Understanding Telecommunications Networks

Understanding Telecommunications Networks
Other volumes in this series:
Volume 9
Volume 12
Volume 13
Volume 19
Volume 20
Volume 21
Volume 25
Volume 26
Volume 28
Volume 29
Volume 31
Volume 32
Volume 33
Volume 34
Volume 35
Volume 36
Volume 37
Volume 38
Volume 40
Volume 41
Volume 43
Volume 44
Volume 45
Volume 46
Volume 47
Volume 48
Volume 49
Volume 50
Volume 51
Volume 904
Volume 905
Phase noise signal sources W.P. Robins
Spread spectrum in communications R. Skaug and J.F. Hjelmstad
Advanced signal processing D.J. Creasey (Editor)
Telecommunications traffic, tariffs and costs R.E. Farr
An introduction to satellite communications D.I. Dalgleish
SPC digital telephone exchanges F.J. Redmill and A.R. Valder
Personal and mobile radio systems R.C.V. Macario (Editor)
Common-channel signalling R.J. Manterfield
Very small aperture terminals (VSATs) J.L. Everett (Editor)
ATM: the broadband telecommunications solution L.G. Cuthbert and
J.C. Sapanel
Data communications and networks, 3rd edition R.L. Brewster (Editor)
Analogue optical fibre communications B. Wilson, Z. Ghassemlooy and
I.Z. Darwazeh (Editors)
Modern personal radio systems R.C.V. Macario (Editor)
Digital broadcasting P. Dambacher
Principles of performance engineering for telecom info systems
M. Ghanbari, C.J. Hughes, M.C. Sinclair and J.P. Eade
Telecommunication networks, 2nd Edn J.E. Flood (Editor)
Optical communication receiver design S.B. Alexander
Satellite communication systems, 3rd Edn B.G. Evans (Editor)
Spread spectrum in mobile communication O. Berg, T. Berg,
J.F. Hjelmstad, S. Haavik and R. Skaug
World telecommunications economics J.J. Wheatley
Telecommunications signalling R.J. Manterfield
Digital signal filtering, analysis and restoration J. Jan
Radio spectrum management, 2nd Edn D.J. Withers
Intelligent networks: Principles and applications J.R. Anderson
Local access network technologies P. France
Telecommunications quality of service management A.P. Oodan
Standard codecs: image compression to advanced video coding
M. Ghanbari
Telecommunications regulation J. Buckley
Security for mobility C. Mitchell (Editor)
Optical fibre sensing and signal processing B. Culshaw
ISDN application in education and training R. Mason and P.D. Bacsich
Andy Valdar
The Institution of Engineering and Technology
Published by The Institution of Engineering and Technology, London, United Kingdom
© 2006 The Institution of Engineering and Technology
First published 2006
This publication is copyright under the Berne Convention and the Universal Copyright
Convention. All rights reserved. Apart from any fair dealing for the purposes of research
or private study, or criticism or review, as permitted under the Copyright, Designs and
Patents Act, 1988, this publication may be reproduced, stored or transmitted, in any
forms or by any means, only with the prior permission in writing of the publishers, or in
the case of reprographic reproduction in accordance with the terms of licences issued
by the Copyright Licensing Agency. Inquiries concerning reproduction outside those
terms should be sent to the publishers at the undermentioned address:
The Institution of Engineering and Technology
Michael Faraday House
Six Hills Way, Stevenage
Herts, SG1 2AY, United Kingdom
While the authors and the publishers believe that the information and guidance given in
this work are correct, all parties must rely upon their own skill and judgment when
making use of them. Neither the authors nor the publishers assume any liability to
anyone for any loss or damage caused by any error or omission in the work, whether
such error or omission is the result of negligence or any other cause. Any and all such
liability is disclaimed.
The moral rights of the authors to be identified as authors of this work have been
asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
British Library Cataloguing in Publication Data
Valdar, A.R.
Understanding telecommunication networks
1. Telecommunication systems
I. Title II. Institution of Engineering and Technology
ISBN (10 digit ) 0 86341 362 5
ISBN (13 digit ) 978-086341-362-9
Typeset in India by Newgen Imaging Systems (P) Ltd, Chennai
Printed in the UK by MPG Books Ltd, Bodmin, Cornwall
An introduction to telephony
Basic telephony
A telephone network
How does a network set up a call connection?
The many networks and how they link
Other forms of telephone networks
Mobile networks
Cable TV networks
Interconnection of networks
International calls
Interconnection of a PSTN and a PNO’s network
Mobile to mobile via the PSTN
The Internet
Access to the Internet
Dial-up via the PSTN and ISDN
Over a cable modem
Leased line access
The specialist networks associated with a PSTN
Operator-services network
Intelligent network
Business-services network
Private-circuit services network
Data services networks
Telex network
A model of the set of a Telco’s networks
Network components
Network topologies
Nodal: concentrator switching
Nodal: route switching
Nodal: packet switching and routeing
Nodal: control (computer processing and storage)
Nodal: multiplexing
Frequency division multiplexing
Time division multiplexing
Code division multiplexing
Nodal: grooming
Nodal: consolidating
3.10 Link component
3.11 Analogue-to-digital conversion
3.11.1 The advantages of digital networks
3.11.2 The A/D process
3.12 Summary
Transmission systems
Transmission bearers
Transmission principles
Transmission media
Multiplexed payloads
The PCM multiplexed payload: the basic building block
of digital networks
The time division multiplexing of digital blocks
Plesiochronous digital hierarchy system
SONET and synchronous digital hierarchy system
The range of transmission systems
Metallic-line systems
Digital subscriber line transmission systems
Point-to-point optical fibre
Dense wave-division multiplex system
Passive optical fibre network
Line-of-sight microwave radio systems
Earth satellite systems
Wireless LANs
Wireless MANs
Transmission networks
Access networks
Scene setting
The copper (‘local loop’) access network
The optical fibre access network
Radio access network
Broadband access
Planning and operational issues
Core transmission networks
Scene setting
PDH network
SDH network
Transmission network resilience
Circuit-switching systems and networks
Circuit-switching systems
Subscriber switching (local) units
Digital telephone switching systems
Digital exchange structures
ISDN exchanges
Network dimensioning
The concept of switched traffic
Call distribution
Traffic flow
Traffic routeing
Exchange capacity planning
Signalling and control
An overview of signalling
Applications of common-channel signalling
ITU common-channel signalling system no. 7
(CCSS7, SS7 or C7)
ITU H323 and session initiation protocol
Call control
Exchange-control systems
Intelligent network
Future network intelligence
Data (packet) switching and routeing
The nature of data
Packet switching
Connection-orientated packet mode
Connectionless packet mode
Comparison of packet switching modes
Asynchronous transfer mode
Internet protocol
The Internet
VOIP over broadband
Network aspects: IP over ATM
8.10 Multi-protocol label switching
8.11 Local area networks
8.12 Wireless LANs
8.13 Summary
Mobile switching systems and networks
Characteristics of mobile networks
Tetherless link
Need for handset identification
Need to track the location of users
Need for a complex handset
Use of complex commercial model
Need for specialised service support
A simple generic model of a mobile system
How does radio work?
Cellular networks
Access mechanisms in cellular networks
The GSM system
GSM system description
Location management in a GSM system
Mobile call in a GSM network
Cell hand-over and power management
GSM frame structure
General packet radio service
Third generation (3G) mobile systems
Universal mobile telecommunications system
Network planning considerations
The wireless scene
9.10 Fixed–mobile convergence (FMC)
9.11 Summary
Numbering and addressing
10.1 Introduction
10.2 Numbering and addressing in telephone networks
10.3 Administration of the telephone numbering range
10.4 Routeing and charging of telephone calls
10.4.1 Numbering and telephone call routeing
10.4.2 Number portability
10.4.3 Numbering and telephone call charging
10.5 Data numbering and addressing
10.6 ATM addressing
10.7 IP numbering/naming and addressing
10.7.1 Internet names
10.7.2 Internet addresses
10.7.3 Translating internet names to addresses
10.7.4 IPv6
10.8 Inter-working of internet and telephone numbering and
10.9 Summary
Putting it all together
11.1 Introduction
11.2 Architecture
11.2.1 Commercial or service model view
11.2.2 Techno-regulatory view
11.2.3 Functional or logical view
11.2.4 Physical view
11.3 A holistic view of a telecommunications network
11.3.1 Logical multi-layered network views of a PSTN
11.3.2 Physical view of the set of a Telco’s networks
Quality of service and network performance
11.4.1 Transmission loss and loudness in the PSTN
11.4.2 Transmission stability
11.4.3 Echo and delay
11.4.4 Digital errors
11.4.5 Apportionment of performance impairments
11.5 Operations
11.6 Network evolution
11.7 Next generation network
11.8 Summary
11.9 Conclusion
Appendix 1
Standards organisations
Appendix 2
List of ITU-T recommendation E.164 assigned
country codes
Writing this book as a personal project has been a great source of enjoyment for me.
But, of course, this was not done in isolation and many people have been kind enough
to give me help in the form of providing information and offering critical reviews.
First, I would like to acknowledge the late Professor Gerry White, who gave
the initial impetus for this book. Although he was able to give early direction
and comments, Gerry’s sad and untimely death in 2004 unfortunately truncated his
However, I have been fortunate indeed in having the willing help of many friends
and colleagues from the telecommunications industry and UCL, whom I am pleased to
be able to acknowledge here. First, I am grateful to Chris Seymour, a veteran of BT and
now a consultant, for his early guidance on the structure of the book. I am especially
indebted to my colleague and long-time mentor Professor Keith Ward, another veteran
of both BT, and more recently UCL, for his support, comments, and permission to
use his ideas and diagrams as a basis for some of the content, in particular, Box 6.1
in Chapter 6. Several other colleagues deserve a special mention for their helpful
inputs and guidance: Tony Holmes and Tim Wright of BT on Chapters 10 and 11,
respectively; and Dr Izzat Darwezah of UCL on Chapter 9. Also, the help from Dr
John Mitchell and Dr Bob Sutherland, both of UCL, is acknowledged.
Furthermore, I have also been lucky enough to have help from several of my
graduate students – notably, David Schultz, Kevin Conroy, Claire Mize and Peter
Weprin – who, despite their studies and full-time work for BT, found time to critique
my draft chapters and I have done my best to incorporate their views. I am also
grateful to Liam Johnston of Fujitsu for his support during the early stages of the
But, this set of acknowledgements must finish with a big thank you from me to
my wife, Su, for not only putting up with me working on the book at all hours, but for
her helpful and diligent reviewing of all the draft text and, of course, her continuous
March 2006
Although everyone is familiar with fixed and mobile telephones and the ability to
dial anywhere in the World, not many people understand how calls are carried, or
how the various networks link together. Similarly, the workings of the Internet and
all the different data and broadband services can be equally mysterious to many, as
is the role of the myriad of underground and overhead cables in the streets. More
importantly, many people now think that in any case all the telephone networks will
be replaced by the Internet! This book aims to address such aspects.
The primary purpose of this book is to describe how telecommunications networks
work. Although the technology is explained at a simple functional level, emphasis
is put on how the various components are used to build networks – fixed (‘wired’),
mobile (‘wireless’), voice and data. I have tried to pitch the explanations at a level
that does not require an engineering knowledge. Indeed, it is hoped that this book will
be helpful to the wide range of people working in the telecommunications industry:
managers who specialise in marketing, customer service, finance, human resources,
public relations, investor relations, training and development, and legal and regulation. However, many of the engineers in the industry may also appreciate an
understanding of aspects beyond their specialisations that this work aims to provide.
I have also written the book with a view to making the broad-based coverage and
emphasis on networks valuable background reading for students on undergraduate
and graduate university courses in telecommunications. The majority of the material
for the book has been based on my lecturing experience over the last 6 years in my role
of visiting professor of telecommunications strategy at the Department of Electronic
and Electrical Engineering of UCL. But my understanding of the business of planning
and operating networks is based on the previous 30 years of wide-ranging experience
at BT.
Finally, this is a rapidly changing industry – one of the reasons why it is so
fascinating – with new technology constantly hitting the scene. But public networks
do not change over night, and the principles of networks and how technology can be
used as put forward in this treatise should endure for many years. So, I hope that this
book has a long and useful life.
Chapter 1
An introduction to telephony
Telecommunications is today widely understood to mean the electrical means of
communicating over a distance. The first form of telecommunications was that of
the Telegraph, which was invented quite independently in 1837 by two scientists,
Wheatstone and Morse. Telegraphy was on a point-to-point unidirectional basis and
relied on trained operators to interpret between the spoken or written word and the
special signals sent over the telegraph wire. However, the use of telegraphy did
greatly enhance the operations of railways and, of course, the dissemination of news
and personal messages between towns. This usefulness of telecommunications on
the one hand and the limitation of needing trained operators on the other led to the
aspiration for a simple means of bi-directional voice telecommunications that anyone
could use. Alexander Graham Bell met this need when he invented the telephone in
1876. Remarkably soon afterwards, the World’s first telephone exchange was opened
in 1878 in New Haven, Connecticut, USA. Since then, telephony has become the
ubiquitous means of communicating for humankind, and telephone networks using the
principles of Alexander Graham Bell have been implemented throughout the World.
This chapter introduces the basic principles of telephony, covering the operation
of a telephone and the way that telephones are connected via a network.
Basic telephony
Fig. 1.1(a) illustrates a basic simple one-way telephone circuit between two people.
The set-up comprises a microphone associated with the speaker, which is connected
via an electrical circuit with a receiver at the remote end associated with the listener.
