field values VPI/VCI and PTIidentify the cell as one that includes user information, the 48-byte data block willbe delivered to the AAL layer. Depending on the connection via which the cell hasbeen received, that is, according to the VPI/VCI value received, a certain SAP isused to deliver the information block to the specific AAL entity associated with theconnection that has carried the data.

3.3.3.2 ATM layer in intermediate nodes, ATM switches

When the ATM layer residing in a switch receives a data flow from the physicallayer through a certain input port, it should first decide if the cells in question carryuser information or whether they include control information, and are therefore as-sociated with the control plane.

In the user plane, ATM switches are in charge of switching and routing inputport cells from a virtual channel or path toward a virtual channel or path in an outputport, as we have seen before.

In the control plane, these switches implement the higher layers where controlfunctions reside; for example, signaling protocols that are in charge of establishingand releasing connections. In this plane, the ATM layer receives control cells anddelivers them by means of a certain SAP to the AAL layer entity that provides accessto the control function that is in charge of processing the received information.

In intermediate nodes, distinguishing between cells belonging to the controlplane and the user plane is carried out by VPI/VCI identifiers. Control plane cellsare transmitted across specific connections, and they therefore use reserved VPI/VCI values (see Table 3.1). For example, signaling information uses reserved virtualchannels, and it is therefore called out-of-band signaling. In a link between a termi-nal node and the first ATM switch of the network, the UNI interface, the signalingprotocols defined by the ATM Forum use the reserved virtual channel VPI=0/VCI=5. Usually, any virtual connection with VCI<32 is reserved for control func-tions and is not available for user data transmission (see Table 3.1).

3.3.3.3 Virtual channel and virtual path connections

As we have seen, VPI/VCI values have only a local significance, as they are modi-fied every time they cross a switch.

• Virtual channel connection (VCC) and VC links: In an ATM network, a VCC isthe concatenation of the VCs used between two VC-type switches to transport


cells between the two extremes where the AAL layer is implemented. Withinthis connection, each span between two VC switches that change the VCI fieldvalue is called a VC link (see Figure 3.13).

• Virtual path connection (VPC) and VP links: On the other hand, in a certain VClink, it might be that cells are processed by various VP-type switches, and thisis why the VPI field value is changing. A VPC is the concatenation of the VPsused between VP switches to transport cells between two nodes where the VCIfield value is modified; that is, between two VC switching nodes.

• Each span between VP switches that modify the VPI field value is called a VPlink (see Figure 3.13).

ATM connections can also be divided into permanent virtual connections (PVCs)and switched virtual connections (SVCs):

• PVCs: When the operator chooses the switches and links that will support theconnection, as well as the identifiers and a manual programming of routing ta-bles for switches, we obtain a permanent connection. The programming is doneby means of management applications that enable us to configure and programswitches. Nowadays this process can be completely automatic, and the opera-tor only needs to define the two ends of the network between which the con-nection is to be established. The network is in charge of choosing the path, aswell as the intermediary channel and virtual path identifiers to use. So, we cansay that a PVC is characterized by being established according to an order giv-en by the operator by means of a management application.

• SVCs: We can talk about an SVC environment when connections are estab-lished and released dynamically by users of an ATM network. To establish andrelease a connection, signaling protocols must be used.

Table 3.1 Reserved VPI/VCI values.

Function VPI Value VCI Value PTI Value CLP Value

Metasignaling cell (UNI) VPI = y VCI = 1 PTI = z -Segment OAM F4 flow cell (VPC) VPI = y VCI = 3 PTI = z -End-to-end OAM F4 flow cell (VPC) VPI = y VCI = 4 PTI = z -Point-to-point signaling VPI = y VCI = 5 PTI = z -User signaling with local switch (UNI) VPI = 0 VCI = 5 PTI = z -VP resource management VPI = y VCI = 6 PTI = 6 -Interim local management interface VPI = 0 VCI = 16 PTI = z -Routing control (PNNI protocol) VPI = 0 VCI = 18 PTI = 0 -Unassigned cells VPI = 0 VCI = 0 PTI = xxx CLP = 0Idle cells VPI = 0 VCI = 0 PTI = 0 CLP = 1


3.3.4 Physical Layer

This is the layer in charge of controlling physical signals, be they optical or electri-cal, and adapting them to the transmission medium and to the coding used.

3.3.4.1 Physical layer structure

The physical layer is divided into two sublayers:

• The transmission convergence (TC) sublayer: Situated together with the highersublayer, TC is responsible for creating a data stream or a continuous cell flow,be they idle cells or cells with contents, to be transmitted later by the physicalmedium. TC is in charge of calculating the HEC before transmitting any cells.Once the flow is obtained and before it is delivered to the lower sublayer, it ispossible to scramble the flow according to a determined polynomial. This is al-most certain to prevent user actions in trying to transport data sequences thatmay affect the physical transmission layer; for example, reproducing certain bitsequences reserved to align physical layer frames. At the destination point, thephysical layer does the opposite: It descrambles the flow, delimiting the cells init. That is, it identifies the edges of each cell within the bit flow received andchecks whether they are correct or not by means of the HEC field. Once incor-rect cells (those whose header has multiple-bit errors) have been discarded, aswell as empty or idle cells, the rest are delivered to the ATM layer.

• The physical medium (PM) sublayer: This takes care of bit transmission andsignal synchronization.

VCVP switch

NNIUNI NNI

VC

VP

NNI

VPVP VP VP VP

VC VC

VC switch VC switch

UNI

TE TEVirtual channel link

Figure 3.13 Virtual channel connection and virtual path connection.

Virtual path linkVCI=b

VPI=C

VCI=c

VPI=E

VCI=a

VPI=A VPI=B VPI=D

Virtual channel connection (VCC)Virtual path connection (VPC)


3.3.4.2 Transmission medium data stream

Both the ITU-T and the ATM Forum have defined and standardized many types oftransmission media or physical interfaces. They can all be grouped into two catego-ries (see Figure 3.14):

1. Cell based: This is a very common type for local networks. It consists of on-line transmission of an ATM cell sequence across the transmission medium,which may be fiber or cable of many types.

2. Frame based:

• T-carrier/PDH frames: Cells are grouped in a plesiochronous frame that in-cludes maintenance functions (see Chapter 1).

• SDH/SONET frames: In this case, cells are packed as synchronous framesknown as STM/STS, and transported at multiple optical rates of 155.52 Mbps(see Chapter 2). These structures also carry synchronization information, to-gether with the overhead essential for transmission. The benefit of STM framesis that they offer a standardized way to carry out multiplexing in channels aslinks increase or decrease their transmission capacity.

The ITU-T chose SDH/SONET as one of the bases for the B-ISDN, for signaltransmission and multiplexing by means of an optical network.

Figure 3.14 The physical level is in charge of transporting valid cells, adapting cell sequences to the structure and to the rate of the transmission infrastructure used.

Fiber

Coaxial

STP

UTPFrames

Slots

53 bytes of ATM Layer

Wireless


3.4 ATM ADAPTATION LEVEL STRUCTURES

As described before, there are different types of AAL layers to satisfy the transmis-sion needs of different services and traffic types.

3.4.1 AAL1 Format

This type of AAL layer provides the functions essential to emulate the service of-fered by a circuit. It therefore offers a transmission service for those applicationsthat generate information as constant bit flows; for example, uncompressed tele-phony or video signals, and so on. The functions to carry out this type of service arethe following:

• Data transmission at a constant, usually low/medium rate between the sourceand the destination;

• Transfer of timing information between source and destination;

• Structured data transmission between originating and destination users;

• Data loss or nonrecovered and corrected error indications. For example, cellloss can be compensated by injecting artificial segments.

The AAL1 layer receives user information as a continuous flow and groups it intoblocks of 47 bytes. To each byte is added a header byte with the following fields(see Figure 3.15):

• The sequence number (SN): This makes it posible to detect loss of cells. This isdivided into two parts: 1. The convergence sublayer indication (CSI): The first bit that indicates thatthere is information available on the convergence sublayer (CS);2. The sequence count (SC): A sequence number of three bits, incrementing foreach transmitted block. Sequence hops in the destination make it possible to

Figure 3.15 AAL1 structure.

SAR-PDU

SNP: indicates sequence number protection. CRC is a redundant code that protects the previous fields.

48 bytes

CSI CS: indication points out if there is information on the convergence sublayer used by the SRTS clock signal

Par: stands for parity of the first 7 bits. This set is able to correct a single-bit error and detect multiple-bit errors.

3

CSI

SN SNP

PCRCSC3 11

bytesbits


detect lost cells. Its range is from 0 to 7, and, logically, this does not enable usto detect big bursts of lost cells.

• Sequence number protection (SNP): A redundant code that protects the SNfield. This is a two-stage mechanism:1. Counting a 3-bit CRC of the SN field;2. Providing an even-number parity bit over the seven previous bits (CSI, SC,and CRC).

3.4.1.1 Transmission of structured information

The AAL1 layer offers a transmission service for structured data (structured datatransmission or SDT). This service delivers information about the structure of thedata transmitted. This information consists of a field situated in the second byte ofthe even-numbered blocks (even-numbered SCs). The field contains a pointer indi-cating the byte of the current block or the following block of 48 bytes that containsthe beginning of the transmitted data structure. The first bit of this field indicates ifthe field is actually transporting a valid pointer value, and the following seven bitsform the pointer value. Note that these seven bits make it possible to address up to93 positions that are enough to indicate the 46 bytes of data of the current block,plus 47 bytes of data of the next odd-numbered block. This service enables the re-ceiver to rapidly recover the synchronization of the transmitted data, even if therewas a significant burst of lost cells.

3.4.1.2 Transmission of timing information between both ends

Another important service offered by the AAL1 layer is transmission of synchroni-zation data between users. A number of applications need to use the same baseclock. In other words, there is a communal clock for both ends; for example, videotransmission applications based on MPEG-2 transport streams, transmission cir-cuits, and so on. The AAL1 layer offers two methods for transporting synchroniza-tion data, adaptive clock, or synchronous residual time stamp (SRTS), described inthe following.

Adaptive clock method

This is the simpler of the two methods. In this method, what is transported is not re-ally explicit source clock information. The source clock is recovered at the destina-tion, as the average of data received at the destination is an indication of thegenerating frequency at the originating point. This average rate softens the effects ofthe cell delay variation (CDV) that occurs in the ATM network. These effects maydistort the estimated originating clock frequency, as they may easily cause oscilla-tion in the cells received at short intervals; that is, in the cell reception frequency.


To obtain the mean value of cells received per unit of time, and in this way es-timate the clock frequency with which the information must be delivered to the des-tination user, a buffer must be used and its level monitored. The information blocksreceived are stored in a buffer. The buffer level is monitored continuously, and it isused to control the phase-locked loop (PLL) that generates the clock in the destina-tion. This clock marks the reading pace of the buffer, to deliver the data to the des-tination. This is to maintain the buffer level environment at a mean level, so that ifthe level of information increases, it indicates that the generation rate has also in-creased at the originating point. That is why this must be done by the destinationclock. On the other hand, if the level decreases, the clock frequency will decrease aswell, generating information at a lower rate (see Figure 3.16).

By means of AAL1, for instance, it is possible to implement an emulation ser-vice for a point-to-point circuit of 2,048 Kbps. With this method of clock recovery,the 2048 Kbps signal obtained in the destination will comply with the requirementsassociated with the jitter of the signal. However, it is not possible to guarantee thatthe recovered signal will meet the wander requirements explained in recommenda-tions ITU-T Rec. G.823 and G.824 (see Section 10.2.5).

Synchronous residual time stamp

In this method, synchronization information is transmitted between two users. Thismethod can only be used if the same base clock is used at both ends. This commonclock may be provided for example by the SDH network that forms the physicallayer of the ATM network.

At both ends, a clock is used that is derived from the clock provided by the SDHnetwork. At the originating point, this clock is compared to that of the data to betransmitted. The relation between these two clocks is formed by a fixed nominal partand a variable residual part, due to frequency variations in both clocks. It is the value

Figure 3.16 Synchronization by adaptive clock. The destination clock is synchronized accord-ing to the level of the input buffer. It is intended that this level remains half full.

Transport Network

Source

Destination Clock

Phase-Locked Loop or PLL

Destination Station


of this residual part that is transmitted across the network in the AAL1-layer CSIfield. At the destination point, this value will be used to modify the SDH-derivedclock, and in this way obtain the same clock as at the originating point. This clockwill be used to deliver the signal to the destination with the same clock as at the orig-inating point (see Figure 3.17).

3.4.2 AAL2 Format

AAL2 offers a multiplexing and transmission service for many low-rate channelsby means of one single ATM connection. When multiple low-rate flows are sup-ported for applications that are sensitive to delays, we can maximize the efficiencyof transmission (see Figure 3.18).

Figure 3.17 Synchronous residual time stamp (SRTS) synchronization. The difference is transported as information.

Network Clock

Differencebetween thetwo clocks

Differencebetween thetwo clocks

Source

DestinationClock-

+Clock

SSCSSSCSSSCS

CID = zCID = x

CID = y

SSCS

CPS

AAL2

CPS

Figure 3.18 Multiplexing many low-rate channels (CIDs) in a single ATM connection.

AAL-SAP

AAL-SAP


The various low-rate AAL channels generate information blocks at differentrates. These blocks, also of different sizes, as generated, are situated side by side inthe 47 bytes available in the payload of the ATM cell. If a part of a block does notfit into the same cell, it is moved to the payload of the next cell to be transmitted. Toeach user-information block of an AAL2 channel (referred to as common part sub-layer service data unit or CPS-SDU) is added a header of three bytes; in this wayobtaining the packet to be multiplexed (referred to as common part sublayer proto-col data unit or CPS-PDU). Besides others, a header has the following fields (seeFigure 3.19):

• Channel identifier (CID): This value makes it possible to identify the channelto which each packet belongs. This field may take a value between 1 and 255channels, with value 0 not being allowed. Values from 1 to 32 are reserved forAAL2 channels with control information, and the rest, from 32 to 255, areavailable for user information.

• Length Indicator (LI): This is a field that indicates the length of the CPS-SDUpacket. This field enables us to know where the current packet ends and thenext one begins inside the cell payload.

CPS: Packet Header (3 bytes)

STF: Start Field (1 byte)

CID: Channel Identifier (0 not allowed)LI: Indicates the length of the CPS-SDUUUI: Indications between end users (User-to-User Indication)

OSF Indicates the offset where the next CPS packet startsSN: Sequence number (varies between 0 and 1)P: Parity Bit

Padding the payload of an ATM cell with zerosif no channel has information available to send

CPS-SDU

Channel 1 Channel 2 Channel 3 Channel 4

Figure 3.19 AAL2 structure and fields.

CID

CPS-SDU CPS-SDU

ATM Cell Payloadheader

ATM Cell Payloadheader

CID=x CID=y CID=x CID=xm, n, o, p, ...

m+3, n+3,....

48

53

Size (bytes)

LI UUI HEC0 7 13 18 23 HEC: Header Control Error

CPS-SDU

OSF SN0

Pbits 6 7

CPS-PDU CPS-PDU CPS-PDUCPS-PDU

header

ATM Cell Payloadheaderheader

bits


CPS-PDU packets are situated side by side inside the payload, which makes it easi-er to identify the boundaries of each packet. On the other hand, the first byte of thepayload of each ATM cell is called the start field (STF), and, besides other fields, itcontains control information:

• OSF: This indicates the offset in the bytes (0-47) where the first user packetstarts (CPS-PDU) inside the payload.

• Padding: If, after inserting the last packet generated by the AAL2 channels,there is momentarily no more information available for transmission, the ATMcell will be completed with 0s. If an ATM cell doesn’t contain the beginning ofany CPS-PDU and it only carries the end of a CPS-PDU and padding, then theOSF field will tag the place where padding begins, provided that no CPS-PDUpacket starts in this cell.

The AAL2 layer is used by real-time services that generate information at lowrates. If several channels were not multiplexed, to minimize the cell filling delay wewould need to send partially empty cells. But multiplexing enables us to take moreadvantage of the available bandwidth, which, naturally, makes transmission moreefficient. Furthermore, to make transmission even more efficient, this layer can usea Service Category as variable bit rate (VBR) (see Section 3.5.3).

The characteristics described above, that is, reducing the cell filling time and ef-ficient use of the bandwidth, together with the general ATM features, such as thequality of service to support real-time applications like voice, make AAL2 the layerthat is used in 3G mobile networks to support voice and data applications simulta-neously.

3.4.3 AAL3/4 Format

First, there were two separate layers, AAL3 and AAL4, but they were later com-bined to AAL3/4.

The AAL3/4 format supports nonsecure transmission of user data frames. Onefunction of internal multiplexing is to make it possible to establish many AAL3/4-type user connections simultaneously on an ATM connection. In each of these con-nections, the integrity of the data sequence is maintained, and transmission errors aredetected.

This layer was designed to offer the functions needed for data transmission.However, its complex operation and the high level of overhead it presents make itnot too efficient. As a consequence, it never became very popular and nowadays itis seldom used.


3.4.4 AAL5 Format

The AAL5 layer also supports nonsecure transmission of user data frames. It main-tains the integrity of a data sequence and detects transmission errors. A typical ap-plication is IP data support.

The AAL5 layer makes it possible to transfer user data units of a variable length,up to 65,535 bytes. To obtain the highest possible efficiency, this layer inserts min-imum overhead. The CS sublayer is divided into two more sublayers:

• The service-specific convergence sublayer (SSCS): This layer does not usuallyexist.

• The common-part convergence sublayer (CPCS): This layer is the only one in-troducing overhead. It adds a trailer of eight bytes with many fields. Amongthese are a field that indicates the length of the user data unit, and a redundantcode that detects errors. This layer also adds a padding of variable length (0-47 bytes) to the user data unit, so that the total resulting packet (CPCS-PDU)would be a multiple of 48 bytes (see Figure 3.20).

By padding, the CPCS-PDU packet can be divided by the SAR sublayer into an in-teger number of 48-byte blocks that will be delivered to the ATM layer. The SARlayer does not insert any overhead. This layer only uses one field of the ATM head-er to transport information. The third bit of the PTI field (see Section 3.2.1) has avalue of 1, if the cell transports the 48 bytes corresponding to the end part of theCPCS-PDU packet, and it takes the value of 0 if it transports the 48 bytes corre-sponding to any other part of the packet.

Figure 3.20 AAL5 service format.

PAD: Complements the PDU up to forming a multiple integer of 48 bytes.UU: allows for transparent data transmission between AAL5 extremes.CPI: Aligns the PDU in 64 bitsLI: Data sizeCRC: G(x) = x32+ x26+ x23+ x22+ x16+ x12+ x11+ x10+ x8 +x7+ x5+ x4 + x2+ x + 1 (Calculated for the entire PDU)

User data CPI LI CRC

1 to 65,535 4

ATM header

CS-PDU:

SAR-PDU:

PAD

BOM EOM

UU

1 2 0-47

. . .

COMPTI=xx0 PTI=xx0 PTI= xx1

PT payloadpayload

1

48

bytes

. . .

PT payloadPT


3.5 QUALITY OF SERVICE

The capacity to guarantee a certain quality of service for each application is proba-bly the most remarkable characteristic of the ATM technology. To be able to meetthe application-specific requirements, the network implements a series of mecha-nisms aimed at traffic control and administration of network resources.

As ATM is a connection-oriented technology, during the process of establishinga connection, switches can reserve resources for traffic associated with this newcommunication in advance. When reserving resources for users, the network canguarantee a certain data transmission quality, QoS.

So that ATM switches can reserve enough resources to satisfy the needs of anew connection, it is essential for the user to specify in the connection set-up mes-sage the kind of traffic to be generated and the kind of QoS required. When theswitches know the traffic and the quality of service, they are able to reserve enoughresources without taking too much, and in this way they can offer service for moreusers and work in a more efficient way. For this, the organizations that have stan-dardized the ATM technology, ITU-T and the ATM Forum, have defined the param-eters needed for a user to determine the traffic generated and the QoS required. Thevalue of these parameters will be provided by the user in the connection set-up mes-sage, and this information will be used by the network to reserve enough resources.

3.5.1 Traffic Characterization Parameters

The following traffic characterization parameters are defined in version 4.1 of thespecification for traffic administration for ATM networks by the ATM Forum.These parameters are used when establishing a connection by using the UNI signal-ing protocol of the forum to describe the traffic that will be transmitted by thesource.

• Peak cell rate (PCR): The maximum rate at which cells can be transmitted in acertain connection.

• Sustainable cell rate (SCR): The mean transmission rate of a connection, ex-pressed in cells per second or bps.

• Maximum burst size (MBS): The maximum number of cells that a burst trans-mitted at a rate specified by PCR may contain.

• Minimum cell rate (MCR): The minimum transmission rate guaranteed by thenetwork.

These parameters are used to describe the traffic transmitted by a source. However,not all of them need to be used to describe it. This set of parameters provided by theuser at connection set-up to characterize the traffic transmitted is called a source


descriptor. The set used depends on the category of service, which will be dis-cussed later in this chapter.

However, another parameter, known as cell delay variation tolerance (CDVT),is also used when establishing a connection. Usually this parameter is not providedby the user, but defined by the network itself.

• Cell delay variation tolerance: Unlike other traffic parameters, this is not a pa-rameter that characterizes the source traffic and forms a part of its descriptor.This parameter represents the randomness of incoming cells at a point, in re-spect to a theoretical time. The ATM layer functions, such as multiplexing ofcells of different users and injection of specific cells for control functions, maychange the traffic characteristics of an ATM connection, as they bring in a ran-dom delay between the moment of generation of a cell and the moment of itstransmission across the link (see Figure 3.21). Cells of a certain ATM connec-tion may be delayed, while cells of another connection are transmitted, or ATMcontrol cells or physical signal overheads are inserted. For this reason, the timebetween consecutive cells has a random component in respect to the referencetime, T (the inverse of PCR). The maximum value for this variation is indicat-ed by the CDVT parameter. The specification of this parameter during the sig-naling process cannot be found, for example, in the ATM Forum, and it isdetermined by the network itself.

The set formed by the source traffic descriptor and the CDVT value applied to theaccess interface form the so-called traffic descriptor of a connection.

a0 a1

b0 b1 b2 b3

c0 c1

a0 b0 c0 b1 a1 b2 c1 b3

Time of arrival of cells transmitted through the same output port

Figure 3.21 Delay variations in flow b due to idle cell injection or insertion of other connec-tions.

Cell streamVariation Variation Variation Variation


3.5.1.1 CLP=0 and CLP=0+1 traffic

In ITU-T specifications Q.2931 and Q.2961, as well as in the UNI 4.0 of the ATMForum for signaling protocols, the traffic of the connection is divided into CLP=0and CLP=0+1 traffic, and you may need to provide a traffic descriptor for bothtypes when establishing the connection. The CLP=0 source traffic is formed by acell flow transmitted with the header value of 0 in the CLP field (see Section 3.2.1),indicating high priority according to the source.

The CLP=0+1 traffic, also known as aggregated traffic, is formed by the wholecell flow transmitted by a source; both those that have the CLP=0 field, and thosethat have the CLP=1 field (low-priority cells that should be discarded by the networkbefore the ones marked as CLP=0, in case of congestion).

3.5.2 Negotiated QoS Parameters

When setting up the connection, the originating user not only supplies the sourcetraffic descriptor, but should also specify the desired QoS, so that the network canreserve enough resources. The parameters that define the QoS required by the user,which are negotiable during the connection set-up, are the following:

• Maximum cell transfer delay (max CTD): The maximum delay that cells mayundergo when transmitted by an ATM network, between their generation at thesource and reception at the destination.

• Peak-to-peak cell delay variation (peak-to-peak CDV): The maximum intervalof variation a delay may undergo. So, if the previous parameter specifies themaximum delay, and if we subtract the value of this parameter, what we obtainis the minimum delay. Cells should experience delays between these two val-ues.

• Cell loss ratio (CLR): The lost cell rate in a connection with respect to the totalnumber of cells transmitted.

These parameters are negotiated by means of a signaling protocol. The user specifi-cation of the independent values for each quality parameter offered by the latestversions of signaling protocols differs from the negotiation procedures used in theUNI 3.1 and recommendation ITU-T Rec. I.351 based on the QoS class. In the firstversions of signaling protocols, each QoS class negotiated implied not only a com-bination of specific quality parameters, but also some defined values for these pa-rameters.

There are other parameters that are not negotiated, nor defined by the user, andthat are used to measure transmission quality or network QoS. These parameters willbe dealt with in Chapter 8.


3.5.3 Service Categories

The ATM Forum has defined five service categories. Each category is intended forapplications with a specific type of traffic, and with certain transmission-relatedquality requirements. This is why each category specifies the parameters used todescribe the traffic, as well as the quality parameters negotiated during connectionset-up. This way, when a connection is established, not only the address of the des-tination equipment is expressed, but also the service category with the correspond-ing traffic and QoS parameters. These parameters must be in line with the requestedservice category. If not, the network may reject the connection.

If network congestion occurs, whether the quality parameters established (espe-cially the CLR) can be met or not depends on the service category used, and if thetraffic offered meets the source descriptor. Such functions as reserving resources fornetwork nodes, routing (switching), and the acceptance protocol for new connec-tions are carried out in a different way, depending on the type of service categoryused.

The service categories defined by the ATM Forum are:

1. Constant bit rate (CBR);2. Real-time variable bit rate (rt -VBR);3. Nonreal-time variable bit rate (nrt -VBR);4. Unspecified bit rate (UBR);5. Available bit rate (ABR);6. Guaranteed frame rate (GFR).

Some of these service categories are equivalent to the ATM transfer capacity de-fined by ITU-T Rec. I.371, although different names are used (see Table 3.2).

Thus the CBR service equals the deterministic bit rate (DBR), and the VBR serviceis called the statistical bit rate (SBR), although for the time being only SBR-nrt is

Table 3.2 Service categories of the ATM Forum and ATM transfer capabilities of the ITU-T.

ATM Forum ITU-T

CBR DBRVBR-rt -VBR-nrt SBRABR ABRUBR -- ABT


defined. The ABR service has its corresponding transfer capacity, also called ABR,but there is no capacity equivalent for the UBR service. However, ITU-T defines atransmission capacity called ATM block transfer (ABT), that does not have a corre-sponding service category.

Each service category has a corresponding conformity definition. This specifiesthe cells transmitted in line with the values of the declared traffic parameters. If auser does not comply with the negotiated traffic parameters, the network will behavedifferently with excess traffic, depending on the type of service used and the statusof the network. Checking the conformity of a connection and the actions carried outwith nonconform traffic are some of the functions included in the usage parametercontrol and network parameter control functions (UPC/NPC) (see Section 3.6.2).

3.5.3.1 CBR service category

The CBR service is used in connections that call for constant bandwidth that isavailable throughout communication. The transmitted traffic is characterized by thepeak cell rate (PCR) parameter. The source may transmit cells constantly at the rateindicated by PCR, or at lower rates, or it may even remain silent.

All the cells transmitted inside the range specified by PCR have the connection-specific QoS guaranteed. This service has some particular quality requirements con-cerning transmission delays, delay variations, and lost cells. This is why the QoS,maxCTD, peak-to-peak CDV, and CLR must be specified.

3.5.3.2 Rt-VBR service category

The rt-VBR service is aimed at those applications that have variable rates and highrequirements concerning transmission delays and the variation of these delays. Inthis case, the traffic transmitted is characterized by PCR, SCR, and MBS parame-ters, as well as the same QoS parameters as the CBR.

3.5.3.3 Nrt-VBR service category

The nrt-VBR is a service for those applications where there is no temporal relationbetween the two ends of the connection, and where traffic is variable in time. Thesame way as in the previous case, the traffic is characterized by PCR, SCR, andMBS parameters. In this case, however, the quality of service does not have any re-quirements concerning response times, but only concerning the CLR.


3.5.3.4 UBR service category

The cell flow transmitted by means of the UBR service does not have any guaran-teed QoS parameters. Cells are not certified to be transmitted correctly, without anyloss, nor is there any assurance of not having delays. In this case, the traffic param-eter used, PCR, does not have any QoS guarantee, and the network uses it to obtaininformation about the maximum bandwidth used. This type of service is offered bythe majority of the current data networks (for example, the Internet, LAN technolo-gies) and it is known as a best effort type of service.

3.5.3.5 ABR service category

In the ABR service, the transmission rate available for the source varies during theconnection time, depending on the network status. This service guarantees a mini-mum transmission bandwidth, although depending on how loaded the network is, itmay provide more bandwidth, up to the maximum indicated by the source. The traf-fic in this service is characterized, therefore, by MCR (the minimum bandwidthguaranteed) and PCR (the maximum bandwidth provided by the network, if the net-work is not too loaded). To implement this service, there must be a flow controlmechanism that monitors the network status and controls the cell transmission rate,depending on the network load (e.g., the explicit rate of the ATM Forum). In casethe source adapts itself to the network rate limits, a minimum cell loss rate is guar-anteed. However, there is no guarantee concerning delays, which means that thisservice is not suitable for real-time applications.

Each service has a corresponding quality-parameter priority as well. This iswhy, in case of network congestion, depending on the service used, the network mayprioritize the maintenance of the quality of certain parameters before others. For ex-ample, in a CBR service, the delay and delay variation values will be maintained,although the cell loss rate may increase, whereas in a nrt-VBR connection the situ-ation is the opposite, and the lost cell rate is maintained at the expense of increasingdelays.

3.5.4 Traffic Contract

When establishing a connection, the user agrees to employ a certain service catego-ry (or transmission capacity, according to the ITU-T), a source traffic descriptor(for both traffic types, CLP=0 and CLP=0+1, whose parameters depend on the cat-egory used), the limits of the QoS parameters in question (or class according to theITU-T), as well as CDVT tolerance. All these parameters constitute the traffic con-tract used by the Connection Admission Control (CAC) and UPC/NPC proceduresin order to operate correctly.


3.5.4.1 The service level agreement in ATM networks

In a contract of this nature, concerning both the user and the service provider, manyparameters must be defined; for example:

• The type of traffic to be sent to the network and the QoS requirements for thistraffic.

• The rates, binary states, or bandwidth required.

So, this contract defines what the client expects from the network, and the docu-ment the service provider will use to price and measure its own traffic (that is tosay, engineer its own network). A service level agreement (SLA) implies that boththe user and the service provider are committed to comply with this agreement; thatis, the user will adapt to it and the operator promises to provide the QoS agreed up-on.

However, for congestion reasons, due to technical problems or bad network en-gineering, the operator might end up in a situation where it cannot provide its clientswith the transmission they need. In this case, it is possible to have a discount or pun-ishment policy for the operator, if the problems are not frequent.

The user in his turn should not generate more traffic than is agreed to, althoughthere might be situations where the user, voluntarily or involuntarily, does so. If auser does not follow the agreement and transmits more cells than the “space” re-served allows, it could happen that the network, when transmitting this data, wouldbe ill treating another user and, in the end, endangering the previous network engi-neering.

For all these reasons, an ATM network has a set of mechanisms to detect whena client sends more traffic than agreed to, and what to do in such circumstances.Once this “fraud” has been detected, be it voluntary or involuntary for the client, theATM cells violating the contract can be tagged or discarded.

To tag a cell means the following: In the header of an ATM cell there is a bitcalled the cell loss priority (CLP), used to establish a first level of priority among theATM cells transmitted. If this value is 0, it means that this is an important cell whileif it is 1, the cell is “not that important.” Therefore, when a client breaks the contractthat has been detected by a series of mechanisms, which we will later discuss in moredetail, the CLP of this client may be put to 1 (that is to say, the cell is tagged). In thisway, the network will continue carrying this cell, but with its CLP changed. If thesecells arrive at a node with congestion problems, they will be among the first cells tobe discarded, just like any other cell with CLP=1.


Where the CLP bit is concerned, the network may operate in two different ways:

• CLP-transparent: The network does not take the CLP bit into account. The aimfor the CLR only refers to aggregate traffic, and therefore the CLR experiencedby cells with CLP=0 or CLP=1 is the same. Tagging does not apply in thiscase; that is to say the traffic conforms or is discarded. For the traffic conformi-ty definition, CLP-transparent applies to such service categories as CBR.1 andVBR.1.

• CLP-significant: The CLR quality objective only applies to the CLP=0 flow,without being specified for the CLP=1 flow, nor for the aggregate CLP=0+1flow. Network tagging is available as an option, and the network may make abest-effort attempt to transmit the CLP=1 flow. To define traffic conformity,CLP-significant applies to VBR.2 and VBR.3 service categories.

3.6 RESOURCE MANAGEMENT

To provide the performance users require, in addition to reserving resources duringthe connection set-up, ATM networks implement a series of procedures to controltraffic and congestion (recommendation ITU-T I.371).

Resource management and traffic control are essential in making an ATM net-work operative and efficient. In line with the ITU-T Rec. I.371, these functions tryto protect the network and its users from traffic violations, and prevent network con-gestion that may prevent the network provider from providing the QoS negotiatedfor existing connections, and limit the QoS offered for new connections.

ATM networks implement the following functions to manage and control the re-sources available (see Figure 3.22):

• Connection admission control (CAC): This procedure is carried out when aconnection is established, to determine if it can be accepted. This is the firstmechanism to protect the network from congestion. The connection will eitherbe accepted or rejected, depending on whether the network has enough re-sources to meet the user requirements without decreasing the performance ofexisting connections. This procedure looks for an acceptable path, that is, onewith switching nodes and links that can support the new connection. In short,the CAC is in charge of routing the call.

• Usage parameter control (UPC): These are functions carried out by the net-work to supervise and control the ATM connections used and user traffic. Theaim, here, is to detect violations of the traffic parameter values negotiated earli-er. The actions to be taken in case of violation depend on network congestionand the type of service used for transmission.


• Network parameter control (NPC): The NPC carries out the same functions asthe UPC, but for links between networks (internetwork interface or INI) in-stead of those between the user and the network (UNI).

• Resource management (RM): Among these functions is the procedure for thenetwork to use special cells to dynamically modify the resources assigned to acertain ATM connection. RM also covers all the design methods and engineer-ing decisions applied to efficiently multiplex and group multiple VCC connec-tions with different types of traffic in VPCs, with the aim of reducing thenumber of resources to manage.

3.6.1 Connection Admission Control

The CAC is responsible for accepting or rejecting a new ATM connection, basingthis decision on the availability of network resources. When calculating the avail-ability of resources, the interests of the existing connections must be taken into ac-count, in such a way that a new connection should never threaten the QoS ofexisting connections.

During the call set-up, an ATM network user must specify all the informationneeded, depending on the requested service category. Service-category-related in-formation, such as traffic parameter values and the QoS provided, are analyzed bythe CAC function, to check if the network has enough resources, and to accept orreject the connection.

Figure 3.22 Location of traffic control functions.

TENT1 NT2

ST

TE NT1 UPCNT2

NPC

S T

INI

Network A

Network B

CAC, RM

CAC, RM

UPC: Usage Parameter ControlNPC: Network Parameter Control

CAC: Connection Admission Control

RM: Resource Management

UNI


When establishing a connection, the user and the network may enter a negotia-tion phase that ends by “signing” a contract whereby the network guarantees a cer-tain QoS for the user. In its new ATM signaling recommendations, ITU-T hasspecified new signaling protocol capabilities that allow for checking the contractduring connection time.

3.6.1.1 Connection process

In an ATM network, the CAC procedure carried out in switches is in charge ofchecking for the resources available and finding an acceptable path. So, the CACprocedures present in each switch make a decision about whether a connection re-quest may proceed, and, if so, toward which nodes. To be able to make this deci-sion, switches must have information about the network status and architecture.

Routing a connection request

Actually, the ATM Forum has specified a protocol called Private network-networkinterface (PNNI) that defines both the routing and the signaling protocol. By meansof a routing protocol, switches exchange information on:

• The network topology, to know the paths to access a certain destination;

• QoS or link-status-related information, to choose the best possible path.

The signaling protocol makes it possible to extend the connection set-up and re-lease signaling used in the UNI to be employed between switching nodes.

However, we must take into account that not all the networks have to implementthis dynamic routing protocol that enables nodes to exchange information on net-work topology and network status. There are other protocols defined by ITU-T withstatic routing, which means that each switch includes programmed information onwhere to address a connection request for each destination; in some cases also takinginto account the QoS information coming from the user.

The first switch chooses the route and sets up the connection. It is in charge ofconnecting with each node chosen to support the connection, and notifying of theconnection set-up. However, it is possible that the status of one of the nodes chosendoes not match the information that the first switch had available, and, in reality, itcannot maintain a connection. This is due to a delay when a routing protocol is com-municating the information in question. In this case, the first switch may decide totry and find an alternative route, or directly reject the new connection. The usershould try to set up the connection again, so that the switch could look for anotherroute inside the network.


Allocating VPI/VCI identifiers

If all the switches accept the connection, the destination user will eventually re-ceive a request for a new connection. If the destination user accepts the call, con-nection is admitted, and it is put into practice in a VC/VP between the originatingand the destination point.

VPI/VCI values are assigned to the opposite direction than the call set-up. Thismeans that the switch farthest away from the originating point indicates to the des-tination user and the preceding switch the VPI/VCI values used to maintain the newconnection. So, successively, from destination to origin, each switch indicates theVPI/VCI values to the switch preceding it on the link that is connecting them. Thisway, each switch can update its switching table with the new entry. When identifiersare allocated this way, we can talk about upstream allocation. In the end, the termi-nal equipment that requested the new connection obtains, from the switch where itis directly connected, the VPI/VCI that must be used to transport cells across the newconnection (see Figure 3.23).

If a user wants more than one connection, which is the case in a multiconfer-ence, for example, the CAC procedures must be carried out individually for eachconnection.

3.6.2 UPC and NPC Policing Functions

The aim is that the source meets the agreement with the CAC, and this is why thisfunction is sometimes called a policing function. Like law enforcement in real life,these functions take care of “law and order” in the sense that they see that users

TETE

VPI/VCI=B/b

VPI/VCI=A/a

VPI/VCI=C/b

VPI/VCI=E/cVPI/VCI=D/bIn VPI/VCI Out VPI/VCI1 A / a 2 B / b

In VPI/VCI Out VPI/VCI1 B / - 3 C / -

In VPI/VCI Out VPI/VCI2 C / - 3 D / -

In VPI/VCI Out VPI/VCI2 D / b 3 E / c

Figure 3.23 Direction of allocating VPI/VCI identifiers.

VPSwitch

VPSwitch

VCSwitch VC

Switch


comply with the contract, since breaking the contract, be it intentionally or uninten-tionally, might endanger the rights of the rest of the users to profit from the agreedQoS.

According to ITU-T Rec. I.371, the first traffic control function is to monitor theconnections established and see that the traffic sent across each of them meets thetraffic parameter values specified during call set-up. The idea is to avoid overload ina connection, as it may produce network congestion that may in its turn endanger theQoS of other users.

The policing function monitors both VCCs and VPCs, and it is carried out at twodifferent interfaces:

1. At the user-network interface, known as usage parameter control. 2. At the internetwork interface, called network parameter control.

It is worth our while to dedicate some words to distinguishing between traffic polic-ing and traffic shaping. The former is used to protect the network, whereas the lat-ter is committed to change the characteristics of a data flow (regarding thewaveform of its bursts, cell arrival times, etc.) so that it meets the agreed parame-ters. Traffic shaping algorithms have been left for later study in international fo-rums. They are intended to be used in ATM user equipment before injecting anycells to the network, which means that they can change a data flow that does notcomply with the contract to another one that does. (For example, an IP router con-nected to an ATM network can cause some short delays in some cells, to reducePCR and jitter without degrading performance.)

Examples of traffic shaping are reducing the peak cell rate, limiting the lengthof burst tolerance (BT), and spacing cells correctly in time. Traffic shaping is an op-tional function, and it may be activated according to the contract with the networkoperator.

3.6.2.1 UPC functionalities

UPC may be considered as a way to control congestion preventively, and it is im-plemented by means of several functions:

1. Checking the VPI/VCI identifier validity and verifying that identifiers notassigned by the network are not actually used.

2. Monitoring the volume of traffic to check that all the parameters conform towhat is agreed. This is done in the terminating point of the first VC or VP link,depending on whether the connection is VCC or VPC. At this point, we shouldhighlight the advantage that VPs offer within the concept of the ATM network:


If there is a user-to-user VP, it is not necessary to monitor each VC individu-ally, but rather just one VP link.

UPCs monitor the data flow throughout the whole connection time, and forcedata sources to not pass the traffic parameters negotiated with the CAC. They mustbe able to distinguish between cell traffic fluctuations and actual violations of the pa-rameters agreed.

3.6.2.2 Traffic control algorithms

There are many techniques to control traffic, most of them based on counting thenumber of cells that cross the control point. However, there are two that are morepopular than any other technique:

1. The window mechanism, which limits the number of cells inside a determinedwindow. This is a simple and efficient algorithm that allows for many methods(see Figure 3.24).

2. The leaky bucket method, which adds a counter for each cell arrival, decreasingit periodically. This is a standardized algorithm and it will be discussed in moredetail later.

Note that neither of these two uses any other type of network information exceptthat of local flow, which is enough to implement methods to prevent congestionacross all the network.

Figure 3.24 Window algorithms to control traffic.

Static Window:

Hop Window:

Continuous Window:

Approved cell

Rejected cell


GCRA algorithm

The generic cell rate algorithm (GCRA) is a standardized traffic control algorithmused by the policing function. This algorithm is described in recommendation ITU-T Rec. I.371, and the ATM Forum has also incorporated it into Traffic Manage-ment 4.1 and earlier. The way in which this algorithm works is very simple. Amongother characteristics, it tries to make a difference between momentary violationsdue to statistical variations, and violations due to generation that does not meet theagreed parameters. GCRA uses two variables, GCRA (I, L) (see Figure 3.25):

• I increment;

• L limit.

This algorithm consists of defining a theoretical cell arrival time. Each theoreticalarrival time is calculated according to the previous arrival time, plus the value ofone I variable. When a cell is received, its arrival time is compared to the theoreti-cal arrival time. If the cell arrives later, it is a conforming cell, and if it arrives be-fore the theoretical time but within an interval specified by L, it is also aconforming cell. However, if it arrives before this tolerance interval, it is a noncon-forming cell.

This L limit is a parameter that enables us to distinguish between statistical vari-ations and violations produced due to traffic generation that is higher than specified.To monitor traffic in a connection and decide whether it is conform or not, the ge-neric values I and L must be identical to the specified traffic values (see Figure 3.26).

Figure 3.25 Generic cell rate algorithm, GCRA (I, L).

Cell k arrives

TAT < ta(k)+L?

TAT < ta(k) ?

at t = ta (k)

Nonconforming CellTAT = ta (k)

TAT = TAT + IConforming Cell

Yes

No

No

Yes

TAT: Theoretical Arrival Timeta (k): Cell Arrival Time

I: IncrementL: Limit


CBR traffic conformity

The PCR parameter is defined in the physical layer of the ATM protocol as the op-posite of the minimum cell arrival time. Therefore, in a theoretical CBR connec-tion, cells would be spaced apart by 1/PCR. This is the value of the parameter Iwhen the GCRA algorithm is applied in a CBR connection.

ATM layer functions may alter traffic characteristics in ATM connections by in-troducing CDV. Here we should remember that when cells of two or more ATMconnections are multiplexed, the cells of one ATM connection may be delayed, dueto cell transmission in the other ATM connection, or even due to OAM cells. CDVmay also occur due to the nature of physical level slots, or due to physical layer head-er transmission (see Figure 3.27).

CDVT is a parameter that describes randomness in the arrival time of two con-secutive cells in a VPC/VCC monitored by GCRA at UNI/INI interfaces. CDVT isthe value of the parameter L when this algorithm is applied to a CBR connection.

Therefore, we may conclude that in a CBR connection, traffic conformity is su-pervised by GCRA (1/PCR, CDVT).

Traffic Conformity to PCR: GCRA (1/PCR, CDVT)

c0 c1 c2 c3

Traffic

Tolerance

T=1/PCR T=1/PCR T=1/PCR

t

Traffic Conformity to SCR: GCRA (1/SCR, BT+CDVT)

c0 c3

Traffic

T=1/SCR T=1/SCR T=1/SCR

t

c0 c0 c0

T=1/SCR

BTBT

Nonconforming cell Nonconforming cell

Figure 3.26 Graphical representation of conformity to PCR and SCR and MBS.

CDVT

Time of arrival of cells

Burst Tolerance (BT): BT = (MBS-1)(1/SCR-1/PCR)


VBR traffic conformity

In a VBR traffic connection, traffic is described by four parameters: PCR, CDVT,SCR, and MBS. Traffic in a VBR connection is conformed by applying the GCRAalgorithm in duplicate.

The first GCRA algorithm supervises conformity to the maximum rate. There-fore, the same GCRA (1/PCR, CDVT) algorithm is applied to supervise that the userdoes not generate a higher rate than the PCR.

The second GCRA algorithm supervises the mean traffic conformity. For this,the theoretical cell arrival time, if cells were transmitted to SCR, would be 1/SCR.However, as even MBS cells may be transmitted to PCR, it turns out that the last cellin a burst will be received quite a while earlier than forecast. This interval is charac-terized by the BT parameter. We can easily deduct the value of this parameter:

BT= (MBS-1)(1/SCR-1/PCR)

This is the tolerance interval value to consider the last cell of a burst as conform.However, if there is an MBS+1 transmitted at 1/PCR, it would already be considerednonconform.

So, the algorithm to supervise cell conformity to the mean rate is given byGCRA (1/SCR, BT).

Figure 3.27 PCR reference model.

Source1

Sourcen

.

.

.

.VirtualShaper

TEFunctions

OtherFunctions UPC

Terminal Model

GCRA: Generic Cell Rate AlgorithmCDV: Cell Delay VarianceUPC: User Parameter Control

ATM Layer

PCR: Peak Cell Rate GCRA (T, 0)

Private UNI Public UNI

GCRA (T, ττττ)

τ*

Generate

SAP

τ, τ*: CDVT of a connection at different interfaces

Physical Layer

GenerateCDV CDV

GCRA (T, ττττ*)

τ

MUX


Granularity of contract parameters

For hardware reasons, it is not possible to ask for UPC/NPC functions that wouldmanage any parameter value. Only a finite and discrete set of PCR, CDV tolerance,SCR, and BT values is discriminated by UPC/NPC functions in the ATM layer of agiven ATM switch.

The increments discriminated in the above-mentioned parameters are called the“grain” or cell rate granularity. They are usually expressed as cells per second.

3.6.3 Other Control Functions

There are other traffic and resource control functions as well. Some of them startworking only if certain service categories are used.

3.6.3.1 Explicit rate

This protocol has been standardized by the ATM Forum for the ABR service cate-gory. By means of this protocol, the network dynamically informs the originatinguser of the rate at which traffic may be generated. This protocol operates in a timescale that is equivalent to the return delay in a connection.

The originating user collates some control cells, called resource managementcells, into the generated user cell flow. In this cell, the originating user requests anincrement in the transmission rate. The cell will be transmitted the same way as datacells across the connection. However, this cell is detected by intermediate nodes, dueto its different PTI field value. When processing this cell, intermediate nodes mayvary the request in line with their congestion status. At the destination point, this cellis returned toward its origin, to find out whether the request has been accepted, re-jected, or modified. ATM switches are also allowed to generate this cell, to informthe originating user of a reduction in its generation rate due to congestion.

3.6.3.2 Explicit forward congestion indication (EFCI)

This procedure enables the network to inform its users about a congestion condi-tion. This mechanism is based on an indication that is transported by the cellsswitched by nodes with congestion. If a user cell is switched by such a node, thenode may change the second bit of the header PTI field from 0 to 1 (see Table 3.3).When the cell arrives at its destination, it receives a congestion indication and mayinform higher layer protocols about this congestion condition, so that they can takeappropriate measures.


3.7 ATM IN ACCESS NETWORKS

ATM has been chosen as the transmission technology in many current solutions, toincrement the capacity of access networks. The ATM technology enables the sameinfrastructure to support the services that are currently more profitable, such asvoice services and leased lines, and at the same time offer high capacity for thoseservices that require more bandwidth, such as Internet access and virtual privatenetworks (VPNs) for office interconnections and remote access to information.

Although there is a continuous debate on the convenience of implementingATM technology in a situation where it looks like most services were based on IP,we can be sure that ATM is found in most access technologies: asymmetric digitalsubscriber line (ADSL), wireless local loop, and the fixed 3G mobile access net-work. The inherent quality of service of ATM makes it the only packet switchingtechnology with stable specifications, and a total compatibility between manufactur-ers, while still being completely functional and allowing for such a variety of servic-es as data, voice, video, and leased lines.

Furthermore, its poor efficiency for data transfer, for which it has often beencriticized, is currently quite relative, as its efficiency for voice transfer is a lot higher.Even if there were more data than voice traffic in the long run, the same argumentsagainst the importance of QoS (due to the greater bandwidth provided by opticaltechnologies) are also valid for reducing the importance of the bandwidth lost byATM due to its poorer efficiency.

3.7.1 ATM as Transport in ADSL

The asymmetric digital subscriber line is one of the transmission technologiesgrouped under the generic name of xDSL that enables us to obtain high transmis-sion rates by means of the copper pair (see Chapter 4).

ADSL defines an asymmetric behavior from the point of view of the bandwidthassigned to each direction of transmission, user-to-network (upstream) or network-

Table 3.3 PTI codification.

PTI Meaning

000 User data. No congestion. SDU=0001 User data. No congestion. SDU=1010 User data. Congestion experienced. SDU=0011 User data. Congestion experienced. SDU=1


to-user (downstream). That makes it suitable for asymmetric services like Internetaccess or video on demand in residential environments.

3.7.1.1 Introduction

ADSL defines a frame with a header, where user data is transmitted. This frame istransmitted to both directions at the rate assigned. According to the ADSL standard,this frame may be divided into a number of channels, which makes it possible to of-fer many communications channels to the user in an ADSL link, by means of timedivision multiplexing, often known as STM service. In any case, in most commer-cial implementations, the frame only has one upstream and one downstream chan-nel, letting the ATM technology take charge of dividing the capacity of the link bymeans of asynchronous multiplexing.

The ADSL link at the customer site (see Section 4.2) is terminated by the ADSLtransceiver unit-remote (ATU-R), a user modem that is one of the ends of the copperpair. In the market, the ATU-R is implemented as internal PC cards, external mo-dems, or as an output port of an interconnection device (router or bridge). The samedevice that implements the ATU-R also implements the ATM layer. A single usermay transmit and receive multiple services by using multiple ATM connections inthe same ADSL link.

The ADSL link at the central office is terminated by the ADSL transceiver unit-central (ATU-C), a network modem forming the other end of a copper pair link. TheDSL access multiplexer (DSLAM) is the network element at the central office thatimplements n x ATU-C in order to offer service to n clients. The ATM connectionsestablished in different ADSL links are multiplexed by the DSLAM in a single ATMlink of a higher rate, and transported by the ATM network to their destinations. TheDSLAM extracts the ATM cells of each ADSL link and multiplexes them in an out-put link, usually in an ATM link at 34, 45, or 155 Mbps. Toward the opposite direc-tion, it demultiplexes the cell flow received across the ATM link toward each ADSLlink (see Figure 3.28).

3.7.1.2 Internet access service

Currently, the main service offered by ADSL is Internet access. This type of serviceis offered by a certain Internet service provider (ISP) that may be a different com-pany than the one offering ADSL service.

In business and professional environments, the client is accustomed to sign acontract with only one ISP, and use a permanent connection. In this case, the ADSLservice provider should only establish an ATM connection between the client andthe service provider. As ATM networks currently only offer permanent connections,


this connection would be a PVC. This type of connection would have an architecturevery similar to the Internet access service, based on the use of a frame relay connec-tion between the client and the ISP, or on the use of a leased circuit between the two.In this case, the only difference is in the use of an ATM-based connection that ispartly supported over a copper pair by means of ADSL (see Figure 3.29). The PVCconnection could use any of the service categories available, depending on the per-formance and the client’s needs, and it should use the AAL5 layer, especially de-signed to transport data.

However, residential users are currently accustomed to having many ISPs avail-able, some of them offering their service free of charge, and the connection is notalways of a permanent nature. In these cases, it is essential for the ISP to continueusing the same characteristics as mechanisms of assigning IP addresses, security atlevel 2, user identification, authorization, and pricing offered by the point-to-pointprotocol (PPP) that was already used in the access by the public switched telephone

Local Loop

ATM

PHY

ADSL

ADSL

DSLAM (ATU-C)

Copper pairs

dC c

aA

bB

ATM Link at34, 45, 155 Mbps

Figure 3.28 Multiplexing ATM in the DSLAM.

ATU-R

NetworkTermination

ATU-R

NetworkTermination

Local Loop

a ABd

Ccd

Copper pairs

User PC ATU-R ATU-C (DSLAM) ISP

Figure 3.29 Structure and protocols for a connection between a company and an ISP.

LLC

10-BT

MAC

LLC

Router NATPAT

10-BT

MAC

IP IP

ATMxDSL SDH802.3 ATM

ATMSDH

AAL5

RFC2684

ATMxDSL

AAL5

RFC2684


network (PSTN) and modem. In short, the Internet access service based on ADSLalso uses the PPP protocol.

On the other hand, the fact that a residential user may want to choose betweenmultiple ISPs, and that ATM networks may only offer PVC services, would make itnecessary to establish a mesh of connection between the user and each ISP. Thiswould imply that as the n number of users grows, with the number of ISPs being m,the number of connections would increase by n*m. For this reason, this solutionwould not be very scalable, due to the high number of connections needed and theirmaintenance. The solution adopted in most cases in commercial environments is im-plementing an intermediate device, a gateway, that makes it possible to group to-gether all the connections coming from different users to one ISP. In this case, all weneed is to establish a PVC connection between each user, together with this interme-diate device, and a connection between the device and each ISP (see Figure 3.30).The ATM connection between each user and the gateway transports IP packets bymeans of the PPP protocol and the AAL5 layer. When a PPP packet arrives at thegateway device, this can multiplex the PPP packets of different users, addressed tothe same ISP in one single connection. The L2TP protocol is used for this. The con-nection between the gateway and the ISP may use an ATM network or any otherconnection, such as frame relay, a point-to-point circuit, or an IP network.

The set of protocols used for the ADSL-based Internet connection depends onthe type of intermediate or interconnecting device where the ADSL modem is im-plemented (see Figure 3.31). The most common cases are implementing an ADSLmodem in an internal PC card for a residential user, or the use of an interconnectiondevice as a router that allows for all the users in a local Ethernet-based network tohave an Internet connection or a connection with external networks. However, thereare also cases where an ADSL modem is implemented in a bridge-type interconnec-tion device, or in an interconnection device with an ATM at 25-Mbps interface at theuser end, so that many users of a private ATM network can access the Internet.

ISP 1 ISP 2 ISP M

User1

User2

UserN

ISP 1 ISP 2 ISP M

ServiceGateway

Figure 3.30 PVCs between users and ISPs with and without an intermediate device. (a) Estab-lishing and maintaining O(N*M) PVCs is not a very scalable solution; (b) With an intermediate network element only one PVC per user and per ISP is needed. The number of PVCs is reduced to O(N+M).

User1

User2

UserN

(a) (b)


RFC 2684, a new version of the RFC 1483 multiprotocol encapsulation over AAL5

Figure 3.31 Structure and protocols involved in ADSL service:(a) ATU-R as an internal PC modem;(b) ATU-R in an interconnecting device with ATM-25 interface (user);(c) ATU-R in a bridge-type interconnecting device;(d) ATU-R in a router-type interconnecting device.

ATMAAL5

xDSL

RFC2684

PPP

ATM ATMAAL5

RFC2684 L2TP

PPP

L2TP

PPP

Gateway NSP

(Internal modem)

LLC

10-BT

MAC

ATMAAL5

xDSL

BridgeRFC2684

10-BT ATMAAL5

SDH/PDH

RFC2684

EtherPPOE

PPP

L2TP

10-BT

Ether

PPPoEPPPIP

ATMAAL5

xDSL

RFC2684

PPPLLC

RouterNATPAT

10-BT

MAC

IP

L2TP

PPP

IP

(c) Bridge-type

(d) Router-type

ATM FR IP

ATM FR

IP...

ATM, FR, IP,...

IPIP

xDSL SDH/PDH SDH/PDH

ATM, FR, IP

ATMAAL5

RFC2684

PPP

ATM ATMAAL5

RFC2684 L2TP

PPP

L2TP

PPP

User DTE xTU-C (DSLAM)

ATM, FR, IP,...

IPIP

xDSL SDH/PDH SDH/PDH

ATM, FR, IP,...

ATMxDSL

PPP

IP

ATMxDSL SDH/PDH

L2TP

ATM FR IP

ATMAAL5

RFC2684 L2TP

PPP

ATM, FR, IP,...SDH/PDH

ATMxDSL SDH/PDH

ATMF 25.6 ATMF 25.6

xTU-R

802.3

802.3

(a) Internal modem

(b) ATM-25 interface


When ATM networks start to provide services based on switched virtual con-nections, the Internet access architecture will be very similar to the current accessarchitecture based on a switched telephone network and a modem. It is only that, inthe first case, instead of establishing a telephone connection with the ISP, an ATMconnection is established.

3.7.1.3 Other ADSL services

As we have seen, the reason to choose ATM as a transmission technology is its ca-pacity to offer QoS to support voice and video services. However, it is also possibleto offer voice services over ADSL (VoDSL), as well as digital video services(MPEG), and they are already commercially available.

Under the name of ADSL voice service, multiple implementations are scram-bled; all of them transported over ADSL in a copper pair. The options are voice overATM (VTOA), voice over IP (VoIP) and ATM, and voice over IP without usingATM as the transmission technology.

VTOA implies that the same device at the user end that implements a data in-terface such as Ethernet, would also implement voice interfaces. Packed and codedvoice in the device is sent by an AAL2-type ATM connection toward the remote end,an interconnection device with the PSTN network. The AAL2 connection can sup-port multiple low-rate subconnections, and therefore the same ATM connection canbe used to offer many voice lines for the user. Coding and packing voice over ATMuses standard ITU-T defined compression techniques, together with such techniquesas deletion of silences that reduce the capacity needed to transport voice in variabletraffic with a peak rate of around 10 Kbps. Therefore, an ATM connection transport-ing voice may use an rt-VBR service category.

On the other hand, video digital service over ADSL is currently also based onATM. This service consists of transmitting a transport stream (TS) or a programstream (PS) over a CBR-type ATM connection in an ADSL link. Due to the limitedcapacity of an ADSL link, a TS may only transport one video program at a time.Therefore, one of the possible implementations consists of transmitting multipleprograms to the central office. However, then only the program selected by the userthrough the electronic program guide (EPG) system carried by the MPEG stream in-forming about all the programs available, and a return channel over the ADSL link,is actually sent to the user.

3.7.2 Wireless Local Loop

The wireless local loop (WLL) access network is based, as its name implies, on theconnection of the user to the network via air. There are many techniques to imple-


ment this type of access network, for example, the local multipoint distribution sys-tem (LMDS).

3.7.2.1 Network elements

The LMDS technique consists of implementing a base station (BS) network, wherethe BSs distribute the signal toward a group of users and customer premises. This isan architecture very similar to mobile networks, but in this case the end-user staysin a permanent location.

Communication between a BS and users is naturally full duplex. User-to-net-work communication is called upstream, and network-to-user communicationdownstream. The network elements that form this access network are (see Figure3.32):

• The BS: This element is situated at the network side, with a link toward the op-erator’s network.

• Customer premises equipment (CPE): As obvious, this element is situated atthe user end. It implements many types of interfaces toward the user, depend-ing on the services offered.

• A radio interface: This establishes the connection between a BS and users towhom it offers services.

EthernetPBX

POTS

T1/E1

nx64 kbps

Router

Data

Residential Users

Business Users

Network Side (BS)

STM-n/OC-m

POTS Ethernet

Figure 3.32 Network elements.

CPE

CPE

ATM


3.7.2.2 Modulation techniques, multiple access, and multiplexing

The technology used in radio communication between a user and a network is notstandardized, so it depends on each manufacturer.

Quadrature amplitude modulation (QAM) and phase-shift keying (PSK) tech-niques may be used as modulations to transfer information.

The access method to multiplex the information between users and a BS is notdefined, either. There may be various implementations combining time division mul-tiple access (TDMA), frequency division multiple access (FDMA), and time divisionmultiplexing (TDM), downstream and upstream (see Section 1.1.4). Usually the twodirections are differentiated by FDMA. The BS-to-user direction uses one frequencyonly, multiplexing (TDM) the information to all the users.

For the network-to-user direction, FDMA may be used, assigning a frequencyfor each user, although some users may share the same frequency, in which caseTDMA is used between them to send the information to the BS (see Figure 3.33).

3.7.2.3 Radio frequency planning

The same way as mobile networks, the LMDS divides the territory into cells, eachcell being associated with a BS. Cell size depends on the maximum distance that aBS may cover, and on the maximum number of users a BS may offer service to.Due to the latter, cells are usually smaller, as the users are concentrated in the samezone, and they therefore cannot be offered service with only one BS.

BS

User1

User2

UserN

f0 + TDM for each CPE

f1

f2

fn

FDMA

BS

User1

User2

UserN

f0 + TDM for each CPE

f1

f1

fn

FDMA

TDMA

Figure 3.33 Combining FDMA, TDMA, and TDM for communication between BSs and CPEs.


To be able to offer services to a maximum number of users, radio frequency(RF) planning divides the cell into multiple sectors. RF systems in base stations areformed by a set of antennas, each of them with a certain beamwidth, from 15º up to90º, offering service to each sector. Depending on the antenna beamwidth, the cellis divided into a certain number of sectors. To maximize the number of users and thebandwidth with a certain number of frequencies assigned to each operator, frequen-cies are reused. Different frequencies are assigned to contiguous sectors, and thesesame frequencies are used in noncontiguous sectors. To optimize this reuse andavoid interferences in border zones between different cell sectors that are using thesame frequency, broadcast is combined with vertical or horizontal polarization.

3.7.2.4 Architecture

Any of the current services can be offered by means of the wireless local loop:voice, point-to-point leased circuits, Internet access, VPN services, and so on. Aswe have seen, this is the reason why in many implementations ATM has been cho-sen as the transmission and switching technology. The use of ATM makes it possi-ble to distinguish between services, offer QoS, and dynamically assign bandwidthsand priorities, which allows for a flexible assignation of resources according to userneeds, as well as for a quick reconfiguration.

At the user end, the equipment has many types of interfaces available, to be ableto provide different services. As these services are usually addressed to businesses,although they might also be addressed to residential users, the most commonly usedinterfaces are Ethernet data interfaces and synchronous n x 64 Kbps interfaces, T1/E1, T3/E3 for voice services (connection of private branch exchange, PBX) orleased lines (see Figure 3.34).

These services are transported by means of an ATM connection, which is usu-ally a PVC, from the user end to another terminate point in the ATM network. Thisother point can be the ISP for Internet access services, or the leased line remote end,or the interconnection device with the PSTN or circuit network for voice and leasedcircuits services. ATM connections are supported by the air interface up to the BS.At the BS, the ATM connections of different users are multiplexed over a single linkat 34, 45, or 155 Mbps, connected to the ATM network. In the opposite direction,the BS receives a set of connections from the ATM network and distinguishes be-tween each user’s connection, sending it to each user through the air interface.

The AAL layer is chosen based on the type of traffic. AAL5 is used for data ser-vices, while AAL1 is used for circuit emulation services. Moreover, the ATM con-nections transporting circuit emulation services are CBR-type connections, with a


bandwidth defined according to the capacity of the circuit, and with CLR, CTD, andpeak-to-peak CDV parameters guaranteed (see Figure 3.34).