A battery provides power for the operation of the microphone and receiver. (The
concept of an electrical circuit is given in Box 1.1.) During talking, variations in
air pressure are generated by the vocal tract of the speaker. These variations in air
Understanding telecommunications networks
Speech voltage
Receiver Listener
Figure 1.1
(a) Simple One-Way Speech Over Two Wires [Ward]. (b) Both-Way
Speech Over Four Wires [Ward]. (c) Both-Way Speech And Alerting
Over Eight Wires [Ward]
pressure, known as sound waves, travel from the speaker to the microphone, which
converts them into an electrical signal varying in sympathy with the pattern of the
sound waves (see Box 1.2 for more explanation). Indeed, if you are to look at the
electrical signal on the circuit leaving the microphone, as illustrated in Fig. 1.1(a),
the level of the electrical signal varies with time, with an average value set by the
voltage of the battery and with modulations above and below this level, representing
the variation in sound pressure hitting the microphone. This electrical signal is an
analogue signal because it is an analogue of the sound wave variations in air pressure.
(Later, in Chapter 3, we consider how an analogue signal is converted into a digital
signal within a telephone network.)
An introduction to telephony
Box 1.1
Electrical Circuits
Consider an electrical circuit comprising a power source, e.g. a battery, and a
length of wire linking both terminals of the power source to some device, say
a lamp. Whilst the circuit is complete the lamp will glow and so a switch is
normally inserted in to the arrangement to control the light on and off. In this
simple example the voltage applied by the battery can be viewed as forcing
an electric current to flow around the circuit from its positive terminal to its
negative terminal. This flow is referred to a ‘direct current’ or ‘DC’. The lamp
contains a coil of special wire that provides an obstacle to the flow of current –
known as ‘resistance’. The greater the resistance of the lamp the less current
the battery can force to flow through the whole circuit. This gives rise to the
simple relationship, known as ‘Ohm’s Law’ in which the resistance (measured
in ohms or ) is given by the voltage (measured in volts or V) divided by the
current (measured in amps or A).
An alternative form of electrical voltage is one which cyclically varies from
zero up to a maximum positive value, drops to zero and then goes to an equal
but opposite maximum negative value and then back to zero. The shape corresponds to the sinusoidal waveform shown in Fig. 1.11. This so-called alternating
voltage creates a corresponding ‘alternating current’, AC. The electrical main
supply is typically at 240 V alternating current (240 V AC), with the cycles
occurring at 50 times per second (50 Hz) in the United Kingdom and Europe
and at 120 V AC cycling at 60 Hz in the United States.
The continuous cycling of the alternating electrical current causes additional
changes to the flow of electricity when passing through a circuit. The first
phenomenon – capacitance – causes the waveform to be delayed; the second
phenomenon – inductance – causes the waveform to be advanced. The results
of these effects, known collectively as impedance, are that the AC current is
out of step with the applied AC voltage. These effects are used throughout
telecommunications and electronic equipment, for example: inductance forms
the basic mechanism exploited in hybrid transformers and loading coils, as
mentioned in this chapter.
At the receiving end of the circuit the analogue electrical signal energises the
receiver (i.e. an earpiece), generating a set of sound waves, which are an approximate
reproduction of the sound of the speaker.
Obviously for conversation to be possible it is necessary to have transmission
in both directions, and therefore a second circuit operating in the opposite direction
is required, as shown in Fig. 1.1(b). Thus, a basic telephone circuit comprises four
wires: one pair for each speech direction. This is known as a basic 4-wire circuit. In
practice, of course, a telephone system would need to include a mechanism for the
caller to indicate to the recipient that they wished to speak. Therefore, we need to add
to the assembly in Fig. 1.1.(b) a bell associated in a circuit with a power source and
Understanding telecommunications networks
Box 1.2
How a Microphone (Transmitter) and Receiver (Earpiece) Work [5]
The microphone used in a telephone handset is really an electro-acoustic transducer, a device for converting acoustic energy to electrical energy. There are
several types of transmitter used for telephones, e.g. carbon granule, electrodynamic and electret. Each performs the conversion of acoustic or vibration
energy to electrical energy in different ways. For example, in the electrodynamic type of microphone a flexible diaphragm is made to vibrate when in
the path of a stream of sound waves. The diaphragm’s movements are transferred to a coil of wire in the presence of a magnetic field, thus inducing a
current in the coil through a phenomenon known as ‘electro-magnetic induction’. This varying electric signal has a voltage pattern that is a replica of the
sound wave pattern impinging on the microphone, i.e. it is an analogue signal.
The telephone receiver (earpiece) is also an electro-acoustic transducer, but
works in the opposite direction to the microphone. For example, with an electrodynamic device at the receiving end of the circuit the analogue electrical signal
is passed through a coil attached to an electromagnet, which is attached to a
permanent magnet. The varying electrical signal in the coil induces a magnetic
field (electro-magnetic induction) to vary similarly, which reacts against the
bias field made by the permanent magnet, thus causing a diaphragm in the
vicinity to vibrate in sympathy. In this way, the diaphragm generates a set of
sound waves, which are a reasonable reproduction of the original sound of the
a switch. The electrical current flowing in a circuit used to ring a bell in a telephone is
known as ringing current. Again, one such arrangement is required in each direction.
This argument brings us to the conclusion, illustrated in Fig. 1.1(c), that a set of eight
wires, four pairs, is needed to provide bi-directional telephony service between two
In a practical telephone network, the most important requirement is to minimise
the amount of cost associated with connecting each customer. Since there are many
thousands or millions of customers on a telephone network, any reduction in the
amount of equipment needed to be provided for each customer would result in large
overall cost savings. Thus, some ingenious engineering has enabled significant economy to be achieved through the reduction of the numbers of wires from eight to two,
i.e. one pair. This is achieved through the use of 4-to-2-wire conversion (and vice
versa), and the time-sharing of functions, as described below.
(i) 4-to-2-Wire conversion. The two directions of speech circuits, shown in
Fig. 1.1(c), can be reduced down to a single circuit carrying speech currents
in both directions, using a device known as a hybrid transformer, as shown
in Fig. 1.2. (See Box 1.3 for a brief explanation of how a hybrid transformer
An introduction to telephony
2-Wire circuit
Figure 1.2
4-Wire circuit
4-Wire circuit
Hybrid transformer
4-to-2 Wire Conversion
Telephone 1
Figure 1.3
Telephone 2
A Simple Two-Phone System
(ii) Time-sharing of functions. The need for two pairs of wires to be dedicated to
ringing circuits can be totally eliminated by exploiting the fact that ringing does
not occur during the speaking phase of a telephone call. Therefore, the single
pair provided for speech can instead be used at the start of a call to carry ringing
current in either direction, as necessary. Once the call is answered, of course,
the single pair is used only to carry the two directions of speech current.
Fig. 1.3 shows that, for our simple two-person scenario, the telephone instrument
at each end needs to comprise a handset with microphone and receiver; a hybrid
transformer; a bell and a means to send ringing current to the far end. The two
telephones need to be connected by a single pair of wires and a battery.
Understanding telecommunications networks
For four phones, no. of links required = 6
Generally, For n phones, no. of links required = n(n−1)/2
Figure 1.4
Direct Interconnection of Several Phones
We can now extend this basic two-person scenario to the more general case of
several people with phones wishing to be able to talk to each other. For example,
the logical extension to a four-telephone scenario is shown in Fig. 1.4. In this case
six links (i.e. 2-wire circuits) are required in total, each telephone terminating three
links. Not only does each telephone need to terminate three links rather than just one,
but it also needs a 1-to-3 selection mechanism to choose which of the links should
be connected to in order to converse with the required telephone. (Not only does this
involve a selection switch within each phone, but the arrangement also needs each
phone to be designated with a name or number – as discussed later in Chapter 10.)
Whilst the arrangement shown in Fig. 1.4 is quite practical for networks of just a
few phones – indeed, many small office and household telephone systems are based
on such designs – it does not scale up well. In general, the number of links to fully
interconnect n telephones is given by n(n − 1)/2. As the number of phones becomes
large, the number of directly connected links approaches n2 /2. Clearly, providing
the necessary 5,000 direct links in a system serving just 100 telephones would not
be an economical or practical design! (In addition, the complexity of the selection
mechanism in each telephone would increase in order for it to be capable of switching
1-out-of-99 lines.) The solution to the scaling problem is to introduce a central hub –
commonly called an exchange or central office – onto which each phone is linked
directly, and which can provide connectivity between any two phone lines, on demand
(Fig. 1.5). With a single exchange serving n phones only n links are required; a good
solution, which in practice scales up to about 50,000 telephone lines with modern
telephone exchanges.
We can now deduce the role of a telephone exchange. Fig. 1.6 gives a block
schematic diagram of the basic functions required to connect two exchange lines.
In the example shown it is assumed that telephone A is calling telephone B. The
first requirement is that both telephones A and B need to have an appropriate power
source. Although a battery could be provided inside each telephone, indeed in the
early days of telephony this was in fact done, it is far more practical to locate the
battery centrally at the exchange, where the telephone company (usually known as a
An introduction to telephony
For n phones, no. of links to each exchange = n
Figure 1.5
Interconnection: Use of an Exchange
Ring Power
Figure 1.6
The Functions of a Telephone Exchange
‘Telco’) can maintain it. When telephones are connected to their exchange by a pair of
metallic wires, usually made of copper, the power for the phones can conveniently be
passed over that pair. This arrangement is convenient for the telephone user because
then they have no need to manage the charging of batteries in their premises and are
also not dependant on the reliability of the local electricity supply. (Although more
recently, of course, the need for users to charge the battery in a mobile phone every
few days has become acceptable.)
However, there are situations where power cannot be passed to the telephone from
the exchange. For example, this is not possible when optical fibre is used to connect
telephones because glass does not conduct electricity! The other notable example is
Understanding telecommunications networks
that of a mobile network, where a radio path is used to link telephones to the exchange,
as described in Chapters 2 and 9.
The first step the exchange has to undertake in managing a call is to detect that
the calling telephone (i.e. telephone A in Fig. 1.6) wishes to make a call. The simplest
method for conveying such an indication from a telephone, and the one that is still
most commonly used today, is for advantage to be taken of the fact that the pair of
wires between the exchange and the telephone can be closed at the telephone end,
thus creating a loop. This looping of the pair by the telephone causes a current to
flow, which operates a relay at the exchange. (A relay is a device that when activated
by an electrical current flowing through its coil causes one or more switches to close.
The latter in turn then pass electrical current on to other circuits or devices – hence
the term ‘relay’.) In the case of the original manual exchanges, the closure of the
relay switch caused a lamp to glow, hence alerting a human operator to the calling
state of the line. For a modern electronic automatic exchange the energising of the
relay by the loop current causes changes in the state in an electronic system, which
is subsequently detected by the exchange control system.
The loop is closed in the telephone by a switch activated by the lifting of the
handset off the telephone casing – this causes a small set of hooks to spring up.
The lifting of the handset creates a condition known as ‘off-hook’. In the case of a
cordless phone this off-hook switch is located in the base unit (attached to the copper
pair termination in the house) and is controlled remotely from the radio handset when
the subscriber presses the ‘dial’ or ‘send’ button – often indicated on the button by a
picture of a handset being lifted.
On the outgoing side of the exchange attached to the line for telephone B there is
a power source, a generator to send ringing current and a relay to detect the ‘off-hook’
condition when the called telephone answers. (For simplicity, Fig. 1.6 assumes that
telephone A is calling telephone B so the ringing current generator is shown attached
only to the line B, but of course any telephone can initiate calls, so in practice all lines
have power supply, off-hook detector (i.e. relay) and a ringing current generator.)
When telephone exchanges were first introduced the method of connecting two
telephone lines together was through a human operator using a short length of a pair
of wires across a patching panel. Each telephone line terminated on the panel in a
socket with an associated small indicator lamp. The human operator was made aware
that telephone A wished to make a call by the glowing of the relevant lamp, activated
by telephone A going ‘off hook’. On seeing the glowing lamp the operator started the
procedure for controlling a call by first talking to caller A and asking them to which
number they wished to be connected. The operator then checked the lamp associated
with the called line, if this was not glowing then that line was free, and the call could
be established. The next step was for a ringing current to be applied to telephone B.
The operator did this by plugging a line connected to a special ringing generator into
the socket for telephone B. It was important for the operator to monitor B’s lamp to
ensure that the ringing was stopped as soon as B answered – otherwise there would
be a very annoyed person at the end of the line! The operator would then make the
appropriate connection using a jumper wire across the patching panel. Finally, the
operator needed to monitor the two lamps involved in the call so that the connection
An introduction to telephony
could be taken down (by removing the connecting cord between the two sockets) as
soon as one of the lamps went out. The call had then been terminated.
In making a call connection, the operator had to follow certain procedures, including writing on a ticket the number of the caller and called lines, the time of day and the
duration of the call, so that a charge could be raised later. It is important to remember
that an exchange needs to serve many lines and that at any time there will be several
calls that need to be set up, monitored or cleared down. The human operators had to
share their attention across many calls; each operator typically was expected to be
able to deal with up to six calls simultaneously.
Generally, today telephone exchanges are fully automatic. However, there are
still occasions when a human intervention is required, e.g. providing various forms
of assistance and emergency calls, and special auto-manual exchanges with operators
provide such services. For convenience, the telephone exchanges considered in the
remainder of this book will be only the fully automatic types. The description above
is based on a fully manual exchange system because it enables the simple principles
of call connection to be explained in a low-technology way – yet, all of the steps and
the principles involved are followed in automatic exchange working [1, 2].
In an automatic exchange, as shown in Fig. 1.6, the role of the operator is taken
by the exchange control (representing the operator’s intelligence), which in modern
exchanges is provided by computers with the procedures captured in call-control software, and the switch or ‘switch block’, usually in the form of a semi-conductor matrix,
which performs the functions of the connecting cords and patch panels. Chapter 6
describes modern exchanges in more detail.
1.3 A telephone network
A telephone exchange serves many telephone lines, enabling any line to be connected
to any other (when they are both free). In a small village all the lines could easily
be connected to a single telephone exchange, since the distances are short. However,
if telephone service needs to be provided to a larger area, the question arises as to
how many exchanges are needed. This is illustrated in Fig. 1.7, where a region of the
country has a large population of telephones to be served. They could all be served by
one central large exchange or by several smaller exchanges. Obviously, the lengths
and hence costs of the telephone lines reduces as the number of exchanges increases,
but this saving is offset by the increase in costs of exchange equipment and buildings.
In addition, a link (known as a ‘junction route’) needs to be established between
each exchange to ensure full connectivity between all telephones in the region. The
trade-off between the cost of the telephone lines and the costs of the switching and
buildings and junction-route costs plotted for various numbers of exchanges, n, to
serve the population of telephones in the region shows a typical bath-curve shape [3].
In this example of Fig. 1.7, the optimum cost is achieved with three exchanges.
Of course, it is not only the number of exchanges that contributes to the optimum costs, but also the location of the exchanges within their catchment areas. The
optimum total cost for the region is achieved when the exchanges are at the centre
Understanding telecommunications networks
Local exchange
Junction route
Total network cost
Total cost
Junction + exchange
+ site and building costs cost
Local loop costs
Number of exchanges
Figure 1.7
Network Optimisation [Ward]
of gravity (i.e. the location where the sum of all the line lengths is the minimum) of
the population of telephones served. In practice, network operators locate telephone
exchanges as close to this centre of gravity as possible, within the constraints of the
availability of suitable sites within a town.
There are also practical limitations on the lengths of telephone lines which constrain the size of the catchment area of lines dependent on one exchange. These limits
are set by the electrical characteristics of the lines, predominantly the resistance of the
loop. (See Box 1.1 for an explanation of resistance.) Typically, this resistance needs
to be less than 2,000 ohms (written as ‘2000 ’) to ensure that sufficient current flows
for the ‘off hook’ condition to be detected by the exchange and also to ensure that
the loudness of the call is acceptable. There are several ways in which the telephone
lines can be kept within the electrical limits, including the use of thicker gauge wire
(more expensive, but having less resistance) on the longer telephone lines. Also, in
the United States, where the terrain requires larger catchment areas, inductors (i.e.
devices comprising tightly coiled wire), known as ‘loading coils’, are added to long
lines to reduce the signal loss. The majority of telephone lines in the United Kingdom
are below 5 km, whereas in the United States there are many lines in excess of 10 km
and exchange catchment areas can be as large as 130 square miles [4]. These aspects
are considered in more detail in Chapter 5.
Thus, the primary design requirement for a network operator is to achieve a
cost optimised set of exchanges, each of which is located at the centre of gravity
An introduction to telephony
26 fully interconnected local exchanges
= 26(26−1)/2 = 325 junction routes
6 fully interconnected trunk exchanges
= 6(5−1)/2 = 15 trunks + 26 junction links
Trunk transit
1 trunk tandem exchanges
= 6 trunk links + 26 junction links
Figure 1.8
Traffic routeing hierarchy
(a) Network Structure-1 [Ward]. (b) Network Structure-2 [Ward].
(c) Network Structure-3 [Ward]. (d) Network Structure-4 [Ward]
for the catchment area, whilst keeping within the electrical limits on telephone line
We can now consider extending this logic to a telephone network serving several
regions, or even the entire country. Given that each exchange needs to be able to
connect to every other exchange (via junction routes), a further network optimisation
issue arises. For example, for an area comprising 26 exchanges the number of fully
interconnected junction routes would be 26(26−1)/2 = 325, as shown in Fig. 1.8(a).
(For a country like the United Kingdom with some 6,000 local exchanges (LEs) the
number of directly connected junction routes would be an impossible 17,997,000!). As
Understanding telecommunications networks
before, this situation is alleviated by the use of a central exchange within each region –
known as a trunk exchange (TE). These trunk exchanges also need to have links (i.e.
‘trunk routes’) between them to ensure full connectivity across all regions. In the
example of Fig. 1.8(b) this arrangement requires just 15 trunk links and 26 junction
links, but with the added cost of 6 trunk exchanges. Further network optimisation is
possible by the addition of a single trunk tandem exchange to provide connectivity
between each of the trunk exchanges, reducing the number of trunk links to six, but
with the added cost of a trunk tandem exchange (see Fig. 1.8(c)).
The arrangement of exchanges and links can be drawn as a traffic routeing1
hierarchy, as shown in Fig. 1.8(d). (The term ‘traffic’, which is used to describe
the flow of telephone calls, is described in more detail in Chapter 6.) The bottom
of the hierarchy is the exchange serving its catchment area of telephone lines. These
exchanges are known as ‘local exchanges’in the United Kingdom (and ‘Class 5 central
offices’ in the United States [4]). Similarly, telephone lines are known as ‘local lines’
or ‘subscriber lines’, and the collection of local lines as the ‘local network’. At the
second level of the hierarchy are the trunk exchanges, and at the top of the hierarchy
is the trunk tandem exchange. This parenting of trunk exchanges on higher-level
exchanges can continue further. However, in practice, telephone networks generally
have no more than three levels of trunk exchanges: primary, secondary and tertiary.
Note that only the local exchanges have subscriber lines attached; the various levels
of trunk exchanges switch only between trunk and junction routes to and from other
exchanges. The significance of this difference between local and all other types of
exchanges will be considered further in Chapter 6.
As an example of a typical public switched telephone network (PSTN), we will
consider the simplified representation in Fig. 1.9. This shows the PSTN for BT (British
Telecommunications), which covers all of the United Kingdom (except for the town
of Hull). The structure is shown as a hierarchy, with the exchanges serving subscriber
lines forming the lowest level. For reasons that will be explored in Chapter 6, there
are three ways of connecting subscriber’s lines to the network:
Type (i)
where subscribers are connected directly to the central local exchange
(known as a ‘processor node’);
Type (ii) where subscribers are connected to a remote unit (known as a ‘remote
concentrator’), which in turn is parented onto and controlled by the central
local exchange;
Type (iii) where subscribers are connected to an autonomous very small rural
exchange, which in turn is connected to the central local exchange.
In BT’s PSTN there are about 800 of the central local exchanges (‘processor
nodes’), about 5,200 remote units (‘remote concentrators’), and some 480 very small
rural exchanges. In total, these exchanges and units serve about 29 million subscribers’
lines [5]. The (central) local exchanges are parented onto their trunk exchange (known
as a ‘digital main switching unit – DMSU’). This hierarchy has only one level of
trunk switching, with all trunk exchanges having direct routes to all others. This
1 The US spelling, ‘routing’, is also in popular use within the UK.
An introduction to telephony
Figure 1.9
BT’s international
BT’s specialised
Other network
Central LE
Very small
BT’s Public Switched Telephone Network
arrangement ensures that any subscriber can be connected to any other on this network
and that there will be a maximum routeing of two trunk exchanges between the
originating and terminating local exchanges. Calls between local exchanges, which
are in the same locality and where there is sufficient demand, are carried over direct
links – known as ‘junction routes’, as shown in Fig. 1.9. The number of junction routes
is minimised in large urban areas, such as London, Birmingham and Manchester by
the use of junction tandem exchanges (known as ‘digital junction switching unit –
DJSU’). Further network optimisation is achieved by offloading the trunk exchanges
of calls flowing just within the region by switching these at regional tandems (known
as ‘wide area tandem – WAT’). BT’s PSTN has some 15 junction tandems (JTs), about
20 regional tandems and around 80 trunk exchanges.
In addition to providing full interconnectivity between all subscribers on its network, BT’s PSTN also provides access to international exchanges within the United
Kingdom for calls to other countries. As Fig. 1.9 shows the international exchange
is above the trunk exchange in the hierarchy. BT is one of several network operators
in the United Kingdom that have international exchanges (known as ‘international
switching centre – ISC’). The routeing of international calls is considered in more
detail in Chapter 2.
The PSTN of BT also acts as a collector of calls from its subscribers to the networks of other operators (fixed and mobile) in the United Kingdom, and similarly as
a distributor of calls from other operator’s networks to subscribers on the BT network
(Fig. 1.9). The points of interconnection (POI) between BT’s PSTN and the other networks are at the trunk exchanges. Similarly, the trunk exchanges also act as collectors
and distributors of traffic between PSTN subscribers and the variety of specialised
BT exchanges covering functions such as operator services and directory enquiries,
Centrex and VPN service for business customers, and network intelligence centres
Understanding telecommunications networks
Brighton (34)
01273 34XXXX
01736 89YYYY
{01273 34XXXX}
{01273 34XXXX}
01273 34XXXX
Brighton (39)
01273 39ZZZZ
Figure 1.10
An Example of a Call Routeing
for the variety of calls that need more complex control than the PSTN exchanges can
provide. Chapter 2 considers how these specialist exchanges are linked to the PSTN,
while the intelligence, data and mobile networks are fully described in Chapters 7, 8
and 9, respectively.
How does a network set up a call connection?
A fundamental requirement of routeing through any telecommunications network –
whether it is to carry a voice call or a packet of data, etc. – is that each termination (or
end point) on the network has an address, which can be used to indicate the desired
destination. In the case of telephony each subscriber’s line has a permanent unique
number, which is used by the exchanges in the network as its address.
We are now in a position to consider a call set up across the PSTN. Fig. 1.10
illustrates the routeing of a national call within the United Kingdom from a subscriber
in Penzance in the South West tip to a subscriber in Brighton on the South coast. As
an example, we assume that the subscriber in Penzance, 01736 89YYYY (i.e. area
code 01736, exchange code 89 and subscriber number YYYY), dials their friend in
Brighton, 01273 34XXXX (i.e. area code 01273, exchange code 34 and subscriber
number XXXX). There are several local exchanges in the Brighton area, two of which –
34 and 39 – are shown in Fig. 1.10.
On receiving the dialled digits from the caller the control system of the local
exchange in Penzance (the ‘originating exchange’) examines the first five digits, as
shown by the underlining of the numbers in Fig. 1.10. The initial ‘0’ indicates to
the control system that a national number has been dialled. Since the control system
does not recognise 1273 as either its own code or that of any other exchanges to
An introduction to telephony
which it has direct routes, the call is switched through to its parent trunk exchange in
Plymouth. This is achieved by the sending of signals between the control systems in
Penzance and Plymouth exchanges (shown by dotted arrows), conveying the required
destination number. The number information sent in the forward direction between
the exchanges is shown between curly brackets in Fig. 1.10.
The control system in Plymouth exchange examines digits 01273 and concludes
that this is not one of its dependent area codes, but is owned by Haywards Heath
trunk exchange. In the UK network all BT trunk exchanges are fully interconnected
and so the control system of Plymouth exchange selects a direct route to Haywards
Heath and signals the full destination number.
Haywards Heath exchange control system recognises 01273 as one of its dependent area codes and so it examines the next two digits, 34, to determine the destination
local exchange. The Haywards Heath trunk exchange then routes the call to the
Brighton-34 exchange. At the destination exchange, the control system on recognising 0123 34 then examines the final XXXX to determine the called subscriber
In the case of a local call from the other Brighton exchange, 39, the set-up sequence
is simpler, as shown in Fig. 1.10. The calling subscriber dials only the local number
since the destination number has the same area code. The control system of Brighton39 on examining the first two digits, 34, recognises that the call is destined for one
of the other exchanges dependent on the Haywards Heath trunk exchange, and so
routes the call accordingly. Notice that, in this case of a local call, the control system
inserts the appropriate area code and signals the full destination national number to
its trunk exchange. Haywards Heath trunk exchange then uses the standard decoding
procedure to determine that the call is destined for Brighton-34 exchange; the call is
then completed as before.
The dialled number is not only used by the exchanges to route the call through to
its destination, but importantly the number is also used to determine the appropriate
charge rate for the call. In both examples above it is the parent trunk exchange that
determines the charge rate for the call and indicates this to the originating local
exchange, which then times the call and records the incurred charge in its control system memory. This is later downloaded to produce an entry on the calling subscriber’s
telephone bill.
1.5 Waveforms
It is important at this point to ensure an understanding of the basic concept of waveforms and how they will be used throughout the remainder of the book. In general,
all communication systems are characterised by the range of frequencies that can
be carried, usually termed the ‘bandwidth’ of the system. Fig. 1.11 shows the classical sinusoidal waveform showing the regular oscillating pattern that occurs often
in nature, e.g. a swinging pendulum. The time between successive peaks or troughs
in the amplitude of the waveform (which might be length, voltage, power, etc.)
is known as the ‘period’. Each period comprises one full cycle of the waveform.
Understanding telecommunications networks
One cycle of wave
(wavelength )
Figure 1.11
Sinusoidal Waveform
The distance travelled in this period of time is called the wavelength, given the
symbol λ. The number of periods per unit time is known as the frequency of the
waveform, measured in cycles per second (cps), or to use the international standard,
Hertz (Hz) [6].
The sinusoidal waveform of Fig. 1.11 in the case of a sound wave would represent
a single tone. For example, a single sinusoidal sound wave of 256 Hz produces the
well-known musical note of middle C. However, speech and music are a mixture of
many different tones, each with different frequencies and amplitudes, resulting in
complex waveform shapes, as shown in Fig. 1.1(a).
Within this book two fundamentally different forms of waves are considered. The
first is that of sound waves, which are moving physical vibrations in a substance.
These can be heard and felt by humans. Thus, sound waves are carried through the air
in the form of vibrating air molecules, e.g. from a person speaking to a listener, or a
guitar string being plucked. Sound waves can also be carried as vibrations through a
wooden door, through water or as vibrations along a railing which someone at the far
end is banging, or as vibrations along the railway lines as the distant train approaches.
Sound is generated by vibrating the air, as is apparent when a finger touches a loud
speaker of a HiFi system at full volume! Vibrations carried along the surface of the
sea (i.e. ‘sea waves’) are another example of sound waves.