3.7.2.5 Advantages

Easy to install, the wireless local loop allows for a quick network rollout. Installa-tion costs are usually lower than those of cable-based access, as there are high laborcosts when installing the cable.

In addition, the WLL has the benefit of simultaneity when installing or extend-ing user equipment and getting a new subscriber. This means that there is no need toinstall user equipment when there is no client demand for the service. By contrast,in cable networks the network must be implemented first, as this requires a lot oftime, and clients are convinced later. With WLL techniques, the infrastructure is notwasted, because if one user unsubscribes, the capacity (the frequencies) is given toother users and in this way reused.

However, the capacity, and therefore also the services available through radiotransmission, are usually more limited than the capacity of cable networks and, ob-viously, optical fiber networks. WLL networks, working at high frequencies of over20 GHz, need direct connection without any physical obstacles. Any small obstacle

EthernetLLC/PHY

Bridge

SNAP

WLL

ATM ATM

SONET/SDH

WLL

ATM

AAL5

E1in/out

PPPEthernet

RFC

CPE1

Base Station

CPE2LLC

2596

Router

n x 64 kin/out

Ethernet

E1in/out

E1in/out

PHY - WLL

PHY

AAL1 AAL1

AAL1 AAL1 AAL5

ATM

Figure 3.34 Structure and protocols for WLL network elements.

Air interface


may interfere with the communication. Usually, wireless networks are more sensi-tive to interference, so noise or extreme weather conditions may easily affect theiroperation.

3.8 CONCLUSIONS

Now that we know the architecture of an ATM network, as well as its basic opera-tion, and the most important procedures involved, we can start to evaluate its oper-ation in a practical environment. In the next chapters we shall describe a series ofprocedures that inform us of the operation and status of an ATM network, togetherwith certain measurements and tests that can be carried out to check for the correctoperation of a connection when it is brought into service.


• ITU-T Rec. I.432.1, B-ISDN user-network interface - Physical layer specification: General character-istics.

• ITU-T Rec. I.432.2, B-ISDN user-network interface - Physical layer specification: 155520 kbit/s and622080 kbit/s operation.

• ANSI T1.102-1993.

• ANSI T1.105-1995.

• ANSI T1.105.02-1995.

• ANSI T1.107-1995, Digital Hierarchy - Formats Specifications.

• ANSI T1.646-1995, Broadband ISDN - Physical Layer Specification for User-Network InterfacesIncluding DS1/ATM.

• ATM Forum af-phy-0016.000, DS1 Physical Layer Interface Specification, Sept. 1994.

• ATM Forum af-phy-0054.000, DS3 Physical Layer Interface Specification, Jan. 1996.

• ATM Forum af-phy-0040.000, Physical Interface Specification for 25.6 Mb/s over Twisted PairCable, Nov. 1995.

• ATM Forum af-phy-0133.000, 2.5 Gbps Physical Layer Specification, Oct. 1999.

• ITU-T Rec. G.804, ATM cell mapping into Plesiochronous Digital Hierarchy.

• ITU-T Rec. I.363.1, B-ISDN ATM Adaptation Layer specification: Type 1 AAL.


• ITU-T Rec. I.363.3, B-ISDN ATM Adaptation Layer specification: Type 3/4 AAL.


• ITU-T Rec. I.365.1, Frame relaying service specific convergence sublayer.

• ITU-T Rec. I.366.1, Segmentation and Reassembly Service Specific Convergence Sublayer for the


AAL type 2.

• ITU-T Rec. I.366.2, AAL type 2 service specific convergence sublayer for narrow-band services.

• ITU-T Rec. I.371, Traffic control and congestion control in B-ISDN.

• ATM Forum af-tm-0121.000, Traffic Management Specification Version 4.1, March 1999.

• ITU-T Q.2100, B-ISDN signalling ATM adaptation layer (SAAL) overview description.

• ITU-T Q.2931, Digital Subscriber Signalling System No. 2 (DSS 2) - User-Network Interface (UNI)layer 3 specification for basic call/connection control.

• ITU-T Q.2630.1, AAL type 2 Signalling Protocol.

• ITU-T Q.2961, Digital Subscriber Signalling System No. 2 (DSS 2) - Additional Traffic Parameters.

• ITU-T Q.2962, Digital Subscriber Signalling System No. 2 (DSS 2) - Connection characteristicsnegotiation during call/connection establishment phase.

• ITU-T Q.2965.1, Digital Subscriber Signalling System No. 2 (DSS 2) - Support of Quality of Serviceclasses.

• ATM Forum af-pnni-0055.000, Private Network-Network Specification Version 1.0, March 1996.

• ATM Forum af-sig-0061.000, ATM User-Network Interface (UNI) Signalling Specification Version4.0, July 1996.

• ITU-T Rec. I.356, B-ISDN ATM layer cell transfer performance.

• ITU-T Rec. I.357, B-ISDN semi-permanent connection availability.

• Reuven Cohen, "Service Provisioning in an ATM-over-ADSL Access Network," IEEE Communica-tions Magazine, Vol. 37, No. 10, Oct. 1999.

• Grossman and Heinanen, "Multiprotocol Encapsulation over ATM Adaptation Layer 5,” RFC 2684,Sept. 1999.

• G. Gross et al., “PPP over AAL5,” RFC 2364, July 1998.

• L. Mamakos et al., “A Method for Transmitting PPP Over Ethernet (PPPoE),” RFC 2516, February1999.

• W. Townsley et. al., "Layer Two Tunneling Protocol (L2TP)," RFC 2661, Aug. 1999.

• M Davison et al., "L2TP over AAL5," Internet draft, July 2000.

• Radia Perlman, Interconnections: Bridges and Routers, Addison-Wesley Professional ComputingSeries.

• José M. Caballero, “ATM, la tecnología de conmutación” in ISDN, Frame Relay and ATM, MadridL&M Data Communications, 2001.

151

Chapter 4

ADSL Technology

4.1 THE ORIGIN OF DSL TECHNOLOGIES

A wide area network (WAN) is usually divided in two, the backbone or transmis-sion network and the access network. Long-distance, high-capacity links are typicalfor a transmission network, and, historically, it has benefited from the technologicaldevelopment in transmission and multiplexing. The access network in its turn hashad a different evolution. While in one part of the WAN, there was a continuoustechnological revolution (first, the introduction of time-division multiplexing withDS1 and E1 signals, then came SONET/SDH, and now DWDM), these changeshad practically no effect on the access network.

The topology of the access network has changed relatively little during all theseyears, and it is almost the same throughout the world. At its most basic, the accessnetwork is formed by copper pairs that connect different users by a local telephoneexchange. Usually, a local exchange has all the means to connect users to the trans-mission network. As for the services offered to users, they have also experiencedvery little change during many years, and basically remained on a 3,100-Hz band-width channel suitable for baseband voice communications.

It is estimated that there are about 700,000,000 copper lines installed in theworld today. Some 95% of these lines are still used for traditional voice services(POTS). The access network has led to very significant monetary investments, dueto its extreme complexity in geographical distribution and its high degree of branch-ing. The size of the access network provides an idea of how difficult it would be tomake modifications that would widely affect the entire network. The difficulty ofbuilding another access network that could offer similar capacity to connect usersworldwide is also obvious. But the problem is that during the past couple of years,the needs of users have changed drastically, and 3,100-Hz voice access is no longerenough. First of all, current applications and services call for digital, not analog ac-cess. Second, there is a growing need for greater bandwidth. As technological inno-vations have merely concentrated in the transmission network, and the access


network has remained as is for the most part, today we are facing a bottleneck at thispart of the network.

4.1.1 The Birth of DSL Technologies: HDSL

An emergency solution to fulfill the new needs of users is to digitalize communica-tion at its source, by means of modems, using the traditional voice channel fortransmission. There have been many innovations in this type of modem during thelast few years. Modems that first offered a 9,600 bps access later increased their ca-pacity to 28,800 bps, and their limit is now 56 Kbps. Digital access through thistype of modem has been very useful to those who look for cost-effective digital ac-cess. However, the capacity of analog modems cannot be increased much under thebandwidth conditions and the signal-noise ratio offered by a normal telephone line(see Section 1.1.2.3).

The ISDN standards define a progressive digitization plan for the whole accessnetwork, starting from the very core of this network, and ending up providing theend-users with a digital communications channel, usually of 128 + 16 Kbps. This isthe so-called basic rate interface (BRI). The ISDN access rollout via BRI had al-ready started in the beginning of the 1980s. It has generally been very complicated,although in some countries it has penetrated the market rather effectively. Anotherreason for the failure in implementing the BRI is the appearance of other technolo-gies that approach the access network problem from another point of view.

Given that the bandwidth assigned to voice band communication cannot providemuch more than 56 Kbps, it is obvious that for higher rates, there should be morebandwidth. The modern DSL access systems take advantage of existing copperpairs. This way, there is no need for a high investment in infrastructure, and the re-sults are positive in that rather high access rates have been obtained.

The idea of using frequency band over voice band in the access network is notnew. For example, the signal bandwidth in a BRI interface is about 80 kHz. Howev-er, there was another application prior to ISDN that used broadband signals in theaccess network.

There is no doubt that one of the most important events in the history of telepho-ny dates back to the 1960s and the introduction of the 1,544 Kbps (DS1) frames inthe U.S.A. and the 2,048 Kbps (E1) frames in Europe. Like all the innovations in thisfield, the digital revolution mostly affected the transmission network. However, withtime, DS1 and E1 services were provided to users who needed a high capacity ofdata exchange. This was done across the same twisted pairs that were used for thePOTS services. In practice, these users were almost always big companies, as theywere the only ones who could afford the installation and maintenance costs of these

ADSL Technology 153

types of circuit. At that time, they were also the only ones who actually needed high-capacity data links.

The pairs that were used for the E1/DS1 service had to be prepared beforehand.For example, no branching could be allowed in cables that were to be used for thiskind of service (see Figure 4.1) (see Section 9.2.6.2). Another problem that appearedwas crosstalk, where electromagnetic coupling between lines may cause the signalof one user to degrade those of other users (see Section 9.2.5). The higher the fre-quency, the more crosstalk, and this is a potential problem that arises from E1/DS1services. However, the biggest inconvenience was that the E1/DS1 signal was ex-posed to serious attenuation in the copper pair, and in general it was not reliable touse, if boosters were not installed between the transmitter and the receiver. In short,it was possible to give high rate services to users, but the preparation and mainte-nance costs of the line made the situation rather complicated.

With time, great progress was made in lengthening the distance between boost-ers. The reason for attenuation in a copper pair is the resistance that the copper wirepresents against the flow of the current. It can be demonstrated that this resistancedecreases when the diameter of the wire increases. The easiest way to increase thereach of a signal is to increase this diameter. The most typically used diameter is0.5 mm, but you can also find diameters in the range of 0.6 mm and 0.9 mm.

In the beginning, signals were transmitted in base band. The line codes werevery simple, the same as the ones used within the transmission network (see Section1.1.3). For the DS1 signal, the alternate mark inversion (AMI) code was used, and

Figure 4.1 Effect of a split pair when transmitting a broadband signal in the local loop. In a split pair the signal is divided into two, one towards the receiving end and the other towards the secondary end. Despite the original signal, a phased and atten-uated version of it is received.

Transmitted signal

Signal reflection (if this is not terminated)

Received signal


for the E1, the high density bipolar three zeroes (HDB3) code. The aim of this typeof line codes was to obtain a signal without the dc component and without too longidle periods that would cause problems for the synchronization-recovering circuitsof the receivers. A big step forward was the introduction of the 2B1Q code (the useof which was already standardized in the ISDN BRI) that made it possible to trans-mit two bits in a line symbol, while the AMI and the HDB3 only allowed transmis-sion of one bit per symbol. The 2B1Q was introduced simultaneously with anotherimprovement that made it possible to transmit one DS1 signal across two pairs in-stead of one. This way, the 1,544 Kbps signal was divided into two signals of784 Kbps. The use of the 2B1Q code over one DS1 signal in two pairs made it pos-sible to offer services to users without the need for boosters. This technology wascalled the high bit rate digital subscriber line (HDSL). An HDSL for the E1 signalappeared soon after this. First across three pairs, but when the technique got moreadvanced, it was possible to do this across two pairs at 1,168 Kbps (see Table 4.1).

4.1.2 New Modulation Technologies

The next step was taken when certain advanced modulation technologies were in-troduced. For example the carrierless amplitude and phase (CAP) modulation thatwas to replace the 2B1Q code. The CAP modulation allows the transmission of be-tween two and nine bits in one symbol. The wide acceptance of the 2B1Q code inthe ISDN world meant that once ETSI and ANSI standardized the HDSL, bothtechnologies were accepted. Even today, for HDSL, the CAP and the 2B1Q existtogether.

Different technologies were also proposed to combat crosstalk. The user infor-mation addressed to the telephone exchange (upstream), as well as the informationcoming from the exchange and addressed to the user (downstream), must be trans-mitted in the same pair where a DSL signal is transmitted. There are two techniquesto multiplex these two ways of transmission. In the first one, two subbands are re-served, one for upstream and the other for downstream. This procedure is called fre-quency-division multiplexing (FDM). In the second one, the whole bandwidthavailable can be shared by upstream and downstream traffic. They will be separatedby an echo canceler when received. This second variant is known as the overlappedspectrum technique.

The overlapped spectrum technique takes more advantage of the available band-width, since, to obtain the same performance, there is no need for such high frequen-cies as with the FDM. Therefore, overlapped spectrum transmission is especiallysuitable for environments where attenuation limits the performance. On the otherhand, in systems based on FDM, it is possible to cancel the crosstalk by frequencyfiltering, as the two methods of transmission do not share the same band. The FDMtechnique is therefore appropriate for environments where crosstalk is the critical

ADSL Technology 155

factor. Since both systems have their advantages and disadvantages, both are used indifferent DSL systems.

As far as the FDM technique is concerned, different studies have shown thatcrosstalk is a more important factor at the exchange than at the user end. The reasonis that in a telephone exchange, a large amount of pairs originating from differentsources come together, and coupling between them is easier. This is why the lowestfrequency band, less sensitive to crosstalk, is occupied in the FDM upstream trans-mission.

The advances in data coding and modulation techniques nowadays make it pos-sible to install DSL services on a copper pair with performance similar to that of theold HDSL across two pairs. This is an important benefit. The DSL method that pro-vides the user with both E1 and DS1 access over a copper pair is called symmetricaldigital subscriber line (SDSL). It has also been standardized in many ways by dif-ferent organizations. In many cases, the lack of exactness in the nomenclature usedto name these systems makes it difficult to know exactly what literature is talkingabout. Of the existing standards, one worth mentioning is the G.SHDSL that offersrates between 192 Kbps and 2.3 Mbps and can be applied all over the world.

4.1.3 Asymmetry

When it was acknowledged that the characteristics of upstream traffic are consider-ably different from those of downstream in a number of applications, room wasmade for a DSL access technique called ADSL.

In many cases, user access to a telecommunications network is of a client/servertype. Here, the user as a client makes requests to a remote server that responds to himin the form of a client-addressed transmission. Statistical studies of the traffic gen-erated by the client and the server show that their nature is different. More specifi-cally, the traffic generated by the server has more bandwidth than that of the user.To benefit from these circumstances, a system was proposed that would dedicatemost of the available bandwidth to traffic generated by the server.

This system was mainly designed in the U.S., with the aim of applying it to avideo service based on the copper pairs used for POTS telephony. The possibility ofoffering video services this way came as a result of American cable operators’ at-tempt to provide telephone services in their coaxial cable networks. This system wasnever introduced, as it was difficult to justify its costs because the ground/earth tele-vision networks, as well as the satellite services, were already used for the intendedpurpose. However, as many other services also share the client/server model, this ac-cess method was not forgotten, and it was later used in data communications, mainlyfor Internet access. Nowadays, this way of accessing the Internet is very popular. In


addition, the interest in offering video services across the copper pair would havedisappeared. These services can be provided over IP or ATM with an ADSL access.

ADSL technology was on the verge of being born in 1992. One of the crucialquestions was which type of modulation to use. There were those who were in favorof the old quadrature amplitude modulation (QAM), used by modems during theprevious 20 years. CAP modulation was also proposed, already used successfully inHDSL; it can be seen as a variant of QAM. The third option was discrete multitonemodulation (DMT), patented but not implemented by AT&T Bell Labs 20 years be-fore, which had several advantages over the first two options.

The use of the 2B1Q code was discarded, as it is a baseband transmission tech-nique, and it was thought that ADSL could be multiplexed with POTS by FDM. Itwas agreed that the separation of upstream from downstream would be done bymeans of FDM, to preserve the system of crosstalk. However, some experts thoughtthat the advantages of FDM were not enough, and, as a result, some methods ofADSL with overlapped spectrum exist even today. Similarly, and although the DMTmodulation system was officially adopted, there are still different versions of ADSLthat use CAP.

Nowadays, with ADSL, it is possible to obtain rates that reach up to 8 Mbps forupstream and 640 Kbps for downstream. These rates depend on the quality of thecopper pair used for transmission, and, more concretely, on its attenuation, as wellas noise level and interference.

One of the recent contributions to the world of ADSL is the so-called standardG.lite (ITU-T Rec. G.992.2). It has been designed to be a high-rate, low-cost access

Table 4.1 Comparison of four access technologies.

Technology Speed Availability Best Application Strengths and Weaknesses

POTS Up to 56 Kbps Yes remote access to consumers

Low cost but low speed, highly avail-able

ISDN 128 Kbps symmet-ric

Highly available in North America and Europe

Video conference and power Internet user

High speed and low error rate but high cost

HDSL 1,544 Kbps or 2,048 Kbps Limited Private corporate

networksLimited availabil-ity and high price

ADSL

Up to 8 Mbps upstream and up to 640 Kbps down-stream

Highly available in North America and Europe, high growth rate

Internet access No new cabling required

ADSL Technology 157

system. The standard G.lite is peculiar in that it offers an ADSL service without thesplitter that is generally used in this technology to separate the baseband POTS sig-nal and the ADSL passband signal. With G.lite, it is possible to offer services of upto 1.5 Mbps downstream and 512 Kbps upstream.

Finally, it can be said that the acceptance of DSL systems has produced a growthin access technologies using twisted pairs with different, intelligent variants (see Ta-ble 4.2). There are many others besides the above-mentioned ones, like ISDN DSL(IDSL) which combines the DSL technologies with ISDN, or very high speed DSL(VDSL) which offers symmetrical bandwidth of up to 52 Mbps (26 Mbps upstreamand 26 Mbps downstream).

4.2 REFERENCE MODELS

4.2.1 ADSL System

ADSL is used to connect user installations with a local exchange that introduces theinformation received in the transmission network (see Figure 4.2).

ADSL is compatible with POTS or ISDN transmitted in the same copper pair.In other words, it is possible to transmit ADSL in a high-frequency band simulta-neously with ISDN or POTS occupying the lower frequencies. These two frequency-multiplexed signals are separated by splitters both at the user end and at the ex-change. To be able to carry out their task, these splitters must be frequency sensitive.

All of the transmission and reception functions of ADSL technology are carriedout by blocks called ATU-C and ATU-R. The ATU-C is in charge of coding andmodulation (as well as decoding and demodulation) of the ADSL signal at the ex-change end, and the ATU-R takes care of the same functions at the user end. It isimportant to note that the ATU-R and ATU-C are transceiver elements, and they donot have any switching functions. Switching takes place in a separate block that ex-

Table 4.2 Comparison between some xDSL variants.

xDSL Variant Speed Typical Reach Symmetric POTS

Symmetric HDSL (SHDSL) Up to 2.3 Mbps 4.4 km (1.5 Mbps) Yes NoHigh bit rate DSL (HDSL) Up to 2048 Mbps 4.3 km Yes NoISDN DSL (IDSL) Up to 144 Kbps 8.0 km Yes NoAsymmetric DSL (ADSL ) Up to 8 Mbps 5.5 km (1.5 Mbps),

1.5 km (7 Mbps)No Yes

Rate Adaptive DSL (RADSL) Adaptive up to 7 Mbps

5.5 km (1.5 Mbps), 1.5 km (7 Mbps)

No Yes


changes information with the ATU-C by means of the standard V-C interface, andwith the ATU-R by means of the T-R interface.

From an OSI point of view, ADSL occupies the lower level of the protocolstack. That is, the physical level. The levels immediately above are usually occupiedby ATM technology (see Section 3.7.1.2). This means that ATM cells are generallycarried above the ADSL frame structure, even though this is not required by the stan-dards. This is why switching is carried out at the ATM level (especially from the ex-change), and the wideband network that acts as a transmission network is an ATMnetwork, typically ATM over SDH, SONET, or PDH.

4.2.2 ADSL Transceivers

Data transfer by ADSL allows two methods, STM and ATM. This transfer is car-ried out by the ATU-C by means of up to seven independent data flows, AS0, AS1,AS2, AS3, LS0, LS1, and LS2, the first four being simplex and the rest duplex (seeFigure 4.3). In the case of ATU-R, it is only possible to transfer across the three du-plex channels (see Figure 4.4). Each independent flow can be assigned a config-urable rate in steps of 32 Kbps. This is done at the initialization and training stagesthat precede normal operation.

The existence of different, separated bearer channels allows for hypotheticalmultiplexing of the information generated for different users or at least for differentapplications. In practical applications, the most frequent solution is to use only onebearer channel. Some variants can be found that use one of the available bearer chan-

Figure 4.2 Reference model for the ADSL access network. The ADSL signal is separated for POTS or ISDN by using frequency-sensitive splitters. The access network ends, in both extremes, at the point where the transmission and reception facilities end in the local loop. The switching facility is already seen as a part of the backbone of the user network.

V-C

xTU-CPhys.level

U-C2

POTS or ISDNNarrowbandnetwork

C Splitter

R Splitter

xTU-R

User installation

network

POTS/ISDN

U-R2

U-C U-R

T/S

T-R

Transmission Network

Access Network

User NetworkTerminal

SM SM

Phys.level

Widebandnetwork

ADSL Technology 159

nels for main data flow, but they can optionally reserve another one for additionalinformation interchange.

Specifically, for transmission using the ATM method, not all of the flows areused, and it is also necessary to add an extra block that carries out the function typ-ical for the transmission convergence sublayer, to adapt the ATM cell flow into theavailable ADSL channels.

As far as ATM cell transmission across ADSL is concerned, it is usually carriedout using the same procedures as in its transmission across SDH or SONET (see Sec-tion 3.2). During transmission, background cells are inserted if necessary, as speci-fied in ITU-T Rec. I.432.1. The HEC is also generated following the proceduredescribed in this recommendation, and the payload is randomized by using the poly-nomial X43+1 as usual.

In reception, to verify the HEC and delimit the cells, the guidelines are set bythe ITU-T Rec. I.432.1. The delineation algorithm, as explained in the recommen-dation, consists of the procedure that is followed to find the limits of the cells insidethe ATM flow. In this algorithm, there are three different stages:

• HUNT: This is to ensure the flow, bit by bit, until a valid HEC is found. Here thedelineation algorithm presumes that it has found an ATM header, and moves onto the next stage.

MUX/synchr.

AS0

AS1

AS2

AS3

LS0

LS1

LS2

NTR

OAM

V-CEO

C/A

OC ib

CRCf

CRCi interlv.

tone

ord

erin

g

cons

tella

tion

IDFT

Out

put b

uffe

rSe

rial/P

aral

lel C

onve

rsio

n

Figure 4.3 Explanatory diagram of the ATU-C transmitter. The separated coding processes for data in the fast and the interleaved buffer, and modulation using DMT.

A B C Zii=1 to 255

511510

480

0

control

DAC andanalog proc.

enco

der

Scram& FEC

Scram& FEC


• PRESYNC: The validity of the HEC cell is verified, cell by cell, starting fromthe delineation data obtained in the HUNT stage. Once DELTA HECs arefound, the algorithm proceeds to the SYNC stage. If an error occurs whensearching for one of the headers, the algorithm goes back to the HUNT stage.

• SYNC: In this stage, even though the HEC is also verified cell by cell, it is notassumed that the synchronism is lost until consecutive incorrect ALFA HECsare found.

Each ASx and LSx channel is multiplexed over a single frame, the contents ofwhich are distributed between two possible buffers:

• Interleaved buffer: The data transmitted by this buffer is submitted to interleav-ing before its transmission. With interleaving methods, the noise detected at thereceiving end appears white, thereby increasing the efficiency of the FEC meth-ods used. The downside of the interleaving methods is an increase of latency inthe transmitted data.

• Fast buffer: Here, there is not too much latency, although the probability ofreceiving erroneous data is higher than with the interleaved buffer.

Two separated buffers can handle each kind of service in a different way to allowadaptation to the appropriate quality requirements for each of them. For example,for interactive services sensitive to transmission delay (voice, interactive video) theuse of interleaving will not be adequate and therefore it is mandatory to use the fastbuffer. For those services very sensitive to transmission errors but with no special

Figure 4.4 Explanatory diagram of the ATU-R transmitter. It is very similar to the ATU-C transmitter. The only differences are the interface (T-R different from V-C) and the number of symbols generated.

MUXsync control

LS0

LS1

LS2

EOC

/AO

C

CRCf

CRCi interlv.

tone

ord

erin

g

cons

tella

tion

IDFT

Out

put b

uffe

rSe

rial/P

aral

lel C

onve

rsio

n

A B C Zii=1 to 255

511510

480

0

DAC andanalog proc.

enco

der

Scram& FEC

Scram& FEC

ADSL Technology 161

requirements regarding delay (file transfer, encoded noninteractive video) the inter-leaved buffer must be used.

The data to be transmitted by the interleaved buffer and by the fast buffer arechosen at the training stage, before transmission.

Once the frame has been formed and FEC, cyclic redundancy check (CRC), andsignaling, among other redundancy bytes, have been added, each bit is assigned asubcarrier frequency. Each subcarrier has a number of bits assigned, and this numberdepends on the transmission conditions in the associated frequency band. As a result,after this operation, each tone has an associated word of bits of a different length ineach case. Once this operation is carried out, each word is assigned a complex num-ber that can be considered as the subcarrier amplitude and phase in each frequency.From another point of view, this sequence of complex numbers could be seen as afrequency representation of the symbol that needs to be transmitted. The switch fromfrequency domain to time domain is done by carrying out an inverse discrete fouriertransform (IDFT). The result is passed on to a digital analog converter (DAC) thattransforms the signal into analog which is suitable to be sent across a copper pair.

The result of an IDFT, when changed into an analog signal, forms a transmis-sion symbol. The rate at which these symbols are sent, or the signaling rate, is4,000 bauds.

4.3 FRAMING

4.3.1 Data and Overhead Buffers

There are many types of framing for ADSL. They are all variants of the same typecalled full overhead framing (see Figure 4.5). There are also versions with reducedoverhead, where a part of the full overhead used is deleted.

Up to four types of frame structure are allowed for the downstream link. Theframings are numbered with the digits 0, 1, 2, and 3. The framing mode that is usedin data exchange between the ATU-R and ATU-C is negotiated between the two atthe initialization stage, before data transmission.

For upstream, there are also four types of frame structures, and they are the sameas for downstream, with the difference that they appear in both types of transmission.The biggest differences between the upstream and the downstream frame are:

• In upstream, there is no A(S) extension byte (AEX) like that used in downstreamto control synchronization. Nor are there any ASx (simplex) signals, which iswhy the maximum number of logical channels is three.


• The maximum depth of the interleaving and the maximum number of FECredundancy bytes vary.

The main framing method is full framing, and both ATU-C and ATU-R must meetit, while it is not necessary for them to comply with the rest of the methods. Fromthis point onwards, we shall only deal with full framing.

The structure of the frame depends on the point of the transceiver that is beinglooked at, since during coding, overhead bytes are added and the information is dealtwith in many ways. The description here refers to the frames at the output of the FECoverhead calculation blocks.

Each frame is composed of two parts, the fast and the interleaved buffer. Theinterleaved part acts as an input for a convolutional interleaver, and the fast part goesdirectly to the DMT modulator.

fast AS0 AS1 AS2 AS3 LS0 LS1 LS2 AEX LEX FEC bytes

1 byte

B (A

S0)

B (A

S1)

B (A

S2)

B (A

S3)

C (L

S0)

B (L

S1)

B (L

S2)

byte

s

byte

s

byte

s

byte

s

byte

s

byte

s

byte

s

f f f f f f f Aby

tesf

Lby

tesf

Rby

tesf

sync AS0 AS1 AS2 AS3 LS0 LS1 LS2 AEX LEX

1 byte

B (A

S0)

B (A

S1)

B (A

S2)

B (A

S3)

B (L

S0)

B (L

S1)

B (L

S2)

byte

s

byte

s

byte

s

byte

s

byte

s

byte

s

byte

s

i i i i i i i Aby

tesi

Lby

tesi

FEC bytes

Rby

tesi

fast LS0 LS1 LS2 LEX

1 byte

B (L

S0)

B (L

S1)

B (L

S2)

byte

s

byte

s

byte

s

f f f Lby

tesf

Rby

tesf

(a)

(c)

sync LS0 LS1 LS2 LEX

1 byte

B (L

S0)

B (L

S1)

B (L

S2)

byte

s

byte

s

byte

s

i i i Lby

tesi

Rby

tesi

(d)

Figure 4.5 ADSL frame models: (a) ADSL frame, divided into two parts: fast buffer and inter-leaved buffer; (b) ATU-C frame model for the fast buffer with a full overhead; (c) ATU-C frame model for the interleaved buffer with a full overhead; (d) ATU-R frame model for the fast buffer with a full overhead; and (e) ATU-R frame model for the interleaved buffer with a full overhead.

Fast Fast data FEC Interleaved data

Fast buffer Interleaved buffer

ADSL Frame

Fast buffer (ATU-C)

(e)

Interleaved buffer (ATU-C)

Fast buffer (ATU-R) Interleaved buffer (ATU-R)

(at the constellation coder input) (at the constellation encoder input)

250 ms

(b)

FEC bytes FEC bytes

ADSL Technology 163

Fast buffer (ATU-C)

The fast buffer basically consists of a sequential grouping of the four logicalchannels ASx and the three LSx, if there are associated data. The frame length is notfixed, and the amount of data to be transmitted across each logical channel of eachframe is decided at the initialization stage.

Before user data, a so-called fast synchronization byte is transmitted. It containsa specific overhead for the noninterleaved data. The contents of this byte depend onthe position of the frame that transmits it inside a superframe.

After user data, the fast buffer transmits the AEX and L(S) extension byte (LEX)used for synchronization, if the input data timing in the ATU-C is not synchronouswith the timing of the ATU-C itself. The AEX and LEX bytes may transport userinformation or not, depending on what the bits reserved for synchronization controlindicate.

Finally, the fast buffer ends with the FEC redundancy bytes that make it possi-ble to correct errors at the receiving end.

Interleaved buffer (ATU-C)

The form of the interleaved buffer is practically identical with that of the fastbuffer. It also consists of sequential grouping of the ASx and LSx channels, with anumber of bytes that depend on the allocation carried out at the initialization stage.It also contains the AEX and LEX synchronization control bytes and the FEC redun-dancy bytes. The sync byte of the interleaved buffer has a meaning similar to that ofthe fast byte. It is either used to transport synchronization control information, ADSLoverhead control (AOC) information, or a CRC to control errors at the multiframelevel.

4.3.2 Superframes

The ADSL superframe is formed by 68 frames (see Figure 4.6). Each of them con-tains a space reserved for both the fast buffer and the interleaved buffer.

Each ADSL frame corresponds to a DMT symbol. Each superframe corre-sponds to 68 symbols that contain information, and one extra symbol that introducesthe modulator and does not contain user information. This extra symbol is called asynchronization symbol, and it is used to establish the limits for the ADSL super-frame.

The contents of the synchronization bytes of the fast (see Figure 4.7) and inter-leaved (see Figure 4.8) buffers depend on the position of the frame in question inside


the superframe. The fast byte of the first frame of each superframe is used to trans-port an 8-bit CRC. Frames 1, 34, and 35 are used to transmit 24 bits of information(ib bits), that is mainly used to carry alarms. The rest of the frames are grouped inpairs of even and odd frames. In each frame group, either embedded operationschannel (EOC) control information or synchronization control information is trans-mitted. The difference between the two is that for synchronization information, thebit of less weight of the fast byte is 0, whereas for EOC information, this takes thevalue of 1.

In the interleaved buffer, the synchronization byte has a structure similar to thatof the fast byte. In the first frame of the superframe, a CRC of eight bits is transmit-ted. The other frames carry synchronization control information and ADSL over-head control information (aoc bits).

In those cases where the timing of input channels is not synchronous with thetransceiver, it is necessary to carry out a synchronization operation of some sort. Ifthe rate of the input channel is faster than the nominal one, the control mechanismallows for extra bytes, AEX and LEX, in the buffers. If the rate is lower than thenominal one, some of the bytes to be transmitted will be eliminated. The bits that are

0 1 2 34 35 66 67 sync

ib 16..23

Figure 4.6 ADSL superframe, formed by 68 frames including a synchronization symbol that does not carry user information. The duration of the ADSL superframe is 17 ms.

ADSL superframe

ib 8..15fast

fast

ib 0..7fast

CRC

fast sync

17 ms

frame

crc7 crc6 crc5 crc4 crc3 crc2 crc1 crc0

ib15 ib14 ib13 ib12 ib11 ib10 ib9 ib8

frame

34

msb lsb

eoc7 eoc6 eoc5 eoc4 eoc3 eoc2 r1 1

sc7 sc6 sc5 sc4 sc3 sc2 sc1 02k

0 ib7 ib6 ib5 ib4 ib3 ib2 ib1 ib0

ib23 ib22 ib21 ib20 ib19 ib18 ib17 ib16

1

35

msb lsb

eoc13 eoc11 eoc10 eoc9 eoc8 eoc7 1

sc 7 sc 6 sc 5 sc 4 sc 3 sc 2 sc 1 02k+1

Figure 4.7 Contents of the fast synchronization byte in each frame that together form the superframe.

eoc12

Fast synchronization bytelsbmsb lsb

sc7 sc6

ADSL Technology 165

used to control synchronization are called sc bits. Some of them can be used to selectthe channel ASx or LSx where synchronization will be carried out, whereas othersgive details of the operation to make (see Table 4.3). They can, for example, enablea byte to carry information, delete some of the bytes that were to be transmitted, ornot carry out any synchronization.

4.4 CODING

To improve the data transmission quality of the ATU-C and ATU-R, many op-erations are performed on the information. For example, data are scrambled, a redun-

Table 4.3 Fast byte format for synchronization.

Bits Function Codes

sc7..6 Designator of the logical ASx channel 00 - AS001 - AS110 - AS211 - AS3

sc5..4 Synchronization control for the assigned ASx channel

00 - no synchronization operations01 - adding of AEX to the assigned ASx channel11 - adding of AEX and LEX to the assigned ASx channel10 - elimination of the last byte of the assigned ASx channel

sc3..2 Designator of the LSx logical channel 00 - LS001 - LS110 - LS211 - no synchronization operations

sc1 Synchronization control for the assigned LSx channel

1 - adding of LEX to the assigned LSx channel0 - elimination of the last byte of the assigned LSx channel

sc0 Synchronization/EOC 0 for synchronization operations

crc7 crc6 crc5 crc4 crc3 crc2 crc1 crc00

msb lsb

Interleaved synchronization byte

sc7 sc6 sc5 sc4 sc3 sc2 sc1 0

aoc7 aoc6 aoc5 aoc4 aoc3 aoc2 aoc1 aoc0

1..67 Signals in the interleaved buffer

No signals in the interleaved buffer

Figure 4.8 Contents of the interleaved synchronization byte in each frame of the ADSL super-frame.

frame


dancy FEC is added, or interleaving is carried out. These coding operations aredescribed in the following.

4.4.1 Error Protection

The ADSL communications channel is usually exposed to a high level of noise andinterference. These interferences degrade the information if no special protectionmeasures are taken. This is because in ADSL some powerful coding mechanismsare used. The CRC and the Reed-Solomon (RS) coding are two different channelcoding techniques that meaningfully increase the roughness of data transmission.The former is used for error detection. The CRC information may be used by upperlayers of the protocol stack to enable a retransmission mechanism to recover infor-mation received with errors. The Reed-Solomon coding implements error correc-tion at the receiver end. This allows correction of the information received withoutretransmissions. The redundancy added by the RS codes is bigger than with theCRC, but its error correcting features serve to compensate for this fact.

In the coding process, the CRC is applied first and then the RS is calculated. Ob-viously, the decoding starts with the RS and then is verified if there are errors in theCRC blocks. That means that errors will only be detected by the CRC decoding al-gorithm if the RS code fails. The RS is the main protection mechanism and the CRCis a complement that serves to prevent any data from passing with errors.

Two independent CRCs are generated for each superframe, one for the fast andanother one for the interleaved buffer. These two CRCs generated in a superframeare placed into the space reserved for this in the following superframe (see Figure4.9). The CRC checks if errors have occurred in transmission at the superframe lev-el.

In regard to the R-S codes, coding and decoding are performed in the Galoisfield . In the case of ADSL, coding algebra is carried out in the field. The field is generated, with the help of the following polynomial ofbinary coefficients:

One feature of the RS codes used with ADSL is that it is possible to set the num-ber of redundancy bytes, R. This makes it possible to select the optimum protectionlevel for each case. The length of the message words and redundancy bytes dependson the initial configuration negotiated between the ATU-C and the ATU-R. maytake the following values: 0, 2, 4, 6, 8, 10, 12, 14, 16. The message word coincideswith the full fast or interleaved buffer in an ADSL frame.

GF pn( ) GF 256( )GF 256( )

q X( ) X8 X4 X3 X2 1+ + + +=

R

ADSL Technology 167

m m0 1 mk-1

1c c c c c c c c2 3 4 5 6 70

frame n-1

frame n

- - - - - - - -

m(X) = m X + m X + ... + m X + m0 1 k-1k-2k-1 k-2

g(X) = X + X + X + X + 18 4 3 2 c(X) = m(X)X mod g(X)8

c(X) = c X + c X + ... + c X + c0 1 767 6

Figure 4.9 (a) This is the operation of the CRC coder. The CRC is calculated over the bits of an ADSL frame, and it is then inserted into the following frame. The result is always eight bits that are obtained as the remainder of the polynomial division between the polynomial message and a standard generating polynomial.(b) This is the CRC decoder. Transmission may change some bits. The CRC code is calculated over the bits of a frame and it is then compared to the CRC received in the following frame. If no bits are altered, all the bits of the CRC should be equal.

m’(X) = m’ X + m’ X + ... + m’ X + m’0 1 k-1k-2k-1 k-2

g(X) = X + X + X + X + 18 4 3 2 c(X) = m’(X)X mod g(X)8

g(X) = c’’ X + c’’ X + ... + c’’ X + c’’0 1 767 6

- m’ m’0 1frame n-1

1c’ c’ c’ c’ c’ c’ c’ c’2 3 4 5 6 70frame n

Tx OK?

-------

(a)

(b)

m m0 1 mk-1

m’k-1

m’ m’0 1 m’k-1

comparisonof c’ with c’’


4.4.2 Scrambling

The data that proceeds directly from the layers that are clients of the ADSL layermay contain a high correlation between symbols. This situation is usually seen asundesirable, for example, it could be harmful for the synchronization recovery cir-cuits. With the aim of decorrelating the data, a scrambler is used. This scrambler isbased on a shift register that acts bit by bit upon the input information. The samescrambler is used for the fast and the interleaved buffer. The function of this scram-bling is:

4.4.3 Interleaving

The only difference between the fast and the interleaved buffer, when it comes tothe process that the information in each of them goes through, is that the interleavedbuffer is subjected to the action of a convolutional interleaver and the fast buffer isnot.

The objective of interleaving is to increase the performance of FEC error cor-rection. The performance of RS coding will be notably reduced, if the autocorrela-tion of transmission errors is high, in other words, if the transmission errors tend toconcentrate at certain points, while other points are practically free of errors. It is de-sirable that the probability of errors in a given bit be statistically independent of theerrors that may have occurred in other bits.

These conditions are often not met in real transmission systems. Errors tend togroup in bursts. Interleaving is a method that is used for receiving error bursts as iso-lated errors.

Interleaving basically consists of varying, in a preestablished way, the order inwhich the symbols are sent in a digital transmission system. This means that when asymbol is ready for transmission, it should wait for the corresponding time gap, and,as a result, there is a delay in transmission. This is the price to be paid for increasingthe efficiency of line coding (in this case, the RS code).

In the case of the convolutional interleaver, the delay in the transmission of eachbit of the RS code word equals to , where is the depth of interleaving.This is equal to writing the symbols as lines in a table of lines, and reading themin columns (see Figure 4.10).

Besides reading in columns and writing as lines, so that the operations can becarried out in the way described before (by simply introducing a delay in each byte

b'n bn b'n 18– b'n 23–⊕ ⊕=

D 1–( ) i× DD

ADSL Technology 169

of every block), the order of all the writing operations associated with the symbolsat the beginning of the block must coincide with the order of the reading operationof the same symbol. For example, if and , the writing operations ofinitial symbols of the blocks are carried out in the following order: 1, 6, 11, 16, 21,...This coincides with the reading order of the same symbols, which means that thebytes of the form stay fixed during interleaving.

It can be demonstrated that to meet this condition, must be a power of 2 and must be odd. The values that the standard allows in the case of ADSL for the

depth of interleaving are 1, 2, 4, 8, 32, and 64 (the two latter values are only validfor downstream).

4.5 MODULATION

The coded digital information is moved to the blocks that form the modulator. Thisfinally constitutes an electrical signal that is transmitted across the local pair. Assaid before, the modulation used is DMT, which makes it possible to take advan-tage of the bands that are in better condition for transmission, while discardingthose that are corrupted by noise and interference.

N 5= D 4=

B1n

1B12B1

3B15B1

4B11B2

2B23B2

5B24B2

1B32B3

3B35B3

4B31B4

2B43B4

5B44B4

1B52B5

3B55B5

4B51B6

2B63B6

5B64B6

1B1X X 2B1X 1B2 X X 2B23B1

1B3X 4B12B3

3B21B4

5B14B2

2B43B3

1B55B2

4B32B5

3B41B6

5B34B4

2B63B5

Interleaved flow

Initial flow

Write

Figure 4.10 This table is used to interleave the information byte by byte. The symbols are intro-duced in columns, leaving one more empty block at the start of each row than in the previous row. The symbols are read in columns.

1B12B1

3B14B1

5B1

2B23B2

4B25B2

1B32B3

3B34B3

5B3

1B42B4

3B44B4

5B4

1B52B5

3B54B5

5B5

1B62B6

3B64B6

5B6

1B72B7

3B74B7

5B7

1B82B8

3B84B8

4B8

Rea

d

N

D1B2

DN


4.5.1 Organizing the Tones

The first operation carried out with the coded information is to assign each bit atone for its transmission (see Figure 4.11). At the initialization stage of the trans-mission system, each DMT subcarrier is assigned a number of bits for their trans-mission. The capacity of each symbol, as the number of bits transported persymbol, is obtained by adding up the bits assigned to each subcarrier. The final stepis to determine which bits exactly of the ADSL frame are modulated in one of thesubcarriers (see Figure 4.12).

The process of assigning bits to subcarriers starts with the fast buffer and withthe tones, which have a smaller number of bits assigned. When the bits of the fastbuffer are used up, assignment of the interleaved buffer is started in the tones thatare not filled yet.

There is a reason to assign the fast buffer bits to tones with few bits, and the in-terleaved buffer bits to tones with many bits. An ADSL signal has a high ratio be-tween the peak power and the mean power. This might result in a possible saturationof the DAC. It has been reported that when saturation occurs, the error magnitude isalmost uniformly distributed in all the tones. The performance of the transmissionsystem increases when the bytes that are more protected against errors are assignedto tones with more dense constellations.

4.5.2 Constellation Coders

The bits assigned to each subcarrier are used to modulate a tone with a QAM mod-ulation. QAM is a modulation that transmits discrete amplitude values modulated ata certain carrier frequency. It transmits two values simultaneously, one being inquadrature in respect to the other. Each pair of values of those transmitted simulta-

0 1 2 3 4 5

Figure 4.11 The process of assigning bits to each DMT subcarrier.

SNR

Carrier

bits

1

2

345

ADSL Technology 171

neously can be represented by a unique complex number. The imaginary part ofthis complex number is the one transmitted in quadrature in respect to the real part.

In the equation presented above, the transmitted symbols meet the following re-lation: .

The finite set of all the values that may take , and that can be represented inthe complex plane (see Figure 4.13), is called the constellation of the QAM signal.

Given that the constellation points are complex numbers, in many cases it is use-ful to identify them by their module and their phase, which furthermore coincideswith the amplitude and the phase transmitted in the real signal.

The function of the constellation coder is to assign each possible set of bits as-signed to the carrier a complex value of ck. Depending on the number of bits as-signed to the subcarrier, one constellation or another is used.

4.5.3 DMT Modulation

The DMT modulation considers the set of symbols associated with each subcarrierat a certain moment as a frequential representation of the signal. The modulationconsists basically of applying the inverse discrete Fourier transform to recover thetemporal representation of the signal (see Figure 4.14).

Figure 4.12 Assigning the contents of the ADSL frame to subcarriers. For simplicity, it is sup-posed that the number of buffer bits is 11.

Interleaved bufferFast buffer

1b 2b 3b 4b 5b 6b 7b 8b 9b 10b 1b 2b 3b 4b 5b 6b 7b 8b 9b 10b0b0b

5 0 3 4 1 2

s t( ) Re ckp t kT–( )ejωt

k ∞–=

∞

∑

akp t kT–( ) ωtcos

k ∞–=

∞

∑ bkp t kT–( ) ωsin t

k ∞–=

∞

∑–= =

ck ak jbk+=

ck


The DMT modulator carries out the following operation:

This is an inverse DFT. The transmitter carries out this operation at a frequencyequal to the symbol frequency, 4,000 bauds.

The constant has the value 512 in the case of downstream and 64 for up-stream, which is twice that of the QAM symbols generated by the constellation cod-

0000 0010

0001 0011

0100 0110

0101 0111

1000 1010

1001 1011

1100 1110

1101 1111

Figure 4.13 Constellation used in ADSL for a tone that transports four bits.

Coefficients of tonesorganized by their index

kci

Frequential representation Fourier transform

ndi

Temporal representation

Figure 4.14 Frequency and temporal representations of an ADSL symbol and the relation between them.

f

t

DFT -1

DMT symbol

dn cke

jωN------nk

k 0=

N 1–

∑=

N

ADSL Technology 173

er. The complexes of index higher than 256 (32 for upstream) are generated in thefollowing way:

The line stands for a conjugated complex. It is necessary to double the numberof QAM symbols by means of hermitian symmetry of those that exist in the begin-ning, to obtain a real signal in the temporal domain, as only real signals can be trans-mitted.

Furthermore, there are some values of for which no information is transmit-ted. The index values 0 and 256 (32 downstream) are not used and do not containenergy. The use of another subcarrier is also reserved for the transmission of a pilottone to ensure synchronization of the receiver. The energy of the pilot tone is con-stant in time.

The set of is converted into an analog signal by means of a digital analogconverter (DAC). The frequency at which the digital to analog converter works al-lows the frequency separation between subcarriers to be 4.3125 kHz. Taken that may reach up to 255, the ADSL signal spectrum will contain energy up to 1.1 MHz(see Figure 4.15).

Besides the pilot, the dc indices and the Nyquist frequency, there are other val-ues of that are not used either. This depends on the transmission mode chosen. Foreach case, the upstream and downstream signals must be adapted to the spectralmasks specified in the ITU-T Rec. G.992.1. Generally, to adapt to the spectral re-quirements, the subcarriers below a preselected index are not used, and only the onesthat meet are used.

ck cN k–=

k

dn

n

k

k Nmin≥

Figure 4.15 ADSL/POTS, and other xDSL technologies, signal spectrum and average reach.

IDSL

HDSL2

2.5

2

1.5

1

.5

3.0 5.0 7.0

Average Reach Achieved (km) Mbps

20 1404 kHz

1100400

SDSL

ADSL up

ADSL down

SHDSL

POTS

Frequency (kHz)

PSD

km


4.5.4 Cyclic Prefix and Synchronization Symbol

The fundamental frequency of the ADSL signal is, as mentioned before,4.3125 kHz, which is superior to the 4.0 kHz that constitutes the signaling rate(symbol frequency). The separation between subcarriers is a little bit higher thanthe signaling rate, so that some extra samples (40 bytes for downstream and fivefor upstream) can be inserted into the 512 (64 in upstream) samples of the signal inthe time domain. In other words, the following is met:

Thirty-two of the 40 remaining samples are used (four of five for downstream)to form the cyclic prefix that is inserted into the DAC input together with the datasamples. This cyclic prefix delimits an ADSL frame.

The remaining capacity is used to form the synchronization symbol that delimitsthe ADSL superframe, and that makes it possible to recover frame synchronizationafter microinterruptions, without the need to start training the modems. A synchro-nization symbol is inserted for each 68 data symbols. The contents of this synchro-nization symbol are formed by a pseudorandom bit sequence (PRBS).

4.6 OPERATION AND MAINTENANCE CHANNEL (EOC)

Some of the bytes of the ADSL frame provide the so-called EOC channel that isused for maintenance in and out of service, and for monitoring the condition of theloop and the remote end.

The EOC channel is used for transmitting information in the form of messages.In the EOC transmission model, the ATU-R works as a slave for the ATU-C. TheATU-C generates messages that are responded to by the ATU-R. This response cansimply be an echo of the received message, or a message that constitutes a responsefor data required by the ATU-C. An exception to this command-response protocolis a dying gasp message that can be automatically generated by the ATU-R.

4.6.1 EOC Message Format

Each EOC message has a length of 13 bits, divided into five different fields (seeFigure 4.16).

The meaning of each of these fields is as follows:

(512 bytes + 40 bytes) x 4.0 kHz = 512 x 4.3125 kHz

(64 bytes + 5 bytes) x 4.0 kHz = 64 x 4.3125 kHz

(downstream)

(upstream)

ADSL Technology 175

• Address (ADDR): This is the addressing field for the device that the message isaddressed to. The 00 value is used by the ATU-C to send messages to the ATU-R. The value 11 is used by the ATU-R to send them to the ATU-C. The otherpossible values (01 and 10) do not have any assigned use.

• A/O: This specifies the contents of the operation code/data (OPCODE/DATA)field. The value 0 indicates that this field contains a data byte, and 1 means thatOPCODE/DATA contains one of the currently existing 58 operation codes.

• P: This is the order bit. The value 0 is introduced in even messages and 1 in oddmessages. This way, exchanging information in multiple reading/writing opera-tions is easier, as they need more than one message to exchange all the informa-tion in question.

• Autonomous message field (AMF): When its value is 1, it means that the mes-sage is of the command/response type. The value 0 stands for autonomoustransfer. For example, the ATU-R gives value 0 for this bit to transmit a dyinggasp message, which is the message transmitted by the ATU-R right beforebeing disconnected.

• OPCODE/DATA: This is a field that contains data (that can be coded) or anoperation code that indicates the type of command to carry out.

4.6.2 EOC Commands

The messages can be used by the ATU-C to vary the internal configuration of theATU-R, or to read or write in information registries assigned to the ATU-R (seeTable 4.4). These registries are standardized. Note that two of them, 6 and 7, aremeasurement results: one for the noise margin of the line and the other for its inte-grated attenuation. This means that the ATU-R itself must have characteristics of atester. Actually, the ADSL transceivers have the capacity to carry out continuousand in-service monitoring in the communications channel, by measuring two pa-rameters defined as follows.

The integrated attenuation (ATN) is defined as the difference in decibels (dB)between the power received by the transceiver and the power of the signal transmit-ted at the other end. This should not be mixed up with the usual parameter for qual-ification of telecommunications systems that can be called a frequency function. Theintegrated attenuation is not a frequency function. It is the difference in decibels be-tween the total transmitted and received power.

EOC message

Figure 4.16 EOC message format.

Address A/O P AMF OPCODE/DATA2 1 1 1 8 bits


Noise margin is the amount of extra power of the received signal, in respect tothe level produced by a BER of 10-7.

The registries can be accessed remotely by using READ commands, and some-times it is possible to write to them by means of WRITE commands. Furthermore,there are many other types of commands. Some of them are used to control the EOCexchange protocol itself, like NEXT, used by the ATU-C to demand the followinginformation byte in data transfer. The end of data (EOD) is used by the ATU-C toindicate that all the data required has been transmitted. The ATU-R uses the unableto comply acknowledgement (UTC) when it cannot carry out a command required bythe ATU-C. See the ITU-T Rec. G.992 for all the commands that can be used.

4.7 INITIALIZATION

The initialization process of an ADSL modem does not only include agreeing onthe parameters for later data transmission with the remote end. This means that theinitialization of the modems is not only a handshake procedure with the remote mo-dem. On the contrary, other procedures are included, such as channel analysis andtraining that are aimed at defining the characteristics of the channel and finding theoptimum transmission parameters in each case (see Figure 4.17).

Table 4.4 Data registers.

Reg R/W Length Description

0 R 8 bytes Identifies the ATU-R manufacturer1 R (manufacturer) Version number minus one2 R 32 bytes Serial number3 R (manufacturer) Autotest results4 R/W (manufacturer) (Depends on the manufacturer)5 R/W (manufacturer) (Depends on the manufacturer)6 R 1 byte Integrated attenuation of the line7 R 1 byte Noise margin8 R 30 bytes ATU-R configuration9-F - - Reserved for future use

Handshake

Initialization

Training Channel Analysis Exchange

Figure 4.17 Model of initialization of an ADSL modem and its different stages.time

ADSL Technology 177

4.7.1 Handshake

The first stage when initializing an ADSL transceiver is constituted by handshakingprocedures described in the ITU-T recommendations G.992.1 and G.994.1. At thisstage, a special type of modulation, a binary differential phase shift keying (DPSK),is used, which is simpler than the DMT. In this modulation, the transmission of abinary 1 means an advance of 180º from the latest transmitted phase. A binary 0does not indicate any turn in the phase. The frequencies used are also special, andthey take the form of or . The values n can take arestandardized in G.994.1.

4.7.1.1 Format of information

A handshake means an exchange of messages between the two transceiving enti-ties. These messages are divided into segments and they are transmitted in frames(see Figure 4.18).

The frame is delimited by groups of bytes with hexadecimal value 7E, and theycoincide with the HDLC frame, as described in the ISO/IEC 3309 standard. The ini-tial part of the frame may contain between three and five bytes of this value, and theframe ends with two or three bytes. The frame check sequence (FCS) field is used totransport a cyclic protection code generated in line with the following polynomial:

n 4.3125kHz× n 4.000Hz×

Figure 4.18 Handshake frame format. Messages: MS, MR, CL, CLR, ACK(1), ACK(2), NAK-EF, NAK-NR, NAK-NS, NAK-CD, REQ-MS, REQ-MR, REQ-CLR.

Flag FCS1Message

(optional)(optional)Flag=7Ex

Identification Standard info Nonstandard info(optional)

Flag Flag Flag Flag FCS1 Flag Flag Flag

Type of message Version number Manufacturer Parameters

G X( ) X16 X12 X5 1+ + +=


Each message is formed by three fields. The last one being optional. The first isan identification field that describes the message type and other types of informa-tion, such as manufacturer and version. The second field transmits information thatrefers to the message that is being sent, in the form of parameters coded in a stan-dardized way. Most of the information in this field corresponds to the operatingmodes and features of the transceiver. The last, optional field also includes informa-tion on operating modes and features.

4.7.1.2 Handshake transactions

The handshake sessions may occur in different ways, depending on the transceiverstatus, the present options, the transmission mode to be chosen, and so on. This sec-tion gives examples of some of the ways the handshake can be carried out. All ofthem have one message in common, which demands a certain operation mode and aresponse whereby this mode is accepted (see Figure 4.19).

Figure 4.19 Handshake sessions.

CLR C

L

ACK(

1)

MS

ACK(

1)

MS

REQ

-MR

MR

MS

ACK(

1)

ATU-R

ATU-C

t ATU-R

ATU-C

CLR: the ATU-R sends a list with all the operation modes it acceptsCL: the ATU-C responds with a list of modes it acceptsACK: the ATU-R has received the list of modes correctly MS: the ATU-R demands an operation mode ACK: the ATU-C accepts itMS: the ATU-R requires a certain operation modeREQ-MR: the ATU-C wants to select the mode and makes a request for an MR messageMR: the ATU-R requires an MS operation modeMS: the ATU-R demands a certain operation mode ACK: the ATU-R accepts the request

ADSL Technology 179

4.7.2 Training

Once the handshake stage is over and the mode to use has been defined, it is neces-sary to carry out training of the equalizers and echo cancelers (if they exist) of thetransceiver. With this objective, a sequence of standard signals is generated. Theyare standardized both in their waveform and their maximum and minimum length atthe two ends of communication.

The training signals are initialized starting from a QUIET (C-QUIET2 and R-QUIET2) stage where the transmitted signal is null. The ATU-C then goes throughdifferent C-PILOT and C-REBERB stages (see Figure 4.20). In the C-PILOT stages,the ATU-C transmits a pure sine-wave signal at pilot frequency. In the C-REBERBstages, it transmits a pseudorandom sequence identical to the one used by the syn-chronization symbol. This sequence can be described by the following equation:

The two first bits of the sequence modulate the subcarrier in continuous fre-quency and the Nyquist frequency, respectively. The remainder of the bits are orga-nized in pairs that act to modulate the rest of the subcarriers with a 4-QAM.

While all this occurs in the ATU-C, the ATU-R also goes through different R-REBERB and R-QUIET stages. It is not necessary to transmit PILOT signals, as theATU-R is the one that gets locked on the ATU-C synchronization signal. At the R-QUIET stages, the transmitted signal is null. At the R-REBERB stages a pseudoran-dom sequence is transmitted and its contents depend on whether the ADSL signal ismultiplexed in the same cable with POTS or ISDN. Furthermore, before 512 DMT

bn1

bn 4– bn 9–⊕

=n 1 … 9, ,=

n 10 … 512, ,=

C-QUIET2C-PILOT1

C-PILOT1AC-QUIET3A

C-REBERB1 C-PILOT2 C-ECT C-REBERB2 C-REBERB3C-QUIET5

C-PILOT3

R-QUIET2 R-REBERB1 R-QUIET3 R-ECT R-REVERB2

ATU-C

ATU-R

Figure 4.20 Sequence of signals generated by modems at the training stage.


symbols of the first R-REBERB (R-REBERB1) phase, the ATU-R should alreadyhave acquired synchronization in a loop with the ATU-C.

By means of these stage transactions, the transceivers may:

• Measure the power in the upstream channel, and, as a consequence, adjust thepower to transmit it in downstream;

• Adjust the automatic gain control (AGC) and synchronization circuits of itsreceiver;

• Train its equalizer.

Besides the described signals, the training process includes C-ECT stages for theATU-C, and R-ECT for the ATU-R, where the transmitted signal can be freely cho-sen by the manufacturer. The ECT signals are used to train the echo canceler inADSL implementations that do not use an FDM multiplexion of the upstream anddownstream channels, but use superimposed spectra instead.

4.7.3 Analyzing the Channel

The third stage of the ADSL initialization process is analyzing the channel. Thisstage is basically to carry out the tasks related to measuring the signal-to-noise-ra-tio (SNR) in the different subbands where the DMT subcarriers transmit informa-tion (see Figure 4.21). This way, the transceivers can generate tables of the ADSLframe bit assignation for each DMT carrier. Furthermore, certain information onthe coding and framing options is exchanged at this stage.

As in the case of training, this stage is formed by successive states where signalsand sequences are transmitted that enable the analysis of both downstream and up-stream. The state from which both ATU-C and ATU-R originate is a SEGUE (C-SEGUE1 and R-SEGUE1) state. This state is characterized by transmission of pseu-do-random sequences that are identical to the ones used at the training stage, butturned 180º with respect to these, in all the carriers except at pilot frequency. Fromthe R-SEGUE state the ATU-R moves to a new R-REBERB state, otherwise identi-cal to the ones used at the training stage, but adding a cyclic prefix. On the otherhand, the ATU-C starts transmitting information on the downstream coding andframing options that will be used in transmission. The states where this transmissionis carried out are called C-RATES1 and C-MSG1. In C-RATES, 992 bits are trans-mitted in total. One bit by a DMT symbol is transmitted in a way that a binary 0 iscoded as a C-REBERB symbol, and a binary 1 as an R-SEGUE. The informationtransmitted in the C-RATES state is basically the following:

• Lists of the number of bits in the ADSL frame fast buffer, assigned to each ofthe ASx and LSx flows;

• Lists of the number of bits in the ADSL frame interleaved buffer, assigned toeach ASx and LSx flows;

ADSL Technology 181

• Upstream and downstream values for interleaving depth and number of paritybytes used by the FEC block.

In the C-MSG1 state, a 48-bit word is transmitted by using the coding describedwith a bit in each DMT symbol (see Table 4.5):

Furthermore, the information transmitted during C-RATES1 and C-MSG1 isprotected by two 16-bit parity words, C-CRC1 and C-CRC2, obtained by polynomi-al coding of the messages, using the generating polynomial:

Table 4.5 Contents of the C-MSG1 message.

Bits Description

47-44 Minimum noise margin in initialization43-18 Reserved17 Trellis coding16 Overlapping spectrums15 Not used (a binary 1 is used)14-12 Reserved 11 Transport of network timing reference (NTR)10-9 Framing type8-6 Transmission of the power spectral density (PSD)

used in initialization5-4 Reserved3-0 Maximum number of bits supported by subcarrier

C-SEGUE1 C-RATES1C-CRC1

C-MSG1 C-MEDLEY C-REBERB4

R-REBERB3 R-SEGUE2 R-MEDLEY R-REVERB4

ATU-C

ATU-R

Figure 4.21 Sequence of signals generated by modems when analyzing the channel.

C-CRC2

R-SEGUE1 R-RATES1R-CRC1

R-MSG1R-CRC2

G X( ) X16 X12 X5 1+ + +=


The C-CRC1 is transmitted immediately after C-RATES1, and the C-CRC2 af-ter C-MSG1.

Once ATU-C has transmitted the C-RATES1 and C-MSG1 messages, the ATU-R proceeds in a similar way, by sending R-RATES1 and R-MSG1 messages (see Ta-ble 4.6) with a similar format. The R-RATES1 message is shorter than the C-RATES1, which has 384 bits.

The sequence that is used to analyze the channel is called MEDLEY (C-MED-LEY for the one generated by the ATU-C, and R-MEDLEY for the one generatedby the ATU-R). The MEDLEY sequence is generated in a similar way to theREBERB, but in the MEDLEYs, a cyclic prefix is always used, and the generatingPRBS is not restarted for each symbol, but continues from each one to the next one.

4.7.4 Exchanging Information

After analyzing the channel, ATU-C and ATU-R know the most adequate frequen-cies for data transmission, and they can generate tables for assignation of bits tosubcarriers. However, the ATU-R only knows the downstream channel and theATU-C the upstream channel. But, to be able to select constellations for each fre-quency, the transceivers need information on each upstream and downstream chan-nel. For this reason, information must be exchanged on the assignation of bits tosubcarriers (see Figure 4.22).

The stage of exchanging information is formed by successive states in the ATU-C and ATU-R, similar to the states described before. The REBERB and SEGUEstates are defined in a similar way to the equivalent states at the training and channel

Table 4.6 Contents of the R-MSG1 message.

Bits Description

47-18 Reserved17 Trellis coding16 Overlapping spectrums15 Not used (a binary 1 is used)14 Support for higher rates13 Support for dual downstream latency12 Support for dual upstream latency 11 Support for NTR10-9 Framing type8-4 Reserved3-0 Maximum number of bits supported by subcarrier

ADSL Technology 183

analysis stage. However, in this case, the cyclic prefix is always used, together withthe PRBS.

The states RATES and MSG are used to transmit configuration options andmeasurement results during the channel analysis. The format of the messages ex-changed in these states is similar to the format of the equivalent messages at the anal-ysis stage. The ITU-T Rec. G.992.1 gives more information about their format. Themain difference is the technique used to transmit. Eight bits for each symbol aretransmitted. Four subcarriers are used to do this. Each of them modulated with a 4-QAM. In addition, as a security measure, four subcarriers more are used to transmitthe same bits simultaneously. The subcarriers used depend on the type of ADSL,whether it is over POTS or ISDN.

Besides the RATES and MSG messages, a third type of message, B&G, is usedto exchange information. At the end of the analysis stage, the ATU-R has informa-tion on the downstream channel and the ATU-C on the upstream channel, but notvice versa. However, the ATU-C needs to know how many bits it must insert intoeach downstream subcarrier, and the gain it must apply to each of them to meet thenoise margin requirements chosen. Naturally, the situation of the ATU-R is similar,but toward the opposite direction. This information is transmitted in the B&G state.The ATU-C transmits a 496-bit message in the following format:

These are the gains (in respect to the gain used when transmitting R-MEDLEY)and the bit assignations for each upstream subcarrier. The ATU-R transmits a mes-

C-REBERB4

C-RATES-RAC-CRC-RA1C-MSG-RA

C-RE

BERB

5

ATU-C

ATU-R

Figure 4.22 Sequence of signals generated by modems when exchanging information.

C-CRC-RA2C-S

EGU

E2C-REVERB-RA

C-SE

GUE-

RA

C-M

SG2

C-C

RC

3C

-RAT

ES2

C-C

RC

4C

-B&G

C-C

RC

5

C-SE

GUE

3

R-S

EGU

E2 R-RATES-RAR-CRC-RA1R-MSG-RAR-CRC-RA2

R-REBERB-RAR-MSG2R-CRC3R-RATES2R-CRC4R-

SEGU

E-RA

R-REBERB5

R-B

&GR

-CR

C5

R-S

EGU

E4

R-RE

BERB

6

R-S

EGU

E5

g31 b31 g30 b30 … g1 b1 , , , , , ,


sage equal to 4,080 bits for downstream. The technique used for transmitting theB&G messages is 8 bits/symbol, as explained before.

Each RATES, MSG, and B&G message is protected by a CRC parity message,the same as during the channel analysis stage.


• Charles K. Summers, ADSL Standards, Implementation and Architecture, Boca Raton: CRC Press,1999.

• Walter Goralsky, ADSL and DSL Technologies, second ed. New York: Osborne/McGraw-Hill, seconded., 2002.

• Dhawan Chander, Remote Access Networks PTSN, ISDN, ADSL, Internet and Wireless. New York:McGraw-Hill, 1998.

• Daniel Minoli, Video Dialtone Technology, Digital Video over ADSL, HFC, FTTC and ATM, NewYork: McGraw-Hill, 1995.

• ITU-T Rec. G.992.1, Splitterless asymmetic digital subscriber line (ADSL) transceivers.

• ITU-T Rec. G.992.1, Asymmetric Digital Subscriber Line (ADSL) transceivers.

• ITU-T Rec. G.994.1, Handshake procedures for Digital Suscriber Line (DSL) transceivers.

• ITU-T Rec. G.996.1, Test procedures for Digital Subscriber Line (DSL) Transceivers.

• ITU-T Rec. I.432, B-ISDN User network interface - Physical layer specification: General characteris-tics.

• Pepe Caballero and Francisco J. Hens, A Testing Time for the Local Loop, International Telecommu-nications, 2001.

• ANSI T1.413, Network and consumer installation interfaces - Asymmetric Digital Subscriber Line(ADSL).

• TS 101 388, Transmission and multiplexing (TM); Access transmission systems on metallic accesscables; Asymmetric Digital Subscriber Line (ADSL), Coexistence of ADSL and ISDN-BA on thesame pair.

• www.paradyne.com, The DSL Sourcebook.

185

Chapter 5

Network Synchronization

Synchronization is the set of techniques that enable the frequency and phase of theequipment clocks in a network to remain constrained within the specified limits(see Figure 5.1). The first digital networks were asynchronous, and therefore didnot call for properly working external synchronization. It was the arrival of SDHand SONET networks that started to make synchronization essential to maintaintransmission quality and efficiency of supported teleservices.

Bad synchronization causes regeneration errors and slips. The effects of theseimpairments vary in different systems and services. Some isochronous1 services,like telephony, tolerate a deficient synchronization rather well, and small or no ef-fects can be observed by the end-user. Others, like digital TV transmission, fax, orcompressed voice and video services, are more sensitive to synchronization prob-lems. In HDLC, FRL, or TCP/IP types of data services, slips that occur force us toretransmit packets, and this makes transmission less efficient.

5.1 ARCHITECTURE OF SYNCHRONIZATION NETWORKS

Synchronization networks can have hierarchical or nonhierarchical architectures.Networks that use hierarchical synchronization have a tree architecture. In suchnetworks a master clock is distributed, making the rest of the clocks slaves of itssignal. A network with all the equipment clocks locked to a single master timingreference is called synchronous. The following elements can be found in the hierar-chical synchronization network:

1. A master clock, which is usually an atomic cesium oscillator with global posi-tioning system (GPS) and/or Loran-C2 reference. It occupies the top of the pyr-

1. Isochronous (from the Greek "equal" and "time") pertains to processes that requiretiming coordination to be successful, such as voice and digital video transmission.

2. Loran-C is an electronic position fixing system using pulsed signals at 100 kHz.


amid, from which many synchronization levels spread out (see Table 5.1).2. High-quality slave clocks, to receive the master clock signal and, once it is fil-

tered and regenerated, distribute it to all the NEs of their node.3. NE clocks, which finish the branches of the tree by taking up the lowest levels

of the synchronization chain. Basically, they are the ones using the clock,although they may also send it to other NEs.

4. Links, responsible for transporting the clock signal. They may belong to thesynchronization network only, or, alternatively, form a part of a transport net-work, in which case the clock signal is extracted from data flow (see Figure5.3).

Figure 5.1 A master clock that marks the significant instances for data transmission. Clocks 1 and 2 are badly synchronized, and the data transmitted with these references is also affected by the same phase error.

Master Clock

Slave Clock-1 with offset

Data-1 with offset

Slave Clock-2 with jitter

Data-2 with jitter

Synchronized Data

t1 t2 t3 t4t0 t6 t7 t8t5

Figure 5.2 Classes of synchronization architectures.

Master

Asynchrony Hierarchical Synchronization Mutual Synchronization

Slaves

Network Synchronization 187

The pure hierarchical synchronization architecture can be modified in severalways to improve network operation. Mutual synchronization is based on cooperationbetween nodes to choose the best possible clock. There can be several master clocks,or even a cooperative synchronization network, besides a synchronization protocolbetween nodes (see Figure 5.2). Bringing these networks into services is more com-plex, although the final outcome is very solid.

Those networks where different nodes can use a clock of their own, and correctoperation of the whole depends on the quality of each individual clock, are calledasynchronous (see Figure 5.2). Asynchronous operation can only be used if the qual-ity of the node clocks is good enough, or if the transmission rate is reduced. The op-eration of a network (that may be asynchronous in the sense described above or not)is classified as plesiochronous if the equipment clocks are constrained within mar-gins narrow enough to allow simple bit stuffing (see Figure 5.2).

General requirements for today’s SONET and SDH networks are that any NEmust have at least two reference clocks, of higher or similar quality than the clockitself. All the NEs must be able to generate their own synchronization signal in casethey lose their external reference. If such is the case, it is said that the NE is in hold-over.

A synchronization signal must be filtered and regenerated by all the nodes thatreceive it, since it degrades when it passes through the transmission path, as we willsee later.

5.1.1 Synchronization Network Topologies

The synchronization and transport networks are partially mixed, since some NEsboth transmit data and distribute clock signals to other NEs.

The most common topologies are:

1. Tree: This is a basic topology that relies on a master clock whose reference is

Table 5.1 Clock performance.

Type Performance

Cesium From 10-11 up to 10-13

Hydrogen From 10-11 up to 10-13

GPS Usually 10-12

Rubidium From 10-9 up to 10-10

Crystal From 10-5 up to 10-9


distributed to the rest of the slave clocks. It has two weak points: it depends ononly one clock, and the signals gradually degrade (see Figure 5.5).

2. Ring: Basically, this is a tree topology that uses SDH/SONET ring configura-tions to propagate the synchronization signal. The ring topology offers a way tomake a tree secure, but care must be taken to avoid the formation of synchro-nizing loops.

3. Distributed: Nodes make widespread use of many primary clocks. The com-plete synchronization network is formed by two or more islands; each of themdepending on a different primary clock. To be rigorous, such a network isasynchronous, but thanks to the high accuracy of the clocks commonly used asa primary clock, the network operates in a very similar way to a completelysynchronous network.

4. Meshed: In this topology, nodes form interconnections between each other, inorder to have redundancy in case of failure. However, synchronization loopsoccur easily and should be avoided.

Synchronization networks do not usually have only one topology, but rather a com-bination of all of them. Duplication and security involving more than one masterclock, and the existence of some kind of synchronization management protocol, areimportant features of modern networks. The aim is to minimize the problems asso-ciated with signal transport, and to avoid depending on only one clock in case offailure. As a result, we get an extremely precise, redundant, and solid synchroniza-tion network.

Clock

Data + Clock

NE

NE

NE

NE

NE

NE

NE

NE

NE

NE

NE

NE

NE

NE

NE

BreakIntra-node

Internode

NE NE

Loop Timing

Master Clock

Figure 5.3 Synchronization network topology for SONET and SDH. This figure does not show links that are for transport only.

Slave Clock (BITS or SSU)

Master Clock (PRS or PRC)

Alternative

NE

Alternative Reference Clock

Slave Clocks

Node or building

Break


5.2 INTERCONNECTION OF NODES

There are two basic ways to distribute synchronization across the whole network:

• Intranode, which is a high-quality slave clock known as either synchronizationsupply unit (SSU) or building integrated timing supply (BITS). These are re-sponsible for distributing synchronization to NEs situated inside the node (seeFigure 5.3).

• Internode, where the synchronization signal is sent to another node by a linkspecifically dedicated to this purpose, or by an STM-n/OC-m signal (see Figure5.3).

5.2.1 Synchronization Signals

There are several signals suitable for transporting synchronization:

• Analog, of 1,544 and 2,048 kHz;

• Digital, of 1,544 and 2,048 Kbps;

• STM-n/OC-m line codes, from which one of the above-mentioned signals isderived, by means of a specialized circuit.

In any case, it is extremely important for the clock signal to be continuous. In otherwords, its mean frequency should never be less than its fundamental frequency (seeFigure 5.4).

5.2.1.1 Clock transfer across T-carrier/PDH networks

These types of networks are very suitable for transmitting synchronization signals,as the multiplexing and demultiplexing processes are bit oriented (not byte orientedlike SONET and SDH), and justification is performed by removing or adding singlebits. As a result, T1 and E1 signals are transmitted almost without being affected by

Figure 5.4 A pure clock signal is continuous, as, for example, the one provided by an atomic clock. A discontinuous signal in its turn could be a signal delivered by a T1 circuit transported in SONET.

Ideal

t

t

Discontinuous

Discontinuity

Clock

Clock


justification jitter, mapping or overhead-originated discontinuities. This character-istic is known as timing transparency.

There is only one thing to be careful with, and that is to not let T1 and E1 signalscross any part of SONET or SDH, as they would be affected by phase fluctuationdue to mapping processes, excessive overhead, and pointer movements. In short, T1or E1 would no longer be suitable for synchronization.

5.2.1.2 Clock transfer across SDH/SONET links

To transport a clock reference across SDH/SONET, a line signal is to be used in-stead of the tributaries transported, as explained before. The clock derived from anSTM-n/OC-m interface is only affected by wander due to temperature and environ-

NE

Figure 5.5 Synchronization network model for SONET and SDH. Stratum 3 has the mini-mum quality required for synchronizing an NE. In SDH the figures indicate the maximum number of clocks that can be chained together by one signal.

BITS

BITS

NE NE NE

Stratum 1

Stratum 2

Stratum 3

Stratum 4

PRS

BITS

BITS

PRC

SSUT

SSUL

G.813 SEC

max.

max. 60

SEC

SEC

SEC+SSU

20SEC

max.10SSU

SSUT

SEC

G.811

G.812

G.812

SEC

SDH

SONET


mental reasons. However, care must be taken with the number of NEs to be chainedtogether, as all the NEs regenerate the STM-n/OC-m signal with their own clockand, even if they were well synchronized, they would still cause small, accumula-tive phase errors.

The employment of STM-n/OC-m signals has the advantage of using the S1byte to enable synchronization status messages (SSMs) to indicate the performanceof the clock with which the signal was generated (see Figure 5.6). These messagesare essential in reconstructing the synchronization network automatically in case offailure. They enable the clocks to choose the best possible reference, and, if none isavailable that offers the performance required, they enter the holdover state.

5.2.2 Holdover Mode

It is said that a slave clock enters holdover mode when it decides to use its own gen-erator, because it does not have any reference available, or the ones available do notoffer the performance required. In this case, the equipment remembers the phaseand the frequency of the previous valid reference, and reproduces it as well as pos-sible. Under these circumstances, it puts an SSM=QL-SEC message into the S1byte of STM-n/OC-m frames, and, if it was generating synchronization signals at1.5 or 2 MHz, it stops doing so.

5.2.3 Global Positioning System

The global positioning system (GPS) is a constellation of 24 satellites that belongsto the U.S. Department of Defense. The GPS receivers can calculate, with extremeprecision, their terrestrial position and the universal time from where they extractthe synchronization signal. The GPS meets the performance required from a prima-ry clock (see Table 5.1). However, the GPS system might get interfered with inten-tionally, and the U.S. Department of Defense reserves the right to deliberatelydegrade its performance for tactical reasons.

Figure 5.6 The S1 byte is used to send SSMs in SDH and SONET.

S1: Clock source

0000 - unknown0010 (QL-PRC) - Primary clock0100 (QL-SSU-T) - Transit clock1000 (QL-SSU-L) - Local clock1011 (QL-SEC) - Synchronous equipment1111 (QL-DNU) - Do not use

SSM (bits 5-8)B2 B2 B2 K1 K2

D4 D5 D6

D7 D8 D9

S1 E2

D10 D11 D12

M1

B2 K1 K2

D4 D5 D6

D7 D8 D9

S1 E2

D10 D11 D12

M1

MSO

H

LOH

SONET SDH 01010101 - invalid clock


5.3 DISTURBANCES IN SYNCHRONIZATION SIGNALS

Since synchronization signals are distributed, degradation in the form of jitter andwander accumulate. At the same time they are affected by different phenomena thatcause phase errors, frequency offset, or even the complete loss of the referenceclock. Care must be taken to avoid degradation in the form of slips and bit errors byfiltering and an adequate synchronization distribution architecture (see Figure 5.7).

5.3.1 Frequency Offset

Frequency offset is an undesired effect that occurs during the interconnection ofnetworks or services whose clocks are not synchronized. There are several situa-tions where frequency deviations occur (see Figure 5.8):

• On the boundary between two synchronized networks with different primaryreference clocks;

• When tributaries are inserted into a network by nonsynchronized ADMs;

• When, in a synchronization network, a slave clock becomes disconnected fromits master clock and enters holdover mode.

5.3.1.1 Consequences of frequency offset in SDH/SONET

To compensate for their clock differences, SDH/SONET networks use pointer ad-justments. Let us think of two multiplexers connected by STM-1 (see Figure 5.8),where ADM2 is perfectly synchronized, but ADM1 has an offset of 4.6 parts permillion (ppm).

ADM1 inserts a VC-4, but as ADM2 uses another clock, it should carry out pointeradjustments periodically, to compensate for the difference between the two clocks.

Figure 5.7 Sources of phase variation.

Error (t)

Error = C + k(t)

Error (t)

t

Frequency offset Clock noise

Error (t)

t1 day

Daytime wandert

Error (t)

t1 year

Stational wander

f2 155.52Mbps 4.6ppm+ 155.52 1 4.6106--------+

Mbps= =

f1 155.52Mbps=


That is to say, a 4.6 ppm frequency in STM-1 equals to:

However, this difference does not affect the whole STM-1 frame, but only the VC-4, and therefore we will only consider the difference of size between the two:

A pointer movement, here, is a decrement of 3 bytes that makes it possible to fit24 more bits from VC-4 in the STM-1 frame. The adjustment period is:

That is, ADM2 decrements the AU-4 pointer every 34.7 ms to compensate for theADM1 drift (see Figure 5.9).

Figure 5.8 Comparison of two reference signals that synchronize two SDH multiplexers. Peri-odical pointer adjustment occurs due to the frequency offset there is between the two signals.

f1

f2

Phase error

STM-1

ADM1

Clock 1

STM-1

Clock 2

GPS

ADM2

VC4

E4

VC4

VC4VC4Pointer decrements

f1 f2

f3 f2 f1– 155.52 1 4.6106--------+

155.52–= = Mbps

f3 155.52 106⋅( ) 1 4.6106-------- 1–+

155.52 4.6× 715.4= = = bps

R VC4( )bytes STM1( )bytes⁄ 261 270⁄ 0.96= = =

fd f3 R× 691.5 bps= =

Tptr 34.7 10 3–× s=

Tptr Decrementbits fd⁄ 24bits 691.5bps⁄= =


5.3.2 Phase Fluctuation

In terms of time, the phase of a signal can be defined as the function that providesthe position of any significant instant of this signal. It must be noticed that a timereference is necessary for any phase measurement, because only a phase relative toa reference clock can be defined. A significant instant is defined arbitrarily; it mayfor instance be a trailing edge or a leading edge, if the clock signal is a square wave(see Figure 5.10).

Here, when we talk about a phase, we think of it as being related to clock sig-nals. Every digital signal has an associated clock signal to determine, on reception,the instants when to read the value of the bits that this signal is made up of. The clockrecovery on reception circuits reads the bit values of a signal correctly when there isno phase fluctuation, or when there is very little. Nevertheless, when the clock re-covery circuitry cannot track these fluctuations (absorb them), the sampling instantsof the clock obtained from the signal may not coincide with the correct instants, pro-ducing bit errors.

When phase fluctuation is fast, this is called jitter. In the case of slow phase fluc-tuations, known as wander, the previously described effect does not occur.

Figure 5.9 The position of the VC-4 container drifts, due to AU pointer adjustments to com-pensate for the differences between the two clocks.

t (msec.)

4.6 ppm

Pointer adjustments

a

a-3

a-6

a-9

a-12

to

Pointer value

to+34.7 to+69.4 to+105.1

Figure 5.10 Phase error of a signal in relation to its ideal frequency.

Clock signal

Significant instants

t0 t1 t2 t3 t4 t5 t6 t7

Bad synchronization


Phase fluctuation has a number of causes. Some of these are due to imperfec-tions in the physical elements that make up transmission networks, whereas othersresult from the design of the digital systems in these networks.

5.3.2.1 Jitter

Jitter is defined as short-term variations of the significant instants of a digital signalfrom their reference positions in time, ITU-T Rec. G.810 (see Figure 5.11). In otherwords, it is a phase oscillation with a frequency higher than 10 Hz. Jitter causessampling errors and provokes slips in the phase-locked loops (PLL) buffers (seeFigure 5.12). There are a great many causes, including the following:

Jitter in regenerators

As they travel along line systems, SONET and SDH signals go through a radio-electrical, electrical, or optical process to regenerate the signals. But clock recoveryin regenerators depends on the bit pattern transported by the signal, and the qualityof the recovered clock becomes degraded if transitions in the pattern are distributedheterogeneously, or if the transition rate is too low. This effect can be countered bymeans of scrambling, which is used to destroy correlation of the user-generated bitsequence. The most commonly used line codes add extra transitions in the pattern,to allow proper clock recovery at the receiving end.

Moreover, this type of jitter is accumulative, which means that it increases to-gether with the increase in the number of repeaters looked at.

Jitter due to mapping/demapping

Analog phase variation in tributary signals is sampled and quantized when these aremultiplexed in a higher-order signal. This is an inherent mechanism in any TDMsystem. In SDH, for instance, every 125 µs, certain bytes of the phase are availablefor adjusting the phase. In short, the phase of tributary signals is quantized.

Also, a tributary signal may be synchronized with a different clock than theclock used to synchronize the aggregate signal that will carry it. The above situationsgive rise to phase justification: Bits of the tributary signal are justified, to align themwith the phase of the aggregate signal frame; that is, creating jitter.

Pointer jitter

The use of pointers in SDH/SONET makes it possible to discard the effects of badsynchronization, but these pointer movements provoke an extensive phase fluctua-tion. Pointer movements are equal to discontinuities in the transported tributaries.


Once the tributary has been extracted, the PLL circuit must continuously adaptitself to bit flows. If the VC-4 pointer has incremented in an STM-1, it will receive24 bits less, and it must slow down to maintain a constant level for its buffer. If bycontrast it has decremented, it will receive 24 bits more and should accelerate. As aresult, the extracted tributary will contain jitter.

5.3.2.2 Wander

Wander is defined as long-term variations of the significant instants of a digital sig-nal from their reference positions in time (ITU-T Rec. G.810). Strictly speaking,wander is defined as the phase error comprised in the frequency band between 0and 10 Hz of the spectrum of the phase variation. Wander is difficult to filter whencrossing the phase-locked loops (PLLs) of the SSUs, since they hardly attenuatephase variations below 0.1 Hz. This is because slow phase variations get compen-sated with pointer adjustments in SDH/SONET networks, which is one of the maincauses of jitter (see Figure 5.11).

Wander brings about problems in a very subtle way in a chained sequence ofevents. First, it causes pointer adjustments, which are then reflected in other parts ofthe network in the form of jitter. This in its turn ends up provoking slips in the outputbuffers of the transported tributary.

t0 t11t5

Figure 5.11 A phase fluctuation of a signal is an oscillating movement with an amplitude and a frequency. If this frequency is more than 10 Hz, it is known as jitter, and when it is less than that, it is called wander.

10

Wander Jitter

t0

t1

t2

t3

t4

t5

t6

t7

t8

t9

t10

t11

Frequency

10310-3 Hz

Amplitude

t


The following are the most typical causes of wander:

Changes in temperature

Variations between daytime and nighttime temperature, and seasonal temperaturechanges have three physical effects on transmission media:

• There are variations in the propagation rate of electrical, electromagnetic or op-tical signals.

• There is variation of length, when the medium used is a cable (electrical or op-tical), due to changes between daytime and nighttime or winter and summer.

• There is different clock behavior when temperature changes occur.

Clock performance

Clocks are classified according to their average performance in accuracy and offset.The type of resonant oscillator circuit used in the clock source and the design of itsgeneral circuitry both add noise, and this results in wander.

5.4 SYNCHRONIZATION OF TRANSMISSION NETWORKS

T-carrier and PDH networks have their first hierarchy perfectly synchronous. In E1and DS1 frames, all the channels are always situated in their own timeslots. Therest of the hierarchical multiplexion levels are not completely synchronous, but fre-quency differences can be accommodated by the bit stuffing mechanism.

T-carrier and PDH nodes do not need to be synchronized, since each of themcan maintain their own clock. The only requirement is that any clock variations must

Loss of clock in PLL

Sampling errors

Input buffer

Empty buffer slip

S&H

PLL

Signal with

Figure 5.12 Jitter and wander affect every stage of data recovery, producing a number of sam-pling errors, clock, losses, and overflow.

Full buffer slip

Network Element

jitter/wander


be kept within the specified limits, so that the available justification bits can be fittedin without problems caused by clock differences.

5.4.1 Synchronization in SONET and SDH

In SONET and SDH, the NEs must be synchronized to reduce pointer movementsto a minimum. Pointer movements, as we have seen, are a major cause of jitter. Thesynchronization network follows a master-slave hierarchical structure:

• Primary reference clock, in SDH, or primary reference source, in SONET: Thisis the one that provides the highest quality clock signal. It may be a cesiumatomic clock, or a coordinated universal time (UTC) signal transmitted via theGPS system.

• Synchronization supply unit, in SDH, or building integrated timing supplies, inSONET: This clock takes its reference from the PRC and provides timing tothe switching exchanges and NEs installed in the same building (it is alsoknown as building synchronization unit) or on the same premises. It is usuallyan atomic clock, although not of such a high quality as the PRC.

• Synchronous equipment clock (SEC): This clock takes its reference from anSSU, although it is of lower quality (for example, quartz). It is the internalclock of all the NEs (multiplexer, ADM, etc.).

Whereas a PRC/PRS clock is physically separate from the SDH/SONET network,an SSU/BITS clock may be a separate piece of equipment, in which case it is calleda stand-alone synchronization equipment, or it may be integrated into an NE (DXCor multiplexer). By definition, an SEC is integrated into an NE. The timing betweenclocks is transmitted by SDH/SONET sections (STM-n/OC-m) or PDH/T-carrierpaths (2 or 1.5 Mbps) that can cross various intermediary PDH/T-carrier multiplex-ing stages, and various PDH/T-carrier line systems. The interfaces for these clocksare 2 or 1.5 Mbps, 2 or 1.5 MHz and STM-n/OC-m, and their presence or absencedepends on the specific implementation of the device.

Table 5.2 Stratum timing accuracy.

Stratum Identifier Accuracy Drift

1 ST1 1 x 10-10 2.523/year

2 ST2 1.6 x 10-8 11.06/day

3 ST3 4.6 x 10-6 132.48/hour

4 ST4 3.2 x 10-5 15.36/minute


5.4.1.1 SONET synchronization network

In a SONET synchronization network, the master clock is called primary referencesource (PRS), whereas slave clocks are building integrated timing supply (BITS)that end up synchronizing the NEs. The GR-1244-CORE specifies the rules andperformance margins for both PRS and BITS.

BITS synchronizes the network equipment, and it is also used by switches. Theperformance required to synchronize a node is Stratum 3 (see Table 5.2).

5.4.1.2 SDH synchronization network

In an SDH synchronization network, the master clock is called primary referenceclock (PRC), whereas synchronization supply units (SSUs) are slave clocks and theNE is a synchronous equipment clock (SEC). All of them must be kept inside theperformance margins defined by the corresponding recommendations (see Table5.3).

5.4.2 Synchronization Models

In SDH/SONET networks, there are at least four ways to synchronize the add anddrop multiplexers (ADM) and digital cross connects (DXC) (see Figure 5.13):

1. External timing: The NE obtains its signal from a BITS or stand-alone syn-chronization equipment (SASE). This is a typical way to synchronize, and theNE usually also has an extra reference signal for emergency situations.

2. Line timing: The NE obtains its clock by deriving it from one of the STM-n/OC-m input signals. This is used very much in ADM, when no BITS or SASEclock is available. There is also a special case, known as loop timing, whereonly one STM-n/OC-m interface is available.

3. Through timing: This mode is typical for those ADMs that have two bidirec-tional STM-n/OC-m interfaces, where the Tx outputs of one interface are syn-

Table 5.3 SDH timing accuracy.

Use Accuracy Drift ITU-T

PRC 1 x 10-11 — G.811

SSU-T 5 x 10-10 10 x 10-10/day G.812

SSU-L 5 x 10-8 3 x 10-7/day G.812

SEC 4.6 x 10-6 5 x 10-7/day G.813


chronized with the Rx inputs of the opposite interface.4. Internal timing: In this mode, the internal clock of the NE is used to synchro-

nize the STM-n/OC-m outputs. It may be a temporary holdover stage after los-ing the synchronization signal, or it may be a simple line configuration whereno other clock is available.

Figure 5.13 Synchronization models of SDH/SONET network elements.

STM-n/OC-m STM-n/OC-m

Primary Reference

External Timing


Other ClockReference

Alternative

Line-external Timing

DerivedClock

BITSSSU

STM-n/OC-m

Other ClockReference

BITSSSU

STM-n/OC-m

Line Timing

STM-n/OC-m STM-n/OC-mSTM-n/OC-m

STM-n/OC-m

Internal Timing

holdover

selector

MUX-DEMUX


PLL

Out1 Ref1 Ref2 Out2

Out1: Alternative ReferenceRef1: Clock Output

Ref2: Derived ClockOut2: Primary Reference

PLL: Phase-Locked Loop

Tributaries O

scil

Filter: Low Pass Filter Clock OutputOscil: Internal Oscillator

Typical model of anADM multiplexer

Through Timing

Loop Timing

filter


5.4.3 Timing Loops

A timing loop is in bad synchronization when the clock signal has closed itself, butthere is no clock, either master or slave, that would autonomously generate a non-deficient clock signal. This situation can be caused by a fault affecting an NE insuch a way that it has been left without a reference clock, and therefore it has cho-sen an alternative synchronization: a signal that has turned out to be the same sig-nal, returning by another route (see Figure 5.14). A synchronization loop is acompletely unstable situation that may provoke an immediate collapse of part of thenetwork within the loop.

The ring network synchronization chain should avoid a synchronization loop(see Figure 5.15).

5.5 DIGITAL SYNCHRONIZATION AND SWITCHING

Digital switching of n x 64-Kbps channels implies that the E1 and T1 frames mustbe perfectly aligned to make it possible to carry out channel exchange (see Figure5.16).

The frames are lined by means of a buffer in every input interface of a switch.The bits that arrive at fi frequency get stored in them, to be read later at the frequencyused by the switch, fo.

But if the clocks are different, |fi - fo| > 0, the input buffer sooner or later endsup either empty or overloaded. This situation is known as a slip: If the buffer be-comes empty, some bytes are repeated, whereas if the buffer is overloaded, somevalid bits must be discarded in order to continue working. That is to say, slips are

Figure 5.14 A synchronization pitfall. The multiplexer A, when left without a reference, should have remained in holdover state, if it did not have another clock signal. Generally, secondary clock references should not be taken in line timing synchronization.

STM-n/OC-m

Multiplexer A

Multiplexer B

STM-n/OC-m

Multiplexer A

BITSSSU

Multiplexer B

1. Normal Operation 2. Timing loop

BITSSSU

Alternative Alternative

GPS GPS


errors that occur when PLLs cannot adapt themselves to clock differences or phasevariations in frames.

where86,400 is number of seconds per dayn: bits repeated or discarded per slipfi = input bit ratefo = output bit rate

When effects are caused by slips:

• In the voice they are usually not noticed; a click may be noticed when voice issent compressed;

• In a facsimile they may damage many text lines;

• In modems they cause microbreaks and may sometimes break the whole con-nection;

• In digital TV, there is loss of color or frame synchronization;

• In data networks like SNA, HDLC, frame relay, TCP/IP, there is loss of perfor-mance.

STM-n/OC-m

PRSPRC

Alternative

Alternative

Figure 5.15 The ring network synchronization chain. “1” is the primary reference, “2” and “3” are alternative clocks, and “0” is to avoid a synchronization loop.

0

STM-n/OC-m

DerivedClock

BITSSSU

1

2

3

0

11 2

1

0

2

1

0

GPS1 2

Alternative21

fd 86 000, fi fo– n⁄×= slips day⁄( )


5.6 SSU IN A SYNCHRONIZATION NETWORK

The SSU is in charge of synchronizing all the NEs of its node. It has many alterna-tive clock inputs or references, to confront possible clock signal losses. It may beintegrated in an ADM or CXC multiplexer, or it can be a stand-alone equipment, inwhich case it is known as SASE (see Figure 5.17).

Figure 5.16 Synchronization of two digital centrals: (a) by signal derived from the PDH chain; (b) by PDH and SASE chain; and (c) across SDH network.

E4

Switch A

Clock

E1

Switch B

Multiplexer A Multiplexer BE1

E4

Switch A

SASE

E1

Switch B

Multiplexer A Multiplexer BE1

Other clockReferenceDerived

ClockBITSSSU

Primary Reference

Switch A

SASE

E1

Primary Reference

STM-n/OC-m

BITSSSU

Switch B

(a)

(b)

(c)


Depending on their performance, there are two types of SSUs:

• Synchronization supply unit transit (SSU-T): These are of higher quality andthey are used to synchronize NEs, or as references for other SSUs.

• Synchronization supply unit local (SSU-L): These are of lower quality, andthey only synchronize the NEs of their own node.

5.6.1 Functions of SSU

An SSU has many functions, and they can be described as follows:

1. The SSU accepts many clock references, tests their performance and selectsone of them, filtering it from noise and other interference.

2. It sends the signal chosen to an internal oscillator that acts as a reference togenerate a new synchronization signal.

3. The new signal is distributed between all the NEs of its node, and it may alsobe sent to another SSU in another node.

4. If the reference chosen starts to degrade or is lost, the SSU should switch to oneof its alternative references.

5. If no valid reference is found, the SSU enters holdover mode, generating aclock of its own that emulates the characteristics of the previous valid refer-ence.

In the case of an SASE, there are other functions as well:

Figure 5.17 Diagram of an SSU function model.

holdover

filter

Ref1 Ref2 Ref2

TMN control

NE

mon

itorin

g

Feedback selector

NE

Feedback


1. It monitors the synchronization status of the NEs of its node by means of returnlinks.

2. It continuously informs the TMN control level of both its own synchronizationstatus and that of the NEs of its node.


• ITU-T Rec. G.703, Physical/electrical characteristics of hierarchical digital interfaces.

• ITU-T Rec. G.781, Synchronization layer functions.

• ITU-T Rec. G.783, Characteristics of synchronous digital hierarchy (SDH) equipment functionalblocks.

• ITU-T Rec. G.803, Architecture of transport networks based on the synchronous digital hierarchy(SDH).

• ITU-T Rec. G.810, Definitions and terminology for synchronization networks.

• ITU-T Rec. G.811, Timing characteristics of primary reference clocks.

• ITU-T Rec. G.812, Timing requirements of slave clocks suitable for use as node clocks in synchroni-zation networks.

• ITU-T Rec. G.813, Timing characteristics of SDH equipment slave clocks (SEC).

• ITU-T Rec. G.822, Controlled slip rate objectives on an international digital connection.

• ITU-T Rec. G.823, The control of jitter and wander within digital networks which are based on the2048 kbit/s hierarchy.


• ITU-T Rec. G.825, The control of jitter and wander within digital networks which are based on thesynchronous digital hierarchy (SDH).

• ITU-T Rec.G.8251, The control of jitter and wander within the optical transport network (OTN).

• José M. Caballero and Andreu Guimerà, Jerarquías Digitales de Multiplexión, PDH y SDH, Sincroni-zación de Redes, L&M Data Communications 2001.

207

Chapter 6

Test and Measurement

There are certain key words that seem to crop up again and again when talkingabout measurement applications: Research and development (R&D), manufactur-ing, approval, acceptance, installation, bringing-into-service, maintenance, andmonitoring.

These key words describe the general areas of application for telecommunica-tions test and measurement instruments. They usually refer to actions carried out onthe equipment in SDH networks, known as network elements: multiplexers, cross-connects, regenerators, and add and drop multiplexers (ADMs).

These areas of application define a series of tests and measurements, some lim-ited to specific areas and others that are used across a wider range. The followingsections describe some of the most important and most common of these measure-ments.

6.1 AREAS OF APPLICATION FOR TEST AND MEASUREMENT

In the area of R&D and the manufacture of NEs, these instruments can be used aslaboratory generators for carrying out measurements on prototypes of the NEs, orparts of elements.

From time to time, telecom operators may wish to approve a specific NE, whichwill be used from then on in its SDH networks (i.e., they will decide which brandand which model to use). In this case, they will compare instruments from a numberof different manufacturers with a view to choosing one. This often requires the helpof specialist external consultants, given the technical complexity of the tests (manysmall operators do not know how to approve instruments) or simply because it is ex-pensive (the operator is not in a position to have its own staff perform these tests).


NEs, once they have been purchased, must undergo a series of tests in order toconfirm that they are working properly before they can be accepted by the buyer.These are acceptance tests, defined by laboratories and carried out by engineersworking for the purchasers. The difference between them and the approval processis that acceptance tests are performed on every one of the pieces of equipment re-ceived, all of the same brand and model.

These acceptance tests, performed by the purchaser of the equipment, usuallygo hand in hand with installation tests, which are performed by the installer. Bothtypes of tests are often defined together, and for this reason the documents wherethey are described refer to them as installation/acceptance tests.

Certain NEs, along with the transmission media to which they are connected(i.e., cables or radio links), go together to make up clearly defined parts of the SDHnetworks known as paths. These paths must be tested when they are brought into op-eration for the first time, and for this reason a series of bringing-into-service tests isdefined. Once an SDH (or PDH) network has been installed and brought into service,it must be maintained to guarantee its operation for its users. This implies carryingout repairs when faults occur, and also performing routine checks on the networkwhen everything is working correctly. For this reason, maintenance tasks are per-formed that also require testers to locate faults and test parts of the network that areworking properly.

One final area of application is that of monitoring. The monitoring of an SDHnetwork is normally carried out using instruments connected to its elements thatmeasure over a long period of time to check that the network is working properly,and these can usually be operated remotely. This means that they must be able towithstand long measurement times (days or even months) without the operating sys-tem failing, and, even if the electrical current fails, the instrument must still be ca-pable of restarting the measurement at the point where it left off.

6.2 TESTS IN THE INTERFACES

An interface could be defined as the “boundary” between two associated systems.When this boundary is defined between two communication devices, it is known asa physical interface. A physical interface is specified in terms of the mechanical,electrical, electromagnetic, and optical characteristics of the connections and inter-actions between two associated pieces of equipment in the interface.

Tests in (physical) interfaces are aimed at checking the characteristics definedby the recommendations for each interface, be it an optical or electrical interface.

Test and Measurement 209

6.2.1 Line Interfaces

The PDH and SDH line interfaces1 under consideration are shown in Table 6.1.

In this section on interface tests, we will also look at 64-Kbps analog and digitalinterfaces, to mention the measurements that affect them and why they operate astributary signal interfaces in some types of multiplexers.

6.2.2 Connection Modes in Electrical Interfaces

Connection mode, in this context, refers to the different ways of electrically con-necting equipment to the network in order to analyze it. The electrical connectionmodes looked at are termination mode, high impedance mode, and mode for con-nection to protected monitoring points, (PMP).

6.2.2.1 Termination

The test and measurement instrument terminates the transmission line, in such away as to adapt impedance and transfer maximum power to the instrument (see Fig-ure 6.1).

6.2.2.2 High impedance

This is an electrical connection method whereby the test instrument is connectedparallel to the line by means of a short stretch of line, in such a way that it does notoverload the main line. This is a common connection mode at 2 Mbps, a connectionto distribution frame with T-connectors (see Figure 6.1).

Table 6.1PDH and SDH line interfaces.

Interface Rate Recommendation

Optical STM-1/OC-1 (155M) G.957STM-4/OC-12 (622M) G.957STM-16/OC-48 (2.5G) G.957STM-64/OC-192 G.691

Electrical 2M G.703, G.823, G.704, G.706, G.80334M G.703, G.823, G.742, G.751140M G.703, G.823, G.742, G.751STM-1 (155M) G.703, G.825

1. Intraexchange interfaces also exist, but here we shall only be observing line inter-faces, from one exchange to another.


6.2.2.3 Protected monitoring points

A protected monitoring point (PMP) provides a digital interface to connect mea-surement devices in order to analyze the digital signal transmitted without causingdegradation. PMPs have been defined for the electrical interfaces of the PDH andSDH STM-1 (155 Mbps) hierarchy. These can be found in the plant equipment orin the digital distribution frame. They introduce attenuation in the signal, dependingon the location and the hierarchical rate being looked at. For STM-1, the attenua-

Transmission line Rx

Tx

Transmission line Rx

Tx

Test instrumentl: length of section of connection line; must be short(If l = 0 m a high Z won’t overload the line)

High impedance (Z>>1 kΩ)(Z of the line is usually 75 kΩ)

(a) Termination mode

(b) High impedance mode

Transmission line

Rx

Tx

Test point (attenuated signal)

High Z

Figure 6.1 Connection modes for signal analysis in electrical interfaces.

(c) Measurement at Protected Monitoring Points

Test instrument

Test instrument


tion for points in the plant equipment can reach 20 dB relative to the output of aG.703 interface or 12 dB in distribution frames (see Figure 6.1).

6.2.3 Measurements in Electrical Interfaces

This section covers measurements pertaining to low-rate electrical interfaces: ana-log voice channel interfaces and digital interfaces under 2 Mbps. These interfacesrepresent input and output ports for tributary signals in multiplexing equipment.Measurements of interfaces at higher rates are dealt with in more detail in separatesections, owing to their greater complexity and the fact that they are not dependenton the type of work signal (i.e., on whether it is optical or electrical).

6.2.3.1 Analog interface

Measurements on analog channels must still be considered, owing to the fact thatanalog subscriber loops and applications, such as ADSL, are still quite widespread,and even the installation of ISDN access points reuses previously installed loops.

Analog channel parameters are quantified in measurements between analog in-put and output interfaces (known as full channel measurements), and in measure-ments between analog and digital interfaces (half-channel measurements). Anexample of half-channel measurement can be performed on a PCM multiplexer bygenerating an analog test signal at its input, and analyzing the 2-Mbps digital signalat its output (the measurement instrument will internally demultiplex the digital sig-nal received, to access the test signal to be analyzed).

Examples of measurements made on analog channels are:

• Signal strength in the interface;

• Peak code (-127 to +127 in line with law A ITU-T Rec. G.711; A/D conver-sion);

• Offset in the A/D coder;

• Distortion in group delay;

• Crosstalk;

• Noise in vacant channels.

6.2.3.2 Digital interface

When dealing with digital interfaces in the multiplexers, BER measurements aremade by inserting and analyzing a test signal in the digital interface, usually calledthe data interface (see Section 6.3.1). In this case the multiplexer output closed in a


loop (either near end or far end, depending on the coverage desired for the measure-ment).

If an excessive error rate is detected, the next step is to locate the source; for in-stance, a test can be carried out on the local multiplexer. When the source is not quiteso apparent, it then becomes necessary to carry out a long-term BER measurement.This way it is possible to determine if the anomalies are occurring continuously orin bursts, for example, a reversal of current in a power source close to the line maycause anomaly bursts.

6.2.3.3 Specifications for electrical interfaces

These specifications are usually measured during design and manufacture, with theaim of ensuring compliance with ITU-T recommendations on interfaces (both net-work node interfaces and user interfaces). The parameters (specifications) to bechecked are:

• Pulse shape: These are usually specified by a mask. The aim of maintaining aspecific form is to avoid intersymbol interference. This measurement is carriedout using an oscilloscope.

• Power and impedance: These are parameters that are usually measured alongwith the pulse form. The impedance is specified as 75Ω with coaxial cable, ex-cept for the primary PDH hierarchy of 2 Mbps, in which balanced connectionsof 120Ω are also allowed.

• Line coding: This is measured implicitly by PDH and SDH measurement de-vices (see Table 6.2).

Table 6.2Line coding test.

Bit Rate (Kbps) Usual Line Coding

64 Codirectional, contradirectional (G.703)2,048 HDB38,448 HDB334,368 HDB3139,264 CMI155,520 CMI or optical622,080 Optical2,488,320 Optical9,953,280 Optical


• Clock frequency: This is a series of tolerances that exist for the clock frequen-cies of the hierarchical signals (see Table 6.3), and are clearly defined in ITU-Trecommendations (see Chapters 5 and 10).

• Framing: This is the extent to which the frame of the signal received conformsto that defined by the corresponding recommendation; it is another of the mea-surements carried out automatically by the PDH or SDH test set. Conversely, asignal can be generated with errors in the frame, to check whether the deviceunder test (DUT) detects this event.

• Jitter and wander: Jitter is the fast phase fluctuation and wander slow phasefluctuation affecting the output interfaces (see Chapter 10).

6.2.4 Measurements in Optical Interfaces

The measurement of optical parameters is required during the bringing-into-serviceand maintenance of sections at STM-4 and higher. Above STM-4, the line interfacespecified by ITU-T recommendations is solely optical (unlike the first level,STM-1, which also allows electrical interfaces). As in the previous section, refer-ence is made here to measurements that are more concerned with the physical prop-erties of the signals in the interfaces, leaving for later sections those that do notdepend on the nature of the signals used, but are more concerned with the transmis-sion itself.

6.2.4.1 Optical power and dynamic range

Essentially, this measurement is carried out by placing a power meter in the aggre-gate output port of the corresponding NE. The measurement of optical output pow-er (the strength of the STM-n optical signal) basically checks compliance with

Table 6.3Clock frequency test.

Bit Rate (Kbps) Tolerance (ppm)

64 1002,048 ±508,448 ±5034,368 ±30139,264 ±15155,520 < 4.6622,080 < 4.62,488,320 < 4.69,953,280 < 4.6


specifications, and is important in relation to what is known as the dynamic rangeof the system.

The dynamic range is the difference between the saturation level and the sensi-tivity of the receiver, that is, between the maximum and minimum power levels thatcan be coped with. The dynamic range must be as wide as possible; otherwise, thepeak-to-peak amplitude of the pulses transmitted will be low and noise will easilyproduce transmission errors. The following section describes a method for determin-ing sensitivity. Between this parameter and that of saturation (see Section 6.2.4.3),the dynamic range of the system will be determined.

6.2.4.2 Measuring receiver sensitivity

Measuring optical sensitivity allows an evaluation to be made of how the NE re-sponds to a fall in the optical power of the line signal. When optical power falls be-low a certain limit, the level of noise above the signal starts to become significant,causing bit errors. In this case, the NE being evaluated is an ADM, and we want tomeasure the optical sensitivity of its STM-n aggregate ports. One procedure forevaluating this parameter is as follows (see Figure 6.2):

An optical attenuator is placed in the line, closing the side to be measured (forinstance, the East aggregate) in a loop. A test signal is then introduced in a tributaryof the STM-n signal, which will move through the loop and be analyzed on recep-tion. Little by little, the attenuation introduced in the aggregate signal is increased,with its value being recorded under three circumstances:

• When the first bit errors appear in the test tributary;

• When the BER is higher than 10-9;

• When the BER is higher than 10-3.

Tributary under test

... ...

OC-m OC-m

W E

Optical attenuator

Figure 6.2 Optical sensitivity of the receiver.

BER measurement


6.2.4.3 Detecting optical overload

An optical transmitter must not be connected directly to an optical receiver, beforethe necessary precautions are taken. This is because the optical receiver has a limitin terms of the maximum power it can receive, and this must be respected. Other-wise, optical saturation or overload occurs, which makes it impossible to correctlydetect the light pulses. When the overload is particularly great, the optical receptiondevice may be damaged. ITU-T Recs. G.957 and G.691 specify the minimum levelof optical power (dBm) for overload to occur in each of the optical interfaces de-fined for SDH. An indication of overload through a light emitting diode (LED) de-vice is often available in test instruments.

Nonetheless, an optical receiver must receive a minimum amount of opticalpower if it is to perform the aforementioned detection correctly. RecommendationsG.957 and G.691 also define this level, which is the sensitivity, for each SDH inter-face. In short, the power that reaches any optical receiver must be at a level some-where between the sensitivity of the receiver and its overload level. When the powerreceived is equal to or greater than the overload level, an optical attenuator must beinserted between the transmitter and receiver. Usually, this will be necessary if theoptical transmitters and the optical receivers are not of the same type, for long-haul,short-haul, or intraoffice signals in accordance with ITU-T Rec. G.957 (see Table6.4).

6.2.5 Measuring Frequency

Low-quality synchronization sources that deviate from the nominal value of thesignal they supply, or badly synchronized clock recovery circuits, can give rise toproblems in the operation of the NEs. For this reason, it is necessary to measure the

Table 6.4Classification of optical interfaces based on application and showing application codes.

Application IntraofficeI n t e r o f f i c e

S h o r t H a u l L o n g H a u l

Source NominalWavelength (nm) 1,310 1,310 1,550 1,310 1,550

Type of Fiber G.652 G.652 G.652 G.652 G.652G.654 G.653

Distance (km) ≤ 2 ∼ 15 ∼ 40 ∼ 80

Leve

l

STM-10 I-10 S-1.10 S-1.20 L-1.10 L-1.20 L-1.30

STM-40 I-40 S-4.10 S-4.20 L-4.10 L-4.20 L-4.30

STM-16 I-16 S-16.1 S-16.2 L-16.1 L-16.2 L-16.3


frequency of the line signal at all hierarchical levels, and check its deviations rela-tive to the margins established by the ITU. Data signal clock frequency and its off-set relative to the nominal value are thus checked to obtain an indication of theprobability of anomalies. To perform a frequency measurement in service, the mea-surement instrument must be connected to a PMP or through a probe in the line (seeFigure 6.3).

The SDH/PDH instrument must be able to measure the frequency of the signalsreceived (and, in some cases, the associated line codes are also measured). The fre-quency of the signal is usually given in Hz, and its deviation relative to the nominalhierarchical value in ppm. Likewise, it shows whether the frequency measured is in-side or outside the range defined by the ITU. The precision of the measurement isdetermined by the synchronization reference used (G.811 or G.812 clock).

6.3 IN-SERVICE AND OUT-OF-SERVICE MEASUREMENTS

An out-of-service (OOS) measurement can be defined as one in which the part ofthe network on which the measurement is to be performed is disconnected from therest of the network, and therefore does not provide a service to the user. In contrast,an in-service (IS) measurement is one that can be performed without disconnectingthe part of the network to be measured from the rest of the network, and the mea-surement is therefore performed alongside the provision of normal service to theuser.

E3E2

E4

OC-3

E1

Attenuating probe

Figure 6.3 Measuring frequency.

ADM

ADM

ADM

ADM

STM-64OC-192

MUX

MUX

MUX


From the above definitions it is clear that for OOS, the traffic of information issimulated (a generator is required), whereas in ISM, the traffic is live (only an ana-lyzer is used). The quantification of the BER forms the basis of OOS, whereas othermechanisms are used for ISM, as will be seen in later sections.

6.3.1 Bit Error Rate

The BER is the relation between the number of errored bits received and the totalnumber of bits received by an analyzer (see Figure 6.4). To obtain this figure, a biterror rate meter is required which must generate a test signal using its generationsection. This signal is a PRBS that is in accordance with the rate of the interface be-ing measured in line with ITU-T Rec. O.150 to O.153 (up to 140 Mbps) for teststructures in PDH, and O.181 for test structures in SDH. This sequence is sent viathe device or system being tested (DUT).

At the receiving end, a bit rate meter receives the signal sent, extracts its clock,and generates the PRBS locally in accordance with this clock. Both sequences (theone received and the one generated locally) are compared bit by bit to determine thenumber of erroneous bits received. Since the length of the sequence is known, thetotal number of bits is also known, and the bit error rate can therefore be established.

6.3.2 Out-of-service Measurements

OOS require the live traffic in a link to be replaced by a known test signal, normallya PRBS, and the correct reception of this signal is then checked at the remote end ofthe communication (see Figure 6.5). This correct reception is quantified by measur-ing the BER of the test signal used. These tests are intrusive, that is, they interruptservice if they are applied to networks in operation, but they provide exact mea-surements since the test signal received is checked bit by bit.

ClockError

counter

DUT

Clock

Data

Figure 6.4 Basic diagram for measuring bit error rate (BER).

PRBS Generator

Local PRBSgenerator

Comparison

COD

DECOD

BER= total of bits received error bits received


6.3.2.1 Test sequences and structures

Mapping between the PRBS used and the transmission rates for data and PDH isdefined by ITU-T Rec. O.150. Each PRBS is identified by its length, expressed as 2x -1 (see Table 6.5).

As the table shows, there are many variants between sequences of the samelength (as is the case of 220-1). This gives rise to different sequences with regard tothe disposition of their zeros and ones, in relation to the different feedback mecha-nisms used in generating these sequences through the offset registers. In the case ofthe SDH hierarchical rate, the test sequences are applied in certain SDH structures,as defined in ITU-T Rec. G.707, thus forming test signal structures (TSSs). These

Table 6.5Test sequences for PDH and ANSI signals and data (O.150).

Length Consecutive Zeros Application

29-1 8 Error measurements on data circuits at bit rates up to14,400 bps

211-1 10 Error and jitter measurements at bit rates of 64 Kbps and n x 64 Kbps

215-1 15 Error and jitter measurements at bit rates of 1,544; 2,048; 6,312; 8,448; 32,064; and 44,736 kbps

220-1 19 Error measurements on data circuits at bit rates up to 72 Kbps

220 -1 14 Error and jitter measurements at bit rates of 1,544; 6,312; 32,064; and 44,736 Kbps

223-1 23 Error and jitter measurements at bit rates of 34,368 and 139,264 Kbps

229-1 29 Special measurements, such us delay at higer bit rates

231-1 31 Special measurements, such us delay at higer bit rates

Figure 6.5 Out-of-service measurement (OOS): Example of application.

Loop

Alternative:end-to-end

measurement

STM-n/OC-m


structures are defined in ITU-T Rec. O.181 for OOS measurements on STM-n sig-nals (see Table 6.6).

The functions for the test applications of NEs from the above table are definedas follows:

• The higher-order path connection (HPC) function allows a flexible allocationof the higher-order virtual containers or VCs, VC3/VC4.

• The lower-order path connection (LPC) function allows a flexible allocation ofthe lower-order virtual containers.

• The lower-order path adaptation (LPA) function allows for a plesiochronoussignal to be adapted to an SDH network by establishing/eliminating the map-ping of the signal inside/outside a synchronous container.

• The concatenation function VC4-Xc or VC2-Xc a bandwidth X times biggerthan VC-4 or VC2 (see Section 2.9).

Table 6.6Test signal structures (O.181)

PRBS Where Structure Application

223-1 To all bytes of a HO C4 TSS1 Test on NE providing HPC and using an AU-4 structure

215-1 To all bytes of an HO C3 TSS2 Test on NE providing HPC and using an AU-3 structure

223-1 To all bytes of an LO C3 TSS3 Test on NE providing HPC and LPC

215-1 To all bytes of a lower-order con-tainer (C-2, C-11, or C-12)

TSS4 Test on NE providing HPC and LPC

223-1 To all the PDH tributary bits mapped in a C-4

TSS5 Test on NE providing only LPA-4 func-tions and using an AU-4 structure

215-1 To all PDH tributary bits mapped in a HO C-3


223-1 To all PDH tributary bits mapped in a LO C-3


215-1 To all PDH tributary bits mapped in LO (C-2, C-11, or C-12)

TSS8 Test on NE providing LPA-m functions (m=11, 12, 2)

223-1

231-1

To all payload bytes of the C-4-Xc concatenated container

TSS9 Test on NE providing VC4-Xc contigu-ous concatenation

215-1 To all payload bytes of the C-2-Xc concatenated container

TSS10 Test on NE providing VC2-Xc contigu-ous concatenation


The TSSs for SDH can be shown graphically by using two-dimensional maps,in line with the usual way of representing SDH frames (see Figure 6.6).

1. TSS1 and TSS5, in which the PRBS 223-1 is applied, with or without mapping,to the C-4.

2. TSS2 and TSS6, in which the PRBS 215-1 is applied, with or without mapping,to the C-3s.

3. TSS3 and TSS7, in which a PRBS 223-1 is applied, with or without mapping,to a lower-order C-3.

4. TSS4 and TSS8, in which the PRBS 215-1is applied, with or without mapping,to a lower-order container (C-2, C-11, or C-12).

5. TSS9 in which the PRBS 223-1 or 231-1 is applied to all payload bytes of the C-4-Xc concatenated container. TSS10, in which the PRBS 215-1 is applied to allpayload bytes of the C-2-Xc concatenated container.

6.3.2.2 OOS modes

The ITU-T Rec. O.181 defines a series of OOS modes. In these, a path is estab-lished through the entity under test (EUT). An appropriate test sequence is then ap-plied to the input on one of the sides of the EUT. The information received isanalyzed in an access point on the same side of the EUT or on the other side. Exam-ples of EUTs are regenerator sections, multiplexer sections, paths, and so on. Tobetter understand the different modes, it is first necessary to remember some basicconcepts:

• End-to-end transparency: Transparency exists between two communicationends when the signal sent from one to the other is transmitted without any bina-ry change. Each bit of the transmitted signal can take on any value at the inputof the EUT.

• Plesiochronous tributary mapping: Mapping is defined as the process by whicha plesiochronous signal is introduced in a synchronous C-n container suitablefor it.

OOS modes are divided into modes with end-to-end transparency, and modes withplesiochronous mapping (see Table 6.7). The TSS to be used and the events (anom-alies: errors, and defects: alarms) to monitor are defined for both of these modes.The following can be put forward as an example for the different measurementmodes.


RSOH

MSOH

AU4

Without mapping: TSS1

POH PRBS with/without

With mapping: TSS5

mapping in C-4

Figure 6.6 Graphical representation of the structures TSSi.

C-4

(a) Structures TSS1 and TSS5

(b) Structures TSS2 and TSS6

RSOH

MSOH

3xAU3

Without mapping: TSS2With mapping: TSS6

C-3

POHPRBS with/withoutmapping in C3

RSOH

MSOH

AU

POHPRBS with/withoutmapping in

TUG2TUG2

TUG2

C2, C11, C12


RSOH

MSOH

AU4C-3

2PRBS with/withoutmapping in C3

TUG3TUG3

TUG3

(c) Structures TSS3 and TSS7

(d) Structures TSS4 and TSS8


C-12

C2c

RSOH

MSOH

AU

(e) Structures TSS9 and TSS10

POH

POH

C4c

PRBS mappedPOHPOH

Fixe

d St

uff

in C4-Xc or C2-Xc


Table 6.7Modes for out-of-service measurements, in line with O.181.

Mode Structure Events to be Monitored

High

er-o

rder

C-4

end

to en

d tra

nspa

renc

y

STM-n regen-erator section

TSS1, TSS5

OOF, B1 errors, TSELOS, LOF, RS-TIM, LSS

STM-n multi-plex section

TSS1, TSS5

OOF, B2 errors, MS-REI, TSELOS, LOF, MS-AIS, MS-RDI, LSS

Higher-order VC-4 path

TSS1, TSS5

OOF, B3 errors, HP-REI, TSELOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, LSS, HPTC RDI, HPTC-LTC

Hig

her-o

rder

C-3

end

to en

d tra

nspa

renc

y


TSS2, TSS6

OOF, B1 errors, TSELOS, LOF, RS-TIM, LSS


TSS2, TSS6

OOF, B2 errors, MS-REI, TSELOS, LOF, MS-AIS, MS-RDI, LSS


TSS2, TSS6

OOF, B3 errors, HP-REI, TSE, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LSS

Lower-order C-3 end to end transparency

TSS3, TSS7

OOF, B3 errors, LP-REI, TSE, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC-RDI, HPTC-LTC, LP-RDI, TU-LOP, TU-AIS, LP-TIM, LPTC-TIM, LPTC RDI, LPTC-LTC, LSS

Lower-order container (C-11/C-12/C-2) end-to-end transparency

TSS4, TSS8

OOF, BIP-2 errors, LP-REI, TSELOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LP-RDI, TU-LOP, TU-AIS, TU-LOM, LP-TIM, LPTC-TIM, LPTC RDI, LPTC-LTC, LSS

Tributary mapping in higher-order C-4

TSS5 OOF, B3 errors, HP-REI, TSE, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC-TC-RDI, HPTC-LTC, LSS

Tributary mapping in higher-order C-3

TSS6 OOF, B3 errors, HP-REI, TSE, LOS, LOF, MS-AIS, MS-RDI, AU-LOP (Note), AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LSS

Tributary mapping in lower-order C-3

TSS7 OOF, B3 errors, LP-REI, TSE, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LP-RDI, TU-LOP, TU-AIS, LP-TIM, LPTC-TIM, LPTC RDI, LPTC-LTC, LSS

Tributary mapping in lower-order container (C-11/C-12/C-2)

TSS8 OOF, BIP-2 errors, LP-REI, TSE, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LP-RDI, TU-LOP, TU-AIS, TU-LOM, LP-TIM, LPTC-TIM, LPTC, RDI, LPTC-LTC, LSS

Contiguous concate-nated VC-2-Xc

TSS9 OOF, BIP-2 errors, LP-REI, TSE, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LP-RDI, TU-LOP, TU-AIS, TU-LOM, LP-TIM, LPTC-TIM, LPTC RDI, LPTC-LTC, LSS

Con

tiguo

us c

onca

-te

nate

d V

C-4-

Xc


TSS10 OOF, B1 errors, TSE, LOS, LOF, RS-TIM, LSS


TSS10 OOF, B2 errors, MS-REI, TSELOS, LOF, MS-AIS, MS-RDI, LSS


TSS10 OOF, B3 errors, HP-REI, TSELOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, LSS, HPTC RDI, HPTC-LTC


OOS of anomalies in a regenerator section

This basically consists of checking the integrity of the test structures when thesecross a regenerator section. The test is performed with a single instrument that is agenerator/analyzer, and for this reason a loop must be set up at the other end.

OOS of anomalies in a multiplexer section

This checks the integrity of the test structures when these are circulating through amultiplexer section that contains several generators. As in the previous case, a sin-gle generator/analyzer is used at one end of the section, and a loop at the far end(see Figure 6.7).

OOS with end-to-end transparency in higher-order containers

This checks the integrity of the test structures that cross various multiplexer sec-tions. The sample shows an example where the sections are defined between dis-tributors-multiplexers (see Figure 6.8). As in the previous examples, themeasurement is performed from one end to the other in a loop.

Line or Multiplexer Section (MS)

Figure 6.7 End-to-end transparency testing between multiplexers and regenerators.

Regenerator

Regenerator Section (RS)

REG

STM-nOC-m

REG

Regenerator

REG

STM-nOC-m

REGMUX

Multiplexer


OOS with end-to-end transparency in lower-order containers

In this case, similarly to the previous one, the NEs that delimit the multiplexer sec-tions are ADMs.

Anomaly measurement with PDH tributary mapping in a higher-order container

This checks the integrity of the test structures used when mapping takes place inhigher-order containers. To include this mapping process, the PDH tributary portsof the NEs (in this case a digital cross-connect or DXC) are set up in a loop thatprovides higher-order access ports for PDH tributaries (see Figure 6.9).

HPCDXC

HPCDXC

TTF TTF TTF TTF

HOI HOI HOIHOI HOIHOIHOI

STM-n/OC-m STM-n/OC-m STM-n/OC-m

Figure 6.8 OOS anomalies: End-to-end transparency in high order and in low order.

LPC

DXC

LPCDXC

HOA HOA HOA HOA

HOI HOI HOILOI LOILOILOI

STM-n/OC-m STM-n/OC-m STM-n/OC-m

HPC HPC

TTF TTF TTF TTF

TTF: Terminal transport function HPC: Higher-order path connectionHOA: Higher-order adaptationHOI: Higher-order interfaceLOI: Lower-order interfaceLPC: Lower-order path connection

(a) Higher Order

(b) Lower Order


Anomaly measurement with PDH tributary mapping in a lower-order container

This checks the integrity of the test structures used when mapping takes place inlower-order containers. To include this mapping process, the PDH tributary ports ofthe NE, in this case an ADM, are set up in a loop that provides lower-order accessports for PDH tributaries (see Figure 6.9).

6.3.3 In-Service Measurements

When an OOS measurement is made, it causes an interruption in the use of the sec-tion or path being measured, thus diminishing, albeit temporarily, the network ca-pacity. This is bad for the operators, who must guarantee not only the quality of thenetwork, but also its availability to the user. The increase in leased lines has accen-tuated this need. For this reason, preventive maintenance must go hand in hand withthe availability of the transmission links. Preventive maintenance is therefore basedon ISM; that is, measurements that do not interrupt network traffic.