The other form of waves are electro magnetic waves. Light is the most obvious example of electromagnetic radiation. Light travels as very high speed waves
of electric and magnetic forces. Other examples of this type of waveform are radio
waves (carrying radio and TV broadcasts, etc.), microwaves and X-Rays. They are all
examples of the phenomenon of electromagnetic radiation, but at different frequencies. Thus, within telecommunications networks, it is electromagnetic waveforms
that are carried as electricity over metallic wires, or as radio waves through the
air from radio masts to a mobile handset, or as light waves through optical fibre
cables. The full range of the electromagnetic radiation spectrum is illustrated in
Fig. 1.12.
The key point is that the sound waves emitting from a speaking person are converted into electromagnetic waves for the conveyance through the
An introduction to telephony
Increasing frequency
Very low
3 kHz
30 kHz
300 kHz
light waves
3 MHz
Long wave broadcast radio
Very high
30 MHz
Ultraviolet waves
Ultra high
300 MHz
FM broadcast radio
Medium wave broadcast radio
Super high
3 GHz
Extra high
30 GHz
300 GHz
Microwave radio, satellite systems
Broadcast TV, mobile phones,
WiFi, cordless phones, Bluetooth
Not to scale
Figure 1.12
The Electromagnetic Radiation Spectrum
Hybrid transformer unit
Figure 1.13
Hybrid Transformer
telecommunications network. (Thus, in Figs 1.1 and 1.2, the sound waves are carried
over the pair of wires as an electromagnetic signal over the wires – and converted
back to sound waves at the far end.) The sound or speech waveform (occupying
the frequency range of 0–4 kHz as described in Chapter 3) can be carried by an
Understanding telecommunications networks
Box 1.3
How a Hybrid Transformer Works [7]
The Hybrid transformer is a system comprising four input terminals, two for
the Go signal and two for the Return signal of the 4-wire circuit, and two output terminals for the 2-wire signal. The role of the Hybrid system is to ensure
that the Go signal is conveyed to the 2-wire circuit only and the signal from
the 2-wire circuit is conveyed to the Return terminal only. This separation is
achieved by the use of a pair of transformers, each with two sets of windings, as shown in Fig. 1.13. The windings are arranged such that the induced
currents from the Go signal, which would otherwise appear at the Return terminals, are equal and opposite and so cancel, with the power dissipated in
a resistance device (the ‘balance’) matching the electrical characteristics of
the 2-wire circuit. The signal from the Return circuit, which would otherwise go into the Go circuit, is similarly cancelled. In this way, the Hybrid
transformer ensures that the two directions of electrical signal on the 2-wire
circuit are transferred to the appropriate Go and Return parts only of the 4-wire
electromagnetic wave at any appropriate frequency level. Thus, if carried straight
over a copper pair the sound waveform is converted to an electrical analogue signal
occupying the same frequencies (0–4 kHz); or they may be carried at higher frequencies within multiplex signal (as described in Chapter 3), or at radio frequencies if
carried over long wave, medium wave or FM radio, or at even higher frequencies in
the electromagnetic spectrum if carried as part of a TV broadcast, and so on up to the
highest frequencies when the speech is carried as light waves over an optical fibre
In this opening chapter we introduced the basics of telephony, in which economies
are gained by the use of a single pair of wires to convey a two-way telephone call
over the link to the serving exchange. We also examined the role of local, trunk and
international exchanges, noting that they can be manual or automatic. This led to an
understanding of the basic structure of a PSTN. A simple example of a call across
the country has been examined to illustrate the sequence of events involved in progressing a telephone call. This highlighted the roles of numbering and addressing,
call control, and the need for signalling between the control systems of exchanges
involved in the call. We also introduced the concept of the two forms of waveforms used in telecommunications networks. Subsequent chapters develop these basic
An introduction to telephony
REDMILL, F. J. and VALDAR, A. R.: ‘SPC Digital Telephone Exchanges’, IET
Telecommunications Series No. 21, Stevenage, 1995, Chapter 1.
ANTTALAINEN, T.: ‘Introduction to Telecommunications Network Engineering’, Artech House, Norwood, MA, 1999, Chapter 2.
FLOOD, J. E.: ‘Telecommunications Switching, Traffic and Networks’, Prentice
Hall, Harlow, 2001, Chapter 1.
BIGELOW, S. J., CARR, J. J. and WINDER, S.: ‘Understanding Telephone
Electronics’, Fourth edition, Newnes, Boston, MA, 2001, Chapter 1.
BIGELOW, S. J., CARR, J. J. and WINDER, S.: ‘Understanding Telephone
Electronics’, Fourth edition, Newnes, Harlow, 2001, Chapter 2.
LANGLEY, G. and RONAYNE, J. P.: ‘Telecommunications Primer’, Fourth
edition, Prentice Hall, Boston, MA, 1993, Chapter 6.
FRANCE, P. W. and SPIRIT, D. M.: ‘An Introduction to the Access Network’,
Chapter 1 of ‘Local Access Network Technologies’, edited by FRANCE, P. W.,
IET Telecommunications Series No. 47, Stevenage, 2004.
Chapter 2
The many networks and how they link
In Chapter 1 we considered how a PSTN establishes telephone call connections
between subscribers on its network. However, this is only part of the story. First,
there are usually several PSTNs in any one country, each owned by different operators, although some may have a limited geographical coverage or serve only certain
(typically corporate business) customers. Many countries also have Cable TV operators who provide telephone service, in addition to broadcasting TV distribution.
Then, of course, nearly all countries now have one or more mobile telephone networks, which are separate from the, so-called, fixed PSTNs. All these networks need
to interconnect so that their subscribers can call subscribers on any of the other networks. This chapter, therefore, briefly describes the various telephone networks and
how they all interconnect.
Although the fixed PSTN and mobile networks are primarily designed for voice
communications, they do also enable subscribers to connect their computers to the
Internet (a network primarily designed for data service). Thus, to complete the interconnect story, this chapter briefly introduces the concept of the Internet and how
access to it is provided by the telephone networks.
The second aspect of the story is the need for a set of specialist networks which
enables a typical national telephone operator to provide a wide range of services
beyond simple telephone calls. This chapter briefly reviews these specialised networks and uses a simple model to help explain how they link and support each
So, clearly there are many networks in each country all linking appropriately in
order to provide a variety of types of call connections between subscribers across the
World. (Not to mention the interconnection of the various non-voice or data networks!
This, for simplicity, is not covered until Chapter 8.) We will now try to understand
this complex picture.
Understanding telecommunications networks
Cell 1
Cell 2
Cell 3
Figure 2.1
station (BS)
Mobile Cellular Network Concept
Other forms of telephone networks
2.2.1 Mobile networks
The concept of a cellular mobile telephone network is shown in Fig. 2.1. The principle
of operation is similar to that of the PSTN, as described in Chapter 1, although there
are some important differences. Obviously, the first difference is that a two-way radio
link is used to connect the mobile handset to its exchange, rather than a copper pair
of wires, so that the user has freedom of movement. This ability of the handset to
move around means that the serving exchange must have a system to identify a calling
handset and another system to keep track of where a particular handset is at any time,
so that it can send or receive calls.
The basis of the ‘access network’ is a set of cells. These cells form an area ranging
from about 1–10 km radius and have a centrally located radio transmitter/receiver
(also known as a ‘transceiver’), collocated with a base station (BS). A two-way radio
link is potentially available to all handsets in the cell area. Groups of radio-channel
pairs (Go and Return), each provided by a separate set of frequencies, are pre-assigned
to the cell, and a pair of radio channels is allocated on demand by the base station
controller (BSC) to a mobile handset wishing to make or receive a call.
A BSC serves a catchment area of several cells with their associated BSs. The
BSs are connected to their BSC by a fixed transmission link carried over either a
point-to-point microwave radio system or an optical fibre cable.
During the call, the handset is free to move within the cell. However, if the handset
travels towards the boundary of its cell during a call the weakening signal from the
The many networks and how they link
BS and the strengthening radio signal from the adjacent cell’s BS is detected by the
BSC and a ‘handover’ between the two is managed without interruption to the call.
This requires changing the send and received channels to those of the new cell. (If a
spare set of radio channels is not currently available within the new cell, the call will
have to drop out.) At the end of the call, or if the terminal moves into another cell’s
area, these radio channels are available for use by other calls.
The mobile switching centre (MSC), which is very similar to a large local PSTN
exchange (but without the subscriber terminations), performs the switching of the
mobile calls between all handsets operating within its catchment area of BSs or to other
MSCs on its network. The MSC is also associated with control systems providing
terminal-location management and authentication [1]. A more detailed account of
mobile networks is given in Chapter 9.
2.2.2 Cable TV networks
Cable TV networks (often referred to as just ‘Cable Networks’) are primarily based
around a cable-distribution system, often deployed in only certain areas of a town,
which delivers many TV broadcast channels primarily to the residential market. In
some countries, for example, the United Kingdom, these networks also provide telephone service to the houses taking their broadcast TV service. Fig. 2.2 illustrates the
principle of such a network. Two separate feeds serve each customer’s premises: a
coaxial cable (similar to TV aerial cable) terminating on a set-top box (STB) linked to
the television set provides TV broadcast service, and a copper pair of wires provides
telephone service to the household. The coaxial cable is part of a tree and branch
arrangement radiating from a street cabinet, serving customers located along one or a
few roads in the vicinity. A separate pair of copper wires back to the street cabinet is
Optical fibre
Coaxial cable premises
(Head end)
Copper pair
Optical fibre
Figure 2.2
Cable TV Network
Understanding telecommunications networks
required for each household served. At the street cabinet the TV distribution and the
telephone lines are combined, i.e. multiplexed (see Chapter 3), onto an optical fibre
cable back to the network centre. Here the TV channels are provided from the TV
source, known as the head end, which may receive the TV channels from a satellite or
other terrestrial links to TV broadcast companies and possibly a TV channel switching
and control centre. At this network centre the telephone channels are connected to the
cable company’s telephone exchange [2].
The telephone portion of the Cable TV network is similar to the standard PSTN.
However, the catchment areas of the exchanges are usually much larger than those of
the PSTN. In addition, the cable companies tend to combine the functions of local and
trunk switching in each exchange. One or more of the exchanges act as the gateway
to the PSTN and possibly mobile and other networks.
Interconnection of networks
2.3.1 International calls
As shown in Fig. 1.9 of Chapter 1, international exchanges (known as ISC) are above
the trunk-exchange level in the PSTN routeing hierarchy. Several network operators
in the UK have ISCs linked to their PSTN. A subscriber wishing to call a number in
another country first dials the international prefix (internationally recommended to
be 00) to indicate to its network that this is an international call. The next 1, 2, or
3 digits, which form the country code indicating the destination country (see Chapter 10), are then examined by the gateway ISC serving that PSTN. Traffic between
ISCs in different countries is directly routed where there is a high volume of calls
between the two countries, otherwise it is routed via one or more transit ISCs located
in intermediate countries. A simple summary of these international routeing options
is shown in Fig. 2.3. This illustrates an example in which Country 1 provides international transit switching for calls between Country 3 and Country 4, and calls between
Country 1 and Country 2 are directly routed.
In addition to determining the appropriate route to the destination country, direct
or to a transit ISC, the originating gateway ISC determines the appropriate rate to be
charged for the call. This charging information is passed back to the originating local
exchange to be associated with the calling subscriber’s billing records.
The cost of building and maintaining the transmission links between ISCs is
shared between the two operators concerned. In the case of a trans-border land cable,
ownership extends from the ISC to the border. Where there is a sea boundary, e.g.
between the USA and the UK, a hypothetical boundary exists mid-way along the
sub-sea cable, thus both operators own a, so-called, half circuit.
Each operator pays the other for the cost of accepting traffic and completing the
calls in the destination country. In practice, this is done by a periodic reckoning for
each ISC (Country A)-to-ISC (Country B) route of the net flow of calls (the difference
between A-to-B and B-to-A), which is then charged at a determined tariff. Thus, the
operator who sends the higher amount of traffic pays the other operator for the net
termination charge.
The many networks and how they link
Country 4
Country 3
Country 2
Country 1
Figure 2.3
International transit and Gateway exchanges, respectively
International Call Routeing
BT’s Network
01736 89YYYY
Brighton (34)
01273 34XXXX
Points of
01736 60AAAA
01273 75BBBB
PNO’s Network
Figure 2.4
Interconnection with a PNO’s Network
2.3.2 Interconnection of a PSTN and a PNO’s network
Fig. 2.4 shows the situation in the UK where BT’s PSTN interconnects with several hundred other network operators, e.g. other PSTN, Cable TV network or mobile
network in the same country. These are generically referred to as public network operators (PNO) or communication providers (CP). We take as our example the situation
described in Chapter 1, Fig. 1.10, but with the addition of two POI with the network of
a PNO. Note that the subscriber on the PNO exchange in Penzance has the same area
code, 1736, as the BT subscriber, but has a different exchange code, i.e. 60; similarly,
Understanding telecommunications networks
the PNO exchange code at Brighton is 75 compared to 34 for the BT exchange. The
two networks are connected at various POI, usually at the trunk exchange level in
BT’s network. All calls from BT subscribers destined for a PNO’s network are routed
to the destination exchange via the designated POI. Similarly, calls originating in the
PNO network are routed to the destination BT exchange via the appropriate POI. As
Fig. 2.4 shows, there may be a choice of POIs to use for a call: the routeing either
staying on the originating network until the last available POI – known as ‘far-end
handover’, or interconnecting at the nearest POI – known as ‘near-end handover’.
As an example, let us consider the routeing of a call from the BT subscriber in
Penzance (01736 89YYYY) to the subscriber on the PNO network in Brighton (01272
75BBBB). The control system of the BT local exchange in Penzance, on examination
of the non-local area code 01273, passes the call onto its parent trunk exchange at
Plymouth. The control system of Plymouth trunk exchange identifies 01273 as an area
code parented on Haywards Heath Trunk exchange and switches the call accordingly.