ISMs are based on checking anomalies in fixed or permitted bit patterns in thelive traffic made up of the user data flow (for instance, the FAS) or in the checksumin predefined data blocks. Some of the measurements are applicable to paths, sincethe parameters are not restarted in an intermediate network interface. Others are onlyuseful at line or section level. These measurements allow for long-term network per-formance monitoring and preventive maintenance without interrupting user traffic(see Figure 6.10).

HPC

Figure 6.9 Examples of out-of-service error performance measurement (a) with plesiochronoustributary mapping in a higher-order container and (b) in a lower-order container.

ADM

TTF TTF

HOIHOIHOI

STM-n/OC-m

LPC

STM-n/OC-m

HPC

(a) Higher Order (b) Lower-order

T3 or E4

HOA HOA

HOILOILOI

TTF TTF

T1 or E1


Another advantage of ISMs is that there is no difficulty in performing them overlong periods of time, which means that the statistics derived from these measure-ments are much more reliable. Periods of several weeks can therefore be analyzed,and, during this time, the ISM parameter values will be stored. For PDH, these pa-rameters may be CRC-4 block errors, FAS errors, HDB3 code errors, E-bit errors(REBE), and alarm histories.

In addition to providing more reliable anomaly monitoring statistics, long-termmeasurements have the advantage of making it easier to capture sporadic bursts ofanomalies. They also make it easier to check compliance with the overall perfor-mance specifications of the circuits being measured. To analyze these long-termmeasurements, it is useful to represent them as histograms.

The mechanisms used by ISMs to evaluate network performance are:

• Detection of line code anomalies: These are detected in each individual section(in the one closest to the receiver). When the signal is regenerated, it is recodedso that the code anomalies are not transmitted to subsequent sections. For thisreason, this mechanism is of no use for reporting on the performance of a com-plete path, unless what is being looked at is line equipment for 2 Mbps (still in-stalled in some cases) that does not recode.

• Detection of anomalies in FAS: These make up a fixed part of the frame that,because it is known, can be checked bit by bit. The checking of anomalies inthe FAS provides an estimate of the BER, when the measurement is performedover a long period of time.

Figure 6.10 In-service measurements.

STM-n/OC-m

......

Attenuatingprobe


• Detection of parity anomalies: These anomalies are detected by calculating theodd or even parity in groups of bits known as data blocks. The method of parityanomaly detection is used in SDH via bit interleaved parity (BIP) codes. In-service performance measurements for PDH and SDH signals are based onITU-T Rec. G.826 and use the BIP codes as a monitoring method. These per-formance measurements will be looked at in detail below (see Figure 6.11).

• Verification of CRC: The integrity of predefined data blocks is checked bysending a checksum to these blocks that is recalculated locally at the receivingend. The result of the calculation is compared to the checksum received, tocheck that there is no discrepancy. If there is, the data block received will be in-terpreted as being erroneous. This is used in PDH signals at 2 Mbps (E1),among other applications.

Other practices also exist for evaluating the performance of systems in service:

• Monitoring events at all levels of the signal: This consists of measuring thetraffic signal using an instrument with a demultiplexing capacity that allows itto identify the events produced and their level (that is, the tributary signal inwhich they are produced, where appropriate). In a PDH signal, for instance, thedefects produced in the tributary could be identified by measurement at pointswhere the only information available is that relative to the transmission rate (asis the case with regenerators).

B1

B2

1 2 3 4 5 6 7 8

Figure 6.11 In-service measurements, BIP-n.

OC-mOC-3SPE

VC11C11

C3

SPEOC-3OC-m

BIP2-LO POH byte V5BIP8-HO POH byte B3BIP24-MSOH byte B2

BIP8

VT15C12

C4

RSOH byte B1

BIP24BIP24

BIP8BIP8 BIP8 BIP8

C11

C3

C12

C4

VT15VC11

STM-n

eight interleaved blocks

1 0 0 1 1 1 1 10 0 0 1 0 1 1 0

1 1 1 1 1 1 0 1

0 1 1 1 0 1 0 0

1 0 0 1 1 1 1 10 1 1 0 0 0 0 0

parity

BIP-8

STM-1 STM-1STM-n VC4VC4

OC-3 / STM-1

B3

V5


• Monitoring voice-carrying digital channels: This basically consists of extract-ing one voice channel, decoding it, and listening to it through a loudspeaker tocheck for possible signal degradation.

• Monitoring data-carrying digital channels: When a digital channel carries da-ta, it is not possible to apply the solution described above, since the informa-tion is unintelligible. For this reason, what is known as a partial OOS test isapplied. This test consists of inserting a PRBS only in the channel being tested,with the rest of the channels carrying live traffic. At the receiving end, the BERof this channel is then quantified. Only one channel is out of service, while therest remain in service.

• As for OOS measurements, ITU-T Rec. O.181 defines (for SDH networks) theevents to be monitored when ISMs are made for each type of measurement (seeTable 6.8).

6.3.4 Connecting a Measurement Device for ISM

During ISMs, the live traffic signal is measured without being disturbed, with a lowamplitude sample being taken from it. In this case, the measurement device doesnot overload the network being measured. The measurement performed is said tobe nonintrusive.

For ISMs, the connection to NEs is made via PMPs. The PMPs have a set atten-uation; typically 20 dB above the traffic signal. The analyzing device must supply

Table 6.8In-service measurements for SDH/SONET.

Measurement On Events to Be Monitored

Regenerating section OOF, B1 errors, LOS, LOF, RS-TIMMultiplex section OOF, B2, MS-REI, LOS, LOF, MS-AIS, MS-RDIHigher-order containerC-4 and VC-4-Xc

OOF, B3 errors, HP-REI, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM

Higher-order container (C-3) OOF, B3 errors, HP-REI, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC,

Lower-order container (C-3) OOF, B3 errors, LP-REI, LOS, LOF, MS-AIS, MS-RDI, AU-LOP (Note), AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LP-RDI, TU-LOP, TU-AIS, LP-TIM, LPTC-TIM, LPTC RDI, LPTC-LTC

Lower-order containerC-11/C-12/C-2 and VC-2-Xc

OOF, BIP-2 errors, LP-REI, LOS, LOF, MS-AIS, MS-RDI, AU-LOP, AU-AIS, HP-RDI, HP-TIM, HP-LOM, HPTC-TIM, HPTC RDI, HPTC-LTC, LP-RDI, TU-LOP, TU-AIS, TU-LOM, LP-TIM, LPTC-TIM, LPTC, RDI, LPTC-LTC


the same amount of gain to compensate for the attenuation that the PMP incorpo-rates, with the aim of ensuring that the measurement results are correct.

The PMPs for electrical interfaces are regulated by ITU-T Rec.G.772, and areusually supplied by the equipment manufacturers, or installed by the network oper-ators in the digital distribution frames (DDFs). When these points are not available,the measurement instruments are connected to the NEs using high-impedanceprobes for the electrical interfaces (see Figure 6.12). The input interfaces of somemeasurement instruments comply with ITU-T Rec. G.703, equalizing automaticallyand compensating for attenuation at the G.772 monitoring points. In order to com-pensate for attenuation at other test points (measurements with attenuating probes),the gain can be programmed (for instance, at 20 dB, for all hierarchies).

If the measurements are being carried out on optical interfaces, splitters are usedto obtain monitoring points that are external to the NE. In some cases, the NEs al-ready incorporate these PMPs for measurements on STM-n optical signals. In thiscase, these points provide nonreturn to zero (NRZ) signals with ECL signal levels.

NE

Optical spliterNE

90%

10%

Optical signals

PMP

NRZ/ECL

Figure 6.12 Setup for measuring optical and electrical signals.

Electrical signals

75W/CMI

T-connector T-connector

75W 75W

R>>75W

PMP

PMP: Protected monitoring pointNE: Network element


6.4 SYNCHRONIZATION OF NE-TEST SET IN SDH

Before starting any test in SDH, the network element must be synchronized withthe test set. This ensures that no uncontrolled pointer adjustments will occur duringthe test. A 2-MHz source, for instance, can be used as the common synchronismreference for the network element and the test set (see Figure 6.13).

There exist two alternative methods of synchronization; namely:

• Configuring the network element to obtain its synchronism clock from theSTM-n signal received (the NE is synchronized with the test signal). Thismethod should not be used when carrying out offset tests in STM-n, since theNE will follow the frequency with the offset generated, and will therefore nev-er provide an offset.

• Configuring the network element to use its internal clock and the measuringdevice so that it is synchronized with the STM-n signal received.

NESTM-n/OC-m

Figure 6.13 Network element and test device synchronization options.

Clock input

Clock input

External reference

NESTM-n/OC-m

Recovered Clock

Internal Clock

(b) NE to the test device

(a) Common external reference

NESTM-n/OC-m

Internal ClockRecovered Clock

(c) Test device to the NE

Clock



• ITU-T Rec. G.711 (09/99), Pulse code modulation (PCM) of voice frequencies ITU-T.

• ITU-T Rec. G.957 (06/99), Optical interfaces for equipments and systems relating to the synchronousdigital hierarchy.

• ITU-T Rec. G.772 (03/93), Protected monitoring points provided on digital transmission systems.

• ITU-T Rec. O.150 (05/96), General requirements for instrumentation for performance measurementson digital transmission equipment.

• ITU-T Rec. O.151 (10/92), Error performance measuring equipment operating at the primary rate andabove.

• ITU-T Rec. O.152 (10/92), Error performance measuring equipment for bit rates of 64 kbit/s andN x 64 kbit/s.

• ITU-T Rec. O.153 (10/92), Basic parameters for the measurement of error performance at bit ratesbelow the primary rate.

• ITU-T Rec. O.181 (05/2002), Equipment to assess error performance on STM-n interfaces.

• Andreu Guimerà, Roger Segura, Puesta en Servicio y Mantenimiento de Redes Avanzadas de Teleco-municaciones, L&M Data Communications, 2001.

233

Chapter 7

SDH/SONET and PDH Roll-Out

Widespread interconnection between telecommunication networks, the develop-ment of new services, and sweeping liberalization policies are fast turning the in-formation age into a reality. In these new surroundings, the transport level isformed by SDH/SONET rings, whereas the access network is constructed by PDHand T-carrier links.

The transmission medium used (fiber, cable, or radio) has to be verified bychecking continuity and signal power. Before connecting the NE, its capacities formultiplexing and add and drop must be tested, along with its defect conditions andits behavior with frequency offset, pointer movements, and synchronization. Whenthe ring is complete, checks must be performed on the BER in the paths, delays, andoverheads.

Operators must also guarantee their clients a certain level of quality for the ser-vices they offer, and provide them with automatic protection when there is no avail-ability due to a breakdown.

7.1 BIT ERROR RATE TEST

7.1.1 BERT of Virtual Container

In order to confirm that an NE does not cause deterioration in the signals it process-es, it is necessary to check the integrity of the information carried by virtual con-tainers. In order to do so, the bandwidth of the container transported by the virtualcontainer is filled with a PRBS test signal. The device under test (DUT) receivesthis signal and processes it. When the signal leaves the device, another measure-ment instrument recovers the test signal from the virtual container received, per-forming a bit error rate test (BERT) on the signal, to check that there are no errors(see Figure 7.1).


7.1.2 Overhead Transparency Test

The aim of this test is to check the integrity of the information sent across the SDH/SONET overhead channels. For instance, the data communications channels(DCCs) carry network administration information that is exchanged between vari-ous NEs (see Section 2.8.2). For this test, the DCCs are taken up by a PRBS. Afterthis signal has passed through the DUT, a BER measurement will be performed onthese channels to check for the absence of anomalies in the transmitted signal. Thesame kind of transparency test can also be performed on order line channels (E1,E2) (see Figure 7.2).

Problems occur mainly due to differences in the programming of overhead byteswhen large networks containing NEs from different manufacturers are connected forthe first time. In such cases, the simulation and analysis of SDH/SONET overheads

Stimulus signal

DUT

BERT


Figure 7.1 BERT on virtual container.

SDH/SONET generator SDH/SONET analyzer

DUTSTM-1

STM-1BERT

Figure 7.2 Overhead transparency test in the overhead bytes like D4-12, E1, E2, N1, N2, or F1.

STM-n

VC4

VC12Pattern in OH channels

STM-n

VC4

VC12

SDH/SONET and PDH Roll-Out 235

are particularly useful, since they allow specific tests to be made on how the systemreacts to different events (when faced with a variation in the overhead byte values).

It is also useful to have instruments that show both the overhead bytes transmit-ted during the simulation and those received on the same screen at the same time.This simplifies problem solving during the simulation and follow-up to the reactionsof the system when it is operating in service.

7.2 STIMULUS-RESPONSE TESTS

These tests are doubly useful. On the one hand, they confirm that the control com-puter for the DUT detects all the potential anomalies and defect conditions correct-ly, while at the same time checking that when faced with such a condition, thedevice generates the appropriate indication. The anomaly and defect indications arestated in ITU-T recommendations. In particular, O.181 includes, in Annex A and B,all the indications of anomalies and defects associated with SDH/SONET networks(classified as relative to paths and sections). The response to the detection by a NEof the indicators is to generate other indicator signals that the NE sends to the otherNEs to which it is connected. ITU-T Rec. G.783 defines how these signals, knownas maintenance signals, interact.

At PDH levels, the frames incorporate a series of bits designed to indicate de-fects and anomalies. This way, defects can be provoked by inserting, for instance, ahigh error rate in the FAS to provoke a backwards response in the form of a remotealarm indication (RAI) and a forwards response in the form of an alarm indicationsignal (AIS). The process is similar to that followed in SDH/SONET, but with muchless capacity for controlling and managing the events produced.

To perform a stimulus-response test, the DUT is subjected to fault conditions(stimulus) to check that it generates the appropriate response or maintenance signals(see Figure 7.3).

DUTBackward response

Figure 7.3 Stimulus-response tests.

Forward response

RDI, or REI, or RFIAIS

Test stimulus


7.2.1 Preliminary Definitions

ITU-T Rec. G.783 defines a series of concepts that relate to the events recorded bythe NEs and used by the stimulus-response tests. SDH/SONET events are classifiedinto anomalies, defects, damage, failures, and alarms depending on how they affectthe service (see Section 2.10.1).

7.2.2 Line and Test Sequence Events

These are events related to test sequences or to the signal itself on the line. Theseevents do not depend on the digital hierarchy, but are only related to the test se-quences used, or to the characteristics of the signal. In the case of bit errors, forSDH/SONET we talk about test sequence errors (TSE) which are related to TSS-istructures (see Table 6.6).

7.2.3 PDH Events

The events to be considered when looking at the PDH hierarchy are repeated at sev-eral hierarchical levels (see Table 1.4). At the first level (2 Mbps), some particularevents are defined relating to the possibility of forming CAS or CRC multiframes(see Table 1.2).

7.2.4 SDH/SONET Events

The layers of the stratified model of the SDH/SONET carrier network defined inITU-T Rec. G.803 (regenerator section layer, multiplexer section layer, higher-or-der path layer, and lower-order path layer) allow the events relating to SDH/SO-NET to be classified (see Section 2.10.3).

7.2.5 Interaction of Maintenance Signals

The test described in the following sections checks the detection and response of anNE when faced with fault conditions, and it is included in the acceptance protocolsfor this type of equipment. In this case, the NE to look at is an ADM.

This test is doubly useful. On the one hand, it checks that the control computerfor the ADM detects each of the potential events correctly, and, on the other hand, itverifies that when faced with such a condition, the ADM generates the appropriateindicators of anomalies and defects.

The response to the detection of such indicators by a multiplexer is to generateother indicator signals that are sent to the network elements to which the multiplexeris connected (see Figure 2.32).


7.2.5.1 Fault conditions: Detection and response

ITU-T recommendations relate the interaction of maintenance signals to the typicalfunctions of the network elements (see Section 2.10), which it describes through afunctional reference model. The corresponding indicator signal will report an eventthat affects the function with which this signal is associated. A map illustrates howmaintenance signals interact (see Figure 2.31); that is, the SDH/SONET layer inwhich each event is detected and the response to this event. For carrying out thistest, the interactions have been classified according to the function they affect.

It will be checked whether when faced with the indicator signals generated bythe measurement instrument, the ADM under test supplies the appropriate responsesignals, both in its aggregate and tributary flows. The corresponding setup for thetest must contain SDH/SONET generators/analyzers, or tester, and a tester capableof detecting AIS signals at 2 Mbps (see Figure 7.4).

Measurement points A, B, C, and D have been defined, and these are referred toby the test procedure. The stimulus is introduced at point A, and the response signals,generated by the ADM under test, are measured at the rest of the points. As well asviewing the response signals by means of the measurement instruments, the ADMis also connected to a control computer, and the control software is able to reflect thesignals detected by the multiplexer.

An example of the operations to perform has been summarized (see Table 7.1).The action column describes the signals to be generated by the tester as stimulus atpoint A. The rest of the columns contain the responses at the measurement points.

Figure 7.4 Connections for testing response to fault conditions.

Forward response

Tributary analyzer

WestSTM-n/OC-m East

Stimulus signal with eventA

BBackward response

C

D

AnalyzerGeneratorand analyzer

PDH tributary

STM-n/OC-m


7.3 STRESS TESTS

Transmission systems operate taking into account the inaccuracies of frequenciesor transmission rates that correspond to their hierarchies. The bit rates in PDH sys-tems have a certain margin of tolerance within which the NEs will work properly.SDH/SONET systems introduce the pointer adjustment mechanism to compensatefor phase offset, or frequency offset in the connection points of two different net-works (see 5.3.1). Bit rates near the limits of the margin of tolerance for PDH sig-nals, or repeated and consecutive pointer adjustments in short spaces of time forSDH/SONET signals, subject the network and PDH and SDH/SONET elements tostress. Stress tests allow us to check how robust the NEs are when faced with situa-tions at the very limit of the conditions that allow them to work properly.

Table 7.1Test summary.

Action (A) Response (B) Response (C) Response (D)LO

S

MS-RDIHP-RDILP-RDI

AIS 2 Mbps TU-AIS

LOF

MS-RDIHP-RDILP-RDI

AIS 2 Mbps TU-AIS

MS-

AIS MS-RDI

HP-RDILP-RDI

AIS 2 Mbps TU-AIS

B2

erro

rs MS-RDIHP-RDILP-RDI

AIS 2 Mbps TU-AIS

AU-AIS HP-RDILP-RDI

AIS 2 Mbps TU-AIS

AU-LOP HP-RDILP-RDI

AIS 2 Mbps TU-AIS

HP-UNEQ HP-RDILP-RDI

AIS 2 Mbps TU-AIS

B3 errors HP-REI - -TU-LOM LP-RDI AIS 2 Mbps TU-AISTU-AIS LP-RDI AIS 2 Mbps TU-AISTU-LOP LP-RDI AIS 2 Mbps TU-AISLP-UNEQ LP-RDI AIS 2 Mbps Loss of pattern and bit errorsBIP-2 LP-REI - -LP-RDI - - LP-REI


7.3.1 Introducing Frequency Offset

Network elements must guarantee a reliable recovery of the clock from the incom-ing signal, provided the transmission rate of the signal received is within the limitsset by the ITU-T Rec. G.703 for each hierarchy (see Table 7.2). This capacity for

clock recovery can be checked by offsetting the clock frequency of the data gener-ated by the tester and delivered to the NE under test. By creating a loop in the NE(see Figure 7.5), the data sequence will be received on the way back and can be an-alyzed, checking for the presence or absence of transmission errors.

7.3.2 Generating Pointer Movements

This is a synchronization test based on analyzing pointer movements. It is aimed atchecking that the pointer-processing circuits of the DUT are working properly. Itbasically consists of generating a signal from a tester that is not synchronized withthe NEs. The frequency of the signal generated deviates from its nominal value inorder to subject the NE (pointer-processing circuits) to stress (see Figure 7.6). As aresult, pointer adjustments are detected in the output of the NE, as is the resultingoffset of the associated virtual container, which can be analyzed with another tester,this time synchronized with the network element. A BER test can be carried out on

Table 7.2Frequency offset tolerance.

Rate Tolerance (better than) Rate Tolerance (better than)

64 Kbps ±100 ppm (±6.4 bps) 64 Kbps ±100 ppm (±6.4 bps)2,048 Kbps ±50 ppm (±102.4 bps) 1,544 Kbps ±32 ppm (±50 bps)8,448 Kbps ±30 ppm (±253.4 bps) 3,152 Kbps ±30 ppm (±95 bps)34,368 Kbps ±20 ppm (±688 bps) 6,312 Kbps ±30 ppm (±189.4 bps)139,264 Kbps ±15 ppm (±2,089 bps) 44,736 Kbps ±20 ppm (±895 bps)155,520 Kbps ±20 ppm (±3,111 bps) 51,840 Kbps ±20 ppm (±1037 bps)STM-n ±20 ppmSTM-n (SEC) ±4.60 ppm

BER test DUT

PRBS in data signalfnom ± ∆ Hz

Figure 7.5 Stress tests: Frequency offset.


the virtual container in order to check its integrity. Likewise, the pointer events canalso be counted by time unit to give an idea of the frequency offset between the testgenerator references and the network element.

7.4 MUX/DEMUX TESTS

These tests are based on checking that no bit anomalies (BER test) arising from themultiplexing and demultiplexing process are present (see Figure 7.7). Basically, themultiplexed tributary signals must maintain their integrity inside the aggregate sig-nal. In the case of PDH multiplexers, this checks that multiplexing and demulti-plexing is being performed correctly, and that the payload is properly allocated andrecovered in SDH/SONET multiplexers.

DUTSTM-n/OC-m STM-n/OC-m

Pointer movements

Figure 7.6 Generation of pointer movement.

with f0nom ± ∆ Hz

PRBS stimulus signal in payload payload offset

BER analysisGenerator

MUXPRBS in tributary

Aggregate

Figure 7.7 Mux/demux test: (a) full-channel test if Tx/Rx have the same bit rate and (b) half-channel test if Tx/Rx have a different bit rate.

BER test


Aggregate signal

BER test

Full channel test Half channel test(a) (b)


7.4.1 PDH Mux/Demux Test

The multiplexer test consists of checking that no errors are introduced in the multi-plexed signals on their way through the multiplexing process. This is done verysimply, by means of a BER test on each of the four tributary ports, that is, by gener-ating a test signal on the tributary to be measured, multiplexing it, creating a loop inthe multiplexer under test, demultiplexing the tributary containing the test signalwith the test instrument, and checking whether or not bit errors have occurred (seeFigure 7.7).

Other measurements for checking multiplexers based on their BER use loopsand tandem connections to run the test signal through all the ports of the devices. Ifanomalies are detected, these can gradually be isolated in the port or ports that gen-erate them. In a first instance, the test is performed on a multiplexer by establishinga local loop in the aggregate port, and successive tandem connections in the tributaryports (see Figure 7.8).

A second variation is an extension of the first that looks at whole paths, in thiscase establishing loops in the tributaries of the remote multiplexer (see Figure 7.9).These measurement figures are valid for both PDH and SDH/SONET multiplexers.

BER test


Aggregate

Tandem Loop

Tx1Rx1Tx2Rx2Tx3Rx3Tx4Rx4

Figure 7.8 Full channel test. This configuration checks all tributary interfaces.

connection

BER test


Tandem Remote loops

MUXtributaries

Figure 7.9 Full channel test. This configuration also verifies the circuit and the remote multi-plexer.

connectionAggregate


7.4.2 SDH/SONET Mux/Demux Test

Plesiochronous signals are carried in synchronous networks as a payload of theSDH/SONET signals. This creates a need for a synchronous multiplexing processat the transmission end, and synchronous demultiplexing at the receiving end. Bothof these operations must be subject to test, that is, it must be checked if the PDHpayload has been correctly allocated to the containers, and if it is correctly recov-ered at the receiving end (see Figure 7.10).

The multiplexing test, sometimes called mapping test, can be carried out by sub-jecting the test signals to a frequency offset, which involves a stress test for the map-ping function of the multiplexer under test.

7.5 MEASURING ROUND TRIP DELAY

Round trip delay (RTD) of a signal is due to two factors: long paths and transit timethrough the NEs. Delays due to long paths are especially evident in intercontinentallinks that include sections via satellite. In these cases, the periodic variation in theposition of the satellite causes frequency offset and RTD variation; both caused bythe associated Doppler effect. In addition to this delay variation, when the delay ex-ceeds a certain value, synchronization and signaling anomalies may occur. There iseven the possibility that the connection is not established, or it may be lost duringdata transmission (see Figure 7.11).

Likewise, when there are several multiplexers and DXCs in the line, a notice-able delay may occur, and this may exceed the maximum advisable limits. The tran-sit delay that these NEs introduce depends on the processing they perform on the

Unframed PRBS

Figure 7.10 Half channel SDH/SONET multiplexer test.

Multiplexing

in tributary

West East

ADM

Unframed PRBS

Demultiplexing

in tributaryPRBS in payload

West East

ADM

STM-n/OC-m STM-n/OC-mSTM-n/OC-m STM-n/OC-m

PRBS in payload


signals they receive. For instance, in an ADM, the delay between the input of anSTM-n aggregate signal and the output of one of its 2-Mbps tributaries through thecorresponding port could be in the region of 75 ms.

ITU-T Rec. G.114 stipulates the tolerable delay between the ends of a connec-tion due to processing time. It excludes the component due to propagation, since thisdepends on the rate, the media, and the distance, which can only be controlled in alimited way (see Table 7.3).

Time longer than 400 ms is acceptable only in exceptional cases, such as doublesatellite hops, the use of satellites to restore land routes, the connection of fixed ser-vice by satellite and cellular telephony system, etc. It is recommended that the pro-cessing delay in one direction be limited to 50 ms. Nevertheless, it tends to be muchless, for instance, in the region of 6 ms for a national system and 3 ms for the inter-national chain.

With the aim of ensuring a correct measurement, the tester must be calibrated insuch a way that the quantization of the delay introduced by its internal circuits iseliminated (see Figure 7.12). For this reason, the delay must be measured after con-necting the input and output in a short loop. The value obtained will be subtracted

Table 7.3Limits for end-to-end transmission time (ITU-T Rec. G.114).

One Direction Delay Application

0 to 150 ms Acceptable for most user applications150 to 400 ms Acceptable if the delay influence on the performance of the user appli-

cations is known (e.g., connections with a satellite hop)> 400 ms Unacceptable

RTD pattern

... LTE

R

RLTE

R

R

Loop

RTD

Figure 7.11 Measuring RTD in a path.


from the delay measured (see Figure 7.12). The loop may be set up between otherpoints (for instance, to eliminate the delay introduced by line terminal equipment),if this is what we want. The test signal used for delay measurements can be a fixedpattern or a PRBS, depending on the instrument, in a PDH or SDH/SONET line sig-nal.

7.6 APS MEASUREMENTS

SDH/SONET networks provide security mechanisms against cutoffs or failure inthe service of a traffic channel. Procedures have been defined for multiplexer sec-tions that automatically switch over from the channel in which a fault has occurredto a back-up channel, with the aim of maintaining service and certain performancelevels.

7.6.1 Network Security: Concept and Classification

A number of security mechanisms can be defined for SDH/SONET networks. Basi-cally, we can distinguish between three: diversification, restoration, and protection(see Section 2.13).

The diversification mechanism divides the route to be followed by the circuitsestablished between two points in a network into two or more different paths. Thisway, a failure in one of the routes will not affect the 100% of the circuits betweentwo exchanges.

The restoration mechanism enables us to analyze failures and look for alterna-tive transmission routes. It requires special multiplexers and an associated restora-

RTD pattern

... NE

Figure 7.12 Generic calibration process and measurement of delay between A and B; (1) mea-suring delay introduced by the NE and the tester; (2) measuring delay between A and B from points C-D.

Network

Delay (A-B) = Delay (C-D) - 2TcalDelay (C-D) = 2Tcal

Tcal

Tcal

RTD pattern

... NE

Tcal

Tcal

(1) (2)

DCDC

A

B

A

B


tion system. The back-up circuits are not allocated beforehand but are used ondemand, independent of the circuits that are down.

The protection mechanism associates back-up circuits beforehand that comeinto operation in a way controlled directly from the NEs. Protection mechanisms areapplied to multiplexer sections, paths, and subnetworks or parts of a path.

The measurements of interest in this section refer to multiplexer section protec-tion (MSP), both of linear and ring topology. The way to implement this protectionis called automatic protection switching (APS). It is based on a signaling protocolthat transmits specific messages by means of the K1 and K2 bytes in the multiplexersection overhead (MSOH). The conditions under which the APS procedure is trig-gered in a NE are as follows:

• Loss of signal (LOS);

• Loss of frame (LOF);

• Multiplexer section alarm indication signal (MS-AIS);

• Signal fail (SF), which occurs when the B2 error rate is higher than 10-3;

• Signal degrade (SD); which occurs when the B2 error rate is higher than 10-6.

Table 7.4K1, bits 1 to 4, indicates a request of a traffic signal for switch action.

K1 (bits 1-4) Condition, State or External Request Ordera

a. An SF condition on the protection section is higher priority than any of the requests that would causea normal traffic signal to be selected from the protection section.

1111 Lockout of protection Highest1110 Forced switch .110111001011

Signal fail high prioritySignal fail low prioritySignal degrade high priority

.

.

.101010011000

Signal degrade low priorityUnused)Manual switch

.

.

.011101100101

Unused Wait-to restoreUnused

.

.

.010000110010

ExerciseUnusedReverse request

.

.

.0001 Do not revert .0000 No request Lowest


The first NE that detects one of these conditions communicates to the rest of theNEs to which it is connected that they need to switch over to protection circuits bymeans of the corresponding messages, which are sent via K1 and K2.

The APS protocols for linear multiplexer sections are described in ITU-T Rec.G.783 and also in ITU-T Rec. G.841, whereas for ring multiplexer sections they aredescribed only by the latter.

7.6.2 Characterizing the Measurement

APS measurements basically focus on:

• Checking the operation of the signaling protocol on the K1 and K2;

• Measuring the total switching time.

7.6.2.1 APS protocol in linear multiplexer section

The K1, bits 1 to 4, contains protection request messages requesting switching ac-tions (see Table 7.4). A request can be:

• A high or low priority condition (SF and SD) associated with a section;

• A state, such as waiting to be reestablished, no overturn, absence of request,overturn request, for the protection function;

• External, such as exclusion from protection, and forced or manual switching.

The K1, bits 5 to 8, indicate the number of the traffic signal or the section for whichthe request is made (see Table 7.5).

The sent K2, bits 1 to 4, shall indicate in null signal (0) if the received K1 byteindicates null signal and the extra traffic is not bridged. For all other cases, K2 indi-cates the number of the signal that is bridged (see Table 7.6). The bit 5 shall indicate“0” if 1+1 architecture and “1” if 1:n architecture. The bits 6 to 8 have some binary

Table 7.5K1, bits 5 to 8, indicates the channels to be switched for APS in linear sections.

Channel Request for Switching Action

0 Null signal (no normal or extra traffic signal). Conditions and associated priority (fixed high) apply to the protection section.

1-14 Normal traffic signal (1-14). Conditions and associated priority apply to the corre-sponding working sections. For 1 + 1 only traffic signal 1 is applicable.

15 Supplementary traffic channel. Only exists in a 1:n architecture.


combinations reserved for future use, and two of them to indicate the MS-RDI andMS-AIS defect indications (see Table 7.8).

7.6.2.2 APS protocol in multiplexer section in a ring

There are a number of different architectures for protection rings (see Figure 7.13).First, we must distinguish between rings with dedicated protection known as multi-plex section dedicated protection ring (MSDPRING) and shared protection knownas multiplex section shared protection ring (MSSPRING) (see Section 2.13.2).

Protocol messages for MSSPRING rings

MSSPRING is a common and widely used protection strategy. As with linear sec-tions, the K1 and K2 bytes will be transmitted in the multiplexer section overheadof the STM-n that carries the protection channels.The protocol rules are appliedconceptually to an APS controller operating in an NE or a node of the ring. Theswitching and signaling actions must be chosen for both sides of the node, based onall the signaling of incoming K1 and K2 bytes in both directions, the failures de-tected on both sides, failures in local equipment and instructions initiated external-ly. Generally speaking, this conceptual controller looks at all the incominginformation, chooses the input with a higher priority, and acts based on this choice.

Bits 1 to 4 of the K1 byte carry bridging request codes, numbered in descendingorder of priority (see Table 7.7). Bits 5 to 8 of the K1 byte carry the identification(ID) of the destination node for the bridging request code shown in bits 1 to 4 of theK1 byte (see Table 7.7). The K2 byte is used for exchanging different information(see Table 7.8).

7.6.3 Measurement Procedure

The procedure is equally valid both when applied to a linear multiplexer sectionand in a ring. In the case of the section in a ring, we will look at the MSSPRINGprotection architecture in a two fiber ring (see Section 2.13.2).

Table 7.6K2, bits 5 to 8, traffic signal number, for APS in linear sections.

Traffic Signal Indication

0 Null traffic signal.1-14 Normal traffic signal (1-14).

For 1+1, only normal traffic signal 1 is applicable.15 Extra traffic signal. Exists only when provisioned in a 1:n architecture.


Figure 7.13 MSDPRING and MSSPRING protection architectures. USHR: unidirectional self-healing ring and BSHR: bidirectional self-healing ring.

x

xy

y

x

xy

y

x

x y

y

x

xy

y

Span

x

xy

y

x

xy

y

x

x y

y

x

xy

y

MSDPRING

MSSPRING two fibers

MSSPRING four fibers Ring Switching

MSSPRING four fibers Span Switching

Circuits under normal conditions Circuits under protection

USHR USHR

BSHR BSHR

BSHR BSHR

BSHR BSHR

Four fibers sharing serviceand protection at 50%.

After a fault the switching isdone in the same section.

Four fibers sharing serviceand protection at 50%.

After a fault an alternativepath is allocated.

Two fibers: sharing serviceand protection at 50%.

After a fault an alternativepath is allocated.

Two fibers: one is active and the other is for protection.After a fault an alternative

path is allocated.


7.6.3.1 MSSPRING architecture in a two fiber ring

Each fiber divides its capacity 50/50 between service channels and protection chan-nels (see Figure 7.14). With the circuit operating under normal conditions, commu-nication between ADM multiplexers is direct. (i.e., it does not pass through otherintermediary multiplexers). When there is a cutoff then communication is reestab-lished by means of an alternative route.

Table 7.7Messages in the K1 byte for APS in MSSPRING ring sections. The reverse request instruction

takes on the priority of the bridging request to which it corresponds.

Bits 1-4 Bridge Request Code Bits 5-8. Destination Node Identification

1111 Lockout of protection (span) or signal fail (protection)

The destination node ID is set to thevalue of the ID of the node for whichthat K1 byte is destined. The destination node ID is alwaysthat of an adjacent node(except fordefault APS bytes).

1110 Forced switch (span)1101 Forced switch (ring)1100 Signal fail (span)1011 Signal fail (ring) 1010 Signal degrade (protection)1001 Signal degrade (span)1000 Signal degrade (ring)0111 Manual switch (span)0110 Manual switch (ring) 0101 Wait-to-restore0100 Exerciser (span)0011 Exerciser (ring)0010 Reverse request (span)0001 Reverse request (ring)0000 No request

Table 7.8Coding of bits of the K2 byte in APS of MSSPRING rings.

Source Node (bits 1-4) Long/Short (bit 5) Status (bits 6-8)

Source node ID is set to the node’s own ID

0 Short path code (S) 1 Long path code (L)

111 MS-AIS110 MS-RDI101 Reserved for future use100 Reserved for future use011 Additional traffic via protection channels010 Bridged and switched001 Bridged000 Idle


7.6.3.2 Steps to follow

A measurement method can be described generically by means of a short series ofsteps as described below (see Figure 7.15):

1. Connect the tester to the multiplexer to be tested by means of a tributary port.2. Synchronize both devices; (e.g., the NE can deliver the clock to the tester).3. Program the tributary signal generated; (e.g., bit rate, line code, test pattern,

etc.).4. Create a loop in the tributary port at the remote end of the section being tested.5. Check that there are no events.6. Provoke APS by forcing it from the network management system; or discon-

necting the work fiber (provoking LOS); or generating B2 errors to obtain SDor SF status, which will trigger the switching when detected.

7. Measure switching time, it should be less than 50 ms.

Frame capture can be carried out with the aim of studying the content of the K1 andK2 bytes to verify the MSP messages exchanged, thus evaluating the operation ofthe APS protocol.

Loop

Tributary

Figure 7.14 Switching test for MSSPRING, 2 fibers.

BSHR BSHR

Loop

Tributary

a) b)


7.7 PERFORMANCE MEASUREMENTS

7.7.1 Introduction to G.821, G.826, and M.2100

7.7.1.1 Concept and aims

The aim of performance measurements is to provide operators and users with infor-mation on the bit errors that occur in their digital links. For this, there is a series ofITU-T recommendations about the performance of digital links, including interna-tional connections.

Anomaly monitoring in digital transmission networks can be carried out on end-to-end connections (paths) or on parts of the network (lines and sections). Measure-ments on paths inform the user of the overall QoS. Measurements on lines and sec-tions are performed during repair, installation, and maintenance, to make sure thatpreviously established transmission performance objectives are met. The main ITU-T recommendations concerning quality and performance measurements are G.821,G.826, and M.2100.

7.7.1.2 Measurements in Line with G.821

ITU-T Rec. G.821 establishes the error characteristics that each direction must havein an international digital connection at n x 64 Kbps that forms part of an ISDN.For this, it defines the performance targets for an n x 64 Kbps circuit forming partof an international hypothetical reference connection (HRX) of 27,500 km in anISDN network. The scope of this recommendation is currently limited to bit ratesbetween 64 Kbps and 2 Mbps (noninclusive). Its appendix D allowed the measure-ment of performance to be extrapolated in line with G.821 for higher rates.

Figure 7.15 APS testing analysis K1/K2 answer to switching events and to APS messages.

ADM

Switching events:

K1, K2backward

forward K1, K2

K1, K2analysis

LOS, LOF, AIS, SF

ADM

APS messages

K1, K2backward

forward K1, K2

K1, K2analysis

bits 1-4 K1


7.7.1.3 Measurements in line with G.826

ITU-T Rec. G.826 provides criteria for monitoring parameters and error perfor-mance objectives for international digital paths of constant bit rate that operate atprimary rate (1,5 Mbps 2 Mbps) or higher. It applies to ISM and OOS measure-ments and is based on the concept of errored blocks, complementing G.821.

7.7.1.4 Measurements in line with M.2100

ITU-T Rec. M.2100 defines the limits of the performance parameters that must bemet by international PDH transmission paths, sections and systems (64 Kbps,2 Mbps, and higher), for their bringing-into-service and maintenance. The aim ofthis recommendation is to define some limits for the performance parameters mea-sured over short intervals of time that guarantee with a certain amount of confi-dence that the quality targets defined are met. It applies to both ISM and OOSmeasurements.

7.7.2 Measurements in Line with G.821

7.7.2.1 Area of application

The quality targets defined in ITU-T Rec. G.821 are given for both directions of aconnection at n x 64 Kbps (1≤n≤31) with an HRX that covers a total length of27,500 km. The same recommendation allocates these overall targets to each of theindividual sections that make up the reference connection as a whole.

7.7.2.2 Events

The parameters for evaluating performance under G.821 are obtained from the fol-lowing events:

1. Errored second (ES): It is a one-second period in which one or more bits are inerror or during which LOS or AIS is detected;

2. Severely errored second (SES): It is a one-second period in which BER ishigher than 10-3 or during which LOS or AIS is detected.

7.7.2.3 Parameters

The performance parameters are measured by observing the defined events over acertain time interval. It should be pointed out that this time interval of observation(Stotal) is divided into two parts; namely, the time during which the connection isconsidered to be available (Savail), and the time is considered unavailable (Sunavail):


• A period of unavailable time starts when the BER in each second is less than10-3 for 10 consecutive seconds, which are considered unavailable time.

• A new period of available time begins with the first second of a period of tenconsecutive seconds that each has a BER better than 10-3. (The definitions rel-ative to availability can be found in the E.800 recommendations.)

The error characteristics, that is, the performance parameters, must only be evaluat-ed while the connection is available:

1. Errored second ratio (ESR): The ratio of ES to total seconds in available timeduring a fixed measurement interval.

2. Severely errored second ratio (SESR): The ratio of SES to total seconds inavailable time during a fixed measurement interval.

There are other performance parameters that are usually supplied by the test equip-ment when measurements are made in line with G.821, although these are not in-cluded in the recommendation itself:

• Degraded minutes (DM): The parameter corresponds to appendix D of the oldG.821, which has been withdrawn. According to appendix D of the old G.821,a minute is considered degraded if the BER is higher than 10-6.

• Background errored bits (BE): The ratio between errored bits and total bits re-ceived during periods of availability, excluding errored bits and totals receivedduring an SES.

• Available time errored bits (AE): The ratio between errored bits and total bitsreceived during periods of availability, including errored bits and totals re-ceived during an SES.

Note that both BE and AE are measured in available time, and what differentiatesthem is the concept of SES. Furthermore, AE will always be greater than BE, sinceit takes SES seconds and non-SES seconds into account.

7.7.2.4 Targets

The quality targets for the overall connection defined by ITU-T Rec. G.821 are asfollows:

SESR < 0.002 ESR < 0.08

The recommendation does not specify the total observation time, although itdoes suggest a reference period of around one month. The targets would be appro-


priate for a connection set up like the reference, but a great many real internationalconnections are shorter, which means they ought to operate at a higher level of per-formance than the limits mentioned above. The small percentage of real connectionsthat exceed the length of the reference configuration may also lower the above-men-tioned limits.

7.7.2.5 Distribution of targets

To distribute and allocate the overall performance targets, a reference connectionmodel has been defined (see Figure 7.16). It divides the total connections into sec-tions, based on the grade of the circuit typically found along the connectionn x 64 Kbps - n x 64 Kbps (see Figure 7.17):

• Local grade: Systems that operate between a user and a local exchange, typi-cally at rates below 1,544 Kbps.

High Medium Localgradegradegrade

MediumLocalgrade grade

Figure 7.16 Distribution of performance targets.

Localexchange

Central Central office office

Localexchange

UserUser

High Medium Localgradegradegrade

MediumLocalgrade grade

1,250 km 1,250 km25,000 km

27,500 km

LE LE

T reference point

Figure 7.17 Hypothetical reference connection model.


• Medium grade: Systems operating between local exchanges within a nationalnetwork. Depending on their geographical location, the length of these sectionsmay vary. Nonetheless, the allocation of targets for local and medium grade isdone as a block, irrespective of length. The sole limitation is that the maximumlength of the route formed by local and medium grade sections is less than1,250 km. This limit in terms of length means that in large countries, the medi-um grade only reaches the primary center, whereas in small countries this sec-tion may reach the secondary or tertiary center, or the international switchingcenter.

• High grade: Includes national and international links that operate mainly athigh bit rates. In this case, the allocation of targets is based on the length of thissegment.

The distribution of the overall quality targets for each section is as follows:

• Local grade, 15% of the overall allowance at each end;

• Medium grade, 15% of the overall allowance at each end;

• High grade, 40%, equivalent to 0.0016% per km, for 25,000 km.

This distribution applies to the overall ESR target and to half of the overall SESRtarget, that is, to 0.001 (see Table 7.9).

If the high grade section includes a satellite system, it is allocated a total allow-ance of 20% of the target and ESR and SESR of the high grade section, irrespectiveof the distance. If the high grade path includes a satellite link, and the rest of the linkcovers more than 12,500 km, the overall targets may be exceeded. The concept ofequivalent distance by satellite that corresponds to the length of a terrestrial path be-tween 10,000 and 13,000 km is useful.

The other half of the overall target of SESR, that is, the remaining 0.001, is anoverall allowance for the medium and high grade section to take into account the ad-verse network conditions that are occasionally experienced in transmission systems.

Table 7.9G.821 performance targets.

Local Medium High Medium Local Global Objectives

Distrib. 15% 15% 40% 15% 15% 100%

ESR 0.012 0.012 0.032 0.012 0.012 0.08

SESR 0.000150.00015 0.0004 0.00015 0.00015 0.002

0.001


ITU-T Rec. G.821 mentions the possibility that checking the compliance of theset objectives will be carried out based on measurements made in paths that operateat rates above or equal to the primary plesiochronous rate. This means that ITU-TRec. G.826 must be used, because defines the targets of the error characteristic forthese paths.

7.7.3 Measurements in Line with G.826


ITU-T Rec. G.826 defines the limits of the performance parameters that must besatisfied by the constant rate digital paths that operate at primary bit rate or higher.The recommendation specifies the events and parameters that define the perfor-mance of the paths. These paths may be based on a PDH, an SDH/SONET or someother carrier network, such as a cell-based network. The performance targets areapplicable in both directions of the path. These values apply from end-to-end to ahypothetical reference path (HRP) of 27,500 km that may include transmission sys-tems via fiber optics, digital radio links, metallic cables, and satellite. The perfor-mance parameters are defined by observing blocks. As a result, ITU-T Rec. G.826gives an ISM of the error characteristic.

7.7.3.2 Definition of block

A block is a set of consecutive bits associated with the path; each bit belongs to oneand only one block. Consecutive bits may not be contiguous in time. Each block issupervised by means of an inherent error detection code (EDC); for example, BIPor CRC. In spite of the fact that this is not specified, it is recommended to use anEDC such that the probability of an anomaly event not being detected is less than10%, supposing a Poisson distribution of the errors. The error detection code itselfforms part of the block bits. Estimating errored blocks in in-service mode will de-pend on the structure of the network employed, and the type of EDC available.OSMs are also block based.

7.7.3.3 Error performance events for paths

The block-based events are errored block (EB), errored second (ES), severely er-rored second (SES), and background block error (BBE) (see Figure 7.18):

• EB: A block in which one or more bits is in error;

• ES: A one-second with one, or more, erroneous blocks or at least one defect;

• SES: A one-second period that contains 30% errored blocks or at least onedefect. SES is a subset of ES.1

≥


• BBE: An errored block not occurring as part of an SES and not belonging to anunavailability period.

7.7.3.4 Available time and unavailable time

• Unidirectional path: A period of unavailable time starts with the first of 10consecutive SES events. These 10 seconds are considered part of the unavail-able time. A new period of availability starts with the first of 10 consecutivenon SES events, and these 10 seconds are considered part of the available time.

• Bidirectional path: A bidirectional path is unavailable if one or both directionsis unavailable.

7.7.3.5 Error performance parameters

These parameters should only be evaluated while the path is available.

• Errored second ratio (ESR): The ratio of ES to total seconds in available timeduring a fixed measurement interval.

1. Periods of consecutive SESs persisting for T seconds, where 2 ≤ T ≤10 (some net-work operators refer to these events as "failures") can have a severe impact on serv-ice, such as disconnection of switched services.

Figure 7.18 Recognition of anomalies, defects, errored blocks, ES, SES, and BBE.

Monitored second

Anomalies Defects

ES SES

?Anomaly Defect

≥ 30 %EB yes (SES and therefore ES)

no

no no

ES=ES+1BBE=BBE+EB

ES=ES+1SES=SES+1

End

yes yes

Availablepath?

Availablepath?


• Severely errored second ratio (SESR): The ratio of SES to total seconds inavailable time during a fixed measurement interval.

• Background block error ratio (BBER): The ratio of BBE to total blocks inavailable time during a fixed measurement interval. The count of total blocksexcludes all blocks during SESs.

• Available time block error (ABE): Block errors registered during a period ofavailable time, including those registered during the SES that take place duringthis available time. This is expressed as a ratio of errored blocks to total blocksreceived. The ABE parameter is not defined by ITU-T Rec. G.826 but is in-cluded in some test procedures.

7.7.3.6 Objectives

End-to-end performance objectives for an HRP of 27,500 km are specified by de-fining the limits of the above-mentioned performance parameters. An evaluationperiod of one month is suggested for measuring these performances (see Table7.10).

For paths at 601 Mbps based on VC-4-4c the block has a size of 75,168 bits, andit is outside the size range recommended for paths at this rate (15,000 - 30,000). Inthis case, the objective for the BBER is of 4 x 10-4.

When a transmission system or a section can transport paths at different bitrates, it is enough just to check compliance with the end-to-end objectives allocatedto the path with the highest bit rate. One example of this would be an STM-1 sectionthat can transport VC-4 paths together with other, lower-rate paths.

Table 7.10End-to-end error performance objectives for a 27,500 km international digital HRX or HRP.

Rate (Mbps)

Connections from

64 Kbps

Paths

1.5 to 5 >5 to 15 >15 to 55 >55 to 160 >160 to 3,500

Bits/block — 800-5,000 2,000-8,000 4,000-20,000 6,000-20,000 15,000-30,000

ESR 0.04 0.04 0.05 0.075 0.16 —a

a. ESR objectives tend to lose significance for applications at high bit rates and are therefore not speci-fied for paths operating at bit rates above 160 Mbps.

SESR 0.002 0.002 0.002 0.002 0.002 0.002

BBER — 2 x 10 -4 2 x 10 -4 2 x 10 -4 2 x 10 -4 10 -4


7.7.3.7 Distribution of overall objectives

A hypothetical reference path can be divided into various frames in such a way thatpart of the overall error performance objectives is allocated to each of these frames(see Figure 7.19). Each frame must comply with its objective, so that the total of allthe frames that make up the whole path also complies with the overall objectives.

International gateways (IGs) are always terrestrial equipment residing physical-ly in the terminating (or intermediate) country. An IG is usually a cross-connect, ahigher-order multiplexer, or a switch, and it defines the boundary between the na-tional and international portions.

Higher-order paths may be used between IGs. Such paths only receive the allo-cation corresponding to the international portion between the IGs. In intermediatecountries, the IGs are only located in order to calculate the overall length of the in-ternational portion to deduce the overall allocation. The overall objectives are allo-cated as follows:

1. National portion: 17.5% + X

The objectives allocated consist of a fixed block allowance and an additional al-location based on the distance between the path end point (PEP) and the IG. The airroute distance should be calculated and multiplied by 1.5 if it is less than 1,000 km,taken to be 1,500 km if it is between 1,000 and 1,200 km, or multiplied by 1.25 if itis longer than 1,200 km. The distance calculated is then compared against the actualdistance, if this is known, and the smaller of the two values is retained. The retainedvalue should be rounded up to the nearest multiple of 500 km, and an allocation of1% per 500 km applied to the resulting distance.

PEPIGIGIGIGIGPEP

International portionNationalportion

Nationalportion

Hypothetical reference connection path27,500 km

Terminatingcountry

Terminatingcountry

Intermediate countries Intercountry(4 are assumed)

IG: International gatewayPEP: Path end point

Intercountry, for example path carried over a submarine cable

Figure 7.19 Distribution of overall objectives.


When a national portion includes a satellite hop, a total of 42% of the overallobjectives is allocated to this section. This allowance completely replaces the 17.5%allowance indicated for all other cases.

2. International portion:

The international portion is allocated 2% per intermediary country, plus 1% foreach terminating country. Furthermore, a distance-based allowance is also added tothe international portion. The air route distance between consecutive IGs should becalculated and each of them should be multiplied by 1.5 if it is less than 1.000 km,taken to be 1,500 km if it is between 1,000 and 1,200 km, or multiplied by 1.25 if itis greater than 1,200 km. The distance calculated is then compared against the actualdistance between consecutive IGs, if this is known, and the smaller of the two valuesis retained. The retained values for each element between IGs are used to calculatethe overall length of the international portion, which should then be rounded up tothe nearest multiple of 500 km, and an allocation of 1% per 500 km applied to theresulting distance.

When there is a satellite hop, this is allocated 35% of the overall objectives.Needless to say, this 35% replaces the percentage that would have been allocated tothis portion for distance.

7.7.3.8 Monitoring PDH paths

For PDH paths, there are two categories of anomalies defined relating to the incom-ing signal:

• An error in the FAS;

• Block error indicated by an EDC.

There are also three categories of defects related to the incoming signal: LOS, AIS,and LOF. Depending on the type of PDH path under consideration, in-service mon-itoring may not be able to calculate the full set of performance parameters. PDHpaths can be divided into four types (ITU-T Rec. G.826, Annex B), as follows:

1. Frame and block structured paths (i.e., structured paths with EDC): The ESR,SESR, and BBER can be estimated in ISM mode; for example, a primary rate,second order framed path with CRC in line with G.704.

2. Frame structured paths without EDC: The ESR and the SESR can be estimatedin ISM mode. In this case, however, an ES is observed when during one sec-ond, at least one error or one service defect occurs in the FAS. In this case, wecannot detect anomalies in the user data, which means that the ESR is muchless exact than in the previous case.


3. Frame structured paths without EDC that do not detect LOF: The ESR andSESR can be estimated in ISM mode.

4. Unframed paths, such as a leased line where only LOS or AIS can be detected;not LOF or any other type of anomaly.

7.7.3.9 Estimating performance at the far end

The available remote in-service indication such as RDI or, if provided, REI, areused at near end to estimate the number of SES occurring at the far end.

7.7.4 Measurements in Line with M.2100


ITU-T Rec. M.2100 provides limits for the performance parameters measured dur-ing short observation periods for bringing into service and maintenance of interna-tional PDH paths, sections, and transmission systems. These objectives include theerror characteristics from Recommendations G.821 and G.826, and also timing andavailability objectives.

This recommendation also describes performance parameters. Carrying outmaintenance on the network requires continuous ISM. Furthermore, bringing intoservice (BIS) tests call for OOS measurements using a PRBS between digital termi-nating points.

7.7.4.2 Event

The criteria are divided in-service (IS) and out-of-service (OOS) measurement.

Events that allow for the in-service evaluation of ES and SES parameters

The evaluation criteria are made based on detecting anomalies and defects in-ser-vice. An anomaly appears when a change occurs in the value of one or some of theoverhead bits with respect to their nominal value, without this affecting the state ofthe signal as a whole; that is, without causing a defect. In-service anomalies in-clude:

• An FAS violation, one or more binary anomalies in a single FAS pattern;

• A CRC code word violation;

• A parity bit violation;

• An interface code violation (as in ITU-T Rec. G.703);

• A controlled slip (ITU-T Rec. G.822);


An in-service defect occurs on a path when there is a change in the state of the sig-nal. A particular defect occurs when the relevant in-service anomalies occur persis-tently. Examples of in-service defects are LOF, LOS, and AIS. Furthermore,indications of a defect or anomaly at the remote end (RDI or E-bit at 2 Mbps withCRC) let us evaluate the performance parameters in both directions of the path bymonitoring just one end. The observation of the above events (anomalies and de-fects) may be processed into ES and SES performance parameters.

The recommendation provides guidance for mapping the defects and anomaliesthat may occur in a wide variety of paths at all PDH bit rates into the ES and SESparameters (see Table 7.11).

It should be noted that in an ISM, anomalies and defects can only be detectedby checking the overhead bits that are known, such as FAS or CRC-4. Additionally,

Table 7.11In-service ES and SES parameter evaluation criteria at the primary level (M.2100).

ES/SES Measurement Criteria in 1 Second

Path Level Overhead Anom/Defects Rx Interpretation Tx Interpretation

1,54

4 kb

ps

(non

CRC

-6)

FAS,

S-b

it

≥ 1 LOF≥ 1 LOS≥ 1 AIS≥ 1 errored FAS≥ 8 frame bit error

ES + SESES + SESES + SESESES + SES

—————

1,54

4 kb

ps(C

RC-4

)

CRC

-6, F

AS

≥ 1 LOF≥ 1 LOS≥ 1 AIS≥ 1 CRC-6 block errors≥ 320 CRC-6 block errors≥ 1 LOF sequence

ES + SESES + SESES + SESESES + SES—

—————ES + SES

2,04

8 kb

ps

(non

-CRC

-4)

FAS,

A-b

it

≥ 1 LOF≥ 1 LOS≥ 1 AIS≥ 1 errored FAS≥ 28 bit errors≥ 1 RDI

ES + SESES + SESES + SESESES + SES—

—————ES + SES

2,04

8 kb

ps (C

RC-4

)

CR

C-4

, E-b

it,

FAS,

A-b

it

≥ 1 LOF≥ 1 LOS≥ 1 AIS≥ 1 CRC-4 block anomalies≥ 805 CRC-4 block anomalies≥ 1 E-bit≥ 805 E-bits≥ 1 RDI

ES + SESES + SESES + SESESES + SES———

—————ESES + SESES + SES


the detection of an event may simultaneously generate an ES and an SES on manyoccasions, while sometimes only an ES is generated.

Events that allow for the out-of-service evaluation of ES and SES parameters

The evaluation criteria are based on detecting anomalies and defects out of service(OOS). An OOS anomaly occurs when there is a fundamental change in the testsignal that does not cause a defect. For example, OOS measurements usually em-ploy a PRBS, and therefore permit resolution to the bit level. As a result, a bit erroris the most basic OOS anomaly that can be measured. An out-of-service defect oc-curs when there is a change in the state of the test signal with respect to its normalstate. Loss of sequence synchronization can occur as a consequence of long dura-tion intense error bursts, long duration AIS, uncontrolled bit slips, and loss of signal(LOS) (see Table 7.12).

7.7.4.3 Performance objectives

The overall objectives for ES and SES performance parameters for a 64-Kbps pathare defined below. The reference performance objectives (RPOs) defined byM.2100 for a 64-Kbps path allow for compliance with the end-to-end objectivesbased on ITU-T Rec. G.821 (see Table 7.13).

Table 7.12Evaluation of ES and SES based on OOS anomalies and defects.

Type of Test Device Anomalies/Defects in 1 sec Rx Interpretation

With PRBS not incorporatedin a standardized path signal

≥ 1 bit error

Integrated BER > 10-3

≥ 1 code violation≥ 1 AIS≥ 1 LOS

ESES + SESESES + SESES + SES

With PRBS incorporated ina standardized path signal(pattern + frame)

≥ 1 bit error

Integrated BER > 10-3

+ anomalies, defects described in ISM

ESES + SESThose that correspond

Table 7.13End-to-end RPO at 64 Kbps.

Parameters End-to-End RPO (Max. % of Time)

Errored Seconds (ES)Severely Errored Seconds (SES)

4.00.1


Performance objectives for paths at PDH rates defined by ITU-T Rec. M.2100maintain alignment with the end-to-end objectives defined in ITU-T Rec. G.826 andmake up 50% of the values in that recommendation (see Table 7.14).

7.7.4.4 Allocation of objectives

The international digital reference path is divided geographically with the aim ofdistributing and allocating part of the RPO to each portion (see Figure 7.21). Theportions that result from this division are called path core elements (PCE). Thereare two types of PCE:

1. An international path core element (IPCE), found between the frontier stations(FS) of a transit country or between an IG and the frontier station in a terminat-ing country (see Figure 7.20).

2. An intercountry path core element (ICPCE) is situated between adjacent fron-tier stations from two different countries. This ICPCE is a higher-order digitalpath that joins the two countries and may be on an earth, based satellite orundersea transmission cable (see Figure 7.21).

For each portion of the international path, whatever type of path core element itmay be, the air route distance must be determined and multiplied by a routing fac-tor (RF) that depends on the value of this distance (RF=1.5 if d < 1,000 km;RF=1.25 if d > 1,000 km). For terrestrial paths, the distance chosen should be theair-route distance or its effective length, if known, whichever is shorter.

Table 7.14Performance objectives for PDH paths.

Network Level Max. ES (% of Time) Max. SES (% of Time)

Primary 2 0.1Secondary 2.5 0.1Tertiary 3.75 0.1Quaternary 8 0.1

2 Mbps

2 %

140 Mbps 2 Mbps 14 Mbps 2 Mbps

1 %

IPCE or ICPCE

1 %

Figure 7.20 IPCE or ICPCE path.


Once the distance is known, each portion is allocated a specific percentage ofthe end-to-end RPOs previously defined for each rate. The sum total of the alloca-tions of all the portions that make up the international path cannot exceed the 40%allocated to this path. If an IPCE or ICPCE path is really the sum total of several sec-tions, it may be necessary to make specific allocations to each section, in such a waythat the total does not exceed the values indicated above (see Figure 7.20).

7.7.4.5 Performance limits

In real measurements, the values of the parameters obtained during these measure-ments, the measurement duration, and the limits used for the procedure need not beidentical to those used for specifying the performance objectives, as long as they re-sult in network performance that meets these objectives. For example, the error per-formance objectives refer to long periods, such as one month. However, practicalconsiderations demand that maintenance and BIS limits be based on shorter mea-surement intervals. Statistical fluctuations in the occurrence of anomalies and de-fects mean that, while the performance values obtained in short measurementperiods meet the objectives, one cannot be certain that these objectives will alwaysbe met in the long run. Limits on the number of events and the duration of measure-ments attempt to ensure that systems or paths exhibiting unacceptable performancein maintenance tasks can be detected, or state with a certain margin of confidencethat the long-term objectives will be met for BIS tasks. Limits are needed for thefollowing functions:

IGFSIG

International portionNationalportion

Nationalportion

IB: International Border

FS FS FS FS FS

IB IB IB IB

IPCE ICPCE IPCE ICPCE IPCE ICPCE IPCE

FS: Frontier StationIG: International Gateway

(a)

Figure 7.21 International digital reference path. The ICPCE crosses two international borders and is typically on a satellite or undersea cable transmission system.

ICPCE: Intercountry Path Core ElementIPCE: International Core Element


• Bringing into service (BIS);

• Keeping the network operational (maintenance);

• System restoration.

The ITU-T recommendations do not provide limits for the installation and accep-tance of transmission systems.

Performance limits for BIS

The derivation of the BIS limits for international paths in all PDH hierarchies is afunction of a given allocation and the duration of the measurement, and depends onthe parameters and objectives of Recommendations G.821 and G.826 (see Table7.15). Two limits, S1 and S2, are defined for use in BIS tests. If performance is bet-ter than the first limit (S1), the entity can be brought into service with some confi-dence. If performance is between the two limits, more testing is necessary, and theentity can only be provisionally accepted. Corrective measures are required if per-formance is worse than the second limit (S2) (see Figure 7.22).

The difference between the RPO and the BIS limit is called the aging margin.This margin should be as large as possible, to keep maintenance interventions to aminimum. Its value is 0.5 times the RPO. The testing duration will obviously be lim-ited to no more than a few days. The BIS limits are calculated as follows:

Figure 7.22 Bringing into service limits and conditions.

No. of events

S2

S1

BIS objective (RPO/2)

The objective

to be satisfied

D

D

is likely

The objective

to be satisfiedis unlikely

uncertainlyuncertainly

bringing into service aborted

further testing

bringing into service accepted

S1 RPO2

------------ 2 BIS objective( )–=

S2 RPO2

------------ 2 BIS objective( )+=


where:RPO = A x TP x PO IS objective = RPO / 2 A: Path allocation (%) TP: Test period in seconds PO: Performance objective

The S1 and S2 limits are rounded to the nearest integer value. The applicablerange for TP goes from a minimum of two hours to several days.

Performance limits for maintenance

Once the paths are in service, supervising the network calls for additional limits tobe specified, as described in CCITT Rec. M.20. Three operational areas are speci-fied for the supervised network with long-term performance limits (more than onemonth) as follows:

• Acceptable level of performance: The level includes the BIS performance lim-its and the performance limits following a repair, with an objective of around0.125 times the RPO.

Table 7.15Allocation of end-to-end RPOs to international and intercountry path core elements.

PCE Classification of Portions Allocation (% RPO)

IPCE national networks

d ≤ 500 km500 km < d ≤ 1,000 km

1,000 km < d ≤ 2,500 km2,500 km < d ≤ 5,000 km5,000 km < d ≤ 7,500 km

d >7,500 km

2.03.04.06.08.0

10.0

ICPCE undersea cable

d ≤ 500 km5,00 km < d ≤ 1,000 km

1,000 km < d ≤ 2,500 km2,500 km < d ≤ 5,000 km

d >5,000 km

2.03.04.06.08.0

ICPCE optical undersea cabled ≤ 500 kmd >500 km

1.02.5

ICPCE terrestrial d< 300 km 0.5

ICPCE satellite Normal operation 20.0


• Degraded level of performance: The limit is obtained from an objective ofaround 0.5 times the RPO for transmission systems, and 0.75 times the RPOfor paths and sections.

• Unacceptable level of performance: The limit is obtained for a given entityfrom an objective of at least 10 times the RPO.

Performance thresholds T1 and T2 are defined for ES and SES in accordance withmonitoring time, 15 minutes, and 24 hours, respectively. T1 is to assist in detectingunacceptable performance, and T2 in detecting degraded performance.

7.7.4.6 Available and unavailable time

A period of unavailable time starts with the first of 10 consecutive SES events.These 10 seconds are considered part of the unavailable time. The period of avail-ability starts with the first of 10 consecutive non SES events. These 10 seconds areconsidered part of the available time. The criterion for a path is that a path is un-available if either one or both directions is unavailable. This definition is intendedto be in line with Annex A of ITU-T Rec. G.826.

7.7.5 Recommendations M.2110 and M.2120

ITU-T Rec. M.2110 refers to procedures to be carried out for the bringing-into-ser-vice of paths, sections, and digital transmission systems. This recommendationspecifies the different measurements, configurations, and processes, together withduration intervals.

In addition, ITU-T Rec. M.2120 specifies the procedures to be carried out fornetwork supervision and maintenance.

7.7.6 Open Network Provision for Leased Lines at 2,048 Kbps

Open network provision (ONPs) are European Telecommunications Standardiza-tion Institute (ETSI) regulations for structured and unstructured circuits includeboth the requirements that these circuits must meet with respect to interfaces, jitter,delay, performance, and so on, and the measurement procedures that check thiscompliance. These tests are designed for BIS and restoration, that is, they are OOSmeasurements.

7.8 TESTS ON ADMS AND DXC

Acceptance tests must be carried out on NEs once they have been purchased, tocheck that they are operating correctly. Today, the divisions between different NEs,is more functional than real, since the equipment sold can usually be configured for


more than one function. An ADM, for instance, could be configured as a terminat-ing multiplexer. Likewise, a DXC, by definition, has add and drop capabilities.

7.8.1 Tributary Continuity Test

This test checks that the signals circulating through the tributary ports have electri-cal continuity through the ADM, and no events (anomalies or defects) occur whenthey pass through the NE. Likewise, it is also useful to check the correct mechani-cal wiring between the tributary ports and the possible additional connection panelsavailable in the frame in which the device has been installed (see Figure 7.23).

A loop must be set up in one of the aggregate ports. Other loops are also set upbetween the input and output of each tributary. This ensures that all the tributariestransmitted in the West-East aggregate flow are received in the East-West aggregateflow, having circulated through the tributary ports. The test must check that the sig-nal transmitted has been correctly received. Events can also be introduced on gener-ation (bit errors in the test signal generated) to check that these are detected onreception. The test can be repeated, changing around the aggregate ports.

A generator/analyzer must be connected to the free aggregate port. The genera-tor will transmit an aggregate signal that contains a test pattern in the tributary to betested. In the rest of the tributaries, background signals are sent which may also bePRBS signals or some other type; for instance, an “all ones” signal. The same oper-ation should be repeated for each tributary. Before carrying out the test, the testermust be synchronized with the ADM, in order to avoid uncontrolled pointer move-ments. For this, either a common synchronism reference source can be used, or theADM can be configured to obtain its synchronization clock from the aggregate sig-nal received from the generator.

STM-n

tributaries

(loop)

Test signal

Figure 7.23 Setup for continuity test.

West EastADM

Distribution panel


When operating with optical signals, an optical attenuator must be connectedbetween an optical output and input to avoid saturating the receiver if the transmitterand receiver are not of the same type (long haul or short haul).

7.9 TESTS ON SDH/SONET RINGS

Ring topologies are widely used at different levels in national networks. There aremany reasons behind this:

• They are the simplest way to connect different NEs;

• They are a good investment in terms of future extension and network growth;

• They make it possible to set up a variety of protection mechanisms (APSs);

• They provide immediate and flexible synchronous paths between users, whichin turn provides immediate access to all the path management facilities.

Rings use optical fiber as their transmission medium, and in some cases radio trans-mission. They basically connect ADMs, although they can also connect to DXCs. Itis therefore necessary to check that these NEs interact correctly within the ring. Ba-sically, the idea is to check the reliability of the rings installed and their compliancewith the appropriate ITU-T recommendations. In order to do so, it is not enough tosimply check each of these NEs, and the transmission media used, one by one tomake sure that they are operating correctly; they also have to operate correctly as awhole when installed in the ring. Bearing the above considerations in mind, the fol-lowing sections describe the tests to be performed when bringing rings into service.

7.9.1 Transparency Tests

These tests check that an NE does not deteriorate the signals it processes. In orderto do so, the integrity of the information transported in the virtual containers mustbe checked, together with the integrity of the information sent through the overheadchannels; for example, the DCCs transport network management information thatthe different NEs exchange between themselves or with the telecommunicationsmanagement network (TMN), (see Section 2.14).

To perform the test, the bandwidth of the container transported by the virtualcontainer is filled with a PRBS test signal, or the DCCs are taken up with a PRBSsignal, depending on whether we wish to check the integrity of the user channels orthe management channels. The BER is then measured, and a G.826 performance testcan also be carried out (see Figure 7.24).


7.9.2 Multiplexers

The aim of this test is to check that the add and drop function of the ADM is per-formed correctly (see Figure 7.25).

The add and drop capacity of each of the ADMs is checked when the signalpasses transparently through the rest. For this, the appropriate loops are set up in thetributary ports and in the far end aggregate port in the section of the ring open for thetest. Once this is done, we must check that there are no bit errors and alarms in thetributary channels using the automatic event scan function usually included intesters.

PRBS in container or in DCC

BER, G.826

Figure 7.24 Transparency tests for the bringing-into-service of rings. The test instrument gener-ates a PRBS in a container (O.181 test structure) and checks G.826 performance.

Loop

ADM

Loop

...

...

...

... Loops intributaries

Figure 7.25 Add and drop test. The test instrument generates a PRBS in a container and checks for the absence of events in the tributaries.


7.9.3 Synchronization Measurements

The usual synchronization system for rings consists of connecting one of the ADMsto an external synchronization source. The synchronization obtained is then trans-mitted from this NE to the other elements in the ring via the STM-n/OC-m aggre-gate signal (see Chapter 5).

When a ring is brought into service, each ADM multiplexer must be pro-grammed according to the table of synchronization priorities established for the ring.This way, primary and secondary synchronization sources are established in a coher-ent manner for each ADM. Once this programming is complete, it must be checkedif it has been carried out correctly. Otherwise, the ring will not be correctly synchro-nized. This would lead to the appearance of high levels of output jitter several uni-tary intervals, in the tributary of the badly configured multiplexers (see Figure 7.26).

The above measurement establishes whether or not the ring is correctly synchro-nized. However, it does not allow conclusions to be made on the quality of synchro-nization in the long run. For this, another test is needed that is based on the analysisof pointer adjustments. This measurement consists of analyzing the traffic in the ringover a long enough period of time (e.g., 24 hours). This measurement can be per-formed in-service or out-of-service. If a high number of pointer adjustments orpointer actions with a sporadic temporal distribution are recorded over this period,the synchronization of the ring is disturbed and the ADM in which it originates mustbe identified. The tester is connected in the section of the ADM with an external syn-chronization reference. When pointer adjustments are identified, pointer analysismust be carried out in the ADMs until the origin of the disturbance is identified.

If a more accurate measurement is required, a wander measurement can be car-ried out using the same basic setup (see Figure 7.27) that will identify the contribu-tion to wander of each ADM, and identify its causes. In this case, the wander

STM-n

Jitter and tributarygenerator/analyzer

2 Mbpstributary

2 Mbps

Externalsynchronization

reference

Figure 7.26 Synchronization test based on jitter in tributary ports.


analyzer needs to be synchronized with a high-quality external reference (e.g., a ce-sium clock).

7.9.4 Protection Switching Tests

When a protection architecture has been set up for the multiplexer sections in aring, it is necessary to check that it is working correctly (see Section 2.13). Basical-ly, this is an APS protocol in which the following must be checked:

• The time taken to switch over from work to protection circuits;

• The correct operation of the APS protocol through K1 and K2 bytes.

You can check the switching time by measuring the time during which the signal ina tributary port shows bit errors or alarms. The second point can be checked by cap-turing frames and analyzing the contents of the multiplexer section overhead bytesK1 and K2 (see Section 7.6).

7.9.5 Defect Indicators in the Network Management System

The TMN provides management functions for telecommunications services andnetworks, and offers communication between itself and the telecommunicationsservices and networks. It enables you to connect different types of operating sys-tems and/or telecommunications equipment for the exchange of management infor-mation by using an agreed-upon architecture and standardized interfaces, includingprotocols and messages (see Section 2.10).

One of the responsibilities of the TMN is to obtain reports on the events detectedby the different NEs. By means of interfaces known as Q interfaces, the NEs com-municate these reports to computers that have the TMN software installed. Detecting

STM-n

Pointer event

Externalsynchronization

reference

Optical

10%

90%

splitteranalyzer

Figure 7.27 Synchronization test based on pointer events.


and reporting on the different defects that occur in the network is a maintenance taskknown as supervision of defects.

The status of defects is communicated by means of messages that circulatethrough the DCC of the section overheads in the transmission frames; in particular,the channel made up of bytes D4 to D12 of the MSOH (see Figure 2.26). One of thebringing-into-service tests for rings checks that the defect supervision system is op-erating correctly. This test consists of generating events to check that these are de-tected, that the corresponding indicators (LOS, AIS, etc.) are reported to the TMN,and that the TMN receives them correctly. The ITU-T recommendations relating toTMN that illustrate these aspects are G.783 and G.784 (see Figure 7.28). The basicTMN concepts are defined in ITU-T Rec. M.3010.

7.9.6 Path Trace Tests

For tributaries to be correctly connected their source must be identified. This is afundamental requirement when managing SDH/SONET networks, in particular forthe routing control search algorithms. Each source, known as network access point,has an access point identifier (API) that starts with the country code defined in bothITU-T E.164 and ISO 3166; the remaining characters are defined in ITU-T T.50. Inthe overhead of an STM-1 frame there is a series of bytes, known as path tracebytes, that define 64-Kbps channels to transport the access point identifiers that in-form the receiver on the source of the signal received. These are transmitted repeat-edly across the channels defined for these bytes.

There are path trace bytes at the following levels:

• J0, for the regeneration section;

• J1, for the VC3, VC4, VC-4-Xc, STS-1 SPE, and STS-Xc SPE paths;

• J2, for the VC-11, VC-12, VC-2, VT-11, VT-12, and VT6.

Tributary

Q interface

ADM

Figure 7.28 A test of TMN function for the supervision of defects.


The source is identified by sending an API across these 64-Kbps channels as astring of characters. This string can either be 64 characters long, or 15 charactersplus one CRC byte over the previous frame. For J0 and J2, a 16-character string isdefined in G.831. For J1, there is the option of a free format string of 64 charactersor the 16-byte string.

Path trace tests are aimed at checking that the NE under test has the correct APIconfiguration. For this, the tester will generate an API that is then compared to theone expected by the NE under test (previously programmed with the same API) (seeFigure 7.29). After checking that no defects are reported (i.e., the API is correctlyconfigured in the NE), the identifier generated by the tester should be changed toprovoke a trace identifier mismatch (TIM) defect in the NE, and an RDI in the op-posite direction. This test should be performed on each element within the ring tocheck the path trace mechanism.


• CCITT Rec. M.20, Maintenance: introduction and general principles.

• ITU-T Rec. E.164, The international public telecommunication numbering plan.

• ITU-T Rec. G.114, One-way transmission time.

• ITU-T Rec. G.703, Physical/electrical characteristics of hierarchical digital interfaces.

• ITU-T Rec. G.704, Synchronous frame structures used at 1544, 6312, 2048, 8448 and 44736 kbit/shierarchical levels.

• ITU-T Rec. G.775, Loss of Signal (LOS), Alarm Indication Signal (AIS) and Remote Defect Indica-tion (RDI) defect detection and clearance criteria for PDH signals.

• ITU-T Rec. G.782, Types and general characteristics of synchronous digital hierarchy (SDH) equip-ment.


• ITU-T Rec. G.784, Synchronous digital hierarchy (SDH) management.


RDI, TIM if the API is changed

NE

STM-n/OC-m test API

from the generatorControl computer(TIM defect detected)

Figure 7.29 Testing the API configuration.


• ITU-T Rec. G.821, Error performance of an international digital connection operating at a bit ratebelow the primary rate and forming part of an integrated services digital network.


• ITU-T Rec. G.826, Error performance parameters and objectives for international, constant bit ratedigital paths at or above the primary rate.

• ITU-T Rec. G.831, Management capabilities of transport networks based on the synchronous digitalhierarchy (SDH).

• ITU-T Rec. G.841, Types and characteristics of SDH network protection architectures.

• ITU-T Rec. M.2100, Performance limits for bringing-into-service and maintenance of internationalPDH paths, sections and transmission systems.

• ITU-T Rec. M.2100.1, Performance limits and objectives for bringing-into-service and maintenanceof international SDH paths and multiplex sections.

• ITU-T Rec. M.3010, Principles for a telecommunications management network.

277

Chapter 8

ATM Performance

8.1 INTRODUCTION

In this chapter, we will describe the parameters used to describe the performanceprovided by the ATM network, as well as control information carried by an ATMnetwork that provides us information about its status. Techniques used to measurethe performance parameters, as well as tests to check the operation and mainte-nance functions, are described. As already introduced in SDH/SONET measure-ments (see Section 6.3), two measurement modes exist in general:

• Out-of-service mode: A dedicated connection used to carry out measurementsby transmitting a special test traffic and analyzing it at the receiving side. Theconnection does not carry actual user traffic.

• In-service measurement mode: In this mode, the measurement is carried outover a connection carrying user traffic. Usually it is based on analyzing controlfields and on special cells generated by the network to ensure the transmissionof user data.

Testing is involved in multiple phases during the network life cycle (see Section8.5). Depending on that, tests of a single device (device under test) will be carriedout, or tests of a portion of the network; for example, an end-to-end connection withmultiple network elements involved (system under test or SUT).

8.2 PERFORMANCE PARAMETERS IN ATM NETWORKS

To define the operation of an ATM network and evaluate its features, a set of pa-rameters has been defined, known as ATM cell transfer performance parameters.The values of these parameters are calculated from cell transfer results (ATM celltransfer outcomes) that are obtained by observing cell-based reference events atmeasurement points.


8.2.1 Cell-Based ATM Reference Events

To evaluate the characteristics of an ATM network, we must monitor user cells asthey are transmitted in a connection. ITU-T Rec. I.356 defines the following refer-ence events for these cells:

• A cell exit event: This event is produced when the first bit of a user cell hasbeen transmitted by the output port of terminal equipment, intermediate nodes,or ATM switch.

• A cell entry event: This event occurs when the last bit of an ATM user cell hasbeen received at the input port of terminal equipment or an intermediate node.

8.2.2 ATM Cell Transfer Outcomes

When a cell has been transmitted between two points, and a cell exit event and acell entry event have occurred, there are the following possibilities for a cell trans-fer outcome (ITU-T Rec. I.356):

• A successful transfer outcome: When a cell entry event occurs within a maxi-mum interval T, after a cell exit event has occurred and the following condi-tions are met: (a) The cell header is valid; and (b) the payload contents of thereceived cell are exactly the same as those of the transmitted cell.

• A tagged cell outcome: This transfer outcome occurs when the same condi-tions are met as in the previous case, but, furthermore, the value of a priority bit(cell loss priority or CLP) has changed during transmission, from 0 to 1.

• An errored cell outcome: When a cell is received within a maximum interval Tfrom its transmission, but the contents of its payload are different from trans-mission payload or header contents, it is considered not valid after completingthe HEC procedure.

• A lost cell outcome: When the maximum time T is exceeded in transmissionand no cell entry event has occurred.

• A misinserted cell outcome: When a cell is received even though there hasbeen no cell exit event. This may be caused by transmission errors that modifythe header and the VPI/VCI values, giving them another combination, valid asregards the HEC. When VPI/VCI values change, if these new values are usedby an existing connection, it causes another piece of end equipment to receivethe cell, although the originating end has not generated it. Misinserted cells arealso generated due to internal errors in the switching matrix of an ATM switch.There is a certain probability that, due to internal errors, a cell is switched to anincorrect output port and addressed to a link that belongs to a connection otherthan the original one.

ATM Performance 279

• A severely errored cell block outcome: A block of cells is a group of n cellstransmitted consecutively across a connection. A cell block is considered as se-verely errored when M or more results of an errored, lost, or misinserted cell isdetected in a received cell block.

8.2.3 ATM Cell Transfer Performance Parameters

Taking into account the cell transfer outcomes described above, we can calculatethe following parameters.

8.2.3.1 Cell error ratio

The cell error ratio (CER) is defined as the ratio between the total number of er-rored cells and the total number of received cells, whether they were transmittedsuccessfully, tagged, or erroneous. The errored or tagged cells or the cells transmit-ted successfully that belong to severely errored blocks must be included in the cal-culation of the CER:

From a measurement point of view, the objective is to know the CER parameterassociated with an ATM connection. Although the definition of errored cell outcome(see Section 8.2.2) considers cells with HEC violation as errored cells, actually,when computing the CER of a connection, only cells with errored payload are con-sidered. The CER parameter is computed by checking only cells received and be-longing to the connection analyses. If an HEC violation is detected, there is a highprobability that the VPI/VCI fields are errored; so, it is not possible to know the con-nection and the cell is discarded. Therefore, for the final user of a connection, a cellwith an errored header will be computed as a lost cell, and only cells with erroredpayload are taken into account to measure the CER (see Sections 8.3.6 and 8.4.2 be-low for a description of CER measurement in-service and out-of-service, respective-ly).

8.2.3.2 Cell loss ratio

The cell loss ratio (CLR) is defined as the ratio between the total number of lostcells and the total number of cells transmitted in a connection. Lost cells and cellstransmitted as severely errored cell blocks must be included in the calculation of theCLR:

CER errored cellscorrectly received cells errored cells+---------------------------------------------------------------------------------------------=

CLR 1 received cellstransmitted cells---------------------------------------– lost cells

transmitted cells---------------------------------------= =


We can define a different loss rate for each type of traffic based on the cell losspriority (CLP) (see Section 3.2.1) of a connection. The traffic can be CLP=0,CLP=1, and CLP=0+1 aggregate traffic:

Cell loss ratio type 0 (CLR0)

The loss rate for CLP=0 traffic is the ratio between the lost cells with CLP=0 plustagged cells, and the number of transmitted CLP=0 cells.

Cell loss ratio type 1 (CLR1)

The loss rate for CLP=1 traffic is the ratio between the lost CLP=1 cells and the to-tal number of transmitted CLP=1 cells.

Cell loss ratio type 0+1 (CLR0+1)

The loss rate for CLP=0+1 traffic is the ratio between lost cells and the total num-ber of transmitted cells. If all the cells transmitted are CLP=1, CLR0+1 coincideswith CLR1.

In the above definitions, all lost cells are registered, and it does not make anydifference if this is due to congestion or a network problem (nonconforming cellsdiscarded by UPC/NPC policing functions). Therefore, these parameters measurethe end-user perception, without really indicating the performance of the connec-tions used. To evaluate the performance or the quality of connections, cells discardedby the network should not be taken into account, as they are not conforming with thetraffic specified when establishing the connection, and do not indicate any malfunc-tion of the connection.

8.2.3.3 Cell misinsertion rate

The cell misinsertion rate (CMR) is the number of misinserted cells observed dur-ing a time interval, divided by the duration of this interval. Misinserted cells andthose observation intervals related to severely erroneous cell blocks are excludedfrom the calculation of this parameter.

CMR misinserted cellstime interval

----------------------------------------=

ATM Performance 281

8.2.3.4 Severely errored cell block ratio

The severely errored cell block ratio (SECBR) is the ratio between severely errone-ous blocks and the number of transmitted blocks:

8.2.3.5 Cell transfer delay

Cell transfer delay (CTD) is the time interval between a cell exit event at a mea-surement point, and a cell entry event at another measurement point. This parame-ter is registered only if the delay is smaller than the maximum delay allowed. Inother words, the cell transmission outcome may be successful, erroneous, ortagged.

The mean transmission loss is the mean arithmetic value of transmission delaysof a certain number of cells.

8.2.3.6 Cell delay variation

The parameters related to cell delay variation (CDV) are defined in the following.

One-point cell delay variation (1-CDV)

This parameter is calculated by first observing arrival time sequences for consecu-tive cells at a measurement point. These arrival times are compared to theoreticalarrival time sequences obtained from the inverse value of the peak cell rate (PCR)T=1/PCR. This parameter measures the variability in the arrival time pattern in re-spect to the agreed PCR. This parameter includes both cell delay variation andsource variation during transmission, be they due to randomness in transmission,CDVT, or a change in transmission rate.

1-CDV for the k cell is defined as the difference between the reference cell ar-rival time, ck, and the real arrival time at the measurement point, ak (see Figure 8.1).

Two-point cell delay variation (2-CDV)

This parameter is calculated by measuring the cell transfer delay between twopoints, and comparing the obtained values. The 2-CDV describes the variability in

SECBR severely errored cell blockstotal transmitted blocks

------------------------------------------------------------------=


the transfer delay, since only the delay introduced by a connection between twomeasurement points is measured.

The 2-CDV for the k cell is defined as the difference between the absolute trans-fer delay of the k cell, and an absolute reference transfer delay defined between thetwo points. Usually, in a real measurement, the reference is the absolute transfer de-lay experienced by the first test cell (see Figure 8.2).

8.2.4 Performance of Permanent Connections

In permanent ATM connections whose service is similar to that of point-to-pointcircuits, besides the parameters that describe transmission performance, there is aset of parameters that describe the performance of the connection. These parame-ters are defined by ITU-T Rec. I.357, and some of them coincide with the ones usedto describe the performance of a circuit.

a0 a1 a2 a3

1-CDV

T=1/PCR T=1/PCR T=1/PCR

tinput

1-CDV1-CDV

1-CDV = c k - a k

Figure 8.1 Representation of the 1-CDV parameter.

Cell arrival

c0 c1 c2 c3

Figure 8.2 Representation of the 2-CDV parameter.

c0 c1 c2 c3

CTD0 CTD3CTD1

c0 c1 c2 c3

CTD2

2-CDV1 = CTD1 - CTD02-CDV2 = CTD2 - CTD0

ttransmission

treception

ATM Performance 283

Severely errored second

We can talk about a severely errored second when one of the following conditionsis met in a connection during one second:

• CLR > 1/1,024;

• SECBR > 1/32;

• The connection is interrupted.

Unavailable time

If a connection enters a period where it is not available during 10 consecutive SESs,this is called unavailable time, also known as down time. These 10 seconds are con-sidered as a part of unavailability time. The connection remains unavailable untilthere is a period of 10 consecutive seconds that are not SES. These 10 non-SES sec-onds will be considered as a part of available time or up time.

There are two more parameters related to availability and unavailability time:

1. The available rate: This is the relation between down time and the total obser-vation time.

2. The mean time between outages (MTBO): This is the mean duration of theavailability period. MTBO= E availability period.

It is important to note that for purposes of performance assessment of a connection,error performance parameters (i.e., CER, CLP, CMR, and SECBR) should only becalculated during available time. Therefore, successfully transferred, errored, mi-sinserted, and lost cells as well as SECBs detected during unavailable time periods,are not accumulated in order to compute error performance parameters.

8.3 OAM FUNCTIONS: IN-SERVICE MEASUREMENTS

Measuring the parameters described in the previous section is not enough to pro-vide the network status and operation. As in any communications network, we alsoneed functions that give us information on the status of the network and its connec-tions, and inform us about any transmission problems. This data is used to managethe network and its services. Furthermore, these functions enable us to make in-ser-vice measurements of the cell transfer performance parameters described (see Sec-tion 8.2.3).

On the other hand, an important characteristic of today’s telecommunicationsnetworks is that they are multivendor networks. If each manufacturer only offered


network management support for its own devices, from a management point of viewthis would create a network structure divided into separate domains. This creates aneed to standardize the network status information, and the interfaces used to accessthe network.

ATM network management, as defined by the ATM Forum, uses the manage-ment model of IP networks based on the management information base (MIB) struc-ture, and on the SNMP protocol used in IP networks, defined by the InternetEngineering Task Force (IETF). It has also standardized a telecommunications man-agement network interface for accessing the information in the MIBs. The ATM Fo-rum has also defined the objects for the specific MIBs of the ATM network thatcontain information on the status of the network and its connection elements andequipment.

The information that the objects belonging to the ATM MIB contain is obtainedby continuously monitoring the links in the network. This also makes it necessary tostandardize the methods used to transport network status information, so that anynetwork equipment of any manufacturer is able to know this status, and generate thecorresponding management information.

8.3.1 Presentation of OAM Functions

To offer greater integration and make it easier to manage ATM networks, the ITU-T defined a series of cells intended to carry network status information. These cellsare inserted in the overall user data flow. The specification of the format of thesecells and their functions make up the OAM protocol for broadband networks (B-IS-DN), as defined in the ITU-T Rec. I.610.

Processing and exchanging information using OAM cells enables the OAMfunctions that reside in the physical and the ATM layer to detect and locate failuresin the network, monitor the integrity of connections, and evaluate the performanceof data transfer through these connections. The fact that part of the network capacityis reserved to carry network management and maintenance information means thatthere is an increase in the overhead, and less user data information can be carried.When the OAM functions of an ATM network are specified, we can carry out thefollowing operations:

• Performance monitoring: Information is added to the user cell flow, to evaluatethe integrity of the transmission being analyzed.

• Defect and failure detection: The connection is monitored continuously, andalarm and status indicators are generated whenever faults and errors are detect-ed in transmission.

ATM Performance 285

• Service protection: This is done by excluding from operation all those entitieswhere faults occur, thus maintaining the service; that is, changing the configu-ration of all the connections and services supported by the faulty element, inorder to maintain the delivery of the service.

• Supply information on defects: When defects occur, this information is passedon to other management entities by providing alarm indications and respondingto requests for situation reports.

• Locate failures: This determines the place where a failure has occurred.

OAM functions are divided into five hierarchical layers that are associated with thephysical layer and the ATM layer of the protocol reference model. The OAM func-tions defined in each OAM layer generate corresponding data flows, and monitorthe connections and the quality of data transmission corresponding to each layer.

8.3.2 Physical Layer OAM Procedures

The first three hierarchical OAM layers are part of the physical layer of the protocolreference model:

• The regenerator section layer;

• The digital section layer;

• The transmission section layer.

The OAM data flows generated to monitor each of the layers are called the:

• F1 flow: Data flow for OAM functions in the regenerator section layer;

• F2 flow: Data flow for OAM functions in the digital section layer;

• F3 flow: Data flow for OAM functions in the transmission section layer.

How to describe and implement the OAM functions and the OAM data flows forthe layers associated with the physical layer depends on the type of transmissionsystem used (SDH, PDH, cell based, etc.).

8.3.2.1 SDH-based physical interface

If cells are carried across SDH, the OAM functions of the OAM layers associatedwith the physical layer will use the capacities provided by the SDH transport struc-ture to transfer information on the network status and performance (see Section2.11). The OAM procedures on this layer monitor and evaluate the quality and in-tegrity of SDH signals.


Table 8.1 summarizes the physical layer OAM functions and the OAM dataflow generated on each layer. When a defect is detected on a specific OAM layer inthe incoming signal in a network element, this may generate an OAM data flow inthe outgoing signal on a higher layer (ITU-T Rec. I.432 for SDH-based interfaces).

Table 8.1OAM functions for an SDH interface.

Layer Function Defects Detected F2 Flow F3 Flow

Reg

ener

atio

n Se

ctio

n La

yer

Supervi-sion of errors and failures in frame alignment and signal detection

LOS or SDH (LOF) detection (A1, A2)

Upstream transmis-sion of RDI signal(K2)

Downstream transmission of path AIS signal (AU-pointer bytes H1, H2)

Upstream transmission of path RDI signal (1 bit in G1)

Checking the incom-ing signal (B1)

Dig

ital S

ectio

n La

yer

Supervi-sion of sec-tion errors

Unacceptable error characteristic (B2)

Upstream transmis-sion of RDI signal of multiplexing sec-tion (K2)

Downstream transmission of path AIS signal (H1, H2)

Upstream transmis-sion of REI signal (M2)

Pointeroperation

AU pointer loss Upstream transmission of path RDI signal (G1)

Tran

smis

sion

Lay

er S

ectio

n

Detection of path AIS

Upstream transmission of path RDI (or FERF) signal (G1)

Supervi-sion of sec-tion errors

Error characteristic (B3)

Downstream transmission of path AIS signals (H1, H2).

Upstream transmission of path RDI signal (G1)

Upstream transmis-sion of REI signal (G1)

Cell rate decoupling

Defect in inserting/deleting unoccupied cells

Cell delin-eation

Loss of cell delinea-tion (LCD)

Upstream transmission of path RDI signal (G1)

ATM Performance 287

Figure 8.3 shows the data sequence corresponding to flows F1, F2, and F3, gen-erated in an ATM network with an SDH-based physical layer. The F1, F2, and F3flows contain both information transmitted to carry out monitoring functions (suchas bytes B1 or B2), and information generated to indicate defects or anomalies in theSDH signal transmission and structure (e.g., RDI or AIS indications).

By analyzing the OAM information of the first three layers, we can detect anydefect or anomaly occurring on the physical layer that prevents the ATM networkfrom operating correctly.

8.3.3 ATM Layer OAM Procedures

The OAM procedures on the ATM layer are divided into two hierarchical layers:

• Virtual channel layer: Those functions intended to monitor a virtual channelconnection that are located in the network elements with termination functionsor with a connection of virtual channel links (network elements where the VCIis processed).

• Virtual path layer: Those functions intended to monitor a virtual path connec-tion that are located in the network elements with termination functions or witha connection of virtual path links (network elements where the VPI is pro-cessed).

AU-AIS

MS-AIS

LOF/LOS LOF/LOSLOF/LOS

B2

MS-RDI

B3

LOP

LCD

B2

HO-RDI

LOS: Loss of SignalLOF: Loss of SDH Frame synchronizationLOP: Loss of AU PointerLCD: Loss of Cell Delineation

Anomaly or defect detectedOAM data generated

Separation between the 3 layers

REI

B1

HO-REI

LOF/LOS

AIS

LOP

Figure 8.3 F1, F2, and F3 flows in a physical interface based on SDH/SONET.

ADM ADMDXCADMRegenerator


We can monitor the virtual connection on each layer either as a whole or just apart. For example, a virtual channel connection is made up of a concatenation of vir-tual channel links. We can monitor the whole connection between the ATM VC su-blayer access points, or just one of the links that make up the connection (see Figure8.4).

The OAM data flows generated to monitor each of the layers are:

• F4 flow for the virtual path (VP) layer;

• F5 flow for the virtual channel (VC) layer.

As said before, these flows are made up of cells especially designed to carry theOAM data that forms the OAM protocol for broadband networks, as defined inITU-T Rec. I.610.

8.3.3.1 F4 flow

The F4 flow cells are designed to carry OAM information to monitor virtual pathconnections (VPCs). OAM cells designed to monitor a specific VPC are inserted inthe user cell flow transmitted in this VPC, which means that both will have thesame VPI. OAM cells are distinguished from user cells because inside the VP theyuse a VC especially intended for them (a preassigned VCI).

The F4 cell flow is bidirectional, and the same VCI is used for both directions.Both directions of the flow must follow the same physical route, so that all the VClayer end or connection points that support the virtual channel connection (VCC) un-

VCI=X VCI=Y VCI=Z

VPI=M VPI=N VPI=P VPI=Q

VC

VP

VC VCVC

VP VP VP VP

VCC (Virtual Channel Connection)

VPC

Channel link Channel link Channel link

Path link Path link

VP layer connection pointVC layer termination or end pointVC layer connection point

VP layer termination or end point

AAL AAL

VPC

Figure 8.4 VC: Layer, link, and connection (VCC). VP: Layer, link, and connection (VPC).

VPC (Virtual Path Connection)

ATM Performance 289

der test can correlate the failure and performance information coming from both di-rections.

There are two types of F4 flow that may coexist in a specific VPC being moni-tored (see Figure 8.5):

• End-to-end F4 flow: A cell flow for monitoring operations between the twoends of a VPC connection, that is, a flow transmitted between two VP-layer ac-cess points (VPC terminal connection points). This flow uses the VCI=4 virtualchannel within the VPC under test.

• Segment F4 flow: A cell flow for monitoring operations for a segment of aVPC connection formed by a single VP link or several VP links connected toeach other. This flow uses the VCI=3 virtual channel within the VP link or VPsegment under test.

Virtual channel with VCI=4 across which the F4 OAM cell flow is transmitted to monitor the VPC from end to end.

Virtual channel with VCI=3 across which the F4 OAM cell flow is transmitted to monitor a segment of the VPC.

Generating end-to-end F4 flowGenerating segmented F4 flow

Receiving and processing end-to-end F4 flowReceiving and processing segmented F4 flow

Virtual channels within the VP across which those user cell flows are transmitted that are delivered by the VC layer.

VC layer

VP layer

Figure 8.5 End-to-end and segmented F4 flows in a supervised VPC.

VPI=X VPI=Z

VCI=3

VP switch

VP/VC switchVP/VC switch

VCI=4 VCI=4

VP layer


8.3.3.2 F5 flow

The F5 flow OAM cells in charge of monitoring a certain VCC are inserted into theuser cell flow transmitted across this VCC. This is why both types have the sameVCI/VPI values. In the cell sequence transmitted across a connection, the OAMcells are distinguished from user cells by the payload type identifier header value.

The F5 cell flow is bidirectional, and it uses the same PTI identifier for both di-rections. Besides, the two ways of the flow must follow the same route, so that allthe end points or the VC layer connection points that support the VCC can correlatethe failure and performance information coming from both directions.

There are two types of F5 flow that may coexist in a certain VCC that is beingmonitored (see Figure 8.6):

• End-to-end F5 flow: A cell flow for monitoring operations between the twoends of a VCC connection, that is, a flow transmitted between two VC layeraccess points (VC layer terminal points). This flow is identified by the code

VP/VC Switch

End device

VP/VC Switch

VPI=X

VCI=n

VP layer

VC layer

AAL layer

Concatenating the VCs that form the monitored VCC. A cell flow formed by usercells and OAM cells differentiated by the PTI value is transmitted through them.

Generating end-to-end F5 flowGenerating segmented F5 flow

Receiving and processing end-to-end F5 flowReceiving and processing segmented F5 flow

Concatenating those VCs that form VCCs established by other users.

Figure 8.6 End-to-end and segmented F5 flows transmitted across the monitored VCC.

VPI=W VPI=Z

VCI=m VCI=m VCI=qVCI=n VCI=q

PTI=100PTI=101

End device

ATM Performance 291

PTI=101, to differentiate it from the user cells (PTI=0xx) of the monitored con-nection.

• Segmented F5 flow: A cell flow for monitoring operations of a segment of aVCC connection formed by one or many interconnected VC links. The cells inthis flow are identified by the code PTI=100, to distinguish them from the usercells that flow across the monitored VCC.

8.3.4 ATM Layer OAM Cells

F4 and F5 cell flows, be they end-to-end or segmented, are formed by the same typeof OAM cells. As described above, it is the information that the cell header containsthat enables us to determine the type of flow the cell belongs to.

8.3.4.1 Types of OAM cells

The different OAM functions, situated at VC and VP layers of the ATM layer, pro-cess each type of OAM cells to obtain information on the network status and theperformance of the connections in question.

The different types of OAM cells are defined in ITU-T Rec. I.610, and their ap-plication for monitoring links at both VC (VCCs) and VP (VPCs) layers (see Table8.2).

Table 8.2ATM layer OAM functions in brief.

Type Function Application

Faul

tM

anag

emen

t AIS Communicates defect indications in the forward directionRDI Communicates defect indications in the backward direction

Continuity check Supervises continuity permanently, to monitor the integrity of a connection when there is no user data flow

Loopback Supervises connectivity and fault location

Perf

orm

ance

M

anag

emen

t Forward performance monitoring (FPM) Feature assessment

Backward performance monitoring (BPM) Backward performance assessment

Act

ivat

e/D

eact

ivat

e

Activating/DeactivatingActivates/deactivates performance assessmentActivates/deactivates the continuity check


8.3.4.2 General structure of OAM cells

The general structure of an OAM cell for the ATM layer, for both VP and VC, isseen in Figure 8.7:

• Header: The flow, F4 or F5, to which the ATM-layer OAM cell belongs isidentified, as mentioned before, by using VCI values predefined for the VP(F4) layer flow, or PTI values defined for the VC (F5) layer flow.

• OAM cell type: This indicates the type of function the cell performs.

• OAM function type: This indicates the type of function the cell performs. TheOAM function is defined both by this and the previous field.

• Function-specific field: The structure and meaning of this field depend on thetype of cell and the type of function. This is specified as follows, in the sectionscorresponding to each function.

• Field reserved for future use: Default values are coded as all zeros.

• Error detection code: This field transports a CRC-10 error detection code cal-culated along the OAM cell information field and excluded from the CRC field(374 bits).

Table 8.3 shows the codes used in the OAM type and function type fields for eachtype of OAM cells.

8.3.5 Fault Management Functions

8.3.5.1 AIS/RDI defect indications

AIS and RDI defect indications, for both VP and VC layers, are used to identify andcommunicate end-to-end defects in the VPC or VCC (respectively).

Header OAMType

Function Type Function-Specific Field Reserved CRC 10

Cell payload (48 bytes)

4 bits 4 bits 6 bits 10 bits

Figure 8.7 General structure of an OAM cell which depends on its function.

45 bytes

OAM structure8 bits

ATM Performance 293

VP-AIS cells are generated by the point that detects a defect in the VP layer(e.g., an ATM cross-connect), and are sent downstream, to the connection destina-tion point. Cells are transmitted across the VP links of all the VPC connections af-fected, which is a way to inform all the nodes that the service is no longer available.VP-AIS cells are generated when transmission path defect indications are receivedfrom the physical layer; loss of cell delineation (LCD) is detected or loss of continu-ity is detected in the VP layer (see Section 8.3.5.2).

The same way, VC-AIS cells are generated and sent downstream from the pointthat detects the defect in the VC layer (for example, an ATM switch), to the connec-tion destination, across all the VC links of all the affected VCCs, informing all thenodes that the service is not available. VC-AIS cells are generated when physicallayer AIS defect indications or VP-layer defect indications (loss of continuity or VP-AIS) are received, or when loss of continuity in the VC layer is detected (see Figure8.8). The AIS cell generation frequency is one cell per second, as long as the defectcondition is valid.

• VP-RDI cells are generated by the destination point of the VPC when it de-clares a VP-AIS status (a VP-AIS cell is received, or an AIS defect is detectedin the transmission path), and are sent to the upstream far end.

• VC-RDI cells are generated by a VCC far end point when it declares a VC-AISstatus (a VC-AIS cell is received, or an AIS defect is detected in the transmis-sion path or in the VP layer), and are sent to the upstream far end.

The function-specific field for AIS and RDI cells is formed by the following fields:

• Type of defect (optional): Indicates the nature of the defect announced;

• Defect location (optional): Indicates, for both AIS and RDI cells, the node thatdetects the defect and generates the cell (see Figure 8.9).

Table 8.3Codes to identify the OAM cell and the function type.

OAM Type Code Function Type Code

Fault management 0001

AISRDIContinuity checkLoopback

0000000101001000

Performance management 0010FPMBPM

00000001

Activation and deactivation 1000Performance management act/deactContinuity check act/deact

00000001


8.3.5.2 Continuity check

In ATM, there may be a connection between two users, even if there is no traffictransmitted across it. This is why detecting loss of connection cannot be based onthe mere interruption of data reception.

The continuity check of a connection, for both VP and VC layers, may be car-ried out simultaneously from one end to another, or in a segment formed by manyconsecutive links. It may be activated during the time the connection has been estab-lished, or any time after.

Figure 8.8 Generating VP-AIS/RDI and VC-AIS/RDI type OAM cells.(1) Detecting a defect in the physical signal in a port implies generation of VC-AIS through all the output VCs of all the VCC connections that are being transmitted across this physical link.(2) Detecting a defect in the physical signal in a port implies generation of VP-AIS through all the output VPs of all the VPC connections that are being transmitted across this physical link. (3) Detecting a VP-AIS signal in a certain VPC of a port implies generation of VC-AIS through all the output VCs of all the VCCs that are being transmitted across this VPC.

A defect or an anomaly is detectedOAM information is generated

VP-AIS

End Device

VP-RDI VP-RDI VP-RDI

VC-RDI

VC-AIS

VP SwitchVP/VC Switch

31

2

End Device VP/VC Switch

Type of defect Locating the defect Not used (6AH)

1 byte 16 bytes 28 bytes

Figure 8.9 Structure of the function-specific field for the AIS and RDI type of OAM cells.

OAM cell

ATM Performance 295

The integrity of the connection is supervised by sending continuity check cellswhen there is no user data flow, although it might also be done by sending these cellscontinuously, interleaving them in the user data flow once it exists.

If this check is carried out end to end, and the destination of the connection doesnot receive a continuity check cell at certain intervals when data is not sent, an AISstatus will be declared due to loss of continuity (LOC). If the check is made on a seg-ment, and the destination does not receive continuity check cells, an LOC defect isdeclared and AIS cells are transmitted downstream, to the destination of the connec-tion.

There are no specific fields for this function, which is why the whole function-specific field is coded as 6AH.

8.3.5.3 Loop-back capacity

This function makes it possible for an operation-related cell flow to be inserted atany VP or VC connection point of a specific VPC or VCC (e.g., any cross-connec-tion element, ATM switch, or network terminal equipment) and returned (or loopedback) at a different downstream location to its originating point. The looped-backcells may be inserted without having to take the connection out of service. The se-quence of user cells is not disrupted by inserting the loopback cells, user cells arenot affected by the function and are transmitted to the other end, without being re-turned to their originating point.

This function is useful for the following tasks (see Figure 8.10):

1. To test, from one single end, the integrity of a connection during its setup;2. To detect failures;3. To estimate the general return delay.

• Loop-back indication: Identifies if the cell travels downstream where loop-back is activated, or, if the loop-back has already been activated and the cell istraveling upstream, the originating point.

Loop-back location ID Source ID Not used (6AH)

16 bytes 16 bytes 8 bytes

Correlation Label0/1Not used (0H)

7 bits 1 bit 4 bytes

Loop-back indication

Figure 8.10 Structure of the OAM loop-back cell-specific field.


• Correlation label: Correlates the transmitted OAM cell with the received one.

• Loop-back location ID (optional): Identifies the point where loop-back is acti-vated.

• Source ID (optional): Enables the origin of the loop to recognize its own cells.

8.3.6 Performance Management: Performance ISM

By analyzing OAM cells, not only is it possible to monitor the status and integrityof the connections established in the ATM network, but also to estimate the perfor-mance the users can obtain in data transmission. The in-service performance man-agement of a VPC or a VCC connection, or of one of the segments that form theconnection, is carried out by inserting OAM performance management cells (PMcells) into a user data cell flow.

PM cells enable us to measure performance parameters of an ATM networkwhile the network is in service. These measurements are known as in-service mea-surements (ISMs), as ITU-T Rec. O.191 indicates. This could be done by means oftest equipment connected to the monitored link, or, alternatively, by means of a net-work management and monitoring system. The results of the OAM cell processingby ATM switches can be stored in their MIB tables as management information, tobe consulted later by the operator. It is advisable to keep continuity check functionsactive during performance management.

8.3.6.1 Description of performance monitoring cells

By analyzing user data cell blocks, we can manage the performance of a connec-tion. After every nth user cell, a PM cell insertion request is made. A monitoringcell is inserted into the first point available after the request. The nominal blocksize, n, may be 128, 256, 512, or 1,024 cells. The real size of the monitored blockmay vary, since the monitoring cell must wait for a free space to be inserted into.However, the monitoring cell must always be transmitted before n/2 user cells afterthe request. This way, the maximum variation in the real block size cannot exceed nby more than 50%. With performance monitoring we can estimate:

• Errors produced in the monitored cell blocks;

• Lost cells;

• Misinserted cells;

• Delay measurement and delay variation.

Each monitoring cell sent by the forward performance monitoring (FPM) functionin the forward direction has a corresponding cell sent by the backward performancemonitoring (BPM) function, to provide information on management results.

ATM Performance 297

The function-specific fields of FPM and BPM cells carry the following infor-mation (see Figure 8.11) to estimate performance parameter values:

• Monitoring cell sequence number (MCSN): This field indicates the current val-ue of a running counter, module 256, of the FPM cell sequence (if it is withinan FPM cell), or the BR cell sequence (if it is within a BR cell). The countervalue for a forward monitoring cell and for its corresponding backward datacell are independent of each other. They can be used to detect misinserted orlost cells.

• Total user cell number (TUC0, 0-1): This indicates, respectively, the total num-ber of user cells (CLP=0+1 aggregate traffic) transmitted until the forwardmonitoring cell is inserted, and the number of these cells corresponding to theCLP=0 flow. The backward data cell contains the same value as the matchingmonitoring cell. The difference between the values of this field for two consec-utive forward monitoring cells indicates the total number of user cells or cellswith CLP=0 transmitted in the last block. If at the receiving end a counter isimplemented for received cells, this makes it possible to count lost cells in themonitored block.

• Block error detection code (BEDC): This field enables us to register block er-rors. It is used only in forward monitoring cells, and it transports the even BIP-16 parity error code counted across the data fields of the cells that form theblock.

• Time stamp (TSTP): This represents the time when the monitoring cell was in-serted. This field is used in time-related performance measurements: delays,delay variations, and so on.

• Total received cell count (TRCC0,0-1): This is only used for backward datacells, and it transports the counter value (65,536 mode) that indicates the totalnumber of user cells received when receiving a forward monitoring cell.

• Block error result (BLER) field: This is used only in backward data cells, and ittells the number of erroneous parity bits detected in the block. This value isonly inserted when it is estimated that a block does not contain misinserted orlost cells. This is explained in more detail below.

4 bytes

TSTP (6AHTUC-0+1 MCSN BEDC-0+1 TUC-0 TRCC-0 BLER-0+1 TRCC-0+1

1 byte 2 bytes2 bytes 2 bytes 29 bytes 2 bytes 1 byte

Figure 8.11 Structure of the performance monitoring OAM cell-specific field (PM cells).

2 bytes


Those OAM functions that monitor the connection or the span of a connection onthe VP or VC layer process the PM cells that form a part of the corresponding flow.With the information the fields contain, we can carry out the following operations.

1. The difference between the TRCC values of two consecutive BPM cells sentfrom their destination to the origin indicates the number of user cells receivedat the destination between two corresponding forward OAM monitoring cells.

2. BPM cells also transport the value of the TUC counters of their correspondingforward monitoring cells. Therefore, the difference between the TUC counterscarried by consecutive BPMs gives us the number of those user cells that weretransmitted between the insertion of two forward monitoring cells.

3. The difference between (2) and (1) enables us to estimate the number of lostcells in a block, if the resulting value is negative, or misinserted cells if thevalue is positive (more cells were received than sent).

8.3.6.2 Estimating performance parameters

To estimate performance parameters, we must analyze the contents of PM-typeOAM cells delivered by end-user equipment or by an NE. The data deriving fromthese contents is compared to the data calculated directly from the received usercell flow. The performance parameters we can estimate by using PM cells are thefollowing:

Errored cells

To estimate errored cells for in-service mode, we can distinguish between two cas-es, depending on whether we analyze forward or backward; in other words, FPM orBPM cell flows. The latter is also known as backward reporting (BR) cells.

If we analyze the FPM cell flow, we will use the BEDC field to determine thenumber of errored cells for a block. The following illustrates how to determine theerror:

1. If an i number of BIP-16 parity violations in the BIP-16 is less than or equal to2 in a block of n user cells, and no lost or misinserted cells have occurred, wecan conclude that the number of errored cells for the block in question is i.

2. If the number of BIP-16 parity violations is greater than 2 in a block of n usercells, and no lost or misinserted cells have occurred, we can conclude that thenumber of errored cells in the block in question is n.

3. Those blocks where lost or misinserted cells are detected will not be counted.

If we analyze the BR cell flow, we will use the BLER field to estimate the numberof errored cells. In this case, the determination is as follows:

ATM Performance 299

1. If the number of this field is i, and 0≤ i ≤2, we can conclude that the number oferrored cells at the far end in a block delimited by an FPM cell matching theanalyzed BR cell is i.

2. If the value of this field is greater than 2, we can conclude that the number oferrored cells at the far end is equal to the size of the block delimited by an FPMcell matching the analyzed BR cell. The size of the block in question is the dif-ference between the TUC field values of the last cell received, and those of thecell right before it.

3. If the value of this field is FF in hexadecimal, errored cells will not be counted,as this value indicates that the far end has detected lost or misinserted cells.

Lost or misinserted cells

For in-service measurements, we can base our estimation of these performance pa-rameters on the PM cell sequence number field (MCSN). However, more accurateresults are obtained if we base our evaluation on the FPM cell TUC fields, andcount the received cells. We can also base the evaluation on the BR cell TUC andTRCC fields, as we mentioned when describing these fields for PM cells.

Delay parameters

PM cells contain a time stamp that enables us to estimate CTD and 2-CDV parame-ters. Usually this time stamp is the value of an internal counter synchronized to aclock or the absolute time value provided by a reference. However, to calculateCTD and 2-CDV values directly, the absolute time value or internal clock countersof both the end and the originating equipment must be the same.

This way, the CTD can be calculated directly as the difference between the endequipment clock value on cell arrival, and the originating equipment clock valuewhen the cell is transmitted and transported, time stamped. That can be achieved, forexample, by using at both ends the same time provided by a GPS. If the clocks ofboth pieces of equipment are just synchronized, it is only possible to calculate the 2-CDV value indirectly, taking into account the time stamp value of different cells; forexample, arrival times.

On the other hand, it is possible to count the 1-CDV parameter by basing it onlyon the user data cell flow, since to estimate this parameter, all we need is to measurecell arrival times and to know the PCR of the connection.

8.3.7 Activation/Deactivation Functions

Monitoring the performance and continuity checks can be activated at the momentof establishing the connection, or any moment after it. Activating can be done ei-


ther by the telecommunication management network (TMN), or an end-user. Afteran activation or deactivation request, an initialization process is started for thepoints between which monitoring functions will be carried out. This process maybe started either by using activation/deactivation OAM cells, or by the TMN.

8.4 TEST TRAFFIC FOR OUT-OF-SERVICE MEASUREMENTS

OOS tests are made when carrying out measurements to evaluate the performanceand the behavior of an ATM network with no user traffic, or to measure the perfor-mance of a new connection to set up. Test traffic is generated out of service, andthis enables us to evaluate performance parameters.

When an OOS measurement is carried out, a connection is established that isdedicated to measuring QoS. A flow of appropriated cells is transmitted across thisconnection, and analyzed on reception. This test traffic is composed of cells with aspecific structure to transport the information needed to measure the parameters thatdefine the QoS. The OOS mode enables us to carry out a more profound test on thenetwork, since all the cells transmitted across the connection will be measured.

8.4.1 Generating Test Traffic

Test cells are those that belong to the traffic that is transmitted across a connectiondedicated to OOS measurements. Data fields situated in the cell payload are codedto transport data that makes it possible to deduce cell transfer results and, startingby observing these transfer results, to measure network performance parameters.

The test cell payload format is composed of the following fields (see Figure8.12):

• Sequence number (SN): The transmission sequence number is a binary counterof 32 bits that is incremented by one unit each time a test cell is transmitted.

• Time stamp (TS): A time stamp is an indication of the time when the cell wastransmitted. It takes the value of a 32-bit accumulator, with a resolution of10 ns.

Cell Header SN TS Not used TCPT CRC16

Cell Payload

4 bytes 4 bytes 37 bytes 1 byte 2 bytes

Figure 8.12 Format of an O.191 test cell.

ATM Performance 301

• Test cell payload type (TCPT): This field defines the test cell payload type. It isused to control the version number, since in future revisions of the recommen-dation there may be additional test functions.

• CRC 16: This field is formed by the CRC-16 error detection code.

The data that these fields contain enables us to estimate cell transfer results, whichare used to calculate error-related network performance parameter values. The celltransfer results defined in Recommendations I.356 and I.357 are:

• Cell transferred successfully;

• Errored cell;

• Lost cell;

• Misinserted cell;

• Tagged cell;

• SECB;

• SES (see Section 8.2.4).

The payload format described above for test cells is adjusted to calculate the QoSparameters related to the lost cell ratio, errored cell ratio, and misinserted cell ratio,and to the calculation of transfer delays and their variations. However, it is not pos-sible to make calculations related to the BER. To measure the BER, the most suit-able payload formats are as follows (see Figure 8.13):

• In the first format, the 48 bytes of the payload are filled with one of the follow-ing PRBSs: 223-1, 211-1, 215-1, 29-1.

• In the second case, the 48 bytes are filled with an up to 16-bit word, program-mable by the user.

Header Pseudo-Random Bit Sequence (PRBS)

Cell Payload

5 bytes

Header Fixed word defined by the user

Figure 8.13 Test cell format to measure BER.

48 bytes


8.4.1.1 Generated traffic profile

Traffic formed by test cells is generated by following a certain programmable pro-file. The type of distribution profile depends on the type of connection or functionthat we wish to measure; for example, continuous traffic, variable bursts, and so on.

8.4.2 Estimating Performance Parameters in OOS Mode

The test cell SN and error-detection (CRC-16) fields are used to estimate cell loss,as well as misinserted and errored cells.

8.4.2.1 Errored cells

A cell is errored if the error detection code, the CRC-16 field, detects one or moreerrors. However, there may have been errors this code has not detected.

8.4.2.2 Lost or misinserted cells

The SN field is used to determine the number of lost cells and the number of misin-serted cells.

8.4.2.3 Decision algorithm

The SN field is protected by the error detection code. However, when facing errorbursts, this protection mechanism may fail, and the sequence number is corruptedand not detected, but counted as a misinserted cell rather than an errored cell. Theprobability of this is very low, but it can be compared to the loss and the misinser-tion rates that we want to measure.

To avoid these possible causes of inaccuracies in measurement, the ITU-T Rec.O.191 suggests the following measurement algorithm. This algorithm does not im-mediately decide if a cell with an extraordinary SN is a misinserted cell, or an er-rored cell, or if loss of cells happened (see Figure 8.14).

• To start this measurement, two valid cells must be received consecutively, withno errors in the payload, and with sequence numbers that are also consecutive.

• There is a reference counter for the sequence number, SNref , initialized withthe value resulting from incrementing the sequence number of the second validcell by one unit. This counter is incremented by one unit every time a new cellarrives.

• When a cell is received, if the protection mechanism detects errors in the pay-load field (S=0?), the cell is considered not valid, and a time counter (E1) is in-

ATM Performance 303

cremented, registering the number of those cells that are received but are notvalid.

• When a valid test cell arrives with a sequence number that is not consecutivewith respect to the sequence number of the cell received before it(Ndiff = Odiff), a counter (Nbreak) is incremented, registering the number of in-terruptions in each sequence.

Figure 8.14 Out-of-service cell transfer outcome decision algorithm defined in ITU-T O.191.

Cell arrival?

Ndiff=SNref-SN

S=0?

Ndiff=0?

Ndiff=Odiff?

Ndiff>0?

E1 Errored cells Ndiff Misinserted cells

SNref=SN

E1=0

Odiff=Ndiff

E1 + Nbreak Errored cells

Nbreak=Nbreak + 1

E1=E1 + 1

|Ndiff| Lost cells

Nbreak=0

SNref=SNref +1

No

No

No

Yes

Yes

Yes

Yes

No

Timer LPACtimeout?

Yes

No

Ndiff = SNref -SN for the current cellOdiff = SNref -SN for the previous cellS = CRC-16 syndromeE1 = Cells received with bad CRC-16, not identified yetNbreak = Consecutive sequence breaks, not identified yet


Our estimation of performance parameters is updated when one of the followingevents occurs:

1. When, again, two valid cells are received consecutively, with consecutivesequence numbers;

2. When a cell is received whose SN coincides with the SNref counter value.

If the first event occurs, the difference between the SNref counter value and the SNvalue of the last cell received is counted as lost cells if it is negative, and as misin-serted cells if it is positive. Furthermore, the E1 value estimates the number of er-rored cells. The two counters are reset.

If the second event occurs, it is estimated that the cells received with extraordi-nary SN really were errored cells. Therefore, the number of errored cells is estimatedas the sum of the two counters, E1 and Ninterr.

At certain significant time intervals, under critical QoS conditions, it might hap-pen that neither of these two events that enable us to update measurements occurs.In this case, it would be impossible to evaluate performance. A maximum period of10 seconds for decision making is defined, within which, if no decision is made, ananomaly known as loss of performance assessment capability (LPAC) is declared.

8.4.2.4 Severely errored cell block

Measuring the number of severely errored cell blocks (SECBs) is based on measur-ing the parameters mentioned above. The cell block size (N) is a function of thepeak cell rate of the connection under test (see Table 8.4). (See ITU-T Rec. O.191.)

Table 8.4Cell block and SECB threshold size.

It is Applied to the PCR of the aggregate cell flow, CLP=0+1.

PCR (Cells/Second) User Data Rate (Mbps) N (Block Size) M (Threshold)

0 < x ≤ 3,200 0 < y ≤ 1.2 128 43,200 < x ≤ 6,400 1.23 < y ≤ 2.46 256 8

6,400 < x ≤ 12,800 2.46 < y ≤ 4.92 512 1612,800 < x ≤ 25,600 4.92 < y ≤ 9.83 1,024 3225,600 < x ≤ 51,200 9.83 < y ≤ 19.66 2,048 64

51,200 < x ≤ 102,400 19.66 < y ≤ 39.32 4,096 128102,400 < x ≤ 202,800 39.32 < y ≤ 78.64 8,192 256202,800 < x ≤ 409,600 78.64 < y ≤ 157.29 16,384 512409,600 < x ≤ 819,200 157.29 < y ≤ 314.57 32,768 1,024

ATM Performance 305

A block of n size will be declared as severely errored if the sum total of lost,misinserted, and errored cells detected with the algorithm mentioned above aregreater than m=n/32.

8.4.2.5 Delay parameters

These parameters are measured by time stamping test cells. Knowing the timestamp that indicates both the moment the cell was transmitted and the cell arrivaltime, we can measure such performance parameters as CTD, 1-CDV, and 2-CDV.To measure CTD and 2-CDV, we must use the same clock for both the measure-ment equipment generating the test traffic, and the tester analyzing this traffic.

One way to measure these parameters would be by setting a loop in the far endof the ATM connection established for the measurement. This way, the same testerwould work as both generator and analyzer, and therefore the same clock would alsobe used to generate and analyze the time stamp field. However, in this case we wouldactually be analyzing the round trip delay and the round trip delay variation (RT-DV).

8.5 MEASUREMENT CYCLE IN ATM NETWORKS

As mentioned in the introduction of this chapter, we can group the measurementsinto several phases during the network life cycle:

• Equipment commissioning: In this phase, we are testing the technical character-istics of the NEs used.

• Network installation: Here, we are making measurements when installing thenetwork and its elements.

• Bringing-into-service or service provisioning: These measurements are carriedout in an ATM connection, when established, before delivering the service tothe user. They are to check the correct operation of the connection.

• Network supervision and service assurance: In this phase, the network is mon-itored to make sure it is working well both globally and for each individualuser of the service.

During equipment commissioning, network installation, and BIS phases, OOS testsare used in order to check network elements behavior, to check correct network in-stallation, and to see that connections are set up correctly and provide the right ser-vice. On the other hand, ISM and monitoring are used during network supervisionand service assurance: The OAM functions, described above, are the key elementthat enables us to check in-service the status of the network and the connections es-tablished, as well as performance.


8.5.1 Properties of ATM Switches

The first set of measurements consists of testing the parameters and characteristicsof the ATM switch equipment. In the following, we will see some examples of thistype of measurement.

8.5.1.1 Latency

Latency is the delay caused by a switch between the arrival of a cell and the trans-mission of the same cell through the corresponding output port. This delay dependson the input traffic level in the ports of the switch. Usually, latency increases as thenumber of connections supported by the switch and its payload increase. This is be-cause cells must share the same switching matrix and output port. If traffic levelsstart to be near congestion, the queues in the switch get longer, and in this way la-tency increases as well. The delay caused by the switch also depends on the type ofservice used for cell transmission. CBR connection cells may experience certainmaximum delays, and therefore have higher priority; this is why they experienceless latency. Latency is measured by generating test traffic and inserting it in theswitch, out of service. For this measurement, an O.191 cell flow is generated, in-cluding a time stamp that enables us to count the delay the cells experience.

Let us look at an example. The minimum latency introduced by a switch can bemeasured by configuring a unique PVC in an ATM switch. In this case, as there isonly one connection, the cells do not compete as regards the switching matrix andthe output port, with cells originating from other connections. This is why we onlyevaluate the time the switch takes to process cells. We will use our measurementequipment to generate test traffic with O.191 cells inserted in the defined PVC. Thetraffic profile does not have any influence on the measurement, as long as it is withinthe bandwidth specified when defining the PVC. By processing the time stamp ofthe cells in the output port, it is possible to evaluate the transmission delay throughthe switch; in other words, its latency. When the amount of traffic in the connectionincreases up to the maximum available in the physical medium, it should not haveany effect on the way the switch responds (see Figure 8.15).

8.5.1.2 Switch throughput

We can carry out this test to determine maximum payload that a switch may pro-cess and switch, without any loss of cells.

This measurement can be made by using many different configurations. How-ever, to test the switch, it is better to choose one that represents the worst possiblecase. If we take a switch of n ports, in each port n-1 connections are configured, andeach connection is addressed to an output port that is different from the input port.The switching matrix is tested when there is traffic from all the ports to all the ports.

ATM Performance 307

Furthermore, in all the output ports there are cells to transmit that originate frommany connections.

For this test, we need a tester that has many output ports and that can generatemultiple flows by each interface. By inserting O.191 traffic in the input ports andevaluating cells at the output, we can detect any loss of cells, as well as monitor thebandwidth. If we do not have a tester with many interfaces available, we can use atester that has only one interface, and connect it to the first ATM switch that can rep-licate the flow in many output ports, and insert the resulting flows into the switchunder test.

By increasing the amount of traffic generated, we can measure the traffic leveland the frames lost at input. When the output traffic level stops from increasing inline with the input and the number of lost cells is significant, the switch has reachedits limit (see Figure 8.16).

8.5.1.3 Architecture of a switch

Another characteristic of a switch is that its output ports are independent. Thismeans that a congestion status in a port does not have any effect on the remainingoutput ports.