The Hayward Heath trunk exchange, on recognising 01273, examines the exchange
code 75 and routes the call to the designated POI for the serving PNO, which in
this case is a route from the BT trunk exchange to the Brighton PNO exchange.
The call is passed to the PNO exchange, the signalling from the control systems
of Haywards Heath giving the full dialled number (01273 75BBBB), so that the
call can be completed within the PNO network – in this case directly to the called
subscriber’s line.
A similar arrangement applies to calls from a PNO to BT’s fixed network.
The flow of calls is measured between the BT trunk exchanges and each of the
interconnected PNO exchanges so that interconnect conveyance charges can be determined. Each network operator must pay the receiving operator for the conveyance
and completion of calls passed over the POI. Normally, this follows the principle of
a receiving operator charging on a call-minutes basis, according to the amount of
their network traversed in delivering the call (either to its subscriber or onto another
network for completion). The rate of charge for traversing a network depends on
the type of network involved, the mobile networks being rated higher than a fixed
network. Like most aspects of interconnect, these call-completion-charge rates are
usually subject to scrutiny or determination by the national regulator [3].
2.3.3 Mobile to mobile via the PSTN
Fig. 2.5 illustrates the general situation for calls between two different mobile networks in the same country. It is interesting to note that there are two basic scenarios for
needing such a configuration. The first is a call between a terminal belonging to (i.e.
registered with a service provider using) mobile network A and a terminal belonging
to mobile network B. The other is a call between two terminals belonging to mobile
network A, but the second terminal has moved temporarily onto (i.e. visiting) mobile
network B – a process known as ‘roaming’.
Although calls between mobile networks in the same country may be delivered
over direct links between MSCs on their respective networks, generally it is more
economical to route calls to other mobile operators, as well as the wide range of fixed
The many networks and how they link
network A
Figure 2.5
network B
Mobile-To-Mobile Call via PSTN
PNOs, via the national incumbent’s PSTN, with its widespread coverage and full
interconnectivity with all the operators in the country. Each mobile network has at
least one MSC that acts as a gateway to the PSTN, providing the required technical
interface between the two networks, as described in more detail in Chapter 9.
Let us consider the case where a terminal, a1 , has roamed from network A to
network B, (Fig. 2.5 refers) and a call is to be made to it from another terminal, a2 ,
on network A. The location-control system of the MSC on network B has already
advised the location-control system of network A that terminal a1 is currently on its
network, so network A knows that it needs to pass the call from terminal a2 over to
network B [1,4]. On receiving a call completion request from network A the control
system of the MSC on Network B allocates on a temporary basis (for the duration
of the call) one of a small pool of PSTN numbers it holds, and advises the control
system of the originating MSC on network A of this number. The originating MSC
then routes the call via its nominated gateway MSC to the PSTN, using the allocated
PSTN number as the destination address. The MSC of network B can then associate
the incoming call from the PSTN with the required BS currently serving terminal a1 .
The call connection between a2 and a1 can then be completed.
Interconnect charging applies across the boundary of network A to the PSTN, and
across the boundary between the PSTN and network B.
2.4 The Internet
The Internet, that widely-known entity, is really a super constellation of many networks around the world – a network of networks. Indeed, the name ‘Internet’is derived
from the term ‘inter-networking’. Each of these networks is based on a special form of
Understanding telecommunications networks
switching designed specifically for handling data, using a standard way of packaging
and addressing the data: the so-called Internet protocol (IP). The enormous utility of
IP is due to the way that it enables data to be easily transferred between programs run
on different types of computers. (The subject of IP and data networks is covered in
more detail in Chapter 8.) However, it was the introduction of the World Wide Web
(WWW), enabling general widespread access to particular sections (i.e. ‘pages’) of
data stored in people’s computers, that really caused the Internet to spread throughout
the World and to become the phenomenon that it is today. A huge range of applications by business and residential users now exploits the Internet. The most common
of these applications are e-mails and other peer-to-peer applications such as gaming,
information gathering from web pages, and Internet shopping.
Fig. 2.6 presents a simplified schematic diagram of the Internet. Each of the
component networks in the Internet constellation are owned and operated by Internet service providers (ISPs). Users and providers of information (in the form of
web pages) subscribe to an ISP for their Internet service. Users access their ISPs
over links provided by one of a range of networks (PSTN, ISDN, mobile, leased
line, ADSL/broadband and Cable TV), as described in the following sections of this
chapter. ISP networks usually comprise a number of IP switching units (known as
‘routers’) located at various points in the country and linked by high capacity transmission circuits leased from a network operator (PSTN), so-called leased lines or
private circuits (see Section 2.6.4 in this chapter). ISP networks link with one another
in order to obtain full interconnectivity between users, the smaller ones (i.e. those that
Backbone ISP
Backbone ISP
Transit ISP
Leased lines
Figure 2.6
The Internet Concept
The many networks and how they link
are less widespread and have fewer links to other ISPs) connect to larger ISPs, who
in turn connect to the largest ISPs, those with a national or international coverage –
the ‘backbone ISPs’. The interconnection of the largest ISPs in the United Kingdom
is achieved by jumpering cables between each of their routers collocated in central
buildings, known as Internet exchange (INX); the largest is located in the City of
London: ‘Telehouse’. This provides participants with full interconnectivity within
the UK and to links going to the United States, Asia and mainland Europe.
2.5 Access to the Internet
2.5.1 Dial-up via the PSTN and ISDN
The original and still a common way of accessing the Internet is through a telephone
call, thus taking advantage of the full interconnectivity of the PSTN, its ubiquity and
its ease of use. The procedure is in two stages: the first is a telephony call to the ISP;
the second stage is a data transfer session between the computer and the Internet via
the ISP. The arrangement is illustrated in Fig. 2.7.
For the first stage, a standard call is set up from the computer, which is
connected to a telephone line in the normal way. A modem (a shortening of ‘modulator/demodulator’) card in the computer makes it behave like a telephone instrument,
providing the equivalent of lifting the receiver (‘off-hook’), detecting dial tone, and
sending out the dialled digit tones. The modem also enables data to be passed over
the standard local telephone line. The ISP is connected to the PSTN over a standard
local line or by an ISDN (integrated services digital network) local line, but in either
case standard telephone numbers are allocated.
Once through to the ISP the second stage begins, the call connection through
the PSTN now just acts as a two-way path between the computer and the ISP.
An interchange of data occurs so that the ISP can verify the identity of the user
Figure 2.7
Dial-Up Access to the Internet
Understanding telecommunications networks
and establish a data session to the Internet. IP is used between the computer and
the ISP’s equipment (an IP router, as described in Chapter 8), passing over the
transparent PSTN call connection path. The ISP then routes the data between the
computer and the Internet for the duration of the session. When the user indicates
that the session is over, the modem card in the computer creates the equivalent of the
‘on-hook’ condition so that the telephone exchange clears the call to the ISP in the
normal way.
Generally, the exchange handles the call to the ISP as standard telephony, in terms
of charging and routeing. However, the users want to be able to run their often-lengthy
sessions to the Internet, involving periodic flows of data as web pages are accessed
or e-mails sent or received, without being concerned about the duration and paying
telephony charges on a timed basis. One solution is for the ISP to take a special
service from the PSTN operator using 0800 numbers, which enables the users to have
free or reduced rate call charges, the ISP being able to recover the charges through
subscriptions from the registered users. In addition, the PSTN operator may offer
special flat rate (i.e. un-timed) charges to subscribers for their calls to the Internet.
Similarly, the incumbent PSTN operator may be required by the national regulator
to make such special flat rate tariffs available as a wholesale interconnect service to
other operators (PNOs) or ISPs.
In addition to needing special tariffs, the carrying of Internet-bound calls causes
a problem of congestion for PSTN telephone operators. This is due to the fact that
Internet sessions tend to be much longer than telephone calls, typically some 45 min
compared to 3 min. Thus, telephone exchanges whose capacity is dimensioned on
the basis of an expected number of 3-min average simultaneous calls can experience severe congestion if the proportion of Internet-bound calls becomes significant.
There are a number of ways of dealing with this problem. The first is, of course,
to increase the exchange capacities. In addition, there may be advantage in routeing
large volumes of traffic to a particular ISP away from busy parts of the network,
say bypassing a trunk exchange and going directly to the exchange serving the ISP
(known as ‘trunk off-load’). This method still does not relieve the local exchanges
because the calls have to be handled by at least one exchange so that the destination
telephone number of the relevant ISP can be detected. However, a fully effective way
of avoiding overloading the PSTN with Internet-bound traffic is the total segregation
of telephony and Internet (data) traffic right at the source, i.e. the user’s computer.
Such segregation is provided by the ADSL (asymmetric digital subscriber line; socalled ‘broadband’) and cable modem access systems – as described in Sections 2.5.2
and 2.5.3 – which provide users with a high-speed Internet access by routeing their
data traffic to the ISP independently of the PSTN, the latter handling just the telephony
Accessing the Internet via ISDN is essentially the same as for the PSTN, except
that the interface of the ISDN line to the computer is digital rather than analogue, so
the modem card does not convert the computer output to analogue and consequently
higher speeds up to 64 kbit/s or even 128 kbit/s are possible. The ISDN service is
discussed in more detail in Chapter 6.
The many networks and how they link
Splitter at the exchange
Figure 2.8
Data network
Splitter on
Access to the Internet via ADSL
2.5.2 Over ADSL
The use of ADSL enables a high-speed data service to be carried in addition to
telephony over the copper pair of a standard telephone line, as described in Chapter 4.
Fig. 2.8 shows how the Internet is accessed via ADSL. At the subscriber premises,
the internal telephone line and the computer are separately connected to the splitter
within the ADSL termination equipment. This device directs the computer output
to the high speed data channel on the line, and similarly in the reverse direction;
while enabling the telephone signal to be carried in the normal way on the local
line. The ADSL splitter and terminating equipment located at the exchange similarly
separate the telephony and high speed data signal [5]. The telephony is carried over
a copper pair to the input of the local exchange switch. The data stream carrying
the Internet-bound traffic is connected to the DSLAM (digital subscriber line access
multiplexor) equipment, which combines the data from many such lines onto a high
speed transmission link to a data network. This network carries the Internet-bound
data in an efficient way, interleaving it with data streams from other sources, to the
connecting points of the various ISPs in the country. A corresponding process applies
in the reverse direction from the Internet to the computer.
Charging for access to an ISP (and hence the Internet) may be purely on a subscription basis, or in the case of high capacity users there may also be usage or throughput
2.5.3 Over a cable modem
Cable TV networks are able to provide a high speed data link to an ISP over one of the
spare TV channels carried through the network. The technique requires a so-called
Understanding telecommunications networks
Coaxial cable premises
Data network
Copper pair
Figure 2.9
Access to the Internet via Cable Modem
cable modem, which converts the output of the computer into a signal compatible
with the TV distribution system. As Fig. 2.9 shows, the computer is connected to a
splitter-like device which segregates the received TV signal from the broadband data
stream at the subscriber’s premises, where both are carried as a composite TV signal
over the coaxial cable from the street cabinet. Since many subscribers are sharing
the data modem channel on the coaxial cable tree, some form of combining of the
many signals is required; the street cabinet therefore performs a similar aggregating
function as the DSLAM, described above [6]. As with ADSL, the Cable TV operator
routes the aggregated data traffic over a data network to the connecting points of the
appropriate ISPs and charging for access to an ISP may be on a subscription basis or
usage basis.
2.5.4 Leased line access
Leased lines, providing high capacity dedicated access from the user’s premises, are
typically used to give Internet access to office blocks and industrial sites. The concept
of leased lines is further explained later in Section 2.6.4.
2.6 The specialist networks associated with a PSTN
Network operators, particularly the national incumbents, need to provide a range of
services beyond the standard telephony of the PSTN. These extra services, which
The many networks and how they link
may be grouped as those available to all customers and those aimed at business
customers, are provided by specialised networks owned and run by the operator
and closely linked to their PSTN. Most of the non-incumbent network operators in
a country also have one or more of these specialised networks, usually aimed at
the business customers. All of these networks comprise special equipment located in
exchange buildings. However, the transmission between the customer’s premises and
the exchange building uses either the copper access network used by the PSTN or use
is made of capacity in optical fibre cables if these already exist or a cable is provided
specially. Links between the exchange-located equipment are provided over common
capacity provided by the core transmission network (coaxial cable, optical fibre cable
and microwave radio), linking all exchange buildings, as described in Chapter 5.
This section introduces the most common specialised networks; Sections 2.6.1
and 2.6.2 provide information on the services available to all, while Sections
2.6.3–2.6.6 on service aimed at business customers.
2.6.1 Operator-services network
This network comprises several specialist telephone exchanges around the country,
similar to those described in Chapter 1, but which support a suite of operator consoles
(typically 10–30), instead of subscriber lines. The operators have more control of
the calls than is provided to normal subscribers; thus, such exchanges are normally
referred as ‘auto-manual’. The operator-services network provides call assistance
(dial 100) and access to the emergency services (dial 999 in the UK , the European
standard 112, USA standard 911, etc.), directory enquiry, and blind and disabled
special assistance services. Calls to this network are routed either directly from the
PSTN originating exchange or from the parent trunk exchange. There are also special
charging arrangements for such services; either free, as in the case of assistance and
emergency calls, or at a special rate, as for directory enquiries.
2.6.2 Intelligent network
The so-called intelligent network (IN) comprises several centres around the country
containing control systems and data-bases that provide a variety of advanced switched
services. Typically, these services are based on the translation of the number dialled
by the subscriber to another number in order to complete the call, together with
some special charging arrangements (e.g. the recipient rather than the caller pays)
and tariffs – as is the case of ‘0800’ services. The translations of the dialled numbers
might be fixed or, for example, vary according to the time of day, day of the week,
the location of the caller, or the availability of assistants in the case of a company
providing a service from various call centres. The IN concept is described in Chapter 7.