To carry out this measurement, we must configure a set of PVCs that can gen-erate congestion in an output port. This set of connections may be composed of URB

Out

Out2

1

Out

Out4

3

In

In2

1

In

In4

3

ATM Switch, VC-type

Physical link

CV32

CV57

CV32

CV57

PVC: Port1 VPI/VCI=2/32, Port2 VPI/VCI=5/57

Generator/Analyzer

Figure 8.15 A simple measurement of the delay caused by a switch when processing cells.

VP2

VP5

VP2

VP5


or VBR type of PVCs. The VBR-type connections allow for peaks of traffic that ex-ceed the total capacity of the output link, and therefore produce congestion. Trafficin URB connections also increases congestion, although the cells transmitted acrossthese connections will be rapidly discarded if there is not enough capacity.

Congestion in the output port 4 (see Figure 8.17) should not have any influenceon the traffic in the remaining ports, and this traffic should be transmitted withoutany effect on its performance parameters.

8.5.1.4 Traffic administration: Differentiating services

One of the main characteristics of an ATM switch is the way it implements trafficadministration, and how it treats each service category. We can carry out differentmeasurements and test its behavior when faced with situations of extreme conges-tion. One possible test consists of analyzing how the switch gives priority to differ-ent services, and maintains their quality parameters during congestion periods in aport due to coincidence of traffic peaks in VBR connections, although all the trafficis conforming with specified PCR, SCR, and MBS traffic parameters.

Let us presume that an output port supports many nrt-VBR and CBR-types ofconnections. By generating conforming traffic that causes congestion in the port, wecan analyze the output flow in a CBR connection. In this case, if the switch works

VP3

VP4

VP7

VP9

VP17

VP41

VP3

VP4

VP3

VP4CV44

CV33

CV33

Out

Out2

1

Out

Out4

3

CV90CV102

CV111

CV38

CV87

CV71

CV93

CV65

CV68

CV44CV33

CV33

In

In2

1

In

In4

3

Figure 8.16 Testing the throughput of a switch by means of a configuration of PVC connec-tions with a structure n to (n-1) full cross. Cells are competing for the capacity of the switching matrix and the output ports.

ATM Switch, VC-type

ATM Performance 309

correctly, it should give priority to the CBR service transmitted, and therefore main-tain the delay performance parameters without any degradation, even if there is acongestion status. However, there are many ways to implement the way in which theswitch deals with this situation:

• The first way consists of giving priority to all the parameters of the CBR con-nection. In this case, the nrt-VBR services may experience a decrease in per-formance, since the number of lost cells grows when they are discarded due tocongestion.

• The second way is to try and give priority to the most restrictive performanceparameters of each type of service. The delay-related performance parametersof the CBR service would not be altered here, but there might be a slight degra-dation when it comes to lost cells. However, the nrt-VBR service would have aconsiderable degradation in performance as regards delays, but the lost cell ratewould remain intact. In short, priority is not given to services, but to the mostimportant performance parameters of each service.

8.5.1.5 Traffic administration: Supervising the UPC policer

The aim of this measurement is to analyze the way the UPC traffic policer works ina switch when traffic that is not conforming to the traffic specified during the con-nection setup is inserted into one of the connections used.

The UPC function monitors the traffic transmitted by the user and decides if itconforms to the traffic described during the connection setup. With nonconformingtraffic, this function may behave in two different ways, depending on the service cat-

VP4 VP4

VP1CV91

CV44

VP3CV33

VP1

VP2CV92

CV91

VP1 CV91

CV44

VP3 CV33

VP1

VP2CV92

CV91Out

Out2

1

Out

Out4

3

In

In2

1

In

In4

3

ATM Switch, VC-type

Out

Out

Congestion

Figure 8.17 Testing the architecture of a switch with independent ports and separate queues for each port.

CV33 CV33


egory used. Some service categories let us define traffic contracts differentiated byCLP=0 and CLP=0+1 flows. In this case, the UPC/NPC function evaluates the con-formity of each type of traffic separately, and acts differently depending on whichone of these flows forms the nonconforming traffic:

• If the CLP=0 flow is nonconforming, the nonconforming cells may be tagged(their value is changed into CLP=1), as long as they are conforming to theCLP=0+1 traffic contract.

• If cells are not conforming to the CLP=0+1 contract, they are discarded.

The definitions of conformity for each service category and for transmission capac-ity are described in the Traffic Management 4.1 specification of the ATM Forum,and in ITU-T Rec. I.371, respectively. The UPC/NPC function is based on applyingthe GCRA algorithm described in the previous chapter.

To carry out this measurement, we must configure the connections in a switch,specify the type of service category, and, depending on the service category, definetraffic parameters for CLP=0, CLP=0+1, or both. Now we can use a tester to gener-ate O.191 flows with CLP=0 across the connection we have established, and varythe profile of the traffic generated. By varying the traffic pattern or the bandwidth,we can generate cells that violate some of the specified traffic parameters (PCR,SCR or MBS). By observing the channel bandwidth for CLP=0 and CLP=1, count-ing tagged cells (CLP=1), and checking O.191 payload to detect lost cells in the out-put port of the switch, we can check if the behavior of the UPC/NPC function iscorrect.

8.5.1.6 Analyzing OAM functions and hierarchical propagation

Another possible test is to check OAM functions implementation, in terms of thehierarchical propagation of defect indications, and see how this is reflected by man-agement platforms.

These tests are commonly used, once the network has been installed, to checkthat when a defect occurs in a certain point of the connection, it is correctly notifiedto the remaining nodes, and the terminal equipment. Moreover, it is useful to checkthat associated alarms are generated and failures are correctly displayed in the man-agement platform. All these tests are based on generating and inserting errors anddefects into the different layers.

Simulating anomalies and defects at the physical layer

Simulating anomalies and faults into the physical layer naturally depends on thetechnology used. If the signal received by a port has a defect on the physical layer,all the ATM connections transmitted across this link will be affected and remain

ATM Performance 311

out of service. In this case, the switch should activate forward defect indications onthe ATM layer, and backward indications on the physical layer.

For example, supposing that cells are transmitted by an SDH/SONET link at155 Mbps, it is possible to generate ATM traffic by inserting errors into the physicallayer; into B2 or B3 bytes. If the error rate is high, the error characteristic is not ac-ceptable. When traffic with unacceptable error characteristics on the physical layeris detected at the input by ATM equipment, it must generate VP or VC-AIS, depend-ing on if the equipment is a VP or a VC switch, through all affected connections (seeFigure 8.18).

Simulating anomalies into the ATM layer

It is also possible to insert errors into the ATM cell header. If the HEC function de-tects many consecutive cells with erroneous headers, the delimitation mechanismthat extracts the cells from the received bit flow will lose its synchronization andcannot operate correctly. The ATM equipment is not able to recover any cells, orany user data, and therefore all the connections transmitted across the link will nolonger be in service. In this case, the equipment defines a loss of cell delineation(LCD) status. It also generates all the defect indications over the affected connec-tions to notify downstream and upstream about the transmission problem (see Fig-ure 8.18).

VP8CV67

VP3

VP4CV44

CV33

VP1

VP2CV92

CV91VP1

VP2 CV92

CV91

VP1

VP2 CV92

CV91Out

Out2

1

Out

Out4

3

In

In2

1

In

In4

3

ATM Switch, VC-type

Physical layer fault

PVC1

PVC2

PVC3

PVC4

Connections affected by a physical fault or an LCD status carrying VC-AIS.

Backward direction of affected connections carrying VC-RDI & VP-RDI.The link also carries backward physical defect indications, such as HP-RDI.

Figure 8.18 Testing the defect indication signals generated by a switch when a physical layer defect is detected, or an LCD status is declared.

or LCD status

CV33


Inserting defect indication signal at the ATM layer

Another possible test is to check if the F4 and F5 flow OAM fault managementcells are processed correctly. By means of a tester, we can generate and transmitthese types of cells across VPC or a VCC configured in the switch.

If a switch receives a VP-AIS indication from a VPC in a certain input port, allthe VCCs that are transmitted by the virtual channels of the virtual path affected willremain out of service. If this happens, the switch will generate VC-AIS cells acrossthe VC links of those VCCs that have been affected by the defect in the VPC. Theswitch will also generate a backward VP-RDI indication across the affected VPC(see Figure 8.19).

On the other hand, if VC-AIS cells are generated across a VC link in an inputport of a switch, the switch will be notified that the connection is out of service, andthe cells in question must be routed to the output port the same way as the user cellsof this connection would be routed. This way, the rest of the nodes supporting theconnection are notified that the connection is out of service.

8.5.1.7 Continuity test

This test is to check that our switches are capable of detecting an LOC defect. Tocarry out this test, we must activate the continuity check function of one of the con-

Defect in a VPCwith a VP-AIS

VP3

VP4CV44

CV33

CV33

VP3

VP4CV33

CV33

PVC1

PVC2

PVC3

VCCs affected by the VP-AIS status, carrying VC-AIS

Backward direction of an affected VPC carrying VP-RDI and VC-RDIs

Figure 8.19 Defect indications detected at the output of a VC-type switch after detecting a VP-AIS in one of the input ports.

VP8CV67

VP1

VP2CV92

CV91Out2

Out

Out4

3

In

In2

1

In

In4

3

ATM Switch, VC-type

Out1

CV44

ATM Performance 313

nections in the ATM switch.This can be done in the switching control or by OAMactivation and deactivation cells. The switch can be configured as a control point tosupervise the continuity check (CC) cell flow on the end-to-end VP layer or seg-mented VP layer.

We can use a tester to generate data across the connection that the switch is su-pervising, with certain periods of silence. During these periods, the equipment willtransmit continuity cells. However, if, during the silence period, the transmission ofcontinuity cells is interrupted as well, an LOC defect may occur in the switch. Theswitch declares an LOC status if it does not receive data cells or CC cells during aninterval of more than 3.5±0.5 seconds.

If an end-to-end VP layer continuity check is activated in a switch, and an LOCstatus is declared in the supervised VPC, all the VCCs supported by the VPC will beout of service. The switch sends a VP-RDI to the transmitting end of the VPC affect-ed, and a forward VC-AIS across all the VCCs that are affected by the loss of con-tinuity (see Figure 8.20).

On the other hand, if a segmented VC layer continuity check is activated in aswitch for a certain VCC, and during a time interval neither user nor CC cells arereceived across this virtual channel, the switch will declare an LOC status in the su-pervised VCC. The switch will send a forward VC-AIS across the VCC that has ex-perienced loss of continuity (see Figure 8.21).

VP3

VCCs affected by the VP-LOC status, carrying VC-AIS

Backward direction of the VPC affected, carrying VP-RDI and VC-RDIs

Figure 8.20 Defect indications detected at the output of a VC switch after detecting LOC in one of the supervised VPCs.

CV33

CV33

PVC1

PVC2

PVC3

VP8CV67

VP1

VP2CV92

CV91

VP3

VP4CV44

CV33

Out2

Out

Out4

3

In

In2

1

In

In4

3

ATM Switch, VC-type

Out1

CV44

An LOC defect inthe supervised VPC

CV33


8.5.2 Installing the Network

Usually, wide area ATM networks use as their physical layer the transmission ca-pacity provided by T-carrier/PDH and SDH/SONET technologies. It is possiblethat the same SDH/SONET network is used to transport ATM services simulta-neously with other types of services, such as voice, leased lines, and so on. In thiscase, circuits are established across the network to connect the different ATM con-nections with each other.

This means that we must carry out all the measurements needed to guaranteethat the circuits established to transport ATM services and to connect all the switchesoffer transparent and error-free communication (see Figure 8.22). Once the ATMnetwork is installed, it is possible to establish permanent connections (PVCs) be-tween different points, to test that communication is working and that all the neces-sary locations can be accessed.

In SVC environments, we can also carry out a set of tests meant to check thatconnections are established and addresses assigned correctly in intermediate nodesand end equipment. In some cases, these installation tests are carried out right beforedelivering the connection to the user. To carry out these tests, we must take into ac-

VP3

VP4CV44

CV33

CV33

VP3

VP4

An LOC defect inthe supervised VCC

VCCs affected by the VC-LOC status, carrying VC-AIS

Figure 8.21 Defect indications at the output of a VC switch after detecting an LOC status in one of the supervised VCCs.

Backward direction of the VCC affected, carrying VC-RDI

CV33

CV33

PVC1

PVC2

PVC3

VP8CV67

VP1

VP2CV92

CV91Out2

Out

Out4

3

In

In2

1

In

In4

3

ATM Switch, VC-type

Out1

CV44

ATM Performance 315

count which protocols are used in the user-network interface to establish and releaseconnections, as well as the type of ATM addresses used.

The measurement is carried out using a tester that implements the signaling pro-tocols needed to generate messages to establish and release connections. The testeremulates the behavior of the terminal equipment, and uses the chosen signaling pro-tocol to establish and release connections with another terminal’s equipment, be thistester or data equipment. Depending on the type of protocol used, we must configurea set of parameters; for example, establish the type of ATM addresses used. In manycases, the ATM address is assigned to the tester by means of the interim layer man-agement interface (ILMI) protocol of the ATM Forum.

By using this protocol, a part of the ATM address assigned to the tester is pro-vided by the ATM switch. If the connection cannot be established, the tester indi-cates the cause provided by the network. In some cases, the cause could be lack ofnetwork resources. Also, the time-outs defined by the protocol are exceeded, as thereare no parameters or the combination is incorrect (supposing that an rt-VBR connec-tion is signaled and no SCR descriptor is provided), or because there is no ATM des-tination address, or it is not accessible. If problems occur, it may be necessary to usea protocol analyzer to capture the messages exchanged during setup in different in-terfaces, and to analyze the whole sequence to locate the source of the problems.

8.5.3 Network Commissioning

Today, most ATM networks provide services based on permanent connections. So,once the ATM network is established, the next step is to supply connections to cus-tomers that have subscribed to a new connection.

Figure 8.22 An SDH/SONET network supporting links between ATM switches.

ADM

ADM

ADM

ADM

STM-64OC-192

1.5 Mbps45 Mbps

140 Mbps

DXC

ADM

ADM

ADM

ADM

STM-16OC-48

ATM

STM-1

OC-3

45 Mbps

155 Mbps

ATM

DWDM

ATM

ATM switches

45 Mbit/s

155 Mbps

155 Mbps

STM-1

140 Mbps

OC-3

OC-3


In permanent connection environments, the connection is tested almost withoutexception before supplying the ATM service to the user, to make sure that it workscorrectly. The operator defines a measurement protocol that is carried out once theconnection is established, before delivering it to the client. The measurement proto-col describes the parameters to measure and the results that must be obtained to ac-cept the connection, together with the measurement time. These tests are naturallymade out-of-service, since the connection has not yet been delivered to the user.They consist of generating test traffic with O.191 types of cells, or with PRBS, de-pending on the parameters analyzed. The first tests consist of checking how the con-nection works. They test that the UPC/NPC function is able to detect if the user isgenerating more traffic than subscribed, and discard any extra traffic. Traffic istransmitted transparently between both ends.

Next, we must check the following parameters:

• Availability of the bandwidth subscribed by the client throughout the wholemeasurement;

• Misinserted cell rate;

• CER or cell bit error rate (CBER), if a PRBS is used as the test cell payload;

• SECBR and SES, together with the unavailability of the connection.

Depending on the type of service contracted, it is also necessary to check QoS pa-rameters: CTD, 2-CDV, and CLR.

We must keep in mind that the measuring time is limited. The most typical mea-surement period is 24 hours. The limit values of the measured parameters defined toaccept the connection are usually more restrictive than the values specified for theuser. The reason for this is that if the defined limit values are met by the acceptanceprotocol during a 24-hour interval, it is supposed that the values for the user connec-tion would be met in the long run. However, it is useful to monitor the parameters ofthe connection in service.

To carry out these measurements, we must use a tester that can generate O.191test traffic, with a traffic pattern in line with the maximum bandwidth that the usercould generate. Usually, this measurement is carried out by setting a loop in the farend that enables us to redirect the test traffic toward the tester. The tester will pro-vide:

• The available traffic;

• Errored, misinserted, and lost cell number and rate;

• The number and rate of severely errored blocks and seconds;

ATM Performance 317

• Total unavailability time, number of unavailability periods, and their mean du-ration;

• The mean, maximum, and minimum values of time parameters: CTD and 2-CDV. The 1-CDV or interarrival time (IAT) (time between the arrival of twoconsecutive cells in a connection) parameters are significant for CBR connec-tions, as they show any inconstancy in cell arrival in respect to the theoreticalarrival sequence T=1/PCR.

If test cells are received back, it means that the connection has been correctly estab-lished, and OOS performance measurements can be made. In SVC environments,the measurement protocol is very similar, although such parameters as SES and un-availability time can be left without analysis, as they are usually related to perma-nent or point-to-point connections. However, some other parameters are significant,and they may even be mentioned in the service level agreement. One of these pa-rameters is the failure rate during connection setup, due to lack of resources. As re-gards the connection setup, we can also measure the time needed from setuprequest to the acceptance of the connection.

For both measurements, we need a tester with signaling that enables us to con-figure the connection. The measurement protocol may consist of trying to establishthe connection a certain number of times during a period of time that follows a ran-dom distribution. The measurement equipment will obtain a rate for the number oftimes the connection was rejected. Depending on the value obtained, we may thendecide whether it is possible to meet the value defined in the SLA, and whether it ispossible to deliver the service to the client or not.

Problems in receiving test traffic

If test traffic is not received, even after checking that it has been correctly generatedby the connection, and seeing that the loop in the far end is configured correctly,there may be two reasons:

1. Switching tables have not been correctly established;2. There is a defect somewhere along the path.

In the latter case, we should receive a VC-AIS/RDI or VP-AIS/RDI type of alarm.If we do not receive any alarm of an anomaly, the problem is that the switching ta-bles have been badly configured.

One way to locate the point where the problem resides is to use loop-back OAMcells. The way to locate the fault is easy: We must generate successive loop-backcells and send them to each switch that supports the connection we have established.


If we do not receive these loop-back cells, we can be sure that the problem is eitherin the last visible switch, in the following one, or in the link between the two.

8.5.4 Bringing-Into-Service in ADSL Environments

The bringing-into-service of an ATM connection in some of the environments de-scribed in the previous chapters has certain peculiarities. For example, when install-ing an ADSL service where ATM is used as the transmission technology, ATMlayer tests only form a part of the whole set of measurements to be carried out;many of them associated with the rest of the layers involved in data transmission.

Furthermore, nowadays there is a tendency to simplify installation tests to min-imize implementation costs. It is quite common that, when provisioning the serviceto the end-user, the tests, if they are made are limited to copper pair qualificationtests (see Chapter 9); testing the synchronization between the user modem and thecentral office modem, and a connection to a Web site to carry out Internet connec-tivity tests; and checking the time needed and the speed reached when downloadinga certain test file.

If these tests give a satisfactory result, the service is delivered to the client. Onlyif the user has communication problems, are more profound tests made and, if pos-sible, all the layers are tested to check for failures. It is very common that these tests,if possible, are made directly from the central office, to avoid sending technicians tothe field; this saves costs.

In any case, there are many ways to carry out the measurements in ADSL envi-ronments, be they during installation or maintenance, and to solve problems, andthese situations usually call for multifunction testers, as we will see in the following.

8.5.4.1 Physical layer measurements

First of all, there is a wide range of measurements and parameters related to thequality of the copper pair and its suitability to transport ADSL services. This set oftests is grouped under the name of copper pair qualification, and it is described inmore detail in Chapter 9.

Copper pair capacity

Once the copper pair has been qualified to transport ADSL services, we can ana-lyze the capacity available for each direction. Depending on the relation betweensignal to noise ratio (S/N) and carrier to interference signal ratio (C/I) in eachband (see Section 9.2), we can estimate the possible number of bits per tone, andtherefore also estimate the total capacity in each direction. We will thereby obtainan estimation of the maximum bandwidth available for the ADSL service.

ATM Performance 319

Testing the ADSL signal

In the following, we will look at a set of tests dedicated to testing the ADSL signalproperly. To carry out these tests, we need a tester that enables us to emulate bothends of the ADSL service; in other words, the behavior of both the ATU-R modemand the ATU-C modem (see Section 4.2.1). In some cases, the tester can substitutethe ATU-R modem and in this way enable us to monitor the different layers in ser-vice, while the connection is being used normally. This is called the golden modemmode (see Figure 8.23).

The first test is to check that both ends of the ADSL link are synchronized cor-rectly. As a result of the synchronization process, we will obtain the working rate inboth directions, as well as the carrier tone allocation, depending on the S/N charac-teristics or interference that occurs in some bands. The tester enables us to show theuser the carrier tone allocation both upstream and downstream, and we can thereforesee the maximum rates available. If the modems are not synchronized, we can obtaininformation on the reasons for faults (see Figure 8.24).

By using two testers, we can carry out a BER test in the ADSL link transmittinga PRBS or a test sequence. In any case, in this type of test, the testers use an ADSLmodem from a certain manufacturer. So, it is not only the link that is being analyzed,but also the capacity of the manufacturer of the modem to detect and correct trans-mission errors by means of techniques used in ADSL, such as interleaving and trelliscoding.

ATU-CATU-CATU-C

Splitter

Splitter

DSLAM

Splitter

Splitter

PSTN

NetworkTermination

ATU-R

NetworkTermination

Figure 8.23 Tester emulating (a) ATU-R; (b) ATU-C; and (c) Golden modem.

ATU-CATU-CATU-C

Splitter

Splitter

PSTN

Local Loop

(a)

(b)

(c)

Central officeLocal Loop

Local Loop

PSTN

DSLAM


In some ADSL modes, the ADSL control channel (aoc) transported in the over-head bytes of the framed structure of the ADSL signal transmits bit swapping mes-sages. These messages enable ADSL modems to change the bits and the powerassigned to a certain subcarrier, without interrupting data transmission, and whilemaintaining the total available bandwidth. With this operation, we can adapt theADSL signal to the changing characters of the SNR, and to interference in subcarri-ers. If supported, we can use a tester to check this functionality in the modems used.

However, ADSL layer tests only guarantee switching between modems, that is,between the user and the DSLAM. This is why problems may continue occurring,regarding the access to the services of the service provider, due to faults on higherlayers.

ATM layer measurements

Problems in trying to access the services may be due to configuration faults inATM PVC connections, or lack of DSLAM buffer capacity to multiplex/demulti-plex the traffic from/to each user into/from the main ATM link. Some manufactur-ers have also introduced some ATM layer tests within their initializationprocedures, such as sending and receiving OAM continuity or loop-back cells.

In ADSL environments, testing the correct configuration of the connections hasan extra problem: Because the connections are asymmetric, it is not possible to carryout the typical measurement of generating ATM test traffic and setting a loop at thefar end. In these cases, it may be necessary to use two testers; one emulating theATU-R equipment, and the other connected to a free port of an ATM switch (seeFigure 8.25). By means of this configuration, it is possible to test the ATM connec-tion between a new user and the switch that the other tester is connected to, whileboth the DSLAM and the link are in service, supporting connections of other users.

Figure 8.24 Graphical representation of bits transported per tone for both directions.

30 138 1104Tone 34 at 146.6 kHz. →2 bits

ATM Performance 321

Both testers are configured to generate O.191 or PRBS cell test traffic, with atraffic pattern and bandwidth in line with that assigned to each transmission direc-tion. Usually, the current commercial services offer bandwidth from 32 to 512 Kbpsupstream, and from 128 to 2,048 Kbps downstream. However, in practice, these val-ues actually correspond to the PCR value of the nrt-VBR connection assigned to theuser. Values for the parameters SCR and MBS are also defined. When test traffic isreceived at the other end, it is analyzed and the following parameters can be evalu-ated:

• Monitoring the received bandwidth;

• The number and rate of errored (CER) and lost (CMR) cells, if the traffic isO.191 type;

• BER and slips, if test cells transport a PRBS.

If the two testers do not receive test traffic, or any defect indication, and the physi-cal ADSL layer in the copper pair has been tested in a satisfactory way, then thefault is in the configuration of the connections. If the testers were configured togenerate the maximum traffic assigned to the user in both directions, there shouldnot be any loss of cells. If cells are lost, this may be due to congestion problems inthe DSLAM or at some point of the connection. In any case, if congestion problemsoccur, it is possible that the cells are received with the EFCI=1 bit. If a notable er-rored cell rate or cell loss ratio is detected in the span, the copper pair link has mostlikely caused these problems, which would mean that the quality of the copper pairis bad.

With the previous configuration, we could only make delay measurements ifboth testers were to use the same absolute time value, such as the time provided by

ATU-R

ATU-C

ATU-C

ATU-C

ATU-C

DSLAM

ATU-R

inoutPHY

VP/VC switch

port 2

port 1 port 3

inou

t PHY

inout

PH

Y

Test portTesting

Figure 8.25 OOS test in an ATM connection with two instruments. Shows ATM link with in-service traffic, and a new ATM connection carrying test traffic.

ATM link

device


a GPS reference. Delay measurements could be carried out, for example, by using atester connected to an ATM switch port and establishing a PVC connection betweenthe tester and the ATU-R modem; this modem should allow a logical loop to be setin the received ATM connections. In this case, by generating O.191 cells with thetester and activating the loop in the ATU-R modem, cells will be received by thetester, and it is possible to evaluate the return delay. We must take into account thatin this measurement, the maximum bandwidth generated is, naturally, the maximumtraffic assigned to the direction with less capacity, since the rest would be lost any-way, due to lack of capacity. So, in this case, the equipment is not tested in extremeconditions of the link.

Measurements on higher layers

Once we have checked that both the ADSL link on the copper pair and the ATMconnection between the user and the service provider work correctly, we know thatproblems that occur when accessing the data are due to faults on higher layers andprotocols.

The problems may be due to faults in IP addressing, or in the configuration ofan intermediary device (such as the gateway), or faults in the way protocols likePPPoE, PPPoA, or L2TP are configured, or in the way they work. Some testers pro-vide PPPoE and PPPoA features, emulating the behavior of the terminal equipment.This makes it possible to check that the service is completely operative, and that theuser can access and register the services of the service provider without problems.

8.5.5 Commissioning in Wireless Local Loop Environments

The most commonly services offered by access networks based on wireless systemsare leased circuits for data or PBX connection, virtual private networks (VPNs), In-ternet access and so on; they are usually meant for corporate and professional use.As said in Chapter 3, ATM technology is used in many of these systems as thetransport and switching technology.

The first requirement when offering these services to a client is naturally the in-stallation or the availability of user equipment at customer premises. Provisioning aservice requires connecting the user equipment to the appropriate CPE interface thatallows for the delivery of the service required. Later on, an ATM connection is con-figured to connect the client to the service provider or to another end-user.

Leased line services

Services to emulate n x 64 Kbps, T1/E1, or T3/E3 circuits, and audio/video trans-mission use the facilities of the AAL1 layer. On the user side, the data received viathe circuit interface is delivered to the AAL1 adaptation functions that convert it

ATM Performance 323

into ATM cells. These cells are then sent across the new ATM connection (PVC)that has been established between the client and the far-end destination (switchedtelephone network or PSTN, Internet Service Provider, customer headquarters, andso forth). The CBR service category is used to emulate a circuit service.

On the network side, the base station node multiplexes the new radio ATM con-nection coming from the new client, with the ATM connections coming from otherusers and established previously in a single link at 155 Mbps or 34 Mbps. In the op-posite direction, the ATM connections received by the fiber or cable link are trans-mitted to each of the users across the radio link.

Similarly to the tests carried out in a leased line before delivering it to the enduser, in this case it is also possible to perform a test to ensure that the circuit is op-erating correctly. This test can be done end-to-end or from the central office:

• An end-to-end measurement is carried out by using a tester connected to theCPE circuit interface, with a loop at the far end (see Figure 8.26). The testergenerates a PRBS or a test sequence, transmitting it across the circuit at binaryrate and receiving it again, since there is a loop established at the far end. ABER test and performance analysis (errored seconds, severely errored seconds,unavailability time, etc.) carried out during 24 hours enable the operator to de-cide whether to accept the new circuit and deliver it to the customer. However,if a high number of errors occurs, there is no way of knowing where exactlythey have occurred. It is logical to presume that radio links are more likely tocause errors, since they are more vulnerable when it comes to interference andnoise.

• To avoid the necessity of sending technicians to the customer’s premises, thesame test can be carried out from the central office. In this case, the whole cir-cuit is tested in two different measurements. A more accurate description of the

New userNetwork side (BS)

OC-1, OC-3c/STM-1or T3/E3 with ATM link

T1/E1n x 64 kbps

ADM

ADM

AD

M

ADM

STM-1 ATM Network

T1/E1 loop

Figure 8.26 End-to-end OOS test in a leased line partially carried across a wireless local loop.


radio link is obtained. The test is carried out by means of a tester connected to afree port in an ATM switch. The new PVC established is switched temporarilyto the port that the test device is connected to. This enables us to perform anOOS test in this new connection, while radio links, the base station link and,ATM switches continue carrying user traffic in other connections (see Figure8.27). The tester generates a test cell flow that is sent to the CPE across the newPVC. A loop is needed in the CPE to send the test cell flow back to the tester.Two types of loops can be set in the CPE.

The first loop is only possible if a logical loop-back can be programmed in the userequipment. The logical loop is a software function that sends back all the cells re-ceived through a specific ATM connection, but it is not available for all network el-ements (see Figure 8.26). When this loop is used, the test flow can be O.191. Thisenables us to measure all the QoS parameters: CER, CLR, CMR, SECBR, SES,UT, CTD, and 2-CDV. The test is usually carried out during a period of 24 hours. Ifthreshold values are not exceeded, the circuit can be delivered to the customer.

The second loop is programmed in the n x 64 Kbps, T1/E1, or T3/E3 interfacein the user equipment (see Figure 8.28). In this case, our test traffic is made up ofcells carrying a framed or an unframed PRBS on the AAL1 layer. The number ofcells per second generated across the CBR connection must meet the following con-dition:

EthernetLLC/

PHYSNAP

WLL

ATM ATM

SDH

CPE1

BS

Ethernet

PDH E1in/out

PDH E1in/out AAL1 AAL1 AAL5

ATM

Bridge

WLL

EthernetLLC/PHYSNAP

CPE2Ethernet

E1in/out

E1in/out AAL1 AAL1 AAL5

ATM

Bridge

WLL

inout

VP/VC

inout

PHY

inoutPHY

Test port

switch

Figure 8.27 OOS test of a new leased line, with a logical loop on the ATM layer.

47 bytes 8 bitsbyte

-------------× N cells/second× Circuit binary rate (bits/second)=

ATM Performance 325

Once test cells have been inserted into the connection to be checked, they are re-ceived by the CPE, and delivered to the AAL1 layer that generates the binary ratefor the circuit by using the 47 user bytes of every cell. The framed or unframedPRBS is then received back by the AAL1 layer, due to the loop that has been set up.AAL1 cells carrying the framed/unframed PRBS are then transmitted to the oppositedirection, and again received by the tester. Analyzing the PRBS received enables usto measure the BER of the connection and detect slips. Moreover, detecting lost andmisinserted cells is made possible by analyzing the SN field of the AAL1 layer. Usu-ally, a 24-hour test is carried out to obtain a good view of the quality of the new ATMconnection.

Services through an Ethernet interface

Some services, such as Internet access or VPN, are offered across an Ethernet inter-face connection. In this case, we must also establish an ATM connection betweenthe client and the opposite end, giving access to the services offered by the networkservice provider. The ATM connection transports AAL5 cells, and generally usesone of the service categories designed for data transmission: nrt-VBR, ABR, oreven UBR. To test the established ATM connection, the CPE must include the fea-ture of programming a logical loop that enables us to send the data received by acertain connection back to the opposite direction. In this case, we can measure the

Figure 8.28 OOS test of a new leased line, with a loop at the user interface.

Loop

EthernetLLC/

PHYSNAP

WLL

ATM ATM

SDH

CPE1

BS

Ethernet

in/out


ATM

Bridge

WLL

EthernetLLC/

PHYSNAP

CPE2Ethernet

E1in/out


ATM

Bridge

WLL

inout

VP/VC

inout

inoutPHY Test port

switch


performance of the connection from a tester connected to a free port in a switch, byinserting an O.191 test cell flow (see Figure 8.27).

If it is not possible to program a logical loop, testing is limited to the use of theIP ping function, to check the access to the other end.

8.5.6 In-Service Measurements

As mentioned with OAM functions (see Section 8.3), when an ATM network is inservice, these are the functions that allow us to detect any anomaly on the physicallayer, indicating if the ATM connection fails and letting us know the performanceof each connection.

8.5.6.1 Interpreting OAM indication signals to locate problems

In current complex scenarios, where networks from multiple operators are used tocarry a service from end to end, OAM defect indications enable us to locate thefault and therefore know which operator is responsible for the problem.

For example, when a network operator providing ATM services uses the trans-mission resources of an SDH network provided by a third player such as a carrier, itis important to locate defects affecting the services, in order to distinguish responsi-bilities. It is very simple, by means of OAM defect indications and OAM functions,to know if the problem is located in the ATM operator network or in the SDH carriernetwork. When ATM switches located at the border of the ATM operator domain,and using carrier-owned transmission resources, detect faults at the physical layer,or there are physical layer defect indications (OAM flows F1, F2, or F3) then theproblem is located at the carrier domain (see Figure 8.29).

Figure 8.29 Service disruption responsibilities in scenarios with multiple players.

STM - 1

STM - 1

REI-L (MS-REI)

ATM

RDI-L (MS-RDI)RDI-P (HP-RDI)REI-P (HP-REI)

SDH network ATM

Carrier domainATM operator domain ATM operator domain

AIS-P (AU-AIS)LOS LOFLOP-P ...

ATM Performance 327

8.5.6.2 Measuring in-service parameters

To keep constant track of the transmission quality obtained by the user, we mustcontinuously measure performance parameters in service. This way, we can makesure that the client receives the agreed service.

This kind of monitoring is only possible by using the OAM performance mon-itoring functions described before (see Section 8.3.6). However, nowadays there arevery few NEs that implement these functions, either by generating or processingcells. This is why in many cases it is impossible to monitor performance parametersin service.

However, there are some possibilities and alternatives to estimate these param-eters, depending on the type of service and the type of data transmitted.

Monitoring connections with ATM adaptation layer type-1 (AAL1) cells

As in any other connection, it is always possible to detect all those cells that haveheader errors, by means of the HEC field. However, when we are monitoring CBRtypes of connections transporting AAL1 cells, the AAL1-specific fields also enableus to obtain a more accurate estimation of transmission quality.

• First of all, the CRC-3 and the parity code transported by AAL1 cells enablesus to detect transmission errors in the AAL1 field.

• Second, hops in the sequence number enable us to detect lost cells. However,since the range of the sequence number is between 0 and 7, we cannot detectbursts with more than eight lost cells.

Monitoring connections with AAL5 cells

Naturally, in this type of connection, it is also possible to detect cells with headererrors. However, if the cells transmitted transport AAL5-layer data, processing thislayer in real time makes it possible to obtain a more precise estimation of transmis-sion quality.

In a connection with cells transporting AAL5-layer data, if the third bit of thePTI field in the header has 1 as its value, this means that the cell is transporting theend part of an AAL5 packet. So, this cell is transporting the trailer of the AAL5packet, where the field that indicates the length of the user data packet (transportedby the AAL5 packet and the CRC32 code calculated over the rest of the AAL5 pack-et) is situated. The following cell already forms a part of the next packet.


We can use a tester to count the bytes transported from the first cell correspond-ing to a packet, until the last cell indicated by the value of 1 in the third bit of the PTIfield. This value, L, is the total length of the AAL5 packet in bytes. Furthermore, thetester also processes the value of the field that indicates the user data packet size inbytes. Let us call this value S. For the AAL5 packet, the following condition mustbe met:

where: 5 is the trailer size of the AAL5 packet,47 is the maximum size the padding added to the user data packet may have in order for the user data packet plus padding plus trailer to be multiples of 48.

If this condition is not met, the reason must be in transmission problems, suchas cell loss. On the other hand, the tester may also calculate the CRC-32 code in realtime, over each packet, and then compare it to the value transported in the trailer ofthe AAL5 packet. This makes it possible to detect AAL5-packet transmission errors.

Therefore, processing the AAL5 layer in real time enables us to detect packettransmission problems and estimate transmission quality, without any need to pro-cess higher levels, or know the data transmitted by the user. This feature is very use-ful, for example, when monitoring performance and detecting problems in datatransmission connections or in Internet access, such as in ADSL environments.

Monitoring traffic

Another thing we must, in many cases, do when maintaining an ATM network is tomonitor the bandwidth of our links. This way we can know the total use and the ca-pacity available for other connections, as well as the bandwidth used by each con-nection that is being transmitted across the link.

Information such as the mean use, the user traffic evolution, and its variation intime may be used to make strategic decisions, such as developing the network, pos-sibly extending it, offering new alternative services, or giving discounts dependingon the time the services are used.

Monitoring SVC-related parameters

In the case of switched connections, the data delivered by the management networkitself or by an alternative monitoring system provide us with the unavailability rateby enabling us to observe the trial/failure statistics obtained by the user. The delaymeasurement can also be made in service, by means of statistics on the delays ob-

S 5+ L S 47 5+ +< <

ATM Performance 329

tained by the user throughout the whole process of attempting to establish a connec-tion. This information is important if we want to obtain information on networkcongestion and the service obtained by the user. This will also give us some indica-tion on whether the user is satisfied with the service or not.

There are not many wide ATM networks offering switched connection services.However, monitoring signaling protocols is very important to solve problems thatmay occur when establishing or releasing a connection. Furthermore, managementsystems dedicated to monitor signaling enable us not only to control the signalingprotocols operation and detect faults, but also deduce such parameters as the use ofthe network and its connections. This knowledge is essential for future strategic de-cisions.


• ITU-T Rec. I.610, B-ISDN operation and maintenance principles and functions.

• ITU-T Rec. I.356, B-ISDN ATM layer cell transfer performance.

• ITU-T Rec. I.357, B-ISDN semi-permanent connection availability.

• ITU-T Rec. O.191, Equipment to measure the cell transfer performance of ATM connections.

• ITU-T Rec. I.381, ATM Adaptation Layer (AAL) performance.

• ITU-T Rec. I.371, Traffic control and congestion control in B-ISDN.

• ATM Forum af-tm-0121.000, Traffic Management Specification Version 4.1, March 1999.

331

Chapter 9

xDSL Qualification

9.1 QUALIFICATION STRATEGIES AND PROTOCOLS

When an xDSL service is installed on a copper pair, certain conditions must be metto obtain acceptable quality. The capacity of a pair is tested by a set of tests thatmay turn out to be both complicated and expensive.

An important factor to keep in mind is who owns the local loop. It usually be-longs to former national operators or incumbent local exchange carriers (ILECs).Copper pairs are very often leased to new operators or competitive local exchangecarriers (CLECs), either by leasing the whole infrastructure (network equipment,copper, and even premises), or just certain frequencies. Shared ownership may havea deep impact on qualification and maintenance strategies. Sometimes, the bringing-into-service of new links is carried out by the xDSL operator, but maintenance is theresponsibility of the owner of the copper infrastructure. In this case, the owner of thecooper should guarantee a certain level of service performance to the xDSL operator.As we will see, the critical factor that compromises the transmission on a copper pairis the crosstalk. Some services are very sensitive to crosstalk while others are provedcrosstalk sources. When different services are provided by different operators an ex-plicit agreement is necessary about how to use and how to share the copper resourc-es, and permanent or periodic monitoring might be necessary to guarantee thequality of each service.

It is also important to decide which quality parameters are to be measured, to-gether with the tests that will be carried out to do this, and, of course, to define ac-ceptance limits for the parameters chosen.

Depending on these decisions, measurement equipment is then chosen and thetesting schedule decided upon.


9.1.1 Prequalification

There are two types of tests for prequalification before implementing the service:analog and service tests (see Figure 9.1).

The first stage consists of checking which pairs are adequate to carry the serviceand which are not, and this is done by carrying out analog tests. The results of thesetests are compared with preestablished limits, and, if they are not met, the pair is ei-ther rejected permanently, or, if the result is close to the minimum accepted, the pairis reserved temporarily.

9.1.1.1 Bulk prequalification

To guarantee the best possible profit for an operator’s copper pairs (be the operatorILEC or CLEC) the use of the pairs must be maximized. But for this, the capacityof each pair must be measured first. This is the aim of the bulk prequalificationstrategy, which has the following advantages:

• Operators have a realistic vision of the possibilities of their network before im-plementing the xDSL service. This enables them to plan their offers optimally.

• Since the capacity of all the pairs is known, if one of them should fail, it canquickly be replaced, minimizing the time the client stays without service.

• Operators can offer very detailed SLAs before installing the service.

However, bulk prequalification also has its downsides:

Analog tests

Tests passed

Doubtful result

Good quality

Marginal quality Service tests

Tests passedREJECTED

ACCEPTED

YesNo

NoYes

Yes

No

Figure 9.1 Copper pair prequalification protocol for the ADSL service.

xDSL Qualification 333

• Costs are high even to begin with, since all the pairs must be tested.

• Carrying out thorough tests slows down the implementation of the service.

9.1.1.2 Selective Prequalification

Selective prequalification consists of estimating the quality of each pair withoutcarrying out any tests. This evaluation covers physical characteristics, age, environ-ment, type of installation, and so on.

We can avoid testing many of the circuits if we take certain situations into account.The status of each pair can be evaluated if we take into account the time elapsedsince installation, the conditions in which the pair was installed, its physical charac-teristics, its maintenance, and so on.

This information that can be obtained without any testing is useful to identifyboth those pairs that cannot be used to install xDSL, due to their length or their sta-tus, and those that can be guaranteed to work at least satisfactorily.

This strategy has the following benefits:

• Operators have a realistic vision on the possibilities of their network before im-plementing the xDSL service.

• Certain levels of the SLA can be offered to the client even before installing theservice.

• The whole process is rather inexpensive, and certainly cheaper than bulk pre-qualification.

The downsides, here, are:

• This method is just an estimation, and some of the pairs accepted will not besuitable for the service.

• In some cases, intermittent failures may occur.

• Some of the pairs rejected could, in fact, be used to offer the service.

9.1.2 Qualification During Commissioning

This strategy consists of checking the pairs during the installation of the xDSL ser-vice, and it has the following advantages:

• Low starting costs.

• Qualification and installation are carried out simultaneously, saving salary andtravel costs.


But this method, too, has its downsides:

• In the beginning, since there is no information available, it is difficult to planhow to exploit the resources and evaluate their profitability.

• It is not possible to offer any level of SLA to the client before installing the ser-vice, unless the operator is willing to change the copper pairs if need be.

9.1.3 Commissioning Without Qualification

Commissioning without qualification is another alternative in this highly competi-tive environment, with often very narrow profit margins. However, maintenancecosts are higher if this strategy is chosen, and it may result in unsatisfied clientswho do not get the quality of service they require.

9.2 COPPER PAIR

Any electrical signal transmitted across the copper pair is subjected to degradation,such as noise, distortion, external interference, crosstalk, and so on, that limit thecapacity of xDSL services.

In the following we will analyze those problems that occur in analog and digitalsignal transmission, without dealing with those problems that arise from the natureof each signal. For this reason, the results have very wide applicability, althoughxDSL broadband signals are especially emphasized.

9.2.1 Attenuation and Distortion

When an electrical signal crosses a transmission line, only one part of the power ar-rives at the receiver. One part of the voltage is reflected to the transmitter; anotherpart turns into ohmic losses; and another part is lost due to outward radiation, espe-cially for very high frequencies (see Figure 9.2).

The most simple way to characterize attenuation is by means of a parametercalled insertion loss, , that is obtained from the relation between the power of thetransmitted, , and the received, , signal:

However, the attenuation experienced by a sinusoidal signal transported acrossa copper pair is a function of the tone frequency. The higher the frequency, the more

LPT PR

LPRPT-------=


attenuation there is. To obtain a description of this, we must define more sophisti-cated parameters.

Mathematically, the attenuation of a sinusoid when it crosses a pair is describedby the copper pair’s transfer function, . The transfer function can be used tocalculate the amplitude and phase of a sinusoidal signal at the output of a section ofa copper pair, and depends on the input amplitude and phase.

To be able to use the transfer function, we must express the input signal (a sinu-soid) as a phasor, and multiply this phasor by the transfer function evaluated at theworking frequency (see Figure 9.3).

Analyzing the copper pair shows us that the transfer function of the line voltageof length takes the following form:

Figure 9.2 Attenuation in a transmission line and its causes. In ADSL, the most considerable cause is ohmic loss in copper pair conductors.

P T P = LR

TX RXTransmission line

P T

Attenuation

Ohmic losses

Radiation

in the conductor

in the dielectric

H ω( )

si t( ) A ωst θ+( )sin=

Si Aejθ= H ω( ) So H ωs( )Si A H ωs( ) ej θ φ H ωs( )[ ]+

= =

so t( ) A H ωs( ) ωst θ φ H ωs( )[ ]+ +( )sin=

Transfer function

Figure 9.3 The transfer function of the copper pair is used to calculate the phase and amplitude of a sinusoid when crossing the copper pair.

l

HV ω( ) e γ ω( )l–=


The phase value, , is:

This is also a complex number with a real part and an imaginary part. The realpart is associated with what really is line attenuation, and the imaginary part repre-sents phase distortion in the receiver, which is caused by the fact that every frequen-cy is propagated at a different velocity. In the expression of we must take intoaccount that and are also frequency functions. Since the effect of is negligiblein most cases, the fact that it depends on is what determines the form of the realpart of , and therefore also the attenuation of the twisted pair.1

When frequencies are high, the charge transported in the electric current tendsto travel near the conductor’s surface. This effect gets stronger as the frequency be-comes higher. As a consequence, there is a net decrease in the useful section of theconductor, and an increase in its resistance to high frequencies. The power dissipa-tion is directly related to the frequency square root, and therefore the attenuation alsoshows the same behavior (see Figure 9.4).

1. The form of the resistance determines the expression of attenuation so that if there arenot many losses and G is negligible, the real part of γ(ω) is directly proportional to theresistance, and in this way both depend on the frequency square root.

γ ω( )

γ ω( ) R jωL+( ) G jωC+( )=

γ ω( )R G G

Rγ ω( )

Figure 9.4 Representation of the attenuation introduced in a cable. This is basically a repre-sentation of the real part of γ(ω). The scale is linear, to make it possible to assess the dependence of the attenuation on the square root of the frequency.

Attenuation

kHz0 300 600 800 1000

0.7

0.6

0.4

0.5

0.3

0.2

0.1

0100 200 400 500 700 900

dB


9.2.2 Return Losses

To achieve optimal signal transfer, the generator output impedance and the loadmust be equal to the reference impedance of the line (see Figure 9.5). The samegoes for transition points between lines.

At those points where the impedance varies, reflections occur, and a part of thetransmitted power is returned to the generator. In the case where a line ends with anopen or short circuit, all the power is returned to the generator, forming a standingwave with no net energy propagation.

The reflection and transmission of a voltage wave at a point where the referenceline impedance varies from to is characterized by means of reflection

Vg Z0

Z0

Figure 9.5 Matched line: The generator output impedance and the load are equal to the reference line impedance.

Z0

Generator LoadLine

ZLZ0 Z0vg

Zg

Z01 Z02

vp1

vn1vp2

vp

vnvp

Figure 9.6 (a) Reflection and transmission of a voltage wave, vp1, at a point where the line impedance varies (reflected wave, vn1. Transmitted wave, vp2). (b) Reflection of a voltage wave, vp, in a resistive load (reflected wave, vn). (c) A wave transmit-ted across a line, vp, depending on the generator voltage, vg.

(a)

(b) (c)

Z01 Z02


coefficients ( ) and transmission coefficients ( ) (see Figure 9.6). These coeffi-cients can be calculated as follows:

Under steady and permanent sinusoidal conditions, the power associated withthe incident wave can be described as a function of the voltage wave amplitude, .

The power of the transmitted and the reflected wave can be described the sameway, using transmission and reflection coefficients:

The energy is conserved and the sum of the power of the reflected and the trans-mitted wave equals the power of the incident wave:

When a voltage wave affects an impedance that ends the line, a part of the en-ergy is reflected and returns to the load (see Figure 9.6). The reflection coefficient isthe following:

ρ τ

ρZ02 Z01–Z02 Z01+------------------------= τ

2Z02Z01 Z02+------------------------=

Vp1

Pp1Vp1

2

2Z01---------------=

Pn1ρVp1

2

2Z01-------------------= Pp2

τVp12

2Z02------------------=

Pp1 Pn1 Pp2+=

ρvn1vp1--------

ZL Z0–ZL Z0+-------------------= =


The power absorbed by the load ( ) is the difference between the power of theincident wave and that of the reflected wave. Under permanent sinusoidal condi-tions, and as a function of the amplitude of the incident wave, :

The latter case is a calculation of the transmitted wave depending on the internalimpedance of the generator (see Figure 9.6). To make it more simple, we can pre-sume that line adaptation is complete, and that its input impedance is equal to thereference impedance, . When it comes to calculating the line impedance, it be-haves the same way as in the concentrated parameter model. Given that the line ismatched, there is no reflected wave, which is why the progressive wave is:

If the signal produced by the generator is a pure tone, the wave will be sinusoi-dal, and the power delivered to the line depends on the phasor associated with thiswave, according to the function:

We are presuming that impedances are complex ( ,). It is easy to see that the progressive wave has the maximum

power, if the impedance of the generator is the conjugated complex of the line im-pedance (see Figure 9.7). In this case, it is said that the power delivered to the lineis the power available in the generator.

It is always recommended that the power the generator delivers to a line be thepower available, to maximize efficiency in transmission.

PL

Vp

PLVp1

2

2Z0---------------

ρVp12

2Z0-------------------–

Vp12

2Z0--------------- 1 ρ 2–( )= =

Z0

vpZ0

Zg Z0+-------------------vg=

PpR0

Rg R0+( )2 χg χ0+( )2+--------------------------------------------------------------

Vg2

2------------=

Z0 R0 jχ0+=Zg Rg jχg+=

PpVg

2

8R0------------=


9.2.3 Noise

Any spurious signal of a random sort that is coupled with the data signal transport-ed by a communications system and that is internally generated by the system iscalled noise. This definition does not cover signals that are generated out of the sys-tem and commonly have an artificial origin, such as noise produced by electric mo-tors, electric shocks, radiated data signals, and so on. All of these are defined asinterference.

The best known and most frequent noise is thermal noise. Produced by all thoseobjects that have a temperature higher than 0 Kelvin, it is a consequence of agitationof charged particles. In an electrical circuit, thermal noise appears as a variable volt-age registered in each terminal of the circuit (see Figure 9.8), and it has the followingproperties:

• It is additive. This means that, in a transmission system, it is added to the sig-nal, thus degrading it.

mW

Z0/Ω

Figure 9.7 Power associated with the progressive wave in a transmission line depends on its typical impedance (presumed to be real).

Figure 9.8 In an electrical circuit, thermal noise appears with a variable and random voltage in each terminal.

vn t( ) ElectricalCircuit


• It can be described as a random Gaussian process. This way, the random vari-ables associated with the process all have a Gaussian probability density func-tion.

• It is white, which means that inside the measurement band it has a constantspectral power density. In other words, the value that the noise takes at a cer-tain moment does not depend on the value it takes any other moment.

A noisy impedance can be characterized as a noiseless impedance in series witha random noise generator with a root mean squared (r.m.s.) voltage, (seeFigure 9.9).

The area below the noise spectral density must be the squared r.m.s. noise volt-age. This property can be used to adjust its value. If the noise is white, its noise spec-tral density must be the following:

The noise spectral density is measured in . Sometimes, noise is also giv-en as an r.m.s. value of noise voltage, referred to one Hertz. That is, no more thanthe square root of . So, the latter value is measured in .

Noise appears in both time domain and frequency measurements. When timedomain magnitude is measured, we will always notice a constant fluctuation in thereading. The closer to the noise level the reading, the more fluctuation there is. In thefrequency domain, the noise is superimposed on the signal spectrum, and makes im-

Figure 9.9 In an electrical circuit, thermal noise appears with a variable and random voltage in each terminal.

Vnrms 4kTRBneq=

Vnrms

ZZ

Z R jχ+=k 138 10 23–× J K⁄=

T Impedance temperature in K

BneqEquivalent to the impedance noise bandwidth

(Boltzmann constant).

Vnrms

SV 4kTR=

V2 Hz⁄

SV V Hz⁄


possible any reading below a threshold given by the noise power spectrum density(see Figure 9.10).

Frequency domain measurements are carried out with a spectrum analyzer thattends to give power results instead of voltage results, as seen up to now. The unitswhere the coordinate axis of a spectrum analyzer is calibrated are either dBm/Hz ordBm. In the first case, the display of the analyzer shows the spectral power densityof the signal it has at its input. In the latter case, the displayed measurement is thepower of the signal filtered by a narrowband filter known as a resolution filter.

To change from noise voltage spectral density measurements to noise powerspectral density measurements, both the output impedance of the noisy DUT and theinput impedance of the tester (not noisy) must be modeled (see Figure 9.11).

The value of the power that the tester measures derives from the impedancepower, , which is the following:

Figure 9.10 Frequency domain magnitude measurement. It is not possible to carry out mea-surements below the noise level.

Signal Spectrum

Noise level

PSD (dBm/Hz)

f

DUTPower

measurement

Zout R jχ+=

Zin RL jχL+=

Pn

Figure 9.11 Noise power measurement.

Vnef

Zout

Zin

Zin

PnVrms

2 RL

R RL+( )2 χ χL+( )2+-------------------------------------------------------=


As we know, this expression is the maximum for the case where the input im-pedance of the tester is adapted to the output impedance of the DUT. The maximumpower that the noisy device can deliver to the load impedance is the noise poweravailable, and its value is:

The available noise power can be predicted rather easily, depending on the tem-perature of the DUT, and the equivalent noise bandwidth of the service. Since thenoise is white, its spectral power density can be calculated by dividing the bandwidthpower by the bandwidth itself:

This is the first reading for the analyzer to obtain spectrums in .

Finally, to carry out a noise measurement over a twisted pair, the impedance val-ue, , will be substituted by the typical line impedance value, , that in mostcases is only resistive.

9.2.4 Longitudinal Conversion Loss

Two transmission modes are possible over a twisted pair (see Figure 9.12):

1. Not balanced: It is typical for this mode that one of the two wires has lowimpedance with ground, while the other wire has a high impedance.

2. Balanced: Both wires of the twisted pair maintain a high impedance withground.

PnavVrms

2

4R------------ 4kTRB

4R----------------- kTB= = =

SPPnav

B------------ kT= =

dBm Hz⁄

Zout Z0

ZL

Figure 9.12 Balanced and nonbalanced modes.

Zg1

Zg2

Z01

ZCVg

Vg

Z02

(a) Nonbalanced line(b) Balanced line

Zg Z0

ZLVg


Even though the balanced mode is more difficult to implement, it has certain ad-vantages. Its capacity to cancel coupled interferences is one of these advantages, andit also has less radiation caused by interference in other communications systems.

In a balanced line, signals are transmitted in differential mode. In other words,there is a phase offset of 180° between the two wires. Ideally, signal generator im-pedances are equal, as are line impedances with respect to ground, and . Inthis way we will obtain that the current that circulates across is null. In practice,it is difficult to obtain equal and impedances, so the current that circulatesacross is not null. Balanced receivers are only sensitive to differential mode, andnot to the common mode. If any interference is added to the common mode, it iseliminated.

To analyze the offset of the and impedances and the current circulat-ing across , we will use a magnitude known as longitudinal conversion loss,(LCL) (see Figure 9.13). The LCL measurement is defined in ITU-T Rec. G.117 asfollows:

Z01 Z02ZL

Z01 Z02ZL

Z01

ZL

Z1

VgZ02

Z2

VT

Figure 9.13 Measurement mode for the LCL standardized by the ITU-T in G.117.

IL

V1

V2

Z01 Z02ZL

LCL 20VgVT------log=


The LCL defined is a good estimation of the and offset. The currentin the impedance must be:

is the equivalent impedance of the impedance network (see Figure 9.13).The current of each wire, and , can be calculated directly by using a currentsplitter:

Now is a function of the stimulus and of the impedances:

or:

This leads us directly to the following equation:

Z01 Z02ZL

ILVgZeq--------

Vg

ZLZ1 Z01+( ) Z2 Z02+( )Z1 Z01 Z2 Z02+ + +

-----------------------------------------------------+

------------------------------------------------------------------= =

ILZ1 Z01 Z2 Z02+ + +( )Vg

Z1 Z01+( ) Z2 Z02+( ) ZL Z1 Z01 Z2 Z02+ + +( )+-------------------------------------------------------------------------------------------------------------------------=

ZeqI1 I2

I1Z2 Z02+

Z1 Z01 Z2 Z02+ + +-------------------------------------------------IL= I2

Z1 Z01+Z1 Z01 Z2 Z02+ + +-------------------------------------------------IL=

VT Vg

VT V1 V2– Z01I1 Z02I2–= =

VTZ2 Z02+( )Z01 Z1 Z01+( )Z02–

Z1 Z01 Z2 Z02+ + +-----------------------------------------------------------------------------=

Z1 Z01 Z2 Z02+ + +( )VgZ1 Z01+( ) Z2 Z02+( ) ZL Z1 Z01 Z2 Z02+ + +( )+

-------------------------------------------------------------------------------------------------------------------------

VTZ2 Z02+( )Z01 Z1 Z01+( )Z02–

Z1 Z01+( ) Z2 Z02+( ) ZL Z1 Z01 Z2 Z02+ + +( )+-------------------------------------------------------------------------------------------------------------------------Vg=


We simplify the numerator:

Now the LCL is only a function of network impedance:

This expression is still too complicated for any valid conclusions to be drawnfrom it. However, we can simplify it if we take the measurement circuit and linecharacteristics into account.

Although it tends to be difficult to measure the impedance of each wire accu-rately and separately, it is much more difficult to calibrate the sum of the two. Thisway, we can define the differential impedance in the following way:

Furthermore, and form a part of a piece of measurement equipment thatis completely balanced. In other words, any internal impedance of the tester can becalibrated accurately so that and will take a value that is exactly equal to thedifferential line impedance.

Finally, we will define the following auxiliary variable:

VTZ2Z01 Z1Z02–

Z1 Z01+( ) Z2 Z02+( ) ZL Z1 Z01 Z2 Z02+ + +( )+-------------------------------------------------------------------------------------------------------------------------Vg=

LCL 20Z1 Z01+( ) Z2 Z02+( ) ZL Z1 Z01 Z2 Z02+ + +( )+

Z2Z01 Z1Z02–-------------------------------------------------------------------------------------------------------------------------log=

Z0 Z01 Z02+=

Z1 Z2

Z1 Z2

Z0 Z1 Z2+= Z1 Z212---Z0= =

∆Z Z01 Z02–=


If we now substitute this in the expression of the LCL, we will obtain:

After the last stage of our simplification, we will obtain:

If we consider that the line is well balanced, ( ), we get:

As we can see, if the line gets more unbalanced, it causes a decrease in the LCL,which is why the LCL is a good unbalance measurement.

Given that line impedance is a frequency function, the LCL is also taken as such.In approval testing, measurements are carried out at many points, to make sure thatthe LCL remains below a frequency mask.

LCL 20Z1Z2 Z01Z02 Z1Z02 Z2Z01 2ZLZ0+ + + +

12---Z0∆Z

--------------------------------------------------------------------------------------------------------log=

LCL 2014---Z0

2 14--- Z0 ∆Z+( ) Z0 ∆– Z( ) 1

2---Z0

2 2ZLZ0+ + +

12---Z0∆Z

----------------------------------------------------------------------------------------------------------log=

LCL 2014---Z0

2 14---Z

0

2 14---∆

2– Z 1

2---Z0

2 2ZLZ0+ + +

12---Z0∆Z

----------------------------------------------------------------------------------log=

LCL 20Z0

2 Z02 ∆2– Z 2Z0

2 8ZLZ0+ + +2Z0∆Z

-----------------------------------------------------------------------log=

LCL 202Z0 4ZL+

∆Z------------------------- ∆Z

2Z0---------–log=

∆Z Z0«

LCL 202Z0 4ZL+

∆Z-------------------------log≈


9.2.5 Crosstalk

The electromagnetic interference caused by a twisted pair over another nearby pairis known as crosstalk. Crosstalk is a coupling produced by induced fields. Thismakes it different from other interferences produced by radiated fields. Crosstalk isrelevant in high-frequency transmission environments, and there are nearby cablesthat are subjected to interference. This is precisely the case of local loops in xDSLservices.

There are many methods to reduce the level of crosstalk, including cable shield-ing, balanced transmission, or the use of closed (for example, coaxial) cables. How-ever, these methods do not eliminate crosstalk entirely, which is why it remains afactor that limits transmission in higher frequencies.

9.2.5.1 Induced coupling in transmission lines

There are two ways in which a twisted pair may affect another pair by coupling:

1. When coupling is caused by an electrical field, we can talk about electrical orcapacitive coupling.

2. When caused by a magnetic field, coupling is said to be magnetic or inductive.

Electrical coupling

Between two nearby cables, a capacity called mutual capacity is created. The sameexpression used to calculate the capacity of a twisted pair is valid here:

is the distance between the axes of the cables, and is the radius of theconductors. This means that the mutual capacity decreases, in line with the recipro-cal value of the logarithm of the distance between the two cables.

Although there is no conduction current between the twisted pairs, there is in-deed a displacement current whose magnitude depends on the mutual capacity. Thisis why one part of the current that circulates on the disturbing line may be passed onto the victim line (see Figure 9.14).

The induced current in the victim pair is divided into two parts that are sent toboth ends. If the ends are perfectly terminated, these two currents are equal. Further-more, the induced current is proportional to the frequency, since coupling is pro-duced by the current that crosses a capacitor.

C12πε

Da----ln

----------≈

D a


Magnetic coupling

The magnetic field generated by a transmission line can interfere in a nearby line aselectromagnetism laws predict. The magnitude of the inductive coupling intensitybetween two lines is called the mutual induction, . Each line may be seen as aspire: The magnetic field generated by a variable current in one of the spires pro-duces a variation in the magnetic flow of another one. This causes an induced elec-tromotive force in the victim line (see Figure 9.15).

The expression of the mutual induction between pairs is the same as the oneused to describe the self-inductance of a pair:

C

Figure 9.14 Mutual capacity in nonbalanced lines.

1

212

I1

In I f Z2

I12Vs

Z1

ZL2

ZL1

Zg2

Zg1

Disturbing pair

Victim pair

M

Figure 9.15 Mutual induction between nonbalanced lines.

1

2

I1

Vn Vf

Z2

M12Vs

Z1

ZL1

ZL2

Zg2

Zg1

V2

Disturbing pair

Victim pair

M12µπ--- D

a----ln≈


is the distance between the two axes of the cables, and is the conductor’sradius.

The same way as in the case of electrical coupling, coupling here, too, is directlyproportional to the frequency: The higher the frequency of the disturbing signal, themore rapidly the magnetic flow of the victim varies, and the stronger is the inducedelectromotive force.

The difference from electrical coupling is that, here, one of the induced wavesis in antiphase in respect to the phase that is propagated in the opposite direction (thephase difference is 180º). The reason is that in this case no effective current injectionis produced in the victim line. On the contrary, the induced voltage in this line pro-duces an electromotive force in the loads that drift away from each other, and thisforce appears in the form of two waves that are propagated in opposite directions. Inorder that the electrical charge be kept, the waves must be in antiphase.

The fact that in electrical coupling the signal coupled in the two directions of thevictim pair is in phase, and in magnetic coupling it’s in antiphase, has an importantpractical consequence: The signal coupled in the direction opposite to the generatoris weaker than the one coupled in the other direction, since the coefficients of elec-trical and magnetic coupling cancel each other out. This means that it is more diffi-cult to measure far end crosstalk (FEXT) than near end crosstalk (NEXT). It wouldeven be theoretically possible to eliminate crosstalk completely at the far end. How-ever, in reality, this is not possible, since mutual capacity and induction cannot becontrolled.

9.2.5.2 Analyzing Crosstalk

Crosstalk is analyzed by means of many parameters in which the crosstalk powerreceived at one end is compared to the interfering signal power. One of the factorsthat must be taken into account when analyzing crosstalk in cables is that, due tothe nature of the coupling that causes it, crosstalk power is received both at thesame end where power is inserted, and at the opposite end (see Figure 9.16).

In line with this, two parameters are defined:

• Near-end crosstalk, which is the relation between the power transmitted by thedisturbing line and the power received by the victim line at the same end wherethe signal is inserted. NEXT is independent of line length.

• Far-end crosstalk which in its turn is the relation between the power transmit-ted by the disturbing line and that received by the victim line at the end oppo-site to where the disturbing signal is inserted. Unlike NEXT, FEXT depends onthe line length.

D a


NEXT and FEXT have one feature in common, and that is that both increase inline with frequency.

NEXT, , and FEXT, , have the following coefficients:

: is the power transmitted by the disturbing pair.

: is the power received by the victim pair at the same end where the trans-mitter of the disturbing pair is.

: is the power received by the victim pair at the same end where the receiverof the disturbing pair is.

As mentioned before, FEXT depends on the longitude of the pairs in question,and the longer the pairs, the less FEXT there is. However, this may be misleading,as the useful signal is also affected by the same attenuation factor. For this reason,ELFEXT, , is defined as the relation between the power received by the disturbingline, and the FEXT power received by the victim pair.

is the power received by the disturbing pair. The constant, , is the powerattenuation of the disturbing pair.

Figure 9.16 Due to electromagnetical coupling between lines, not only the disturbing line receives the signal power at the far end, but the victim lines as well, both at the same end where the disturbing power is inserted (NEXT) and at the opposite end (FEXT).

Transmittedpower

Receivedpower

Disturbing line

Victim lines FEXT

PRPT

PN PF

Near-End Crosstalk

Far-End CrosstalkNEXTpower power

p t

pPNPT-------= t

PFPT-------=

PT

PN

PF

e

ePFPR-------

PFLPT---------- t

L---= = =

PR L


The parameters defined are enough to analyze the NEXT produced on only onevictim line, and caused by only one disturbing pair. In real-life installations, thingsare more complicated, and it tends to be quite normal to work with a large numberof pairs coupled electromagnetically with each other. For example, it is common towork with 25 to 100 pairs. In these cases, each pair may be coupled with the remain-ing pairs. On some occasions it is even necessary to consider the coupling effect thatoccurs between different groups.

To analyze environments with many pairs, NEXT and FEXT matrixes are used.

In this example, the presence of a signal in just one of the pairs causes reception ofpower in all the remaining ones, at both near and far ends (see Figure 9.17). TheNEXT and FEXT coefficients between the ith disturbing pair and the jth victim pairare:

The NEXT ( ) and FEXT ( ) matrices are defined as follows:

PT1 PR11

2

3

4

5

6

7

8

PF12

PF13

PF14

PF16

PF17

PF18

PF15

PN12

PN13

PN14

PN16

PN17

PN18

PN15

Figure 9.17 The presence of a signal in just one of the pairs causes reception of a certain signal level in the remaining pairs. This is at both near and far ends.

pijPNjPTi---------= tij

PFjPTi---------=

P T

P

p11 p12 … p1np21 p22 … p2n… … … …

pn1 pn2 … pnn

= T

t11 t12 … t1nt21 t22 … t2n… … … …tn1 tn2 … tnn

=


The matrices are always square. The coefficients of the main diagonal, , inthe case of NEXT coincide with the return loss, and in the case of FEXT, , withthe attenuation of the twisted pair.

Another question is the symmetry of these matrices. For reciprocity, the effectof an interference caused by the ith pair over the jth pair must be identical to the effectcaused by the jth pair over the ith pair and, therefore, the matrices will be symmetri-cal.

Although the and matrixes fully characterize the set of pairs, from thepoint of view of NEXT and for one transmission direction, they are not able to im-mediately quantify the effect that the interference caused by NEXT would causeover a certain pair. In the same way, it does not allow us to ascertain which one ofthe pairs in a group is the most suitable for transmission, and which one of them isthe most sensitive to NEXT. For this reason, two new parameters are defined, theNEXT power sum loss (NPSL) and the FEXT power sum loss (FPSL). For the jth pairin a group, NPSL ( ) and FPSL ( ) are defined as follows:

In other words, they are the result of the sum total of all the NEXT and FEXTcoefficients, except for those of the main diagonal, maintaining the receiver con-stant. So, the NPSL and the FPSL are results of summing the and matrices bycolumns, ignoring those values that are above the main diagonal. If a particular valueof the or is too high, then the isolation between pairs ith and jth will be bad,thus degrading the transmission over both pairs. The values of the NPSL and theFPSL take into account isolation of a pair and all the possible disturbers. That meansthat the smaller the values of the NPSL and the FPSL, the more quality the cable has.

9.2.6 Other Reasons for Defects

9.2.6.1 Inductive load of the pair

We desire that pairs do not distort the signal, and that they transmit all signal fre-quencies equally. However, a pair is a distorting element for both the phase and theamplitude (see Section 9.2.1). In spite of this, if certain restrictions are set for line

piitii

P T

pj tj

pj piji 1=

n

∑ pjj–= tj tij tjj–i 1=

n

∑=

P T

pij tij


parameters ( , , , and ), phase distortion may disappear completely. Thetransfer function of the pair can be simplified by setting the following condition:

In the new transfer function, the phase of the complex exponential varies lineal-ly from the frequency, which means that the signal will not be affected by phase dis-tortion at the destination, but only by a delay. Distortion in the amplitude cannot beeliminated, since the ohmic losses ( and ) are also a frequency function:

If the described condition is met, it is said that the pair is equalized. The equal-ization condition can be expressed in a simpler way:

or, similarly, the equalization condition is:

What happens is that in real-life lines this equality is usually not verified in anatural way, and lines tend to be rather capacitive than inductive, and losses in theconductor are greater than those in the dielectric. To present this in the form of anequation:

R G L C

G jωC+ K R jωL+( )=

R G

γ ω( ) K R jωL+( ) HV ω( ) e γ ω( )l–=⇒ e R K– e jωL K–= =

G jωC+ K R jωL+( ) G⇒ KR= = C KL=,

RG---- L

C----=

RG---- L

C----<


To compensate for this deficiency, the line is sometimes loaded inductively withcoils, and in this way higher quality is obtained for transmission (see Figure 9.18).

This procedure has been used occasionally to improve the features of voice-band transmission lines. From this we will see that the inductive load of a line is notcompatible with broadband signal transmission.

There is a decrease in voltage in the serial inductances that appear (see Figure9.19):

This is a lowpass transfer function, and in practice it filters broadband signalhigh-frequency components, and weakens reception severely. For this reason, beforeinstalling a service such as ADSL, we must check that the line does not have anytype of coil.

l 1

L1 L2

L1 L2

l2 l 3

Figure 9.18 Inductive load of a transmission line, for equalization reasons.

LL

Figure 9.19 Analyzing an inductance between two spans of a line.

V1 V2

Z2

V2Z2

jωLL Z2+-------------------------V1=


To demonstrate this, we can take a special case where is purely resistive,with a value 120Ω, and is a load of only 1 nH. In this case, the cut-off frequencyat -3 dB is:

This is enough to carry the voice signal that only contains frequency compo-nents up to 3,400 Hz. However, it would attenuate a part of the ADSL signal that stillcontains power at 1.1 MHz.

9.2.6.2 The effect of bridged taps

Pairs that form a local loop very often contain line spans added in parallel to themain span, due to maintenance activity on the access network. If these uselessspans are badly terminated, they cause a capacitive or an inductive effect that filtersand attenuates the main signal in certain frequency bands (see Figure 9.20).

Under steady sinusoidal conditions, the line input impedance, depending on itslength, , is variable, and can be described by the following equation:

is the coefficient of reflection at the end of a line, and it depends on the loadimpedance, :

Let us analyze a special case. Let us suppose that the auxiliary line ( length)does not end with any impedance, and that it is in an open circuit. The reflection co-efficient in this case is:

Z2LL

fc1

2π------

Z2LL------ 55kHz≅=

Z l( )

Z l( ) Z0e2γl ρ+

e2γl ρ–-------------------=

ρZL

ρZL Z0–ZL Z0+-------------------=

l2

ρ 1=


So, the equivalent bridged tap impedance will be reduced to:

Let us analyze this expression supposing that losses are negligible. In this case,the real part of the propagation constant disappears:

The expression for the equivalent impedance in the parallel line is:

We can see that the equivalent impedance is purely reactive. If the lateral line isvery short compared to the wavelength, reactance is very high, which makes it im-possible for the current to flow on the auxiliary line. But as the length increases, re-actance starts to decrease, and current starts to flow on the line, even if this occursonly at the far end of an open circuit (see Figure 9.21). On the other hand, reactancefor short lines is capacitive. This means that from the point where the bridged tap islocated, each of the two cables that forms the line can be regarded as the face of a

l1l’1

Figure 9.20 The effects an auxiliary line has on the main line can be seen as an impedance in parallel. This impedance will have an inductive or a capacitive value, depending on the length of the lateral line.

Zp

Z0Z0

Z0

Main pair

Bridged tab

l 2

Zp Z l2( ) Z0e

2γl2 1+

e2γl2 1–

---------------------= =

γ jβ=

Zp Z0e

j2βl2 1+

ej2βl2 1–

----------------------- jZ0–2βl2sin

1 2βl2cos–-----------------------------= =


capacitor. Reactance caused by a line decreases as the line gets longer. The line be-comes resonant when the length is:

This is equal to 50m at 1 MHz, assuming that the propagation velocity is of twothirds at the velocity of light. At the bridged tap, the reactance is null. For longerlengths, the inductive effects start to dominate, which means that from the bridgedtap, the lateral line is seen as a spire. Finally, the behavior of the whole system isperiodical, and, each time it is repeated, the total length of the lateral line gets incre-mented by .

Obviously, the most damaging situation occurs when lateral lines are in reso-nance, and this should be avoided. We have already seen that for a 1-MHz line, thefirst resonance length is about 50m. With a lateral line of this length, higher frequen-cies of the discrete multitone modulation (DMT) used generally in ADSL would beseverely attenuated. For shorter lateral lines, from 10 to 20m for instance, the reso-nance frequency increases by many MHz, and this does not put transmission in dan-ger.

9.3 ANALOG MEASUREMENTS

In previous sections, we have described parameters that are significant when con-trolling transmission quality in a copper pair. In this section, we will look at mea-surement strategies for each of the quality parameters described. These

l2π

2β------ λ

4--- c

4f-----= = =

λ 2⁄

Reactance (Ω)

l/m

Figure 9.21 Reactance at the output of a line, depending on its length. The wavelength, here, is 200 m.


measurements can be classified into three groups, according to the way they arecarried out (see Figure 9.22):

• From one end only;

• From the two ends simultaneously;

• With a bridged connection.

9.3.1 One-End Measurements

Measurements from one end only are the easiest to carry out, since the only require-ment is to have a tester connected at one end. All we need is a person to do the job,and there is no need for this person to move from one point to another to obtain re-sults.

9.3.1.1 Return loss measurements

Too high return losses make installing an xDSL service nonviable. By measuringthis parameter, we can evaluate any mismatch near one of the ends of a pair. Themain causes for unacceptable return loss are:

• Defects in the connectors used to transmit the signal;

• Diameter variations in the copper cables that form the pair;

• Twist defects in the twisted pair.

Analyzer

ADSL modem

Figure 9.22 Analog measurements: Measurement is (a) carried out from one end only, (b) from the two ends at the same time, and (c) with a bridge connection.

Generator

DSLAM

(c)

(a)

(b)


The measurement is made from one end of the line by transmitting a test tone ora series of tones at different frequencies, if the frequency behavior is to be analyzed.The test is usually carried out by transmitting a signal with a fixed power level, forexample 0 dBm, and measuring the signal reflected at the same transmission end(see Figure 9.23).

To obtain a more exact result, it is recommended to terminate the remote endwith a typical line impedance, especially if the line is short. For longer lines, the sig-nal reflected from the furthest end becomes very attenuated, ending up almost neg-ligible compared to any reflection of a nearby mismatch.

9.3.1.2 Noise measurement

This measurement must be carried out in the bandwidth that the xDSL signal occu-pies. For example, in the case of ADSL, where the downstream and upstream chan-nels are frequency-multiplexed (ITU-T Rec. G.992.1), the frequency bands inwhich the measurement is to be carried out would be from 20 kHz to 138 kHz forthe upstream channel, and from 140 kHz to 1,100 kHz downstream. Any noise out-side these bands is filtered by receivers and is not significant. Furthermore, it mustbe noted that it is not meaningful to measure the upstream noise near the modem or,the downstream noise next to the central.

It is recommended that the cable be terminated with the typical impedance at theopposite end to where the measurement is carried out. In any case, in long pairs, theincidence of possible reflections at the remote end is small, since any signal deriv-ing from that point would be received very attenuated anyway.

The simplest way to make this measurement is by means of a power tester thatenables us to adjust the measurement bandwidth to suitable values. In this case, wewill directly obtain a value in dBm (see Figure 9.24).

In practice, it is not possible to measure separately the noise generated internallyby the system like thermal noise, and the externally coupled interferences, becauseboth are present at the same time in the measurement. The different components ofthe final result can be recognized with a tester that has the capacity of a spectrum

Z0

Z0Test signal

Reflected signal

Figure 9.23 Return loss measurement.


analyzer. In this case, we can obtain a frequency representation of the received noise.This way we can differentiate between:

• White noise, with a plain frequency spectrum;

• Radiated interference, which are spurious radio frequency signals with a spec-trum with peaks at certain frequencies;

• Induced interference, which is cross talk deriving from nearby pairs, with aspectrum with peaks at certain frequencies.

9.3.1.3 NEXT measurement

By the procedure described for noise measurements, we can also measure the cou-pled NEXT power for in-service2 pairs. The measurement described in this sectionhas the aim of calculating the module of the transfer function between the terminalsof the interfering pair, and that of the coupled pair.

2. The condition for the result to be significant is that the rest of the contributions (noiseand radiated interference) to the result are negligible compared to FEXT.

Z0

Z0

Noise

Figure 9.24 Noise measurement.

Interference

Analyzer

Z0

Figure 9.25 NEXT measurement.

Generator

Z0

Analyzer

Z0

Z0


For this, we must use a sinusoidal signal generator and measure the coupled sig-nal at the coupled-pair end. If the frequency of the signal generator can be adjusted,we can calculate the module of the transfer function across the entire frequency band(see Figure 9.25).

To avoid reflections that may mask the result, it is recommended to terminatethe lines with the reference line impedance. The generator and analyzer terminalsmust also be adapted to the line.

In NEXT measurements, the signal transmitter is usually physically close to thereceiver, and sometimes the transmitter and the receiver form a part of the sameequipment. Furthermore, the signal received will generally have a level much lowerthan the transmitted one. To avoid direct coupling (without crossing the pair) of thepower of the transmitting device with the receiving device, the two devices must bewell isolated.

9.3.1.4 Measuring loss of longitudinal balance

The same recommendations for the measurements described earlier are also validfor this measurement, in particular, that of terminating the line under test with refer-ence impedance. To carry out this measurement, the tester must have access to theearth/ground voltage of the pair. The ground is usually connected to the shield ofthe pair (usually the shield is shared by more than one pair). We must take into ac-count that many pairs share the shield (see Figure 9.26).

9.3.1.5 TDR measurements

The time domain reflectometer (TDR) measurements are very important, not onlyat the copper pair qualification stage, but also to locate faults during maintenanceand repair.

The reflectometry is based on an echo measurement (return signal) derivingfrom the line where a test pulse is introduced. TDR is a tester that uses reflectometryto carry out measurements.

Z0

Figure 9.26 Isolation measurement.

Z0


A TDR transmits a sequence of pulses across the line under test. It contains acircuit that is able to separate the transmitted pulses from the received ones. By eval-uating the time elapsed between transmitting the test pulse and receiving the echo,we can also evaluate the distance at which the reflection has occurred (see Figure9.27).

Usually, a TDR can do much more than the simple time count of echoes. TDRshave graphical displays for graphical representation of the received pulses. Some-times, the echo contains pulses received from different impedance discontinuities.Analysis of these pulses gives very important information about the status of the pairunder test.

The simpler measurement to make by using a TDR is to estimate the length ofthe line. If the time it takes to receive the echo is called , the cable length would be:

where is the propagation rate and the cable length.

The polarity of the received pulse depends on the line. If the line ends in a shortcircuit, the polarity of the reflected pulse is inverted. If the circuit is open, there isno change in the polarity of the signal. Generally speaking, we can say that the po-larity of the reflected signal depends on the sign of the reflection coefficient (see Fig-ure 9.28).

This conclusion can be extended to any situation where the transmitted pulsemust cross a zone on the line where the impedance is different from nominal. Whenthis happens, the test pulse is divided. One part of it is transmitted, and another partreflected with inverted polarity, if the new impedance is less than the nominal one.

Pulses

Figure 9.27 Start of a TDR measurement.Chronometer

Start

EndPulse generator

td

l ct2----=

c l


9.3.2 Two-End Measurements

Basically, measurements from the two ends simultaneously are more difficult tocarry out than one-end measurements. Two testers are required, one to generate asignal and another one to analyze it. It is essential for these testers to interact in away that the graphical user interface of one enables us to configure the other. Thisway, we will not need a person per instrument to carry out these tests successfully.

9.3.2.1 Attenuation measurement

Line attenuation is the most typical parameter for the two testers to measure. Nor-mally, one of these testers is a sinusoidal signal generator, and the other a powertester. The insertion loss measurement, in the case of a sinusoidal test signal, is re-lated to the transfer function module. If the sinusoidal generator is able to scan forfrequencies, the receiver can estimate the transfer function module or even repre-sent it graphically (see Figure 9.29).

TDR

TDR

TDR

Figure 9.28 (a) Pulse reflected at the end of a line terminating with an open circuit. (b) Pulse reflected on a line terminating in a short circuit. (c) Pulse reflected on a line with a lateral line. Echoes also appear, caused by open-circuit endings in principal and auxiliary sections.

Open circuit

Short circuit

Lateral line

a)

b)

c) Open circuit


When we are measuring the attenuation of a cable, we must minimize othercauses of power loss, in other words return losses, at reception. For this, it is neces-sary to adapt the impedance of the generator and the analyzer with the line, as wellas possible.

9.3.2.2 FEXT measurement

The FEXT measurement is very similar to the NEXT measurement. The only dif-ference is that in this case, the generator and the analyzer are physically far awayfrom each other (see Figure 9.30).

9.3.3 Bridged Measurements

The power spectrum density measurement is an example of a measurement thatmust be carried out bridged. Its purpose is to analyze the power spectrum transmit-ted in an xDSL system, to see that the power meets the masks defined by standards.

This is an in-service measurement carried out with an xDSL signal present. Ob-viously, these measurements must intrude as little as possible, so as not to damagethe measured signal. For this, high-impedance connections are used (see Figure9.31).

Sinusoidal generator

Z0

Figure 9.29 Attenuation measurement.

Zout=Z0

Power tester

Zin=Z0

Z0Z0

Figure 9.30 FEXT measurement.

Sinusoidal generator

Z0

Z0


Very often, the analyzer input is of low impedance (for example, 75Ω or 120Ω),which is why we must use baluns (devices that allow to connect a balanced with anunbalanced interface) and impedance converters. This way, we can extract a smallpart of the signal power transmitted, without varying line impedance significantly,and avoiding undesired reflections. When a sample is extracted from the line signal,all the results are affected by a scaling factor. Measurements must be multiplied bythis factor to obtain the real power density values for the line.

9.3.4 Digital Measurements

Contrary to analog measurements, which are based on measuring parameters, digi-tal measurements are based on testing the service directly. Digital tests are used tocheck that the service works, but in most of the cases they do not provide any pa-rameters to quantify the quality of the channel.

The information an in-service test provides is in many cases just a pass/fail in-dication. In the case of a short circuit, the polarity of the reflected pulse changes.Digital tests should be considered as a complement to analog tests which confirm thevalidity of these tests; however, they are not a substitute for them.

DSLAM

Figure 9.31 Power spectrum density measurement.

DownstreamUpstream

Frame

Loop

Modem

Test Head

Remote Test

Switching matrix

Loop and CPE test

IP

DownstreamUpstream


Usually, tests that measure the transmission bit error rate directly are difficult tocarry out, and rather time consuming. The tests carried out are based on checkingone of the ADSL services directly. For example, in the case of Internet access, thesimplest tests are a file transfer protocol (FTP) file transfer test or an IP ping test toa remote host (see Figure 9.32).


• Alan Keen, Qualifying Copper, Trend Communications, 2001.

• Francisco Hens and José M. Caballero, xDSL Service Integration, Trend Communications, 2001.

• ITU-T Rec. G.117, Transmission aspects of unbalance about earth.

• ITU-T Rec. G.992.1, Splitterless asymmetic digital subscriber line (ADSL) transceivers.

• ITU-T Rec. G.961, Digital transmission system on metallic local lines for ISDN basic rate access.

• A. Bruce Carlson, Communication Systems. McGraw-Hill International Editions, 1986.

Figure 9.32 Service verification by means of remote testing.

ISP

ATM/IP

DSLAM

Test Unit

Control Center

CPE

Service availavity

Service Verification

AuthenticationLogical connectivityTCP/IP Bandwidth

WiringNumber assignationLogical connectivityBandwidth

Core ATM throughputPerformanceVCC connection

ATU-R emulation

IP PingIP

Switching matrix

369

Chapter 10

Jitter and Wander Control

10.1 DEALING WITH JITTER

Jitter is one of the three traditionally considered effects that disturb the timing ofnetworks: jitter, wander, and slips. The first two are variations of the same physicaleffect: phase fluctuation. Slips are a different phenomenon, although related to theabove.

10.1.1 Phase Fluctuation

In terms of time, the phase of a signal can be defined as the function that providesthe position of any significant instant of this signal relative to its origin in time. Asignificant instant is defined arbitrarily; for instance, it may be a trailing edge orleading edge, if the signal is a square wave (clock signal).

Here, when we talk about a phase, we think of it as being related to clock sig-nals. Every digital signal has an associated clock signal in order to determine, on re-ception, the instants at which to read the value of the bits this signal is made up of.The clock recovery on reception circuits reads the bit values of a signal correctlywhen there is no phase fluctuation or when there is very little. Nevertheless, whenthe phase fluctuation, presented by the signal received is fast enough, due to techno-logical limitations, the circuits may not be able to trace these fluctuations (absorbthem). It is in such cases that the sampling instants of the clock obtained from thesignal may not coincide with the correct instants, producing bit anomalies.

When phase fluctuation is fast, this is called jitter. In the case of slow phase fluc-tuations, known as wander, the previously described effect does not occur. Randompointer adjustments do occur, however, and may be the cause of jitter in the tributarysignals carried by the synchronous signals.


10.1.2 Jitter Metrics and Measurement

The parameters that characterize the jitter of a digital signal are amplitude and fre-quency. The amplitude quantifies the extent to which a significant instant deviatesfrom its ideal reference position. The frequency tells us how quickly this significantinstant is moving relative to its ideal position in time.

If we look at the amplitude of phase fluctuation with time as a periodic signal,when its frequency is higher than 10 Hz, the fluctuation is said to be fast, and this isjitter. Phase fluctuation is not usually a periodic signal in real cases, and for this rea-son we analyze the presence of frequency components in its spectrum above or be-low 10 Hz, to determine if what we have is jitter or wander.

0 1 11 00

Data signal

Clock signal

jitter

Sampling instants (example: in leading edges of clock)

Signal sampling points

tb

tb: bit time; it is 1 IU to the signal frequency

Figure 10.1 Definition of unitary interval.

Figure 10.2 The jitter measurement does not need an external reference clock because it is recovered from the incoming signal to be tested.

phase detector

VCO

measurementfilters

signal with jitter

clock with jitter phase error

Jitter measurementlowpassloop

PLL

internal reference clock(without jitter)

filter

clock recovery

Jitter and Wander Control 371

10.1.2.1 Unitary interval

It is usual to measure jitter amplitude in terms of a relative unit. This unit is normal-ized in respect to the signal rate, and is called a unitary interval (UI). A unitary in-terval is defined as the time equivalent to the bit time for the work rate in question(see Figure 10.1). Thus, a unitary interval for a 2-Mbps signal corresponds to ap-proximately 488 x 10-9 s, whereas a UI for an STM-1 signal corresponds to 6.4 ns.

10.1.2.2 Jitter measurement filters

The simplest jitter measurements have the goal of obtaining peak-to-peak ampli-tude values in UI within a specific frequency band over a specific measurement in-terval (see Figure 10.2). This means that any instrument capable of measuring thesefluctuations must have a bank of weighting filters that limit the band of the signalmeasured. These filters (in terms of their frequency cut-offs and slopes) are definedby ITU-T recommendations.

ITU-T Rec. G.823 establishes the levels of jitter that can be found in PDH in-terfaces, from 64 Kbps to 140 Mbps (see Figure 10.3). In the same way, TelcordiaGR-499 deals with the jitter related to T-carrier 1.5 and 45-7Mbps interfaces. Twomeasurement filters are specified:

• Wideband: This filter measures jitter over the whole band of frequencies onwhich phase fluctuation is thought to exist, and the band depends on the specif-

Figure 10.3 Jitter measurement filters for PDH and maximum jitter values in line with G.823.

LowpassHighpass

Wideband (B1)

Highband (B2)

HP1 HP2 LP

jitter pk-pk (B1)

jitter pk-pk (B2)

dB

f

Rate B1 B2

64 Kbps 0.25 0.052 Mbps 1.5 0.28 Mbps 1.5 0.2

34 Mbps 1.5 0.15140 Mbps 1.5 0.075

Rate HP1 HP2 LP64 Kbps 20 Hz 3 kHz 20 kHz2 Mbps 20 Hz 18 kHz 100 kHz8 Mbps 20 Hz 3 kHz 400 kHz

34 Mbps 100 Hz 10 kHz 800 kHz140 Mbps 200 Hz 10 kHz 3,500 kHz


ic hierarchical interface. This filter is specified between the frequencies HP1and LP (highpass filter and lowpass filter, respectively).

• Highband: This filter allows us to characterize the spectral distribution of thehigh frequency jitter, which is the jitter most likely to cause problems in clockrecovery circuits.

As with PDH or T-carrier, measurement filters are established for SDH and SO-NET; this time according to ITU-T Rec. G.825 (SDH) and ANSI T1.105.03, ANSIT1.102, and Telcordia GR-253 (SONET). In this case, a highband and a widebandfilter are also defined (see Figure 10.4).

In short, weighting the measurement using filters serves to determine the spec-tral content of jitter in each frequency band (pass bands of the programmed filters).These weightings allow conclusions to be drawn when problems appear, or even en-able us to predict them. For instance, a concentration of energy in a specific low fre-quency band may result in specific synchronization problems, or problems inoperating the terminal equipment.

10.1.2.3 Measurement interval

As mentioned before, a jitter amplitude measurement must be carried out over agiven measurement interval. The usual measurement period is 60 seconds, althoughwhen measuring jitter, longer periods are required due to pointer adjustments(phase quantization), as these occur sporadically (see Figure 10.5).

HP1-LP:

HP2-LP:

1.5 UIpp

0.15 UIpp

Figure 10.4 Jitter measurement filters and maximum jitter values for SDH, in line with G.825.

Highpass

Wideband

Highband

UIpp

HP1 HP2 LP

1.5

0.15

f

Lowpass

Rate HP1 HP2 LPSTM-0 100 Hz 20 kHz .4 MHzSTM-1 500 Hz 65 kHz 1.3 MHzSMT-4 1 kHz 250 kHz 5 MHz

STM-16 5 kHz 1 MHz 20 MHzSTM-64 20 kHz 4 MHz 80 MHz

STM-256 80 kHz 16 MHz 320 MHz


10.1.3 Measuring Jitter in Output Interfaces

This measurement attempts to obtain the jitter amplitude (expressed in UIs) presentin the output port of a specific NE. The ITU-Tspecifies and limits the maximumamount of jitter allowed in a network. In particular, ITU-T Rec. G.825 (SDH) andG.823 (PDH), and Telcordia GR-253 (SONET) and GR-499 (T-carrier), limit themaximum amount of jitter in the NE output ports. This output jitter may be generat-ed by the NE itself, or may result from the transfer of jitter from one of the inputs ofthe element, either the data input or the synchronization input. The result of themeasurement is the jitter amplitude in a specific bandwidth, specified by the above-mentioned recommendations for each rate in the PDH and SDH hierarchies (seeFigure 10.6). Should we wish to measure the jitter generated by the NE, we need toconnect an input signal free of jitter.

Figure 10.5 The integration period is the time interval measuring the jitter.

Jitter modulationAmplitude

Integration period

Jitter for t=titi

T

t

UIpp

Jitter analyzer

...

Network

DUTInput signalfree of jitter

(a)

(b) Intrinsic jitter

Jitter

Figure 10.6 (a) Jitter in output port: General case.(b) Instrinsic jitter. DUT is the NE being examined.


10.1.4 Measuring Jitter Tolerance

Network elements are designed to tolerate a certain amount of jitter in their inputswithout losing synchronization or introducing anomalies. This amount is specifiedin Recommendations ITU-T Rec. G.823 (PDH), and G.958 and G.825 for SDH,and Telcordia GR-499 (T-carrier), and GR-253 (SONET).

Jitter tolerance is therefore defined as the maximum jitter amplitude in the inputof an NE that does not produce bit anomalies or synchronization anomalies. Theseamounts are specified by the recommendations in the form of masks. In these masks,jitter amplitude is specified in UI versus frequency. It is recommended that the mea-surement configuration be synchronized with a reference clock common to both theNE and the tester, to avoid occasional pointer adjustments. The input signal in whichjitter is to be introduced must be a pattern that is suitable for the frame rate of thesignal (see Figure 10.7). The pattern depends on its rate (for instance, those stipulat-ed by ITU-T Rec. O.150 for PDH, or O.181 test structures for SDH).

The type of NE examined determines which input and output interfaces comeinto play for the measurement being performed:

• Regenerators: Measurements are made in the output line interface correspond-ing to the input line interface where jitter is inserted.

• Multiplexers: Measurements are made on the channel, in the output aggregatesignal corresponding to the input interface where jitter is inserted.

• Demultiplexers: Any tributary interface is representative of the aggregate inputinterface where jitter is inserted.

DUTSinusoidal jitter

Reference

Anomalies or excessive jitter

Figure 10.7 Jitter tolerance measurement configuration.


More precisely, the measurement is carried out by inserting sinusoidal jitter inan input port. The amplitude of the jitter tone introduced is increased until events aremeasured at the output port. This is repeated at set frequencies. There are two waysto carry out the tolerance measurement: onset of errors and BER penalty.

10.1.4.1 Onset of errors

This is the usual method for checking that the buffering and clock recovery func-tions are working properly in the NEs. It consists of increasing the jitter amplitudeuntil we can observe the signal deteriorating to the point where it reaches a certainthreshold (for instance, the threshold recommended by O.171 for PDH signals istwo seconds with anomalies in a period of 30 seconds). At this point, the jitter am-plitude is registered, and this gives us the tolerance to this frequency. The test isthen repeated for a set range of frequencies.

10.1.4.2 BER penalty

This method is more appropriate for line systems (regenerators), normally usingoptical interfaces. In this case, the tolerance level is established when the deteriora-tion of the signal is the same as that produced by lowering the transmission powerby 1 dB. In other words, if the anomaly rate produced for a set amplitude of sinuso-idal jitter coincides with that measured when the power is lowered by 1 dB, thisamplitude is the jitter tolerance for that frequency. The test is then repeated for a setrange of frequencies.

10.1.4.3 Tolerance masks

Jitter tolerance measurements are aimed at checking that certain limits of jitter am-plitude, preestablished by ITU-T and Telcordia recommendations for a set range offrequencies, are not exceeded. These limits are shown in masks or amplitude-fre-quency graphs in Recommendations ITU-T Rec. G.823 (PDH) and G.825 (SDH),and Telcordia GR-499 (T-carrier), and GR-253 (SONET). The masks show the fre-quencies at which tolerance should be measured (see Figure 10.8).

For optical regenerators in SDH networks, ITU-T Rec. G.958 likewise definestolerance values by means of two masks: one for A-type and another one for B-typedevices. The mask for A-type devices complies with ITU-T Rec. G.825, while themask for B-type devices is much more restrictive (see Figure 10.9).


10.1.5 Measuring Jitter Transfer

Network elements have a limited capacity to eliminate the jitter that may occur intheir input ports. To evaluate this filtering capacity, the function of jitter transfer isto define the relation between the jitter amplitudes on input and output for a setrange of frequencies. As usual in transfer functions, this relation of amplitudes is

A0

A1

A2

A3

f0 f1 f2 f3 f4 f5 f6 f7Frequency

Input jitter and wander tolerance G.823 (PDH)

A0

A1

A2

f1 f2 f3 f4 f5

Input jitter tolerance G.825 (SDH)

Figure 10.8 Input jitter and wander tolerance, G.823 and G.825

Wander

Amplitude (UIpp)

Frequency

Jitter

Amplitude (UIpp)

Rate Ao (µµµµs) A1 (µµµµs) A2(UI)))) A3(UI)))) fo(µµµµH) f1(mH) f2(mH) f3(Hz) f4(Hz) f5(Hz) f6(kHz) f7(kH)

64 Kbps 18 - 0.5 0.05 12 - - 4.33 20 600 3 202 Mbps 18 8.8 1.5 0.2 12 4.88 10 1.7 20 2400 18 1008 Mbps - - 1.5 0.2 - - - - 20 400 3 400

34 Mbps 4 1 1.5 0.15 10000 32 130 4.4 100 1000 10 800140 Mbps 4 1 1.5 0.075 10000 32 130 2.2 200 500 10 3500

Rate Ao (UI)))) A1 (UI)))) A3(UI)))) fo(H) f1(H) f2(kH) f3(kHz) f4(kHz) f5(MH)

STM-1e 38.9 1.5 0.075 10 19.3 0.5 3.3 65 1.3STM-1o 38.9 1.5 0.15 10 19.3 0.5 6.5 65 1.3

STM-4 see * 1.5 0.15 9.65 1000 1 25 250 5STM-16 622 1.5 0.15 10 12.1 5 100 1000 20STM-64 2490 1.5 0.15 10 12.1 20 400 4000 80

f0

(*)1500 f -1 UI


expressed in decibels. As with tolerance, the jitter introduced at the NE input to car-ry out the measurement is sinusoidal jitter.

One important aspect to consider is intrinsic jitter; that is, jitter generated insidethe NE. To perform a correct transfer measurement, this intrinsic jitter must be sub-tracted from the output jitter; that is to say, the measurement must be calibrated. Intransfer measurements, there is also the possibility to apply a filter to the NE output(as with output jitter). This depends on the measurement being performed (see Fig-ure 10.10).

The function of jitter transfer J(f) is:

10.1.5.1 Jitter transfer in PDH and T-carrier

Recommendations G.742 and G.751 establish the jitter transfer requirements forplesiochronous multiplexers and demultiplexers. These requirements are set out inline with a set mask, in such a way that the performance specified for these NEswill be according to the tolerable jitter levels in an interface, as specified in ITU-TRec. G.823 (see Figure 10.11). For T-carrier interfaces, the maximum tolerable jit-ter is defined by Telcordia GR-499, ANSI T1.403 (1.5 Mbps) and ANSI T1.404(45 Mbps).

1.5 UIpp

f0 ft

Tolerance mask

25 kHz

f0

250 kHz

5 kHz

STM-4

STM-16

6.5 kHz

ft

1 MHz

65 kHzSTM-1

Tipo A

1.2 kHz

f0

12 kHz

1.2 kHz

STM-4

STM-16

1.2 kHz

ft

12 kHz

12 kHzSTM-1

Tipo B

0.15 UIpp

Figure 10.9 SDH optical regenerators, tolerance masks; ITU-T Rec. G.958.

Type BType A

Amplitude (UIpp)

frequency

J f( ) 20 OutputJitter IntrinsicJitter–InputJitter

----------------------------------------------------------------------- dB( )log⋅=


10.1.5.2 Jitter transfer in SDH and SONET

To check jitter transfer between SDH/SONET synchronous interfaces, the NE un-der test must be synchronized with the input interface where jitter is generated,since the reference synchronization of the NE is actually what determines the tim-ing of its STM-n/OC-m outputs. With this prior condition, it is established that theNE cannot amplify jitter above 2.3 % (0.2 dB) of the passband, which is deter-mined by its clock recovery filter. This bandwidth typically reaches 1 Hz for NEswith G.813 category clocks (ETS 300 462-5); (see Figure 10.12). The equivalentStratum 3 clocks are defined in ANSI T1.105.09 and Telcordia GR-1244.

Obviously, it makes sense to check the jitter transfer between the NE PDH trib-utary ports. In this case, the performance of the device must meet RecommendationsG.742, G.751, and G.823.

DUT

Intrinsic jitter

Sinusoidal jitter + intrinsic jitter

J(f)

Figure 10.10 Setup for measuring jitter transfer for each frequency.

Sinusoidal jitter

Ji(f) Jo(f)

Ji: Input JitterJo: Output JitterJ(t)= Jitter transfer

0.5

f0 f5

dB

f

Jitter transfer in PDH

-19.5

f6 f7

Figure 10.11 Jitter transfer mask for PDH. The numbering of the frequencies in the graph is not correlative, but no significant frequency is missing.

f0 f5 f6 f7

< 20 Hz 40 Hz 400 Hz 100 kHz 2 Mbps< 20 Hz 100 Hz 1 kHz 400 kHz 8 & 34 Mbps< 20 Hz 300 Hz 3 kHz 800 kHz 140 Mbps


10.1.6 Mapping Jitter and Combined Jitter

Jitter from phase quantization and desynchronization results in mapping jitter andcombined jitter.

10.1.6.1 Mapping jitter

Mapping is the process through which PDH/T-carrier signals are introduced inSDH/SONET signals for transport. This takes place in multiplexers. The signalclock source is independent of the SDH clock source, so the PDH/T-carrier data isasynchronous from the SDH/SONET signals. In other words, the clocks of the trib-utary systems have no fixed relation with the multiplexer, or even between them-selves. The tributary signals are asynchronous, that is, they allow deviations withina tolerance margin relative to their nominal clock value. This is where stuffing bitscome in. They are used to resolve this asynchronicity, and in this way the PDH/T-carrier signal becomes part of the SDH/SONET payload which has a greater capac-ity. This excess capacity is filled with stuffing bits to obtain the constant rate speci-fied for the container.

At the transmitting end, the bits of the tributary signals are continuously record-ed in elastic memories, but are read discontinuously, because in order to proceedwith their reading (and later transmission), these memories must first be filled; oth-erwise they would end up emptying completely. Reading is carried out at the highestpossible rate, since the clock adaptation process performed by the multiplexer pro-

level fc

Type A STM1 130 kHzSTM4 500 kHzSTM16 2000 kHzhigher tbd

Type B STM1 30 kHzSTM4 30 kHzSTM16 30 kHzhigher tbd

dB

f

0.2

dB

f

Optical Regenerators

0.1

fc

Add and Drop Multiplexer

Figure 10.12 Jitter transfer masks for SDH on ADMs and optical regenerators.


vides a transmission channel with a capacity higher than the sum total of the tribu-tary rates plus the permitted tolerance (deviations relative to the nominal value).Since we want the reading clock to stop at times, but the output rate of the multiplex-er aggregate signal must remain constant, stuffing bits are sent when there are no bitsof information to be transmitted.

At the receiving end, these stuffing bits must be extracted to recover the tribu-tary signals correctly. For this reason, these bits are not written in the elastic memo-ries of the receiver. The frames contain indications that let them decide whether a bitis a stuffing bit or not. If the bit received is a stuffing bit, the writing clock is stopped.On reception, therefore, writing is discontinuous, whereas reading is continuous(since everything the memory contains forms part of the message).

On reception, the reading clock is derived from the writing clock by means of aphase-locked loop (PLL). Since the low pass filter of the PLL is not able to com-pletely eliminate the discontinuities in the writing process, a residual phase modula-tion remains, and this is known as mapping jitter (also stuffing jitter).

10.1.6.2 Combined jitter

Mapping jitter is measured when there are no pointer adjustments in the aggregatesignals. The pointer adjustment mechanism is one of the causes of jitter (pointer jit-ter), but this cannot be separated from mapping jitter, which is inherent in the gen-eration of SDH/SONET signals. For this reason, pointer jitter cannot be measuredseparately from mapping jitter, and the fluctuation measured is known as combinedjitter (see Figure 10.13).

Pointer jitter is the most important type of jitter found in SDH/SONET net-works. It is the main cause of disturbance in hybrid PDH/SDH or T-carrier/SONETnetworks and, compared to mapping jitter, it is a component of much greater impor-tance when it comes to measuring combined jitter. The cause of this type of phasefluctuation is the pointer adjustment mechanism, and it appears in those tributariesthat, once disassembled (extracted from their virtual container), have undergonepointer changes along their path.

The pointer mechanism forms the basis of the SDH/SONET signal structure.For example, the essential difference between an SDH frame and a PDH frame isthat in the former, the overhead information from the higher-order signals is enoughto determine the overhead position of the lower-order signals. Pointers are values(divided into various bytes) that contain the position of the overheads. For example,for the mapping of 2-Mbps signals in SDH, two pointer levels are used. The higherlevel (AU-4 pointer) identifies where the VC-4 virtual containers start inside theSTM-1 frame. The lower level (TU-12 pointer) identifies the start of a VC-12 virtual


container relative to the VC-4. In one STM-1 frame there will therefore be one AU-4pointer and 63 TU-12 pointers.

When there are clear differences in the clock signals from two different net-works or two different elements in the same network, it is necessary to compensatefor these differences by offsetting the lower-order signals into higher-order (for ex-ample, VC-4 virtual containers in the STM-1 frame). This is achieved by increasingor decreasing the pointer value by one unit (depending on the appropriate adjustmentat that time). The value of this offset depends on the pointer being adjusted. Return-ing to the case of mapping a 2-Mbps signal in SDH, an AU-4 pointer adjustmentmeans an offset of 24 bits, whereas if it is a TU-12 pointer, the offset is of 8 bits. TheAU-4 pointer adjustment contributes more to jitter, due to the fact that it appearsmore commonly than TU-12 pointer adjustment.

In any case, these offsets cause abrupt phase variations in lower-order signals(tributaries). As with mapping jitter, the reading clock at the receiving end is derivedfrom a PLL circuit. The lowpass filter in the control loop of the PLL tries to smoothout these phase shifts, but they nevertheless cause residual phase modulation to re-main. This is pointer jitter, and it is the main contributor to combined jitter.

10.1.6.3 Measuring combined jitter

To check how effectively an NE compensates for the effects of pointer adjustments,some pointer adjustments are generated that have been specially designed to subjectthe element to stress, in such a way that the situations simulated might occur undernormal conditions. A pointer adjustment is a change in its value, such as a unitaryincrement or decrement, for instance. Pointer sequences are defined in Recommen-dations ITU-T Rec. G.783, ANSI T1.105.03, and Telcordia GR-253. The measure-ment consists of checking the output jitter against this input stimulus. Combinedjitter (which is largely pointer jitter) appears in the tributaries of a synchronous sig-nal when these are extracted from the signal (see Figure 10.13).

G.783 pointer sequences

Combined jitter

Figure 10.13 Combined jitter measurement, which includes pointer jitter.

ADM

... ...

STM-n/OC-m

Tributary

STM-n/OC-m


10.1.7 Jitter in Leased Lines

Although tolerance to jitter is usually checked by means of sinusoidal stimuli, thereare certain cases when this test is not performed this way; and such is the case withleased lines. In this type of system, it is necessary to have a method that character-izes and supplies more reliable results about the real jitter tolerance level than theresults obtained by using frequency tones. The reason for this is that although thesetones are useful for checking that the buffers are working correctly, they are not ex-act enough when it comes to reproducing the random characteristics of the jitterfound in “real” systems.

For example, the ETSI has defined a series of broadband signals for checkingjitter tolerance in leased lines. These take advantage of the random characteristics ofthe PRBS patterns, to modulate the phase of the data signal of the line (2 Mbps inthe example), following the appropriate filtering. The jitter tolerance of the line isevaluated at the remote end, depending on the anomalies that appear, and the amountof output jitter.

10.2 DEALING WITH WANDER

To guarantee correct operation for SDH/SONET networks, the elements that makeup these networks must be synchronized, that is, they must share a common refer-ence clock. The common reference signal to which the clocks of the NEs them-selves are synchronized, usually comes from high-quality clocks that act as aprimary reference clock (PRC). Based on these clocks, the signal is distributed in anetwork of subsidiary clocks until it reaches the NEs. So, there is a network ofclocks that synchronizes the SDH/SONET network (see Figure 5.3).

Here, wander is a critical type of phase fluctuation, since it builds up in the syn-chronization chain. This slow phase fluctuation can often be observed on the bound-ary between two different SDH/SONET networks, each with its own PRC, and onthe international boundaries where there are networks using different referenceclocks. The causes of this wander are:

• Changes in the propagation delay of cables (temperature);

• Drifts in the PLLs of the clocks;

• Phase fluctuations due to reconfiguration of the synchronization chain, eitherby the operators themselves or by the automatic protection switching mecha-nism;

• Differences in frequency resulting from a loss of synchronization in a networknode (limit of 4.6 ppm).


10.2.1 Synchronization of SDH/SONET Networks

A functional separation can be established between the SDH/SONET network andthe network of clocks that synchronizes it (in some cases, even if this separation isnot physical, as will be seen). Owing to the problem of wander in synchronizationchains, the ITU, ETSI, ANSI, and Telcordia have produced some recommendationsfor limiting this phase fluctuation in all these clocks (PRC/PRS, SSU/BITS, andSEC), and guaranteeing that the SDH/SONET network operates correctly (see Sec-tion 5.4.1).

10.2.2 Measuring Relative and Absolute Wander

Given our understanding of wander as the difference of phase (or of time) betweentwo clock signals, it is important to distinguish between relative and absolute wan-der measurement (see Figure 10.14).

The measurement of absolute wander at a given instant is equal to the phase dif-ference that exists, at that moment, between the clock of a signal and UTC, as de-fined in ITU-T Rec. G.810, ETS 300 462-1, and ANSI T1.101. The effective qualityprovided by the synchronization network can be seen by carrying out this measure-ment which requires a high-quality clock source, derived directly from the UTC(such as that provided by a GPS receiver).

The measurement of relative wander at a given instant is the phase differencethat exists, at that moment, between any two clocks in the network. It is very usefulto carry out this measurement in two interfaces, when we want to check such aspectsas wander generation in a synchronous element (checking the gap between input andoutput interfaces), the possible appearance of pointer adjustments between twoSTM-n/OC-m signals that converge in a single NE, and so on. In short, the differencebetween both types of measurement depends on the reference clock chosen: In thecase of absolute measurement, the clock being measured is compared against themost stable reference that exists.

Figure 10.14 Wander measurement needs an external reference; otherwise it will take a lot of time to get a good recovered clock.

clock phase detector

TIEanalysis

signal with wander

clock with wander TIE (wander amplitude)

Wander measurementexternal reference clock

(without wander)MTIE, TDEV

recover


10.2.3 The Metrics of Wander: TIE, MTIE, and TDEV

Given that wander is a slow phase fluctuation (with spectral components below10 Hz), wander measurements require long periods of time. It is also necessary todetect phase transients during these measurements, which calls for high temporalresolution, and, as a result, there is a considerable accumulation of data. With theaim of summarizing all this information, three parameters are defined below thatare fundamental in measuring wander: TIE, MTIE, and TDEV.

10.2.3.1 TIE

The time interval error (TIE) is the slow phase fluctuation amplitude, which meansthat it indicates the phase variation of the clock to be measured, relative to thephase of an ideal reference clock during each instant of the measurement. Usually,TIE=0 is taken as a reference at the start of the measurement. The TIE can be andusually is expressed in absolute time (ns, µs, ms), but it can also be expressed rela-tive to the signal period (unitary intervals).

10.2.3.2 MTIE

The maximum time interval error (MTIE) is the maximun value of peak-to-peakTIE (TIEpp) in a certain observation time, t. This means that in order to calculatethe MTIE, a time window must be scrolled over the function TIE(t), recording themaximum peak-to-peak value of the TIE: TIEpp. This can be repeated for differentvalues of τ, thus obtaining a graph of MTIE (τ) (see Figure 10.15).

10.2.3.3 Application

The MTIE helps in obtaining realistic information on the buffer size of synchro-nous instruments. The buffers of digital instruments associated with clock recovery(PLL) allow frequency fluctuations to be absorbed, but their size must be limited, toavoid increasing latency. This size is calculated by using the MTIE (see Figure10.15).

10.2.3.4 TDEV

Time deviation or TDEV is a measurement that characterizes the spectral content ofa TIE(t) signal. This means that it measures the energy of wander frequency com-ponents. As is the case with MTIE, the TDEV is a function of the integrationtime τ. The functional diagram of a TDEV measurement circuit is shown in Figure10.16.


The first block, H(f), is a filter with its passband (0, 1/τ) centered on the value0.42/τ. The analysis is therefore limited to the passband mentioned before. The sec-ond block calculates the value of the root mean square (r.m.s.), which evaluates theenergy of the components in the band being analyzed. By varying the value of τ, wecan then analyze the different frequency bands that are of interest to us. The aboveconsiderations have ITU-T Rec. G.810 and ANSI T1.101 as their source.

For a correct calculation of the TDEV, it is recommended that the length of themeasurement be 12τ, although 3τ is enough; that is, we must have samples of TIEin at least the time interval 0,3τ, with t=0 being the starting instant of the measure-ment. Given that the TDEV is an r.m.s. value, it is always positive, as a sum ofsquare (see Figure 10.16).

10.2.3.5 Application

The TDEV lets us evaluate the short-term stability of the clock signal. We can char-acterize the transfer of wander in the NE used, in order to limit the buildup of thisphase fluctuation (ETSI specifications on the transfer of wander between the portsof a synchronization source - clock - specify this transfer in terms of TDEV).

Furthermore, the TDEV converges for many types of phase noise, which makesit possible to identify the source and eventually correct the causes of degradation intransmission (see Table 10.1).

Figure 10.15 MTIE (τ): maximum peak-to-peak amplitude of the slow phase fluctuation or TIE in an observation window τ.

TIE (ns)

t20 6040 80 120100 140

TIE

Ref

Clock

MTIE (ns)

10 3020τ

MTIE (τ)

10

20

30

40

10

20

30

40

τ

MTIE

TIE


10.2.4 Measuring Output Wander

The aim of this measurement (as with jitter) is to quantify the amount of wanderpresent in a network interface or a device. This quantification is usually made interms of MTIE and TDEV (see Figure 10.17).

Synchronization signals are distributed through SDH/SONET sections (STM-n/OC-m) and PDH/T-carrier paths (2 or 1.5 Mbps). The ETSI RecommendationETS 300 462-3 (PDH) and Telcordia GR-499 (T-carrier) establish masks showingthe limits of MTIE and TDEV for these rates, where applicable. The synchronizationinterfaces at these rates in which output wander is measured are:

• PRC/PRS clock outputs;

• SSU/BITS clock outputs;

• SEC clock outputs;

10

203040

TIE (ns)

t (s)20 6040

0.42 /10f

0.42 /30f

H(f)dB

0.42 /20f

dBdB

observation time

80 120100

x x x

TDEV (ns)

τ10 20 30

Figure 10.16 Calculation of the TDEV step by step. (1) for each window τ swept; (2) low pass filter; (3) measurement of r.m.s value; and (4) new TDEV point.

H(f)H(f)

TIE(t)TDEV(ττττ)H(f) r.m.s.

Band (0, 1/ττττ) Energy evaluation

(1) (2) (3) (4)

τ=10 τ=20 τ=30

τ

(4)

(3)

(2)

(1)

r.m.s. r.m.s. r.m.s.


• PDH/T-carrier synchronism distribution outputs.

The masks refer directly to each type of interface, that is, the TDEV and the MTIEare represented relative to the output wander of a PRC, an SSU, an SEC or a PDHoutput.

10.2.5 Measuring Tolerance to Input Wander

Tolerance to wander is based on the same concept as tolerance to jitter: How muchphase fluctuation can an NE withstand at its input without seeing any degradationin its operation? ITU-T Rec. G.823 and G.825, ANSI T1.105.03, and T1.102, andTelcordia GR-499 and GR-253 specify the tolerance to wander (and jitter) for theinterfaces of NEs (see Figure 10.8). When we are dealing with tolerance to wanderin synchronization input interfaces, ETS 300 462-4 and ETS 300 462-5 apply toSDH, while ANSI T1-101 and T1.105.09 apply to SONET, specifying this toler-ance in terms of MTIE and TDEV by means of the corresponding masks for anSEC and an SSU/BITS. In any case, the tolerance specifications for SEC and SSU/

Table 10.1The TDEV converges for some types of noise, which makes

it possible to investigate the causes of phase noise.

Noise Slope of the TDEV

White noise phase modulation (WPM) -1/2Flashing phase modulation (FPM) 0White noise frequency modulation (WFM) 1/2Random walk frequency modulation (RWFM) 3/2

Wander analyzer

Network

DUT

(a)

(b)

Wander

Wander

Reference clock

Figure 10.17 Measuring output wander: (a) in a network interface, the analyzer is synchronized with a clock free of wanderm; and (b) in the interface of a device, the device and the analyzer are synchronized using a common reference clock.

Output wander


BITS are in line with the maximum values for amplitude of phase fluctuation underITU-T Rec. G.823 and G.825, and Telcordia GR-499 and GR-253.

Evaluating the tolerance of an SSU/BITS or an SEC consists of establishing theminimum values in terms of MTIE and TDEV in the input of the NE at which theSSU/BITS or SEC operates correctly, and comparing them with the correspondingmasks (the values must be above or at least equal to those of the masks when oper-ating correctly). By “operating correctly,” we understand that the following condi-tions are met:

• The clock remains within its margins of correct operation;

• No defects are shown;

• No switching of reference source takes place, or, if it does, this is without en-tering holdover mode.

These tolerance requirements are independent of the interface in which they aremeasured: 2 or 1.5 MHz, 2 or 1.5 Mbps or even STM-n/OC-m synchronous inter-faces will do.

To measure tolerance to wander, the stimulus signal at the input (which may be2 or 1.5 MHz, 2 or 1.5 Mbps or STM-n/OC-m) is usually subjected to a set wandermodulation (see Figure 10.18). It may be sinusoidal although, for example, ETSIrecommends the alternative of using broadband spectrum signals for this measure-ment, which should last at least 1,000 seconds.

For an SEC, tolerance is expressed in terms of TDEV in accordance withETS 300 462-4 and ANSI T1.101, which define a mask. The phase modulation sig-nal that is provided by the TDEV defined by this mask is a special (broadband) noisesignal; the generation of which is also specified in its Annex C.

Device under test

Figure 10.18 Measuring tolerance to input wander.

SEC Test signal with wander

SSUor


With regard to an SSU/BITS, tolerance to wander is expressed in terms ofTDEV, and also of MTIE, and for special test signals that are adapted to these masks.Since the way of generating these special modulations is under study, a tolerance towander mask is given as a valid alternative measurement when the phase modulationof the stimulus signal is sinusoidal.

The reference clock used in the measurements must always be of higher qualitythan the one in the DUT. For instance, if the tolerance of an SSU/BITS is being mea-sured, the reference clock must be of PRC/PRS quality.

10.2.6 Measuring Wander Transfer

When the synchronization reference signal at the input of an NE contains wander,this is transferred to the STM-n/OC-m outgoing traffic signal, or to the output syn-chronization signal (see Figure 10.19). This transfer leads to an accumulation ofwander in a chain of devices, and must therefore be limited. RecommendationETS 300 462-4 and ANSI T1.101 define the limit of wander transfer, for an SSU,in terms of output TDEV.

In order to measure wander transfer in an SEC, Appendix B of the ETS300 462-5 recommends five different methods, all of them valid and characterizedby the different phase modulation in the stimulus signal at the input.The signal atthe input to the DUT is usually a synchronization reference (see Table 10.2).

10.2.7 Response to Phase Transients

There are a number of different sources from which an SSU/BITS (be it integratedor SASE) or an SEC clock can take its synchronization input signal. This allows itto be programmed in such a way that if this synchronization signal should fail insome way, it will switch over to a secondary synchronization source, and reestab-lish its normal operating status. This switchover may be caused by:

Test signal with wander Output with wanderSEC

Reference clock

Figure 10.19 Measuring wander transfer.

orSSU


• Interruption of the reference signal;

• Phase skipping in the synchronization input due to reconfiguration in the syn-chronization chain;

• Frequency deviation in the reference signal;

• An AIS defect in the 2 or 1.5-Mbps reference signal (when used).

During switchover, a transient period occurs within which the phase of the out-put clock of the SSU or the SEC undergoes shifts. In short, a phase error or excursionoccurs in the synchronization output signal relative to the input signal at the instantwhen the reference is lost. Since the wander in the output of a synchronization chainis limited, it is necessary to limit this phase error as well (see Figure 10.20). For ex-ample, in SDH there is a phase error limitation curve (ETS 300 462) (see Figure10.21). In this case, switching to a secondary reference source is considered to beshort term, which means that it takes place quickly.

Table 10.2Phase modulations for the measurement of wander transfer in an SEC,

in line with ETS 300 462-5.

Phase Modulation Application

Phase step Devices that accept large phase transients or have a low level of inter-nal noise.

Frequency step This measurement method prevents the SEC from entering holdover or overflow occurring in the phase detector.

Sinusoidal Found on devices with linear designs; its advantage is that the output can be selectively measured in terms of frequency.

Phase white noise(frequency domain)

Found on devices with nonlinear designs; also allows SEC to be char-acterized in frequency.

Phase white noise(time domain)

Response of SEC can be characterized in the time domain using a set measurement plan.

Measuring output wander

clock

SEC

Figure 10.20 Transient due to reference switching.

or SSU

connection

disconnection


The curve limits the maximum phase error (Emax) and the maximum durationfor the transient (Tmax). For instance, in the case of an SEC, Emax = 1 ms andTmax = 15 s for the configuration of measuring the phase error (TIE) (see Figure10.22).

In the case of an SEC, then, the response to the transient is measured by evalu-ating the TIE from which the reference is disconnected until it connects again 15slater, and comparing this with the corresponding curve in ETS 300 462 5.

10.2.8 Operating in Holdover Mode

When the reference source is lost and the switching transient goes beyond a certaintime (15 seconds for an SEC), reference switching is said to be long term. In thiscase, the clock enters holdover mode. The restrictions with regard to the phase errorare different from those presented in the section above.

10.2.8.1 Holdover

In holdover mode, the SSU/BITS or the SEC use statistics from the source withwhich they have been synchronized (frequency, frequency drift, etc.) to provide atiming that is as similar as possible to the one they had been supplying. As this isextended, the phase error on output will increase, owing to such factors as varia-tions in temperature, aging of the unit, and so on.

10.2.8.2 Measuring phase error in holdover

The measurement configuration is the same as for short-term switching measure-ment. In this case, it is necessary to compare the measurement results with the cor-responding phase error restrictions for this mode. For example, as with short-term

Tmax

Emax

0 t

Phase error (ns)

Loss of reference

Interlocking to new reference

Figure 10.21 Limit-curve of transient due to short-term reference switching; ETS 300 462-4 for SSU and ETS 300 462-5 for SEC.

SSU SECEmax 240 ns 1 ms

Tmax Ty 15 s


switching, Recommendations ETS 300 462-4 (for SSU) and ETS 300 462-5 (forSEC) give limit-curves, which can be expressed by the following function:

DT(S) = (a1 + a2)·S + 0.5·b·S2 + c;

where:S: holdover time (S > 15s or Ty);DT(S): phase error;a1: maximum value of initial frequency offset;a2: offset due to variations in temperature, if these occur;b: frequency drift due to aging of the unit; c: phase skipping that occurs when entering holdover.

The values for an SSU and an SEC are different (see Table 10.3).

The corresponding phase error curves are shown. Here, it is assumed that thereis no offset due to temperature variation (see Figure 10.22).

Table 10.3Parameters of phase error due to long-term reference

switching (holdover) for an SSU and an SEC.

SSU SEC

a1 0.5 ns/s 50 ns/s

a2 2.0 ns/s 2,000 ns/s

b 2.3 x 10-6 ns/s2 1.16 x 10-4 ns/s2

c 60 ns 120 ns

Time elapsed since loss of reference (s)

Phase error (ns)

SEC

Figure 10.22 Limit-curve of transient due to reference switching in holdover; ETS 300462-4 for SSU and ETS 300 462-5 for SEC.

1E+51E+1 1E+2 1E+3 1E+4

1E+1

1E+5

1E+3

1E+7

SSU


10.3 TESTS ON ADMS AND DXC

Acceptance tests must be carried out on NEs once they have been purchased, tocheck that they are operating correctly.

10.3.1 Measuring Jitter

In this example of acceptance, mapping jitter and combined jitter are measured inan ADM. Basically, these are measurements of jitter amplitude in the multiplexertributary ports (output jitter).

10.3.1.1 Measuring mapping jitter in an ADM

Measuring mapping jitter consists of generating an aggregate signal with a PDH orT-carrier test payload that has a frequency offset with respect to its nominal value.This payload is unmapped in the ADM, and analyzed to quantify the jitter it pre-sents. The conditions under which the measurement is made are as follows (seeFigure 10.23):

1. Three frequency offset values are generated for the test signal: 0 ppm,+50 ppm, and -50 ppm.

2. For each offset generated, two jitter amplitude values are obtained: one corre-sponding to a band of frequencies from 20 Hz to 100 kHz, and another corre-sponding to a band of frequencies from 18 kHz to 100 kHz.

3. For both frequency bands, the measurement interval is 60 seconds.4. The measurement is repeated regularly for a tributary from each tributary card

in the multiplexer, and for both aggregate sides, although this depends on thestipulations of the acceptance protocol for each operator.

10.3.1.2 Measuring combined jitter in an ADM

Measuring combined jitter (pointers plus mapping) consists of generating TUpointer adjustment sequences over the TU pointer associated with the virtual con-

Frequency offset in test payload

Figure 10.23 Setup for measuring mapping jitter.

2 Mbps + mapping jitter... ...


tainer of the PDH/T-carrier test signal in the aggregate signal. The response to theseadjustment sequences (defined by ITU-T Rec. G.783, ANSI T1.105.03, and Telcor-dia GR-253) is the jitter measured in the associated tributary signal.

The conditions under which the measurement is carried out are:

1. Two types of adjustment sequences are generated:

• Isolated adjustments of alternate TUs1or VTs;

• Single TU adjustments with double TU or VT adjustment2.2. For each offset generated, two jitter amplitude values are obtained: one corre-

sponding to a band of frequencies from 20 Hz to 100 kHz, and another corre-sponding to a band of frequencies from 18 kHz to 100 kHz.

3. For both frequency bands, the measurement interval is 60 seconds.4. As with mapping jitter, the measurement is repeated regularly for a tributary

from each tributary card in the multiplexer, and for both aggregate sides,although this depends on the stipulations of the acceptance protocol for eachoperator.

The measurement setup is similar to the one used for measuring mapping jitter (seeFigure 10.24).