Calls are routed to an IN centre when the PSTN local or, more usually, trunk
exchange detects that the number dialled requires translation. This might be on the
basis that the number is from a special range (e.g. beginning 08 in the United Kingdom)
or it is in the standard range but listed as special in the exchange-control system (e.g.
for number portability purposes – see Chapter 10).
Understanding telecommunications networks
2.6.3 Business-services network
There are a range of telephony services designed for businesses which give direct deskto-desk dialling using private numbering schemes and special charging arrangements
between offices and factory sites, etc., e.g. Centrex and VPN (see Boxes 2.1 and 2.2).
These services can be provided by the local PSTN exchange. However, it may be more
economical to use separate dedicated special exchanges for these services, which form
a business-services network. Since the majority of calls on this network are within
the businesses, only a few calls need to ‘break out’ to (or ‘break in’ from) the PSTN.
Interconnection is provided between one or more of the business-services exchanges
and a trunk exchange in the PSTN.
Access to these services is separate from standard PSTN, using private or combined public/private numbers (known as direct dialling in – DDI, as discussed in
Chapter 10) and possibly using special telephone instruments. Subscriber lines from
the business customers sites (e.g. an office block, factory or warehouse) are provided
over the copper pairs of the local telephone network or over specially provided optical
fibre cables or microwave radio links.
2.6.4 Private-circuit services network
The private-circuit service network provides point-to-point un-switched links, known
as ‘private circuits’ or ‘leased lines’, between business premises (e.g. offices, factories or warehouses) [7]. Although the private-circuit service provided to the business
customers is non-switched, the network may use automatic digital cross-connection
units (DXC), a form of switching equipment for private circuits, as a way of automating the connection of circuits between transmission links – conceptually a form of
electronic jumpering. The DXCs are technically exchange switch blocks but without
the exchange call-control system. Connections once set up are held for the duration
of the lease of the circuit, typically several years. Such private circuits are used for
linking private telephone systems (i.e. voice) or alternatively, linking data terminals
and computers (i.e. data) in different business premises.
Fig. 2.10 illustrates the outline arrangement for the private-circuit services network. Two examples are given. Fig. 2.10(a) shows how a private circuit is provided
simply by the manual jumpering at the local exchange between the copper pairs
from premises A and B. Longer private circuits are obtained by jumpering of transmission links in concatenation from the customer’s premises at A to their premises
at B. Fig. 2.10(b) gives a block schematic diagram of a private circuit provided using
automatic cross-connection units to establish the required path across the network.
The connection across the DXCs is controlled and monitored from a central network
management centre [8].
2.6.5 Data services networks
There are a variety of specialised networks providing a range of data services using
the packet switching technique, rather than the circuit switching technique used for
telephony described in Chapter 1. The most common of these data services are: ATM
The many networks and how they link
(a) Manually Jumpered Private Circuit
premises A
Local exchange
premises B
(b) Private Circuit Formed With Digital Cross-Connection Units
Trunk exchanges
premises A
NTE: Network termination equipment
DXC: Digital cross-connection unit
Figure 2.10
premises B
A Private-Circuit Services Network
(asynchronous transfer mode), Frame Relay, SMDS (switched multi-megabit data
service), IP and MPLS (multi-protocol label switching). Typically, each of these data
services is provided from a separate network, i.e. five data networks would be needed
for the above list of services.
Business customers use the data services provided by the network operators to
connect terminals and computers from one set of premises to another. In addition,
businesses now use data networks within their premises, such as LANs (local area
networks); the network data service then acts as a link between distant LANs, forming
wide area networks (WANs). Data services and networks are described further in
Chapter 8.
2.6.6 Telex network
Telex is a basic text messaging service which was used extensively throughout the
World by businesses. It is now very much a legacy service, being replaced by the
use of fax and e-mails. However, it is still used extensively within many developing
countries and internationally to their trading partners, e.g. in the United Kingdom.
The Telex Network is quite separate from the PSTN, with its own numbering scheme
for the telex terminals and its own exchanges. However, these are very similar to
telephone exchanges and the Telex network is structured with a routeing hierarchy
similar to the PSTN. Also, the Telex subscriber lines are connected to their exchanges
(which are located in the same buildings as the PSTN exchanges) over the local copper
access network [9].
Understanding telecommunications networks
2.7 A model of the set of a Telco’s networks
It is useful to picture the above-described range of specialised networks associated
with a PSTN operator as a set of building blocks, as shown in Fig. 2.11. This simple
model portrays the various types of customers at the top and a stack of networks,
shown as building blocks, arranged as two horizontals with a set of verticals sandwiched in between. The top horizontal block is the access network, which provides
the means of linking customers (or subscribers) to the various service networks. The
access network is made up of the copper lines (local loop) as described in Chapter 1,
providing links to the PSTN and several of the specialised networks. The specialised
networks, arranged as vertical blocks, comprise a range of switching and intelligence equipment. The access network also contains optical fibre cables providing
broadband access to business premises, and point-to-point microwave radio providing an alternative means of linking subscribers to many of the specialised networks.
Some of the specialised networks cannot be directly linked to subscribers through the
access network, but are linked instead via one of the other vertical networks, usually
the PSTN. For example, calls to operator services are first switched at the PSTN
exchange and then routed over the core transmission network to the operator services
network (shown as OSN in Fig. 2.11).
All of the above service networks (vertical blocks) are supported by the (horizontal
block) core transmission network, which comprises a set of transmission links, usually
carried over optical fibre cables and microwave radio paths, together with automatic
cross-connection and add/drop terminal equipment forming the nodes. (The technology for the core transmission network is described in Chapter 4.) As its position in
the model suggests, the core transmission network acts a common resource for all
Access network
Core transmission network
Figure 2.11
A Model of the Networks Associated with the PSTN
The many networks and how they link
the service networks. Since the service networks, i.e. the verticals in the model, do
not contain any transmission links all their nodes need to use the core transmission
network for connectivity.
In practice, the exchange buildings would house not only a local switching unit
(‘local exchange’), but also potentially a trunk switching unit (‘Trunk Exchange’), and
several of the non-PSTN specialised units. To illustrate this point we will refer again to
the example used in Chapter 1 where the transmission capacity between the exchange
buildings in Haywards Heath and Penzance would also provide the required links
between the private circuit network, business service network or operator services
network specialised units, respectively, located in these buildings.
Other models of the various telecommunications networks are introduced in
Chapter 11, where the concept of network architecture is examined in more
This chapter introduced the concept of the many telecommunications networks that
exist today and how they all interlink. First, we looked briefly at the two other
types of network, in addition to the PSTN, that carry telephone calls, namely: mobile
networks and Cable TV networks. It is noted that all three networks use essentially the
same type of switching system, but the mobile networks use radio links and mobility
management in place of the fixed-wire access of the PSTN and Cable TV networks.
exchange A
exchange B
exchange C
(a) Virtual private network
Customer’s premises
Private branch exchange (PBX)
exchange A
exchange B
(b) Centrex service at Two Sites
Figure 2.12
Centrex and Virtual Private Networks (VPN)
Understanding telecommunications networks
The necessary interconnection of all these types of networks within a country and to
networks in other countries was described.
We also introduced the concept of the Internet and how subscribers gain access to
it via the PSTN, cable modems over Cable TV networks, ADSL broadband or over
private circuits using optical fibre.
Box 2.1
PBXs and Virtual Private Networks
Businesses which have more than a few telephones use a private branch
exchange system, known as a PBX, to provide call connections between each
telephone (which become ‘extensions’) and links into the PSTN [7]. The PBX
is really a small version of the PSTN exchanges, typically ranging in sizes from
10 up to 5,000 extensions. A private numbering scheme is required to enable
extension to extension dialling, also special codes (e.g. ‘dial 9’) are required to
enable calls to be made to the PSTN. Incoming calls from the PSTN have to be
answered by a receptionist or operator at a manual console so that the appropriate (privately numbered) extension can be contacted. Alternatively, the extension numbers can form part of the public numbering scheme (see Chapter 10)
so that calls from the PSTN can be directly switched by the PBX to the required
extension (known as DDI), so avoiding the need for manual intervention where
the caller knows the number of the wanted extension. Only the calls to the PSTN
are charged. The corporate customer owns and pays for its PBX.
In the case where a company extends over two or more sites (e.g. office
or factory buildings) the PBXs on each site can be linked by private circuits,
thus enabling calling between all the extensions. This is known as a ‘private
corporate network’ (or just ‘private network’). In this case the private numbering scheme extends across all the PBXs and usually each PBX is linked to
the PSTN. Charging only applies to calls leaving the private network for the
PSTN, although, of course, a rental charge is made by the network operator for
the lease of the private circuits.
A virtual private network (VPN) provides an alternative to the use of private
circuits between each PBX, as shown in Fig. 2.12(a). The VPN exchange
switches calls between the PBXs connected to it (e.g. between PBXs ‘a’ and
‘b’) as well as to trunk links to the other PBXs in the VPN corporate network.
However, the VPN is provided over public exchanges, either as special business
exchanges or as part of the PSTN. Each VPN exchange switches the private
network calls of several private corporate networks, although each operates
in isolation, using its own numbering scheme. Thus, each corporate network
appears to have the benefits of a private set of links between their PBXs, even
though connectivity is provided over public exchanges – hence the use of the
word ‘virtual’ in VPN. The VPN customer is charged a subscription for the
VPN service based on a certain level of inter-PBX calls (traffic); there are
usually charges made when the level of calls between any two PBXs exceeds
the agreed threshold.
The many networks and how they link
Box 2.2
Centrex Service
Centrex is the generic name of a service in which extension-to-extension calls
within a customer’s site are switched by the public exchange, thus eliminating the need for a PBX [7]. This requires that each extension from the
building be carried over the access network to the centrex exchange, which
serves many centrex customers, each with their own private extension numbering scheme. Again, calls between the different centrex groups are kept
isolated within the exchange, and charging only relates to calls that go out to
the PSTN. The service can extend across several centrex exchanges (known
as ‘networked centrex’), as required to serve the company’s private corporate network, as shown in Fig. 2.12(b). (Note that networked centrex is
not the same as a VPN since the role of the PBXs is taken by the centrex
Finally, the various specialised networks associated with a PSTN were introduced,
Operator services network;
Business services network;
Intelligent network;
Private circuit-services network;
Frame relay network;
ATM network;
IP network;
MPLS network;
Telex network.
The chapter concluded with a simple model, which aimed to help position the
specialised networks with the common access and core transmission networks.
SCHILLER, J.: ‘Mobile Communications’, Second edition, Addison-Wesley,
Harlow, 2003, Chapter 4.
FRANCE, P. W. and SPIRIT, D. M.: ‘An Introduction to the Access Network’,
Chapter 1 of ‘Local Access Network Technologies’, edited by FRANCE, P.W.,
IEE Telecommunications Series No. 47, Stevenage, 2004.
BUCKLEY, J.: ‘Telecommunications Regulation’, IEE Telecommunications
Series No. 50, Stevenage, 2003, Chapter 5.
ANTTALAINEN, T.: ‘Introduction to Telecommunications Network Engineering’, Artech House, Norwood, MA, 1999, Chapter 5.
Understanding telecommunications networks
FOSTER, K. T., COOK, J. W., CLARKE, D. E. A., BOOTH, M.G. and
KIRKBY, R. H.: ‘Realising the Potential of Access Networks Using DSL’,
Chapter 3 of ‘Local Access Network Technologies’, edited by FRANCE, P.W.,
IET Telecommunications Series No. 47, Stevenage, 2004.
TASSEL, J.: ‘TV, Voice And Broadband IP Over Cable TV Networks’Chapter 15
of ‘Local Access Network Technologies’, edited by FRANCE, P.W., IET
Telecommunications Series No. 47, Stevenage, 2004.
BELL, R. K.: ‘Private Telecommunication Networks’, Chapter 11 of ‘Telecommunications Networks’, Second edition, edited by FLOOD, J. E., IET
Telecommunications Series No. 36, Stevenage, 1997.
MARSHALL, J. F., ADAMSON, J. and COLE, R. V.: ‘Introducing Automatic
Cross-Connection into the KiloStream Network’, British Telecommunications
Engineering, Vol. 4, Part 3, 1985, pp. 124–128.
DREWE, C.: ‘BT’s Telex Network: Past, Present – and Future?’, British
Telecommunications Engineering, Vol. 12, Part 1, April 1993, pp. 17–21.
Chapter 3
Network components
The first two chapters have considered the way that a telephone call is set up and how
the various networks providing telephone service are interconnected. Also, Chapter 2
introduced the roles of the other types of network: specialised voice networks and
data networks. This chapter considers the basic components that go together to make
up these networks. In fact these components can be treated as part of a set of building
bricks, each of a different shape and size; all of the telecommunications networks
described in this book are made up using a suitable mix of these bricks. Subsequent
chapters describe the networks and systems in more detail, assuming the reader has
an understanding of the basic components described in this chapter.
Network topologies
Any network is a system of nodes and links. There are many examples of networks
which are encountered in everyday life. We talk of the road network, where the junctions form the nodes and the stretches of road in between are the links of the network.
Similarly, the rail network comprises the rail links joining the station nodes. A further
example is that of the airlines network, where airports provide the nodal functions
and the airline routes provide the links. Networks can also be on far smaller scale, for
example the electrical circuitry in a television set with its physical wires or printed
circuit rails linking the nodal electronic components. The various telecommunications
networks are similarly made up of a variety of nodal functions and transmission links.