1. Isolated adjustments of alternate TUs: This sequence provokes a single pointeradjustment every 10 seconds with alternate polarity. A positive adjustment increasesthe pointer value by one unit, and a negative adjustment decreases it by the sameamount; in both cases the phase is offset by 8 bits (8 UI) (see Figure 10.25).

2. TU with double TU adjustment: This sequence provokes cycles of four pointeradjustments with 0.75 seconds between them, except for the third, which is a doubleadjustment with a separation of 2 ms. All the adjustments have the same polarity (seeFigure 10.25).

Figure 10.24 Setup for measuring combined jitter.

G.783 pointer adjustment sequences

2 Mbps + combined jitter... ...


10.3.2 Synchronization Tests

Synchronization can be distributed to the SDH/SONET network elements in twoways, either from the synchronization network (separate from the SDH transportnetwork), or by using the transport network itself. When a loss of external synchro-nization occurs, the SEC of an NE can act independently, either in holdover or infree-running mode. In the first case, the synchronization signal is reconstructedbased on the data from the synchronization source previously connected. In the sec-ond case, the NE obtains the synchronization signal from its own internal clock.

In free-running mode, the NE clock must maintain a certain level of exactitude,as stated in ITU-T, ANSI, and Telcordia recommendations. This test checks whetheran NE is operating correctly in free-running mode, and when it receives its synchro-nism reference signals through the transport network.

10.3.2.1 Free-running test

This situation comes about, as stated above, when the NE does not have an externalsynchronization signal, and has to obtain its signal from its internal clock. ITU-TRec. G.813, ANSI T1.105.09, and Telcordia GR-1244 determine that the exactitudeof output frequency of an SEC with respect to PRC must not be greater than4.6 ppm.

The free-running test verifies this requirement, as well as checking that tributar-ies have been correctly extracted in the corresponding ports of the element undertest. To carry out this test, the tributary input must be disconnected. This will pro-voke the switchover to free-running mode, and the element under test (ADM in this

T1=10s

Figure 10.25 ITU-T Rec. G.783 sequences. Single alternate adjustments and regular adjustments with double adjustments.

Positive pointer adjustment

Negative pointer adjustment

T2=0.75sT3=2ms

Single pointer adjustment

Double pointer adjustment

Single alternate adjustments

Double adjustments


example) will generate an AIS signal that will be received by the analyzer from thetributary output at the free-running frequency within the tolerance margin,f0nom+4.6 ppm, f0nom-4.6 ppm. One possible procedure for carrying out this test isthe following (see Figure 10.26):

1. Offset the frequency of the tributary signal generated f1 to a value within thetolerance margin, f1nom+50 ppm, f1nom-50 ppm, and check that the frequencyof the tributary signal recovered on extraction coincides with the value withinthe margin (2,048,000±102 Hz).

2. Check that the frequency of the signal in the ADM synchronization output doesnot vary when f1 varies, and that it is within the tolerance margin,f0nom+4.6 ppm, f0nom-4.6 ppm, that is, 2,048,000 ± 9 Hz, in accordance withthe recommendations.

3. Disconnect the tributary signal generated at the input port. As a consequence ofthis loss of signal, the ADM will have to generate an AIS that the 2-Mbps ana-lyzer will receive at the tributary output. The frequency value of this signalmust remain within the margin, f2nom+4.6 ppm, f2nom-4.6 ppm, that is,2,048,000 ± 9 Hz.

10.3.2.2 Synchronization reference switching (I)

This test checks the switching from the primary reference source (in this example,the SDH generator) to a secondary reference, which in this case will be the internalclock of the NE (free running).

f1 : (f1nom+50ppm, f1nom-50ppm) then 2048000 ± 102 Hz

f2 : (f2nom+4,6ppm, f2nom-4,6ppm) then 2048000 ± 9 Hz

f0 : (f0nom+4,6ppm, f0nom-4,6ppm) then 2048000 ± 9 Hz

Figure 10.26 Free-running test.


ADM

f1 f2

SEC

WestSTM-1

EastSTM-1

f0


The ADM takes its synchronization from the West aggregate signal (STM-1 W)that the generator of the tester provides as a primary reference, disabling its synchro-nization input for a secondary reference if the first one fails. The internal clock ofthe ADM will then take on the role of secondary reference. Likewise, the primarysynchronization output will be enabled to follow the primary synchronization input.

The test is performed by disconnecting the West aggregate input. The ADMthen switches over to free-running mode, that is, secondary synchronization. Thetributary analyzer must not register anomalies or defects, and the tributary frequencymust remain constant.

As in the previous case, the capacity of the testers is used to measure the fre-quency of the tributary signal analyzed, checking that it remains unchanged duringthe switchover from one reference source to another.

One way of performing the test could be the following (see Figure 10.27):

1. Offset the frequency of the tributary signal generated, f1, to a value within thetolerance margin, f1nom+50 ppm, f1nom-50 ppm, and check that the frequencyof the tributary signal recovered on extraction coincides with the value withinthe margin, 2,048,000±102 Hz. This checks that the tributary has been recov-ered correctly. For the rest of the test, program the frequency:

f1 = f1nom+25 ppm.2. Vary the frequency fSTM1 a few ppm and see that the frequency of the ADM

synchronization output, f0, varies by the same amount. You must also check

WestSTM-1

EastSTM-1


SEC clock

f0

f1 f2

Synchronizationoutput

(secondary reference - free running)

Primary reference

fSTM1

Figure 10.27 Synchronization reference switching.


that the frequency f2 is not affected by these variations.3. Program the frequency of the West aggregate to the value of fSTM1nom+5ppm.

The frequency meter applied to the synchronization output should measureapproximately 2,048,010 Hz, that is, 2,048,000+5 ppm, the synchronizationoutput moves the same amount of ppm as the synchronization reference.3

4. Disconnect the West aggregate input. This way, the ADM loses its primarysynchronization source and, because it does not have a secondary synchroniza-tion source programmed, it enters into free-running mode. In this mode:

• The 2-Mbps analyzer must not register anomalies or defects;

• The f2 frequency of the tributary analyzed must not vary only because it hasswitched from one synchronization reference to another;

• f0 takes the value of free running frequency;

• The corresponding defects appear in the user interface (UI) of the ADM con-trol computer.

5. Connect the aggregate input to the SDH generator again. The following shouldbe observed:

• The defect indications disappear from the UI of the control computer;

• The synchronization reference switching from the internal clock of the ADMto the aggregate signal does not affect the value of the f2 frequency;

• No anomalies or defects are registered in the 2-Mbps signal analyzed;

• The ADM synchronization output signal takes on the frequency value of2,048,010 Hz again, and follows the frequency variations of the West aggre-gate signal.

10.3.2.3 Synchronization reference switching (II)

This test checks the switching from the primary reference source to a secondary ref-erence; both of these being aggregate signals (see Figure 10.28). In this example,the NE takes its synchronization from the East aggregate signal (STM-1 E) as itsprimary reference, and from the West aggregate (STM-1 W) as secondary refer-ence, should the first fail. The internal clock of the element will therefore take onthe role of a tertiary reference, coming into action (free running) if the other twofail.

3. The difference between this value and the frequency value obtained for free runningin the previous test (2,048,000±9 Hz). This makes it easy to distinguish between theWest aggregate acting as a synchronization reference and operation in free-runningmode.


Likewise, the primary synchronization output will be enabled to follow the pri-mary synchronization input, and the secondary synchronization output will followthe secondary synchronization input. To carry out the test, it will be checked that thesynchronization output follows the frequency variations of the primary source(East).

The primary source will also be disconnected to provoke the switchover to theWest reference, checking that the synchronization output now follows the variationsof this aggregate. Finally, the West aggregate will be disconnected so that the ADMenters into free-running mode, checking that an AIS is received in the tributary an-alyzed at the free-running frequency (which has been previously measured).

As in the above measurements, the capacity is used to measure the frequencyover the signals analyzed, checking their expected values at each stage of the mea-surement, in line with the operative reference sources.

Procedure for carrying out the reference switching

1. The primary synchronization reference is supplied by the East side tester.Check that the primary synchronization output (f0) of the ADM follows the fre-quency variations of the East aggregate. Next, introduce an offset of +5 ppm:fSTM1E=fSTM1nom+5 ppm (155,520,000 Hz + 778 Hz = 1,555,520,778 Hz).Check also that the frequency f2 varies with the frequency of the TU 2.4.1 trib-utary by changing the offset of this tributary. Set this offset at +30 ppm:f2 = f2nom + 30 ppm (2,048,000 Hz + 61 Hz= 2,048,061 Hz). This part of the test checks that the synchronization is correctly extracted fromthe East aggregate.

2. When the East aggregate input is disconnected, the following must beobserved:

West East

Figure 10.28 Synchronization reference switching (II).

f0

f1 f2

Synchronization outputSecondary reference

fSTM1O

Primary reference

Tx=Rx and autoconf.offset: +5 ppmoffset 2M: +30 ppm

Tx=Rx and autoconf.offset: -5 ppmoffset 2M: -30 ppm

fSTM1E

PRBS15 in tributary

PRBS15in tributary

TU 3.2.1 TU 4.2.1


• No anomalies have occurred in the TU 3.2.1 received in the West side;

• The 2-Mbps analyzer receives an AIS. Note its frequency value (f2);

• The primary synchronization output frequency (f0) has changed to the value as-sociated with the secondary reference (West signal with offset of -5 ppm), thatis, 2,047,990 Hz;

• The control computer displays the corresponding synchronization defects.

This stage of the test checks that the switching between primary and secondary syn-chronization sources is correct.

3. The analyzer section of the 2-Mbps device should now be connected to the out-put of tributary port. Check that:

• The synchronization output frequency of the ADM (f0) follows the frequencyoffsets of the West aggregate (fSTM1O=fSTM1nom + offset), and then set the val-ue at fSTM1O= fSTM1nom - 5 ppm.

• The frequency of the signal from the tributary (f2) follows the variations intro-duced from the West side SDH generator in the TU 3.2.1. Set this offset at -30 ppm:f2 = f2nom -30 ppm (2,048,000 Hz - 61 Hz = 2,047,939 Hz).

4. When the West aggregate input is disconnected, the following must beobserved:

• No anomalies are detected in TU 3.2.1;

• The 2-Mbps device receives an AIS signal. Check that the value of f2 coincideswith the value noted in step 2;

• f0 changes to the free-running frequency value measured in step 2 of the test onthe ADM in free-running mode;

• The control computer displays the corresponding synchronization defects.5. Connect the West aggregate input again and check that:

• The ADM switches its reference to the STM-1 West signal (secondary). To doso, measure f2 = f2nom -30 ppm (2,048,000 Hz - 61 Hz = 2,047,939 Hz) and f0 = f0nom -5 ppm (2,047,990 Hz) to check that this switching has taken placecorrectly;

• The control computer displays the corresponding synchronization defects.6. Connect the East aggregate input again and connect the output of tributary port

to the 2-Mbps analyzer. Check that:

• The ADM switches to its primary reference by measuring f2 = f2nom + 30 ppm(2,048,000 Hz + 61 Hz= 2,048,061 Hz) and f0 = f0nom +5 ppm (2,048,010 Hz);

• The control computer no longer shows any synchronization defects.



• ITU-T Rec. G.742, Second order digital multiplex equipment operating at 8 448 kbit/s and using pos-itive justification.

• ITU-T Rec. G.751, Digital multiplex equipments operating at the third order bit rate of 34 368 kbit/sand the fourth order bit rate of 139 264 kbit/s and using positive justification.


• ITU-T Rec. G.798, Characteristics of optical transport network (OTN) hierarchy equipment func-tional blocks.


• ITU-T Rec. G.810, Definitions and terminology for synchronization networks.

• ITU-T Rec. G.811, Timing characteristics of primary reference clocks.

• ITU-T Rec. G.812, Timing requirements of slave clocks suitable for use as node clocks in synchroni-zation networks.




• ITU-T Rec. G.825, The control of jitter and wander within digital networks which are based on thesynchronous digital hierarchy (SDH).

• ITU-T Rec. G.958, Digital line systems based on the Synchronous Digital Hierarchy for use on opti-cal fibre cables. (Its content has been absorbed by ITU-T Rec. G.783 and Corrigendum 1 to ITU-TRec. G.798.)

• ITU-T Rec. G.8251, The control of jitter and wander within the optical transport network (OTN).

• ITU-T Rec. O.171, Timing jitter and wander measuring equipment for digital systems which arebased on the plesiochronous digital hierarchy (PDH).

• ITU-T Rec. O.172, Jitter and wander measuring equipment for digital systems which are based on thesynchronous digital hierarchy (SDH).

• ETSI ETS 300 147, Transmission and Multiplexing (TM); Synchronous Digital Hierarchy (SDH);Multiplexing structure.

• ETSI ETS 300 462, Transmission and Multiplexing (TM); Generic requirements for synchronizationnetworks; Part 1: Definitions and terminology for synchronization networks. Part 2: Synchronizationnetwork architecture. Part 3: The control of jitter and wander within synchronization networks. Part4: Timing characteristics of slave clocks suitable for synchronization supply to Synchronous DigitalHierarchy (SDH) and Plesiochronous Digital Hierarchy (PDH) equipment; Implementation Conform-ance Statement (ICS) proforma specification. Part 5: Timing characteristics of slave clocks suitablefor operation in Synchronous Digital Hierarchy (SDH) equipments; Part 6: Timing characteristics ofprimary reference clocks; Implementation Conformance Statement (ICS) proforma specification.


• ETSI ETS 302 084, Transmission and Multiplexing (TM); The control of jitter and wander in trans-port networks.

• ETSI ETS 417-6, Transmission and Multiplexing (TM); Generic requirements of transport function-ality of equipment; Synchronization layer functions.

• Telcordia GR-499, Transport Systems Generic Requirements (TSGR): Common Requirements, Dec.1998.

• Telcordia GR-253, Synchronous Optical Network (SONET) Transport Systems: Common GenericCriteria, Sept. 2000.

• Telcordia GR-1244, Clocks for the Synchronized Network: Common Generic Criteria, Dec. 2000.

• ANSI T1.101-1999, Synchronization Interface Standard.

• ANSI T1.102-1993, Digital Hierarchy Electrical Interfaces.

• ANSI T1.105.03-1994, Synchronous Optical Network (SONET) Jitter at Network Interfaces.

• ANSI T1.403-1999, Network and Customer Installation Interfaces-DS1 Electrical Interface.

• ANSI T1.404-1994, Network-to-Customer Installation-DS3 Metallic Interface Specification.

• ANSI T1.105.09-1996, Synchronous Optical Network (SONET) Network Element Timing and Syn-chronization.

• Stefano Bregni, “A Historical Perspective on Telecommunications Network Synchronization.” IEEECommunications Magazine, June 1998.

• T. S. Brown, M. J. Gilson, M.G. Mason, “Synchronisation in Data Networks.” BT Technol J, Vol. 16No. 1, January 1998.

403

Appendix A

Error Detection and Correction Techniques

To improve the features of a communications system, some channel coding tech-niques are used. These techniques serve to protect the information of transmissionerrors. All those codification schemes are based on the addition of some redun-dance in the information to be transmitted. That means that the bandwith availableto transmit user information will decrease. The compensation is an improved im-munity to errors.

A.1 CYCLIC REDUNDANCY CHECK

The cyclic codes are a special class of channel codes. This family of codes is one ofthe most used in practical applications because some very efficient coding and de-coding algorithms can be used. The CRC and also RS and BCH codes that are dis-cussed in the next section are special cases of cyclic codes.

The mechanism of the cyclic redundance check is commonly used for error detec-tion (not correction). The CRC adds parity bits to a message of bits.Thanks to its versatility, it is an omnipresent technique in all kinds of digital tele-communications systems.

The CRC mechanism is based on the computation of the remainder of the divi-sion between a polynomial related to the message, and a special standard polynomialthat has certain features of message protection. More precisely, for encoding, a poly-nomial, , with binary coefficients and degree , is chosen. The message, ,can be understood as a finite sequence of bits:

But it can also be defined like a polynomial with binary coefficients:

k n 1+

g X( ) k m

m mn mn 1– … m0, , ,( )=


For the encoding, a third polynomial, , is calculated in the following way:

Given that is the remainder of a quotient with , the following condi-tion will be the case:

The polynomial , is understood as a parity check polynomial. This polyno-mial is added to the original message in such way that the bit sequence that is finallytransmitted is:

The polynomial is called a generator polynomial because all the codewords [all that can be written as ] are its multiples. In other words,all the code words are divisible by . It is precisely this fact that allows error rec-ognition. If the received polynomial is not divisible by it can be concluded thata transmission error has occurred.

A.2 RS AND BCH CODES

The RS codes were invented by Reed and Solomon in 1960. They were used inNASA missions to Mars, and today can be found in many applications: CD, DVD,digital TV and wireless communications. Their main feature is that they allow veryefficient correction of transmission errors in the receiving end of a digital comuni-cation (forward error correction).

The RS codes are cyclic codes like the CRC codes. This is why the messagewords and the code words can be seen as polynomials. Coding is done the same wayas in CRC:

m X( ) m0 m1X … mnXn+ + +=

r X( )

r X( ) m X( )Xkmodg X( )=

r X( ) g X( )

deg r X( ) deg g X( ) < k=

r X( )

c X( ) m X( )Xk r X( )+=

g X( )

g X( )g X( )

r X( ) m X( )Xkmodg X( )=

m X( )Xk r X( )+

Appendix A 405

The notation used is similar to that of CRC. is the check polynomial, is the message, the generator polynomial, and is the number of re-

dundancy symbols. The level of protection given to the message depends on the val-ue of . As obvious, the more redundancy bits there are, the more possibilitiesthere are for error correction, and the message has more protection.

In the case of the RS codes, the coefficients of the polynomials do not need tobe binary. Generally, they belong to a Galois field .

A Galois field can be thought of as field whose elements are polyno-mials and the algebraic operations are modular. In other words, the result of a sumor a product is performed modulo a given polynomial, . This is the reason be-cause to calculate the sum and the product within a field, the following ex-pressions will be used:.

The polynomial must be irreducible to be sure that all the elements butthe 0 have multiplicative inverse and therefore the set configures a field.

The basic characteristic that distinguishes the RS codes inside the group of cy-clic codes is the expression of their generator polynomial, generalized easily for dif-ferent values of :

The element is a primitive element of . In other words, the succes-sive powers of must generate the elements of , except for 0.

The particular expression of the generator polynomial serves to assure that theminimum distance of the code, (parameter directly related with the detecting andcorrecting features of the code) will accomplish:

r X( )m X( ) g X( ) k

k

GF pn( )

GF pn( )

q X( )GF pn( )

s X( ) p1 X( ) p2 X( )+=

t X( ) p1 X( ) p2 X( )⋅=

mod q X( )

mod q X( )

q X( )

k

g X( ) X α i+( )i 0=

k 1–

∏=

α GF pn( )α pn GF pn( )

d

d k 1+≥


BCH codes, are named after Bose, Chandhuri, and Hocquenghem, are also ex-tensively used. Coding and decoding is based in polynomials like the RS case.

The difference between RS and BCH codes is the expression of the generatorpolynomial. Both RS and BCH codes are constructed over a field but BCHcodes have the extra requirement that the coefficients of the generator polynomial(and the coefficients of the message and check polynomials) must be in .The generator polynomial of BCH codes contains a sequence of roots thatare consecutive powers of a minimal element of . This allows us to assurethat the minimum distance of the code will be but in order to get a generator poly-nomial with elements in , it must be completed with other different roots.This is similar to a polynomial with real coefficients. Such a polynomial may containcomplex roots, but if it contains a particular complex root, it also must contain itscomplex congugate to get a polynomial with real coefficients.

RS and BCH codes are good for finding and correcting errors if the error haslow rate poisson statistics but they fail against bursty errors. The error correction ca-pabilities of RS and BCH codes are exceeded if the number of errors in a block isabove the minimum distance of the code. To improve the protection features of RSand BCH codes interleaving techniques are commonly used. In this way, the bits ofa single block are spread and the effect of an error burst is distributed over manyblocks.

A.3 BIT INTERLEAVED PARITY

The bit interleaved parity (BIP) is a simple algorithm that provides with error de-tection capabilities a digital communication. The mechanism for generating BIPwords is made up of two phases. The first of these involves dividing the portion ofsignal into blocks. The second phase involves laying out these blocks in a tablemade up of n columns. Each bit of the code word corresponds to one column/blockand has a value of “1” if the number of ones in the column is odd and “0” if it iseven, thus defining a method of checking even parity (see Figure A.1).

A discrepancy between the value received for the ith block monitoring bit andthe value calculated on reception implies that one or more bit errors have occurredinside this block, (i.e. the block is considered as a block error).

Note that BIPs cannot correct errors because it is not possible to find out whichis the errored bit in the block. In addition, an even number of errors in the same blockcannot be detected.

GF pn( )

GF p( )δ 1–

GF pn( )δ

GF p( )

Appendix A 407

By construction, the bits of the same block are not consecutive. This is the rea-son for the strength of BIP codes against a bursty error distribution.

As a conclusion, it could be said that BIP has opposite characteristics of CRC,RS and BCH codes. BIP has poor detection capabilities but is stronger against errorbursts if no interleaving is used for CRC, RS, and BCH.

Figure A.1 Bit interleaved parity (BIP-n) enables error monitoring. A transmitter performs the XOR function (even parity) over the previous block. The value computed is placed in the n bits before the block is scrambled.

1 0 0 1 1 1 0 10 0 0 1 1 0 1 11 0 1 1 0 1 1 11 0 0 1 0 1 0 0

10100101

XOR

= BIP-n

110010100110

XOR table

++

1 2 3 ... ... ... ... ndata

409

Appendix B

Masks for Copper Qualification

B.1 ANSI MASKS

Table B.1 Voice service.

Frequency (Hz) 200 300 400 600 1,020 2,000 2,800 3,000 3,400

Noise (dBm max) -84.80 -84.80 -84.80 -84.80 -84.80 -84.80 -84.80 -84.80 -84.80Return loss (dB max) - -6.00 -6.50 -7.00 -9.00 -9.00 -9.00 -9.00 -Ins. loss (dB min) - -10.00 -11.50 -14.00 -20.00 -23.50 -24.50 - -LCL (dB max) -45.00 -45.00 -45.00 -45.00 -45.00 -45.00 -45.00 -45.00 -45.00NEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00FEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00

Table B.2 Modem 56 Mbps.

Frequency (Hz) 200 300 400 600 1,020 2,000 2,800 3,000 3,400


Table B.3 ISDN.

Frequency (kHz) 1 5 10 20 30 40 45 50 50

Noise (dBm max) -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 - -Return loss (dB max) -2.40 -5.00 -8.20 -12.00 -15.00 -15.00 -15.00 - -Ins. loss (dB min) -24.90 -25.20 -26.00 -29.20 -32.00 -33.00 -34.00 - -LCL (dB max) -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 - -NEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 - -FEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 - -


Table B.4 HDSL One pair.

Frequency (kHz) 5 10 20 40 100 150 200 300 402

Noise (dBm max) -70.80 -70.80 -70.80 -70.80 -70.80 -70.80 -70.80 -70.80 -Return loss (dB max) -4.50 -6.50 -9.00 -13.50 -14.00 -14.00 -14.00 -13.50 -Ins. loss (dB min) -13.40 -14.00 -17.00 -20.60 -22.00 -24.00 -27.00 -32.00 -LCL (dB max) -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -NEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -FEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -

Table B.5 HDSL Two pairs.

Frequency (kHz) 5 10 20 40 100 150 200 245 245


Table B.6 T1.

Frequency (kHz) 102 200 300 400 500 750 900 1200 1400

Noise (dBm max) -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 -68.80Return loss (dB max) -12.00 -13.50 -13.80 -14.00 -14.50 -14.50 -14.50 -14.50 -14.50Ins. loss (dB min) -12.00 -14.00 -16.00 -19.60 -22.00 -24.20 -28.20 -30.60 -32.30LCL (dB max) -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00NEXT (dB max) -60.00 -60.00 -60.00 -60.00 -60.00 -60.00 -60.00 -60.00 -60.00FEXT (dB max) -60.00 -60.00 -60.00 -60.00 -60.00 -60.00 -60.00 -60.00 -60.00

Table B.7 ADSL G.lite.

Frequency (kHz) 20 40 100 150 200 300 400 450 500


Appendix B 411

Table B.8 ADSL 2 Mbps.

Frequency (kHz) 20 40 100 150 200 300 400 600 800



Frequency (kHz) 20 40 100 150 200 300 400 600 800


Table B.10 SDSL.

Frequency (kHz) 10 18 30 56 100 150 240 300 420


Table B.11 SHDSL.

Frequency (kHz) 10 18 30 56 100 150 240 300 420



B.2 ETSI MASKS

Table B.12 Voice service.

Frequency (Hz) 200 300 400 600 1,020 2,000 2,800 3,000 3,400


Table B.13 Modem 56 Kbps.

Frequency (Hz) 200 300 400 600 1,020 2,000 2,800 3,000 3,400


Table B.14 ISDN.

Frequency (kHz) 1 5 10 20 30 40 45 50 50


Table B.15 HDSL One pair.

Frequency (kHz) 5 10 20 40 100 150 200 300 402


Appendix B 413

Table B.16 HDSL Two pairs.

Frequency (kHz) 5 10 20 40 100 150 200 245 245

Noise (dBm max) -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 -68.80 -Return loss (dB max) -4.50 -6.50 -9.00 -13.50 -14.00 -14.00 -14.00 -14.00 -Ins. loss (dB min) -16.40 -17.00 -20.70 -25.20 -27.00 -29.50 -31.50 -34.20 -LCL (dB max) -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -40.00 -NEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -FEXT (dB max) -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -65.00 -

Table B.17 E1.

Frequency (kHz) 102 200 400 600 800 1024 1250 1500 1750


Table B.18 ADSL G.lite.

Frequency (kHz) 20 40 100 150 200 300 400 450 500



Frequency (kHz) 20 40 100 150 200 300 400 600 800




Frequency (kHz) 20 40 100 150 200 300 400 600 800


Table B.21 SDSL.

Frequency (kHz) 10 18 30 56 100 150 240 300 420


Table B.22 SHDSL.

Frequency (kHz) 10 18 30 56 100 150 240 300 420


415

Appendix C

Two-Wire Transmission Line Model

The two-wire line is known to support transverse electromagnetic mode (TEM)waves. In other words, it creates a perpendicular electromagnetic field to the propa-gation direction. Studying TEM structures is simpler than studying the generalcase, as we can define currents ( ) and voltages ( ) in a plane perpendicular tothe propagation axis (see Figure C.1). This is not possible in transmission structuresthat do not support TEM modes, such as wave and fiber guides.

The circuit model used to describe transmission lines is called the distributedparameter model. It is very suitable for describing electromagnetic phenomena inhigh-frequency lines. The basic magnitudes used are voltage and current, which area function of axial coordinates. These voltages and currents are described as and . This model may be considered as a generalization of the low-frequencymodel called the concentrated parameter model.

I V

I

IV

Figure C.1 In two-wire transmission lines, we can define transversal current and voltage.

Plane perpendicular to the propagation axis

Propagation direction Z

V z( )I z( )


C.1 CHARACTERISTIC PARAMETERS OF THE LINE

For a small longitudinal fragment, with respect to wavelength, we can use a con-centrated parameter circuit model (see Figure C.2). This model includes a serial re-sistance, , that represents the ohmic loss in conductors, due to the finiteconductance that these have. The parallel conductance, , also represents ohmicloss, in this case in the dielectric medium that separates the two conductors. In cop-per pairs, dielectric loss is much less than loss in the conductor. The line parameters

and are responsible for attenuating the signal progressively as it crosses theline.

A transmission line is always formed by two separate conductors, by a dielectricmedium which in itself forms a capacitor. The parameter represents the capacityof this capacitor. The line also has a self-inductance, , associated with it. This isdue to the electromagnetic field that appears around it when current flows across it.Both and are in charge of transmitting signals across the line as waveforms ofvoltage and current.

Since , , , and depend on the length, they are measured in ohm/meter(Ω/m), farad/meter (F/m), and henry/meter (H/m).

This model is valid for nonbalanced lines, that is, for those where one of the con-ductors has high impedance with earth/ground, and the other one, low impedance. Ina balanced line (where the two conductors have the same impedance with earth/ground), it would be necessary to include the parameters of each connector into themodel.

If the line is short enough for ohmic losses to be ignored,1 we can use a closedexpression to describe the voltage and current:

1. Note that this is not normally the case for local loop applications.

RG

R G

C

L

G

R

∆

I(z) I(z + dz)

V(z) V(z + dz)

Figure C.2 Concentrated parameter circuit model for a transmission line.

z

∆z ∆z

∆z ∆z

CL

L C

R G L C

Appendix C 417

These represent arbitrary waves that move in the two possible directions withina transmission line (subindex p to the positive direction, n to the negative direction).The constant is the propagation rate, and it has the following value:

If the line has no losses, the rate depends only on the physical characteristics ofthe line, and not on the signal that is propagated. This is an important parameter, asit is related to signal latency and propagation time.

The mathematical analysis of the transmission line model also shows that volt-age and current are both related:

This means that in all points of the line, voltage and current are related by meansof a simple quotient. The factor that relates these magnitudes is transmission line ref-erence impedance ( ):

The reference impedance depends only on the physical configuration of the line,and not on the nature of the transmitted signal. The capacity of a line to absorb anelectrical signal without it being reflected again at its origin depends on the imped-ance.

Analyzing the response of a line to a purely sinusoidal signal is important, asthis enables us to deduce its frequency behavior. When it comes to sinusoidal stim-uli, we can find simple expressions of signals on a line, including when there areohmic losses that cannot be ignored. In this situation, reference impedance is nolonger real, and a reactive part appears that causes phase offset between the line cur-rent and voltage (see Figure C.3). The following mathematical expression describes

V z t,( ) Vp z ct–( ) Vn z ct+( )+=

I z t,( ) I= p z ct–( ) In z ct+( )+

c

c 1LC

-----------=

Vp Vn–Ip In+---------------- Z0=

Z0

Z0LC----=


the real (resistive) and the imaginary (reactive) part of line impedance when thereare losses that cannot be ignored:

However, the reactive part is very often negligible, and it is quite usual to referto reference impedance by means of a real number, although this is only true in a linewhere there is no loss.

Figure C.3 Resistance and reactance as frequency functions. This is line with losses that can-not be ignored.

Reactance (Ω)

Resistance (Ω)

Hz

390

350

270

310

230

190

150

110101 102 103 104 105

Hz101 102 103 104 105

10

-30

-110

-70

-150

-190

-230

-270

-310

-350

Re Z0( ) LC---- 1

2--- 1 1 R

ωL-------

2++

=

Im Z0( ) LC----– 1

2--- 1– 1 R

ωL-------

2++

=

Appendix C 419

Furthermore, when there are losses, the reference impedance is no longer con-stant, and it becomes a decreasing function of the frequency. This is why, for exam-ple, in voice services the standard reference impedance is 600Ω, and lower referenceimpedances, for example 120Ω, are used for broadband services.

421

About the Authors

José M. Caballero is the international marketing manager for Trend Communica-tions Ltd., High Wycombe, Bucks, England. He has an M.B.A., and holds a B.S. and an M.A. in computer communications from the Universitat Politécnica de Cat-alunya, Barcelona, Spain. From 1984 to 1994, he worked for IBM in Charlotte, North Carolina, and in Barcelona, Spain, as an SNA/SDLC developer. From 1993 to 1997, he was a computer networks professor at the Universitat Politécnica de Catalunya in Barcelona.

Francisco Hens is a product manager with Trend Communications in Barcelona, and holds a B.Eng. and an M.A. in telecommunications from the Universitat Politécnica de Catalunya, Barcelona, Spain.

Roger Segura is a product manager with Trend Communications in Barcelona, Spain, and holds a B.Eng. and an M.A. in telecommunications from the Universitat Politécnica de Catalunya, Barcelona, Spain.

Andreu Guimerá is a marketing manager with Trend Communications in Barce-lona, Spain, and holds a degree in marketing from Camara de Comerç de Barce-lona, as well as a B.Eng. and an M.A. in telecommunications from the Universitat Politécnica de Catalunya, Barcelona, Spain.

423

Index

AAAL 115

AAL1 115AAL2 118AAL3/4 120AAL5 121CS 108SAR 108types of 108, 115See also ATM

AAU 102A-bit 23–24, 28Acceptance tests 268, 393Access network, see NetworkAccess point identifier (API) 274Access technologies, comparison of 156Adaptive clock 116Add and drop multiplexer (ADM) 41, 199,

207, 268tributary continuity test 269See also NE

ADM 272Administrative unit group (AUG) 49Administrative unit (AU) 45ADSL 139, 141, 151, 155–156, 158–159,

166, 170, 177, 403, 411, 415analyzing the channel 180ATM 139–140, 142, 318, 322coding 165error protection 166frame structure 161framing 161handshake 176–178initialization 176interleavingmodulation 169MPEG 144origin of 151overhead control 163rates 156scrambling 168superframe 163, 166

training 179transceiver 158use of 157VoDSL 144 See also ATM, ATU-C, ATU-R, G.lite,

STMADSL 2 Mbps 411ADSL control channel (aoc) 320Aggregate signal, see Signal and East aggre-

gateAging margin 266AIS 19, 22, 29, 84, 235

AU-AIS 85MS-AIS 85PDH-AIS 85TU-AIS 85

Alarm 76Alarm indication signal (AIS), see AISAlarms

PDH 28SDH 78

Aligning 45, 53Alternate mark inversion (AMI), see Line

codeAnalog signal, see SignalAnomaly 76, 261

parity 227ANSI T1.101 383, 385, 388–389APS 87, 244–246, 270

meausrements 244Asymmetric digital subscriber line (ADSL),

see ADSLAsymmetry 155Asynchronous transfer mode (ATM), see

ATMATM 99–100, 140, 158–159, 294, 329

AAL 107access networks 139ADSL 139–140, 142, 318, 322CAC 129–131cell

definition 101

Index424

format 101tagging 128transfer outcomes 278

commissioning 315IS measurement 277layers 106, 109, 111, 113measurement cycle 305network architecture 106NPC 130OOS measurements 277, 283, 300, 321operating mode 104performance parameters 277physical layer 113policing functions 132QoS 122, 129, 131RM 130routing 131service categories 125–126signaling 131switching 104–105, 158test traffic 320UPC 129, 133virtual channel 111virtual path 112voice 144 See also OAM functions

ATM adaptation layer, see AAL 106–107ATM-user-to-ATM-user (AAU), see AAUAttenuation 3

integrateddefinition 175

Attenuatoroptical 214, 270

ATU-C 140, 157, 161, 163, 165–166, 174–176, 179–180, 182–183, 319

See also BufferATU-R 140, 157–158, 161, 165–166, 174–

176, 179–180, 182–183, 319, 322Automatic gain control (AGC) 180Automatic protection switching (APS), see

APSAvailable time 253, 257, 268, 283Available time block error (ABE)

definition 258Available time errored bits (AE)

definition 253

BBackground block error ratio (BBER) 258

definition 258Background block error (BBE) 256–258Background errored bits (BE)

definition 253Backward performance monitoring

(BPM) 296Bandwidth, see ChannelBER 7, 17, 19, 82, 214, 217, 228, 252, 301

CBER 316definition 217measurements 211penalty 375

Bidirectional path 257Binary DPSK 177BIS

definition 208performance limits for 266

B-ISDN 35, 114, 284Bit

AAU 102aoc 164sc 165

Bit eight-zero suppression (B8ZS), see Line code

Bit error 16, 194, 214, 369Bit error rate test (BERT) 233, 239Bit error rate (BER), see BERBit interleaved parity (BIP) 80, 227, 406Bit rate 125Bit stuffing, see StuffingBlock, definition 256Bose-Chaudhuri-Hocquengheim (BCH) 82BRI interface 152Bringing into service (BIS), see BIS 261Broadband-ISDN, see B-ISDNBuffer

fast 160, 162–163, 168, 180interleaved 160, 163, 168, 180

Building integrated timing supply (BITS) 189, 199

CCAP modulation 154, 156Carrier to interference signal ratio (C/I) 318CAS 20

LOM 22multiframe alignment signal 21nonmultiframe alignment signal 21signaling multiframe 20

C-bit 32channel signal codes 32parity 31–32parity framing 32

CCS, see SignalingCDV 116, 136, 281

Index 425

peak-to-peak 124tolerance 123

Celldefinition 101entry event 278exit event 278PM 296resource management (ATM) 138tagging 128transfer outcomes, see ATMSee also Payload

Cell bit error rate (CBER) 316Cell delay variation (CDV), see CDVCell error ratio (CER) 316

definition 279Cell loss priority (CLP) 102Cell loss ratio (CLR) 124

definition 279Cell misinsertion rate (CMR) 280Cell transfer delay (CTD) 281Channel

2 Kbps 214 Kbps 3056 Kbps 3164 Kbps 14, 29bandwidth 5, 310

definition 7digital

monitoring of 228DS0 30E1 15embedded 94full channel measurement 211half-channel measurement 211

Channel associated signaling, see CASChannel coding 7Ci-bit 32Circuit 7, 21

demultiplexing 26signaling circuits 20

CLEC 331Clock

frequency 213master 185, 187, 199output signal 26performance of 187, 197PLL 382, 384PRC 198–199, 382PRS 198–199recovery 195, 239reference 382sampling 194, 369

SEC 198–199, 387–388, 395signal 194, 369slave 186transfer

SONET/SDH 190T-carrierr/PDH 189

Coded mark inverted (CMI), see Line codeCode-division multiplexing access (CD-

MA), see MultiplexingCoding

Trellis 181–182, 319Combined jitter, see JitterCommon channel signaling (CCS), see Sig-

nalingCommon management information protocol

(CMIP) 94Communications network, definition 1Competitive local exchange carriers,

(CLEC) 331Concatenation

contiguous 72definition 71virtual 52, 71–74, 86, 96

Concentrated parameter model 415Connection admission control (CAC), see

ATMConstellation coder 170Continuity check (CC) 294, 313Continuous signal, see SignalCopper pair 334Coupling 348

electrical 348magnetic 349

C-PILOT 179CRC 16, 18, 30, 161, 163–164, 166, 184,

261, 328, 403, 405advantages of 17coder, operation of 167multiframe alignment 17verification of 227

Crosstalk 6, 153–154, 211, 348Cyclic prefix 174Cyclic Redundancy Check (CRC), see CRC

DDamage 76Data block 227Data communications channel (DCC) 234,

270, 274Data rate, definition 7Defect 76

definition 83

Index426

supervision of 274Degraded minutes (DM)

definition 253Delineation algorithm

HUNT 159PRESYNC 160SYNC 160

Demultiplexer, see NEDemultiplexing

devices 19synchronous 242

Desynchronization 379Device under test (DUT), see DUTDigital analog converter (DAC) 161, 173Digital cross-connect 268, 393Digital cross-connect (DXC) 41, 199, 224Digital distribution frames (DDF) 229Digital signal, see SignalDigital video broadcasting (DVB) 52Discrete multitone modulation (DMT) 156Discrete signal, see SignalDistortion 3–4Distributed parameter model 415Diversification

definition 244DMT 163, 169–172, 180Doppler effect 242DS1-DS3, see FrameDSL 155, 157

ADSL 151G.SHDSL 155HDSLorigin of 151SDSL 155VDSL 157

DSL access multiplexer (DSLAM) 140DUT 213, 217, 233–234, 239Dying gasp message 174–175Dynamic range

definition 213

EE1, E2, E3, E4 26East aggregate 214, 398–400

See also West and West-East aggregateE-bit 18Embedded operations channel (EOC), see

EOCEnd of data (EOD) 176End-to-end transparency 220Enhanced remote defect indication (E-

RDI) 85

Entity under test (EUT) 220EOC 174

ADDR 175AMF 175A/O 175commands 175message

fields of 174format of 174

OPCODE/DATA 175P 175

Error 17, 27–28, 168block errors 18digital transmission 6monitoring 17–18

definition 80onset of 375PDH 260PDH events 22, 29See also Bit error

Error detection code (EDC) 256, 260Errored block (EB) 252, 256

definition 256Errored second ratio (ESR), see ESRErrored second (ES) 252

definition 256ES 256–257, 262–263

definition 252ESR 253, 255

definition 257Ethernet 52, 72, 96, 325

GbE 64, 95LAPS 52SDH 52

Events 78, 261block-based 256cell-based

entry 278exit 278

monitoring of 78, 227PDH 22, 29

Extended superframe (ESF) 30–31

FFailure 76Far-end alarm and control (FEAC) 32Far-end block error (FEBE) 32Far-end crosstalk (FEXT) 350

measurement of 365FAS 16–18, 22–24, 28–29, 225, 235, 260–

261detection of anomalies 226

Index 427

Fast buffer, see BufferFast phase fluctuation, see JitterFast synchronization byte, see Synchroniza-

tionFault management 292F-bit 30FDM 10, 154–156, 180FDMA 10FEC 82, 160–163, 166, 168FEXT 350–351

measurement 365Forward performance monitoring

(FPM) 296Fourier, see Inverse Fourier transformFrame 114

alignment 15alignment signal (FAS), see FASDS1 29, 31–32, 152–153

extended superframe 30superframe 30

DS2 32DS3 32E1 15, 19, 21, 152–153, 201E2 26E3 26E4 26PPP 52SDH/SONET 114STM 54–55, 57, 69–70STS 54T1 201T-carrier/PDH 114 See also Signal

Frame bit, see F-bitFrame check sequence (FCS) 177Framing 161, 213

full 161–162Frequency

measuring 215Nyquist 173, 179

Frequency division multiple access (FDMA) 146

Frequency division multiplexing (FDM), see FDM

Frequency offset 192, 239, 242, 393Frontier stations (FS) 264Full channel 240Full channel measurement 211

GGalois field 166, 405Generic cell rate algorithm (GCRA) 135

Generic flow control (GFC) 102Generic framing procedure (GFP) 52Global positioning system (GPS) 191Golden modem 319G.114 243G.691 215G.703 229, 239, 261G.704 19–20G.7041 52G.7042 52G.707 45, 52, 218G.711 211G.722 229G.741 23G.742 378G.751 24, 378G.783 235–236, 246, 274, 381, 394G.784 274G.803 236G.810 196, 383, 385G.813 378, 395G.821 251–252, 256, 261, 266G.822 261G.823 371, 374–375, 377–378, 387–388G.825 372, 374–375, 388G.826 227, 251–252, 256, 261, 264, 266G.841 246G.957 215G.958 374–375G.992 176G.992.1 173, 177, 183, 360G.992.2 156G.994.1 177G.lite 157

HHalf-channel measurement 211Handshake 178

ADSL 176–178definition 177

HDSL 152, 154–156Header error control (HEC) 103, 159Header, definition 101High bit rate digital subscriber line (HDSL),

see HDSLHigh grade 255High impedance 209High-density bipolar three zeroes (HDB3),

see Line codeHigher-order path connection (HPC) 219Higher-order path (HP) 43Holdover mode 191, 204, 391

Index428

definition 187operating in 391phase error 391

Hypothetical reference connection (HRX) 251–252

Hypothetical reference path (HRP) 256

IILEC 331Impairment 3–4

attenuation 3distortion 3–4noise 3

Impedance 212high 209

Incumbent local exchange carriers (ILEC) 331

Indication 76Induction

mutual 349In-service measurement (ISM), see ISMIntegrated attenuation, see AttenuationIntegrated services data network (ISDN), see

ISDNIntercountry path core element (ICPCE) 264Interface 208

analog 211BRI 152digital 211electrical 229

connection modes 209measurements in 211specifications 212

Ethernet 325input 374network-to-network (NNI) 101optical 229

measurements in 213output 374PDH 209

jitter 371physical 114

definition 208Q 94, 273SDH 209sensitivity 215T-carrier 377tests 208user-to-network (UNI) 101X 94

Interim layer management interface (ILMI) 315

Interleaved buffer, see BufferInterleaving 319International gateway (IG) 259International path core element (IPCE) 264International portion 260Internet access 140Internode 189Intersymbol interference (ISI) 5, 212Interval, see Unitary interval (UI)Intranode 189Intrusive test

definition 217Inverse discrete fourier transform

(IDFT) 161ISDN 152, 154, 157, 179, 183

See also B-ISDNISM 217, 225–226, 228, 252, 256, 261–262,

296, 326–327ATM 296definition 216network performance 226OAM functions 283

I.351 124I.356 278, 301I.357 301I.371 129, 133, 135, 310I.432 286I.610 284, 288, 291

JJ-bit 24, 27, 60Jitter 61, 194, 213, 369–370

ADM 393amplitude 375

masks 374combined 379

definition 380measurement 381

definition 195demapping 195digital signal 370intrinsic 377leased lines 382mapping 195, 379, 393

definition 380measurement filters 371metrics 370output 373, 377pointer 195, 380–381regenerators 195sampling errors 195SDH 195

Index 429

sinusoidal 375SONET 195stuffing jitter

definition 380tolerance 374–375

definition 374masks 375

transfer 376PDH/T-carrier 377SDH/SONET 378

See also PhaseJustification

phase 195positive 27R-bit 27See also J-bit

LLatency

definition 306Layer

ATM adaptation layer (AAL) 327physical 44, 106, 113, 310, 318seven-layer model 94sublayer 108

PM 113TC 113

Line code 1542B1Q 154AMI 8–10, 30, 153B8ZS 9, 30CMI 8, 10error detection 226HDB3 8–10, 154, 226

Line coding 212Line section 44Line terminal multiplexer, see LTMUXLink capacity adjustment scheme

(LCAS) 40, 52, 74Local area network (LAN) 52, 95Local grade 254Local multipoint distribution system

(LMDS) 145Longitudinal conversion loss (LCL) 344LOS 19, 28, 32, 245Loss of cell delineation (LCD) 311Loss of continuity (LOC) 295Loss of frame (LOF) 28, 245Loss of multiframe (LOM) 22Loss of performance assessment capability

(LPAC) 304Loss of signal (LOS), see LOS

Lower order path (LP) 43Lower-order path adaptation (LPA) 219Lower-order path connection (LPC) 219Lower-order path, see LPLTMUX 41

MMaintenance

definition 208signals 235–236

Management entity 93Management information base (MIB) 94,

284Mapping 45, 53, 55, 220

asynchronous 54, 56–57definition 379higher-ordercontainers 224jitter 379lower-orderr containers 225synchronous 57See also ADM

Masks for copper qualification 409Master clock, see ClockMaximum burst size (MBS) 122Maximum cell transfer delay 124Maximum time interval error (MTIE), see

WanderMeasurement

filtershighband 372wideband 371

interfacesoptical 213

interval 372Measurement device 228Measurements

APS 244ATM layer

320attenuation 364BER 217bridged 365combined jitter 381digital 366FEXT 365G.821 252G.826 256, 261in-service 216interfaces

electrical 211jitter and wander 371jitter in output interfaces 373

Index430

jitter tolerance 374jitter transfer 376loss of longitudinal balance 362mapping jitter 393NEXT 361out-of-service 217output wander 386performance 251relative and absolute wander 383round trip delay (RTP) 242synchronization 272TDR 362two end 364wander transfer 389

Medium grade 255Message

B&G 183–184dying gasp 174–175EOC 174MSG 183RATES 183

MFAS 21Minimum cell rate (MCR) 122Modulation 169

CAP modulation 169discrete multitone modulation (DMT) 156DMT 171phase modulation 390quadrature amplitude modulation

(QAM) 156technologies 154

Monitoringdefinition 208errors 80events 78performance 78See also Protected monitoring point (PMP)

MSDPRING 247–248MSSPRING 90, 247–249MTIE, see WanderMultiframe alignment signal (MFAS) 21Multiframe CRC-4 16Multiple access 10Multiplex section dedicated protection ring

(MSDPRING) 90Multiplex section dedicated protection ring

(MSDPRING), see MSDPRINGMultiplex section linear protection

(MSLP) 90Multiplex section overhead, see SOHMultiplex section protection (MSP) 89Multiplex section shared protection ring

(MSSPRI NG), see MSSPRING 247Multiplex signaling device, see SignalingMultiplexer

mapping 379SDH rings 271 See also ADM

Multiplexer section alarm indication signal (MS-AIS) 245

Multiplexer section overhead (MSOH) 245, 274

Multiplexer section protection (MSP) 245Multiplexer section (MS) 44, 223Multiplexer, see NEMultiplexing 10, 31, 45, 53–55, 57, 158, 195

CDMA 10definition 10devices 19FDMA 10level 23–24level 4 24PDMA 11SDMA 11synchronous 242TDMA 10testing of 242types of 10See alsoFDM, TDM, QAM 10

Multiprotocol label switching (MPLS) 95Mux/demux tests 240

PDH 241SDH 242

M.2100 251–252, 261, 264M.2110 268M.2120 268M.3000 93

NNational portion 259NE 40, 81, 145, 198, 234–235, 378, 387,

398acceptance tests 208, 268, 393ADM 224–225, 236

measuring jitter 393synchronization reference

switching 396synchronization tests 395tests on 393

ADM, see Add and drop multiplexerapproval 207bringing into service tests 208delays 242demultiplexer 374

Index 431

DXC 224, 242free running mode 395holdover 187input and output interfaces 374installation tests 208jitter 374, 376multiplexer 374pointer adjustments 381regenerator 40, 374regenerator section 41SDH rings 270STE 40stress tests 238test applications 219types of 268

Near-end crosstalk (NEXT) 350measurement of 361

Networkaccess network 151asynchronous 187circuit switching 100connection-oriented 99packet switching 99–100plesiochronous 187PSTN 142security 85synchronization 188, 199, 203

definition 185synchronous 37See also Topology

Network architecturecontrol plane 106management plane 106user plane 106

Network element (NE), see NENetwork node, see NodeNetwork parameter control (NPC), see ATMNetwork security 244Network-to-network interface (NNI) 101Network, elements of 1New data flag (NDF) 61, 63NEXT 176, 350–351

measurement 361NFAS 19–20, 28NMFAS 21Node 42, 60, 99, 104, 110, 197

base station 323definition 1interconnection 189internode 189intranode 189terminal

destination 111originating 110

Noise 3, 5, 340atmospheric 6crosstalk 6impulse 6intermodulation 6thermal 5, 340white 361

Nonframe alignment signal, see NFASNonintrusive measurement 228Nonmultiframe alignment signal

(NMFAS) 21Nonreturn to zero (NRZ) 8, 229Nyquist frequency 173, 179Nyquist sampling rate, see SamplingNyquist sampling theorem, see Sampling

OOAM functions 93, 305

analysis of 310ATM layer 291cells 291data flows 285ISM 283–284layers 285SDH 286

OOS 217, 219, 223–225, 252, 261, 263, 268ATM 321definition 216end-to-end 323leased line 323, 325modes 220, 222performance parameters 302test traffic 300

Open network provision (ONP) 268Open systems interconnection (OSI) 94Operation and maintenance channel, see

EOCOperation support system (OSS) 93Operation, administration and management

(OAM), see OAM functionsOptical attenuator 215Optical power 213, 215Optical receiver 215Optical sensitivity 214Out-of-service measurement (OOS), see

OOSOutput aggregate, see SignalOutput jitter, see JitterOverhead 99

byte 235

Index432

Overhead addition 45Overhead block 46Overload 215

measuring 215O.150 217, 374O.153 217O.171 375O.181 217, 219–220, 222, 228, 374O.191 296, 302, 304, 316

PPacket 99

over SDH/SONET (PoS) 52routing 100

Packet switching 99Padding 120–121Parity

anomaly 227See also C-bit and Error

Pathbidirectional 257layer, definition 43unframed 261unidirectional 257

Path core elements (PCE) 264Path end point (PEP) 259Path overhead, see POHPath terminal equipment (PTE), see PTEPath trace bytes 274Path trace tests

SDH ring 274Payload 103, 119

BER 301block 46maximum of 306synchronization 60test cells 300–301, 316

Payload type identifier (PTI) 102, 290PCM 10–12, 20–21, 26

encoding 13quantization 13sampling 12

PDH 1, 13, 22, 227ATM 314bringing into service 208clock transfer 189events 22, 29, 236hierarchy 25–26jitter transfer 377limitations 36multiplexer 27synchronization 197

transport of tributaries 53Peak cell rate (PCR) 122, 304Performance

acceptable level of 267ATM 279BPM 296clock 187definition 7degraded level of 268FPM 296LPAC 304management of 296measurements of 251monitoring of 78, 296OOS 302parameters 279, 298unacceptable level of 268

Permanent virtual connection (PVC) 112Phase

error 392fluctuation 194, 369–371, 380, 382

slow 385See also Jitter and Wander

justification 195modulations 390quantization 12–13, 372, 379See PLLtransient 384, 389See also PLL

Physical interface, see InterfacePhysical layer (PL), see LayerPlesiochronous

definition 37synchronization 26See also Signal

Plesiochronous digital hierarchy (PDH), see PDH

Plesiochronous tributary mapping 220PLL 195–196, 202, 380–382, 384POH 47, 64–66

4 bytes 679 bytes 67higher-order 67lower-order 67

Point to point protocol (PPP) 52Pointer 47, 60, 71

adjustment 60, 272, 369, 374, 380–381decrement 62definition 380format 61generation 61increment 62

Index 433

movementsgeneration of 239

procedures 61sequence 381See also Jitter

Polarization division multiple access (PD-MA), see Multiplexing

Policing functions 132, 280GCRA 135See also Traffic 133

POTS 151–152, 155–157, 179, 183PRBS 174, 217–218, 228, 233, 244, 263,

270, 319, 325PRC, see ClockPrimary reference clock (PRC), see ClockPrimary reference source (PRS), see

Clock 199Protected monitoring point (PMP) 209–210,

216, 229Protection

1÷n 881+1 88active and passive 87automatic protection switching 87bridge/switch signals 89dedicated and shared 87definition 245dual-ended or single-ended 89line 89multiplex section 89revertive or nonrevertive 89ring-switching or span-switching 88subnetwork 92trail or subnetwork 89types of 87–88unidirectional or bidirectional 88virtual container path 91working or protection resources 87See also Test 87

Pseudorandom bit sequence (PRBS), see PRBS

Pulse 10, 212Pulse code modulation, see PCM

QQoS 122, 125, 129

ATM 122, 129, 131negotiated parameters 124service categories 125traffic characterization parameters 122traffic contract 127

Quadrature amplitude modulation

(QAM) 146, 156, 170, 172Qualification strategies 331Quality of service, see QoSQuantification 386Q.2931 124Q.2961 124

RR-bit 24, 27REBE 226Reference clock 382Reference performance objectives

(RPO) 263Regenerator section overhead, see SOHRegenerator section (RS) 41, 44, 223Regenerator, see NERemote alarm indication (RAI) 20, 235Resilience

protection 87restoration 86

Resilient packet ring (RPR) 52Resource management (RM), see ATMRestoration

definition 244Round trip delay variation (RTDV) 305Round trip delay (RTD) 242, 305RS coding 166, 168, 404

SSampling 12–13, 194, 369

errors 195, 197frequency 12Nyquist rate 12Nyquist sampling rate 12Nyquist sampling theorem 12period 14Shannon sampling theorem 12theorem 12

S-bit 23–24SDH 35, 52, 60–61, 64, 69, 75, 96, 207

bringing into service 208clock transfer 190defects 85Ethernet 52events 76, 78, 83, 236

analysis 83layers 42malfunction 74monitoring 208multiplexing map 49OAM functions 286performance monitoring 78

Index434

ringmultiplexers 271path trace tests 274protection switching tests 273synchronization measurements 272tests on 270transparency tests 270

synchronization 185, 199, 230, 383timing 60

SDH/SONETATM 314clock transfer 190pointers, use of 195protection 87restoration 86security 85synchronization 198

models 199–200SDSL, see DSLSection overhead, see SOHSection terminating equipment, see NESection, see Regenerator section (RS)SEC, see ClockService bit 24Service level agreement (SLA) 317SES 253, 256, 258, 262–263, 283, 301

definition 252, 256SESR 253, 255

definition 258Seven-layer model 94Severely errored cell block ratio

(SECBR) 281Severely errored cell block (SECB) 301, 304Severely errored second ratio (SESR), see

SESRSeverely errored second (SES), see SESShannon sampling theorem, see SamplingSHDSL 411Signal

ADSL 170, 319aggregate 240, 269, 380analog 2, 189attenuation 3broadband 153, 382clock signal 189, 194continuous 2delay distortion 5demultiplexing of 23, 29diameter 153digital 2, 9, 189digitalization of 13discrete 2

distortion 4DS1 30, 32DS2 32E1 14, 22, 154, 189–190, 227frequency 216higher-order 381information and 1input signal 27intraoffice 215line code 8

2B1Q 154AMI 8, 10, 153B8ZS 9, 30CMI 8, 10HDB3 8–10, 154, 226

long haul 215lower-order 381mapping 219noise 3, 5OAM indication 326offset 381optical 270output aggregate 27output clock 26PDH 244plesiochronous 22, 26, 219propagation 3SDH 239, 242, 244See Maintenance signal 235short haul 215significant instant 194, 369spurious 5–6square wave 369synchronization 60, 189, 192T1 14, 189–190tributary 27, 195, 240, 380types of 2See also AIS, LOS, NFAS, Phase, and

SamplingSignal degrade (SD) 245Signal fail (SF) 245Signal to noise ratio (S/N) 318Signaling 20–21, 31, 111

APS 64ATM 131CCS 20channel 20circuits 20end-to-end 31in-band 31loss of incoming 21multiframe 21

Index 435

multiplex signaling device 21rate 161words 20See also CAS

Simple network management protocol (SNMP) 94

Slave, see ClockSlip 185, 202, 261, 369Slow phase fluctuation, see WanderSOH 67

LOH 69MSOH 54, 67, 69, 83RSOH 44, 54, 67, 83

SONET 35, 39, 75, 372comparison of SDH and 39events 78formats and procedures 45frame structure 46hierarchy 69jitter 195, 378layers 42line section 44malfunction 74multiplexing map 49path layers 43pointers 61section 44synchronization 185, 383synchronization tests 395virtual concatenation 72virtual tributaries 47See also SDH 74

Space division multiple access (SDMA), see Multiplexing

Splitter 157–158, 229Stimulus-response test 235STM 46, 55, 57, 69–70, 114, 140, 213, 398

See also FrameStress tests 238, 242STS 54, 61, 67, 69–70, 79, 94STS-1 46Stuffing 24, 27, 32, 45, 187, 197, 379

bits 380jitter, definition 380

Subcarrier 161, 170–171, 173, 179, 182–183Subnetwork 42, 92, 245Subnetwork connection protection (SNC-

P) 92Superframe

ADSL 163, 166Superframe (SF) 30–31Supervision bits 18

Sustainable cell rate (SCR) 122Switch

architecture of 307ATM 105VC 105VP 105

Switch throughput 306Switched virtual connection (SVC) 112, 144Switching

digital 201Symmetrical digital subscriber line (SDSL),

see DSLSynchronization 239, 398

bad 185, 195byte

fast 163–165interleaved 165

definition 185desynchronization 379digital 201hierarchical 185models 199–200mutual 187network 188

elements of 185SDH and SONET 188, 190, 383topologies of 187

payload 60PDH 197plesiochronous 26SDH/SONET 198–199signal 189, 192symbol of 163, 174T-carrier 197tests of 395–396

Synchronization measurementsSDH rings 272

Synchronization of NE and test set 230Synchronization status messages (SSM) 191Synchronization supply unit (SSU) 189,

198–199, 203, 387, 389Synchronous digital hierarchy (SDH), see

SDHSynchronous equipment clock (SEC), see

ClockSynchronous optical network, see SONETSynchronous residual time stamp

(SRTS) 116–117Synchronous transmission module (STM),

see STMSynchronous transport signal (STS), see STS

Index436

TTandem connection monitoring (TCM) 81T-carrier 1, 13, 26, 198, 371, 377, 387

ATM 314clock transfer 189hierarchy 14, 29jitter transfer 377mapping 379synchronization 197transport of tributaries 53

TDM 10, 14, 27, 146, 195asynchronous 99standards 14synchronous 99See also multiplexing

Telecommunications management network (TMN) 270, 284

Testoverhead transparency 234stimulus-response 235stress 238

Test cells 300–301, 316Test sequence errors (TSE) 236Test signal structures (TSS) 218, 220, 236Test traffic 277, 300, 305–306

generation of 300, 316, 321reception of 321reception problems 317See also Test cells

TestsADM and DXC 268free-running 395in the Interfaces 208path trace 274PDH mux/demux 241protection switching 273SDH mux/demux tests 242SDH/SONET rings 270synchronization 395tributary continuity 269

Threshold 304, 324, 375TIE 385, 391

definition 384wander metrics 384See also MTIE

Time deviation (TDEV) 388application 385calculation 386definition 384wander 384

Time division multiple access (TDMA) 146Time division multiplexing (TDM), see

TDMTime domain reflectometer (TDR) 362Time interval error (TIE), see WanderTime slot, see TS0, TS16Time-division multiplexing, see TDMTiming

external 199internal 200line 199through 199transparency 190

Timing loop 201Topology 42

distributed 188hub/star 42meshed 41, 188mixed 41ring 41, 188tree 187

Trace identifier mismatch (TIM) 275Traffic

administration of 308–309characterization parameters 122contract 127control algorithms 134monitoring of 328policing 133, 309shaping 133VBR 137

Training 179Transmission

aggregate transmission rate 14definition 3digital 8media 1, 114

conductive 2definition 2dielectric 3

quality 18security 85symbol 161

Transparency tests 270Transverse electromagnetic mode

(TEM) 415Trellis coding 181–182, 319Tributary continuity test 269Tributary signal, see SignalTributary unit groups (TUG)Tributary unit (TU) 45TS0 16TS16 20–22T.50 274

Index 437

UUnable to comply acknowledgement

(UTC) 176Unavailable time 253, 257, 268, 283Unframed paths 261Unidirectional path 257Unidirectional path switched ring

(UPSR) 93Unitary interval (UI) 373

definition 370–371Upstream

frame structure 161Usage parameter control (UPC), see ATMUser-to-network interface (UNI) 101

VVariable bit rate (VBR) 120, 125–127, 129,

144traffic conformity 137

Very high speed DSL (VDSL), see DSLVirtual channel connection (VCC) 111Virtual channel identifier (VCI) 102

allocation of 132upstream allocation 132

Virtual channel (VC) 103–104link 111switches 105

Virtual circuit (VC)definition 100

Virtual container 46, 54–55, 233higher-order 55, 219, 223–224lower-order 55, 219, 225lower-orderr 224POH 66protection 91

Virtual path connection (VPC) 112, 288Virtual path identifier (VPI) 102

allocation of 132upstream allocation 132

Virtual path (VP) 103–104link 112switches 105

Virtual private network (VPN) 95, 139, 322Virtual tributary path (VT Path) 43Voice services, see POTS

WWander 61, 194, 213, 369–370, 382–383

absolute 383clock performance 197definition 196input tolerance 387MTIE 384

application 384definition 384

phase fluctuation 382phase modulations 390phase transient 384relative 383TDEV 384

definition 384temperature changes 197TIE 384transfer

measurement of 389See also Phase

West aggregate 397–398, 400West-East aggregate

See also East aggregate 269Wide area network (WAN) 151Wireless local loop (WLL) 144, 148

XxDSL 139, 331

comparison table 157prequalification 332qualification 331qualification during commissioning 333selective prequalification 333thorough prequalification 332

X.86 52

Index438

[Jose M. Caballero, Francisco Hens, Andreu GuimerÃ(BookFi.org)

Documents

Transcript of [Jose M. Caballero, Francisco Hens, Andreu GuimerÃ(BookFi.org)