The pattern of links between the nodes in a network, i.e. its topology, determines
the possible routeings between any two nodes – either direct if a link between the two
nodes exists or connected through one or more intermediate nodes, known in general
as a ‘tandem routeing’. Fig. 3.1 shows the possible set of network topologies. A star
configuration provides a direct link between all nodes and a hub node, all routeings
between nodes need to transit via the hub. A typical example of a star topology is
Understanding telecommunications networks
Figure 3.1
Network Topologies
the local network with dependent exchanges linked to the parent exchange. With the
ring configuration, routeings between all non-adjacent nodes must transit around the
ring. The local and regional transmission networks tend to use this topology because
it enables a dispersed set of exchanges (nodes) to be linked by cables which follow
the road network and usually requires the least overall length of cable. In contrast, the
mesh configuration uses a direct link between each node, thus no tandem routeings are
required. The top level of a routeing hierarchy (see Chapter 1) is configured as a mesh.
The grid or matrix arrangement is useful within switching and similar equipment (see
Sections 3.3 and 3.4 in this chapter) because it enables any horizontal link to be
connected to any vertical link. The delta is shown for completeness, although it is
really a special case of a mesh. In practice, telecommunications networks usually
comprise a combination of topologies in order to minimise the cost, while producing
an appropriate level of resilience. For example, a star network may be used to minimise
link costs, but with some meshing added between certain nodes to give the network
a degree of resilience to link failures.
The main service features are provided in the nodes of telecommunications networks. It is for this reason that network operators (other than virtual network operators)
need to own and control all their nodes, while none of the transmission links needs
to be owned and can instead be leased from other network operators. The following
sections introduce the most common component functions that are used to provide
the service features in the network nodes.
Nodal: concentrator switching
The concept of telephone switching is introduced in Chapter 1, where the switch
is shown in Fig. 1.6 providing connections between a particular input circuit to a
required output circuit for the duration of the call. Actually, the single box marked
Network components
per channel
= 40%
per channel
= 4%
per channel
= 40%
Figure 3.2
per channel
= 30%
(a) Concentration Switching. (b) Route Switching
‘switch’ comprises two types of switch, i.e. concentrator and route (also known as
‘group’ or ‘interconnect’), as described in more detail in Chapter 6. The purpose of
concentration in a network is to achieve cost savings in equipment by taking calls (or
traffic) from lightly loaded subscriber lines – over which only a few calls a day may be
made – and switching these on to fewer, but consequently more highly loaded lines.
Fig. 3.2(a) illustrates the case of the calls from 2,000 lines on the input side of the
switch, each line having an average occupancy of 4 per cent, being concentrated onto
200 lines with 40 per cent occupancy. This concentration ratio of 10 to 1 is typical
for a telephone concentrator switch.
Although the concentrator switch described above has improved the occupancy of
the lines by a factor of 10, it has introduced the possibility of subscribers not being able
to get through since calls from only 200 of the 2,000 subscribers can be carried at any
one time. For the time that any calls are unable to get through, the switch is deemed to
be in ‘congestion’ and these calls are considered as ‘lost’. Clearly, the probability of
lost calls may be determined statistically. Normally, this probability is low, typically
only about 1 or 2 per cent during the busiest time of day. The network operator has to
make an economic trade-off between keeping the acceptable probability of lost calls
as low as possible against gaining as much cost saving in equipment and network
capacity by having as high a concentration ratio as possible.
Nodal: route switching
Route switching provides the interconnection of calls from the set of input circuits to
an equal number of circuits on the output side, as shown in Fig. 3.2(b). The role of
Understanding telecommunications networks
the switch is to provide connectivity at an exchange between either internal lines with
high occupancy (40 per cent in the example of Fig. 3.2(b)) or external lines, also with
high occupancy, going to other exchanges. Trunk, junction tandem and international
exchanges comprise only route switches, since they do not have any subscriber lines
with their low occupancy, and therefore no concentration switching is needed.
Route switches, which are sometimes known as ‘group switches’, introduce no
possibility of losing calls (or traffic) if working within the dimensioned load, i.e.
normal conditions.
Nodal: packet switching and routeing
Data is fundamentally different to voice in its nature and origin. The primary difference is that data is information represented in the form of letters and numbers which
are generated, stored and displayed by devices, e.g. a computer; whereas, voice is
generated as a sound by humans. The consequence of this as far as telecommunication
networks are concerned is that voice originates as an analogue signal (i.e. a varying
waveform that copies the sound pressure pattern of the speaker) and data originates as
a digital signal (i.e. a string of numbers). We look at the important distinction between
analogue and digital signals and their attributes in more detail later in this chapter.
The need to convey data may be on the basis of a slow trickle, as in the case of a
telemetry signal (e.g. the remote measurement of the height of water in a reservoir), or
a very high-speed voluminous flow, as in the case of the transfer of a file between two
computers. The need for data communication between two points may be continuous
during a session, or intermittent and in bursts, or a single one-off enquiry. Furthermore,
much of the data is not time critical, and it can therefore tolerate some delay before
it is delivered to the far end, e.g. e-mails.
Data can be conveyed over a telephone network, provided that appropriate
modems are used to convert the digital signal to an analogue signal for transfer
over the copper access lines. Indeed, the first form of public switched data services
did use the PSTN, in which a telephone call was established between a terminal and a
computer. This service was known in the United Kingdom as ‘Datel’, meaning ‘data
over telephone’. Similarly, as described in Chapter 2, data sessions with the Internet
are also provided over the PSTN between the user and the ISP. However, a continuous
call connection as provided by the PSTN, in which a circuit is held for the duration
of the call (defined as ‘circuit switched’), is not an efficient way of conveying data
that might be bursty and intermittent. Also, the call set-up time of circuit switching
may be proportionately too long in the case of short rapid-enquiry type data transfer.
These issues, together with the fact that the charging and numbering arrangements of
the PSTN are optimised for telephony voice service, mean that the alternative use of
packet networks for data is usually preferred.
In a packet network the data to be conveyed is split into conveniently sized
segments, each associated with an address, to form packets. These packets, which
being digital are in the form of a string of 1’s and 0’s (i.e. ‘binary’), may be of fixed
or variable length and sent on a regular or on an as-and-when basis, depending on the
Network components
Continuous circuit connection
Payload (contents)
Figure 3.3
Packet and Circuit Switching
Output packet streams
Input packet stream
Figure 3.4
Packet Switch or Router
type of packet system. Fig. 3.3 illustrates the difference between a circuit and a packet
connection. Note that the circuit is held for the duration of the call, irrespective of
whether either person is speaking, while in contrast the packets need be sent only when
data is present during a session between two terminals/computers. The advantage of
a packet system compared to the continuous call connection system of a PSTN is that
packets from different users can be interleaved on each circuit between two packet
switches, thus giving improved utilisation and hence lower cost than the equivalent
single occupancy of a circuit-switched (i.e. PSTN) connection.
Fig. 3.4 illustrates the principle of a packet switch. The input line to the
switch, which comes from another switch A, carries a stream of interleaved packets.
The addresses in the header of each packet are read in turn and the packet is passed
Understanding telecommunications networks
to the appropriate output line. In the example of Fig. 3.4 all packets with addresses
‘x’ and ‘z’ are sent out to the line to switch C, while all packets with address ‘y’ are
passed to the line to switch B. The way that the packets are sent through the network
differs according to whether a predetermined path is set up for all packets belonging
to a session between two end points, known as a virtual path, or whether each packet
follows an independent routeing. The former is usually referred to as ‘packet switching’, the latter as ‘packet routeing’. Chapter 8 describes data networks and the roles
of packet (or cell) switches and routers more fully.
Nodal: control (computer processing and storage)
There is a category of nodal functions associated with the control of calls or the
provision of additional service features which can be used by callers. These functions
are provided by systems using computer processing and storage. Such systems, which
are referred to as ‘intelligent networks’, are usually located at several centres in a
national network, generally housed in exchange buildings. Calls requiring the use of
such facilities are handled in the normal way in the PSTN until one of the control
systems of the exchanges detects, e.g. from the number dialled, that more advanced
call control and service features are required. That exchange will then request further
information and instructions from the network intelligence so that the call can be
completed. Chapter 7 describes the role of these systems in more detail.
Nodal: multiplexing
One of the most common nodal functions of a telecommunications network is that
of multiplexing, i.e. the carrying of several tributary streams over one high capacity
bearer. Fig. 3.5 illustrates the example of ten tributaries, each with one channel’s worth
of capacity being multiplexed onto a 10-channel bearer. In general, a multiplexed
system comprises an m:1 multiplexing node, a multiplexed bearer (link) and at the
far end a 1:m de-multiplexing node. The multiplexed bearer may be a transmission
link between two cities, extending to several hundreds of kilometers or a bus (i.e.
a common highway) within an exchange switching system, extending just a few
metres. Either way, the multiplexing system provides important reduction in the total
network cost due to the economies of scale of higher capacity bearers. The important
characteristic of a multiplexed system is that the capacity of the multiplexed bearer
equals that of the input and output tributaries, so there is no concentration and hence
no loss of traffic.
There are many ways in which multiplexing may be achieved. The most commonly used methods in telecommunications are described below, using the example
of telephony. The specification of any multiplexing system depends on the frequency
characteristics of the signals to be carried (e.g. voice, video, music-quality sound,
broadcast colour television, data). The concept of frequency and waveforms is introduced in Chapter 1. In the case of telephony, the sound power of the human voice
Network components
10 channels on one bearer
A multiplex system
Figure 3.5
300 Hz
Figure 3.6
3,400 Hz 10,000 Hz
Waveform Analysis of Speech
throughout the frequency range is characterised in Fig. 3.6. The sound wave power
spectrum is low at frequencies below 400 Hz, and rises to a peak around 3–4 kHz, and
progressively drops off with increasing frequency. Of course, the exact shape varies
between individuals, with louder speakers having more power in their voice and hence
the curve is higher up the Y axis. Also, female speakers have their frequency-power
Understanding telecommunications networks
curve shifted to the right due to the higher pitch of their voice, while male speakers
have their curve shifted to the left of the average.
It has been agreed worldwide that a band of frequencies between 400 Hz and
3.4 kHz should be allocated within a PSTN to each telephone connection. This band
represents a compromise between capturing sufficient of the voice power spectrum
to ensure acceptable reproduction of the speaker’s voice on the one hand, and the
cost of providing the capacity on the other. In practice an absolute band of 4 kHz
(0 Hz–4 kHz) is actually allocated to each channel, with the 3.4 kHz point representing
the level where the output power is halved and the attenuation rapidly approaches
cut off at higher frequencies. Whilst passing this relatively narrow band of the voice
frequencies manages to capture the bulk of the power spectrum, the absence of the
frequencies above 4 kHz causes not only some loss to the quality of the voice, but
also more importantly some of the consonants become indistinct. Thus, speakers over
the telephone are often required to say ‘S for sugar’, etc., in order to overcome the
lack of precision caused by the loss of the higher frequencies.
In describing the commonly used types of multiplexing in telecommunications,
it is useful to consider how the dimensions of time and frequency are allocated to the
channels within the multiplexed system.
3.7.1 Frequency division multiplexing
With frequency division multiplexing (FDM) the frequency dimension is divided
into 4 kHz bands, each channel occupying one band. Fig. 3.7 illustrates the FDM
concept, showing that each of the five channels occupies just its allocated band of
frequencies for all time. The bandwidth of this FDM system is 5 × 4 kHz = 20 kHz.
Inter-exchange transmission systems introduced from the 1950s onwards (and still in
use in many parts of the World) used FDM systems based on a 12-channel format,
known as a ‘group’. The FDM technique is also used for radio and TV broadcast,
with each radio or TV channel allocated to a different carrier frequency; the tuning
mechanism on the radio or TV set enables the user to select the carrier frequency and
the associated band of frequencies for that channel.
4 kHz channel
Figure 3.7
Frequency-Division Multiplexing (FDM)
Network components
TDM frame
Ch1 Ch2 Ch3 Ch4 Ch5 Ch1 Ch2
Time slot
Figure 3.8
Time-Division Multiplexing (TDM)
3.7.2 Time division multiplexing
With time division multiplexing (TDM) only small slices of time are allocated to
each channel and during these times the full bandwidth of the TDM highway is made
available to that channel. The concept is illustrated in Fig. 3.8. Each of the five
channels is sampled sequentially and passed through the multiplexed system, and
thereafter the five channels are repeatedly sampled in sequence. The time period for
all five samples is called the ‘time frame’, and the time allocated to each sample is
known as a ‘time slot’. Slicing up a voice call into these periodic samples, to be carried
over their respective time slots, will not introduce any impairment to the quality of
the voice heard at the receiving end provided that the samples are taken frequently
enough. The rule for the required sampling rate, attributed to Nyquist and also known
as the ‘Sampling Law’, states that the minimum rate of sampling is twice the highest
frequency of the signal to be carried on each channel [1].
For telephony, where each channel is frequency-band limited to an absolute top
frequency of 4 kHz, as described above, the necessary Nyquist rate of sampling is
8 kHz, (i.e. 8,000 times/s) or once every 125 microseconds (millionths of a second,
written as ‘μs’), i.e. 1/8,000th second. Thus, for the 5-channel TDM system shown in
Fig. 3.8, the second sample of channel 1 needs to be made 125 μs after the first sample;
and subsequent samples need to be sent at 125 μs intervals thereafter. Similarly, the
samples for channels 2, 3, 4 and 5 also need to be sent at this rate. The time frame,
therefore, for all telephony TDM systems, irrespective of the number of channels
carried, is always 125 μs long. In this example of a 5-channel TDM system each time
slot occupies a fifth of the time frame, i.e. 25 μs.
The principle of sampling and TDM is illustrated in more detail in the three
picture sequence of Fig. 3.9 in which, for visual clarity and simplicity, a 3-channel
Understanding telecommunications networks
Sampling gate
Ch1 samples
TDM highway
Sampling gate
Ch1 samples
TDM highway
Ch2 samples
Sampling gate
Ch1 samples
TDM highway
Ch2 samples
Figure 3.9
Ch3 samples
Principle of TDM
Network components
TDM system is used as an example. The voice signal on channel 1 is taken through a
low-pass filter to eliminate all frequencies above 4 kHz before entering the sampling
gate, as shown in Fig. 3.9(a). A string of pulses, P1, each separated by 125 μs opens
the sampling gate at the sampling intervals and the stream of samples of the bandlimited voice on Ch1 pass onto the TDM highway. Notice how each of the samples
on the highway equals the height of the Ch1 waveform at the respective instant of
sampling. At the output of the TDM highway a second sampling gate is opened at the
appropriate instants by another stream of P1 pulses to allow the sequence of samples
of Ch1 to pass. These samples are then taken through a low-pass recovery filter which,
in effect, smoothes the sequence of samples into a continuous waveform – a replica
of the input to the sampling gate at the beginning of the TDM system.
Now we can consider adding the second channel to the system, as shown in
Fig. 3.9(b). The Ch2 input waveform after being band-limited to 4 kHz is sampled
under the control of a string of pulses P2, which are 125 μs apart but positioned a third
of a time frame (i.e. 41.7 μs) later than the P1 set of pulses. The samples from Ch2 are
then passed to the TDM highway, interleaved in the gaps between the samples from
Ch1. Application of the second stream of P2 pulses then opens the output sampling
gate at the appropriate instants so that the samples of Ch2 may be taken through the
output low-pass filter in order to recover the Ch1 waveform.
Finally, Fig. 3.9(c) shows how the full 3-channel TDM systems works. The stream
of P3 pulses, also 125 μs apart, but staggered by 41.7 μs from the P2 stream of pulses,
operates the sampling gate to allow the appropriate samples of the band-limited Ch3
waveform onto the TDM highway interleaved between the samples of Ch2 and Ch1.
At the end of the TDM highway the second stream of P3 pulses opens the output
sampling gate allowing the Ch3 samples to pass to the recovery filter.
Clearly, for this TDM system to operate without interference between channels the
input and output pulse trains must be precisely timed. Thus, for example, the output
set of P2 pulses must occur at the same instants as the input P2 pulses otherwise some
portion of the samples of either the preceding Ch1 or the following Ch2 will pass
through the CH2 output sampling gate and corrupt the final result of the Ch2 output.
TDM systems in general, therefore, need a mechanism to synchronise and align
the corresponding streams of input and output sampling pulses. TDM transmission
systems are described in more detail in Chapter 4.
TDM is employed by the majority of transmission systems used in today’s
telecommunications networks (with the TDM multiplexors at the network nodes)
as well as within the exchange switching (nodal) equipment. A derivative of TDM,
known as time division multiple access (TDMA), is used by telecommunications satellite systems and GSM phone networks to allocate calls to available radio channels,
as described in Chapter 9.
3.7.3 Code division multiplexing
With code division multiplexing (CDM) the full frequency and time capacity of the
CDM system is made available to all the channels, as illustrated in Fig. 3.10. This
technique is more difficult than either FDM or TDM to explain. The best way to
Understanding telecommunications networks
Figure 3.10
Code-Division Multiplexing (CDM)
understand the principle of operation is to consider the analogy of a cocktail party.
At such parties there are many people talking together in small groups around the
room. Anyone at the cocktail party can join a group and, despite the general high
level of background chatter, they would be able to concentrate on the conversation
of their group by tuning in to the particular voices involved. During the conversation
the person would also be able to monitor what is being said elsewhere in the room by
momentarily tuning in to the various groups within earshot – just in case something
more interesting is being said! Similarly, CDM allows all the channels to talk at once
on the multiplexed highway; separation is achieved by applying a different binary
code to each channel at the input and using the appropriate code at the far end to select
each channel – all other channels appearing as unintelligible background noise. Just
as with the cocktail party, there is a limit to the number of channels a CDM system can
convey because the increasing background noise will reach a point where it swamps
any individual channel (i.e. the ‘signal-to-noise ratio’ becomes unacceptable).
Like the other multiplexing systems, the CDM system needs to synchronise the
binary codes used at the input and outputs of each channel. Also, each input channel
is band limited to 4 kHz before being converted to digital (see Section 3.11) prior
to entering the CDM system. Multiple access (MA) systems based on CDM, i.e.
CDMA, are used extensively in second generation digital mobile phone systems [2],
particularly in the United States, and a wideband variant is used by the new third
generation phone systems around the World, as described in more detail Chapter 9.
Nodal: grooming
The grooming function is an important component in minimising network transmission cost where several services are carried over lines from many different sources,
e.g. as in the subscriber access network. Fig. 3.11 illustrates the principle of grooming.
Network components
Figure 3.11
Multiplexed bearers from nodes i, ii, iii and iv all terminate at a node at which the
grooming system is located, each bearer carries circuits (service A) or (service B)
which are destined for either v or vi nodes. Typically, nodes i–iv would be big customer sites and service A could be private circuits dealt with by node v, and service
B might be a data service dealt with by node vi. The grooming system permanently
assembles all the channels carrying service A (21 channels in total) and passes them
to the output line to v and assembles all the channels carrying service B (29 channels in total) and passes them to the line to vi. Note that the total input capacity is
equal to the total output capacity, so there is no traffic loss. Groomers operate on a
semi-permanent basis, where the reconfiguration of channels is controlled by a local
equipment-management system within the exchange building.
Nodal: consolidating
The consolidation function, as the name suggests, is the packing of the occupied
circuits on a number of tributaries onto one more-fully filled output line transmission system. As Fig. 3.12 illustrates, the consolidation of the occupied capacity of
three transmission systems onto a single output system carrying all 22 circuits would
provide network economies. The core transmission network of a typical telecommunications operator undertakes consolidation wherever possible at the transmission
nodes, as described in Chapter 5. This act of consolidation is also common in many
of the other types of networks, for example, the consolidation of containers on cargo
ships, which are re-stacked onto other ships at strategic ports around the World.
Link component
All of the components described so far in this chapter relate to functions undertaken
at the nodes of a telecommunications network. We will now consider a generalised
Understanding telecommunications networks
Figure 3.12
Transmission system
conversion of
input signal
to transmission
signal format
Figure 3.13
Transmission medium:
propagation of signal
over medium (e.g. line code,
conversion of
received signal
to output signal
Link (Transmission) Functions
description of the link transmission component. The link comprises three parts, as
shown in Fig. 3.13. The transmitter (‘Tx’) converts the electrical (usually) output from
the sending node into the format required for the type of transmission link. Examples
of such formats are special digital signals designed for propagation over a copper line
or coaxial cable, the digital pulsing of light for transmission over optical fibre or the
generation of a radio signal for a radio (mobile or fixed) or satellite transmission link.
The transmission medium part comprises the copper cable, transverse-screen cable,
coaxial cable, optical fibre or in the case of a radio system, the air, over which the
telecommunications takes place. This part may also include repeater and regenerator
equipment (located below ground, above ground or even in outer space) along the
route of the transmission medium which boosts and corrects the signal, compensating
for the loss and distortions incurred. At the far end, the receiver (‘Rx’) performs the
reverse of the Tx, converting from the format required for the transmission medium
to the format of electrical signal required by the terminating node.
Network components
Figure 3.14
A Transmission System Incorporating the Nodal Function of Multiplexing
Since the cost effectiveness of a transmission link system is higher the larger its
capacity – in terms of the unit cost per circuit (provided the capacity is used) – practical
transmission systems are usually associated with multiplexing equipment. Fig. 3.14
illustrates the generalised arrangement for a transmission system incorporating the
nodal functions of multiplexing. It should be noted that because electronic devices
are unidirectional, the combined transmission/multiplexing arrangement is a 4-wire
system, with separate Go and Return paths.
The configuration of Fig. 3.14 can be optimised by using a further stage of multiplexing so that the Go and Return transmission signals are sent over the same
transmission medium. Hence, for example, a single optical fibre (which is a bidirectional medium) could carry both the Go and Return transmission signals by using
FDM. However, it should be noted that if this optical fibre system had a repeater using
unidirectional electronics, the Go and Return signals would need to be separated and
recombined at the entry and exit of the repeater.
3.11 Analogue-to-digital conversion
3.11.1 The advantages of digital networks
Chapter 1 introduced the concept of the analogue signal generated by the microphone
in a telephone handset receiving air pressure variations from a speaker and the subsequent conversion back to air pressure variations by the earpiece at the receiving
telephone handset. Fig. 3.6 considered earlier in this chapter also shows the characteristics of an analogue signal. Originally, the PSTNs around the World were fully
analogue; they comprised analogue transmission systems in the access and core parts
Understanding telecommunications networks
of the network, using a variety of FDM-based systems to gain economies, together
with analogue switching in the local and trunk exchanges. Digital transmission systems (PCM, as described later) using TDM, were first introduced around the mid 1960s
into the shorter-distance inter-exchange routes (so-called junction and toll routes), as
a way of providing multiple circuits over copper pairs. Later, digital transmission was
introduced onto the long distance coaxial-cable routes.
Digital transmission not only provides a cost-effective multiplexing and transmission technique, but it also greatly improves the quality and clarity of the circuits –
providing immediate benefit to the telephone users. This quality improvement is due
to the ability of digital systems to regenerate the binary 1’s and 0’s where necessary
in the network to compensate for the deterioration of the conveyed signal, unlike
analogue systems where use of amplification to boost the signal equally affects the
background noise, causing the impairments to increase with distance. We are familiar in every day life with the merits of the quality of digital systems compared to
the analogue equivalents, e.g. (digital) CDs and (analogue) audio cassette tapes, and
(digital) DVDs and (analogue) VHS video cassette tapes.
However, it was not until the late 1970s and early 1980s that digital switching was
used to replace the analogue switching systems, initially the trunk then later the local
exchanges. The primary driver for introducing digital switching was to be able to
make fully semiconductor electronic exchanges, with the consequential reduction in
manual operational costs, lower capital costs, less accommodation required, improvements in clarity of the connections and the potential for new services and features.
Progressively, the analogue switching and transmission systems have been replaced
by digital systems so that since the mid-1990s nearly all PSTNs in the developed
World are entirely digital.
A PSTN with digital transmission and switching is known as an integrated digital
network (IDN). There are several major advantages of an integrated digital PSTN,
compared to an equivalent analogue PSTN, as summarised below [3].
(a) Capital cost savings due to the all-electronic nature of the equipment enabling
the cost and performance benefits of semiconductor computer technology to
be exploited. In addition, the integration of the switching and transmission
technology (all digital) eliminates the need for costly inter-working equipment.
(b) Operational cost savings resulting from the significant reduction in floor space
taken up by digital electronic equipment compared to that required for the
bulky electro-mechanical and semi-electronic analogue switches. This leads to
a reduction in the accommodation costs for the network operator.
(c) Operational cost savings, primarily in manpower, due to the virtual elimination of manual adjustments to the equipment and the reduction in the overall
maintenance load.
(d) A high level of transmission quality, in terms of clarity for the user, lack of
background noise and a constant loudness irrespective of the distances involved
or the number of exchanges in the connection.
(e) The ability to easily mix different types of services on the same line by virtue of
the use of digital conveyance for both voice and data, i.e. ‘integrated services’.
Network components
Key :
Digital SPC Telephone Exchange
Digital transmission (PCM for speech)
Figure 3.15
Analogue line
Integrated Digital Network
Fig. 3.15 shows the concept of an IDN. Despite the merits of digital working, the
ubiquity, cost effectiveness and adaptability of the copper local loop means that the
majority of PSTN subscriber lines are still analogue. Thus, the IDN begins at the
entry to the digital local exchange (DLE) where the analogue-to-digital conversion
(‘A/D’) is provided and extends to the exit of the local exchanges, where the analogue
line is connected via a D/A conversion.
The provision of ISDN service adds a digital line system over the local loop
together with a digital termination at the subscriber’s premises, thus in effect extending the IDN to cover the end-to-end connection. The term ISDN means ‘integrated
services digital network’ – a name derived from ‘IDN’ and the attribute (e) above.
The advantages of the IDN listed above then extend directly from the customer’s
terminal equipment, e.g. a computer connected digitally to the ISDN termination at
the desktop.
3.11.2 The A/D process
As previously discussed, an analogue electrical waveform or signal is, by definition, an analogue of an input waveform, i.e. the air pressure variation caused by a
speaker. An analogue waveform also has another defining attribute, namely, that it
may take any value within the permitted range. In contrast a digital signal is a series
of numerical representations of an input waveform and, importantly, the digital signal can take only a limited number of values, i.e. it has a fixed repertoire – in the
case of a binary digital signal this number is two. (It is the use of binary values that
enables digital transmission to achieve such good quality performance irrespective of
distance, since the signal can be periodically cleaned up and regenerated by comparing the impaired signal with just one threshold, unlike the analogue signal in which
impairments accumulate with distance and cannot be corrected.)
Fig. 3.16 shows a simple illustration of the process of A/D conversion. First, the
analogue input signal must be passed through a low-pass filter to limit the maximum
Understanding telecommunications networks
Speech samples (still analogue)
Binary representation of samples
Figure 3.16
Analogue-To-Digital Conversation
frequency of the waveform, as described earlier for the process of TDM. The waveform is then sampled at or above the Nyquist rate to produce a series of analogue
samples or pulses, whose heights represent the waveform at the sampling instants.
(This stream of pulses is known as ‘pulse amplitude modulation’, PAM.) These analogue samples are then converted to digital by the comparison of each sample with a
digitalising ruler. The digital representation is the numerical value of the graduation
on the ruler at or just above the PAM pulse height, as shown in Fig. 3.16. This coding
process generates a binary number for each PAM pulse, the digital output of the D/A
Conversion back to analogue from digital operates in reverse to the D/A described
earlier. However, in converting each of the binary numbers to the PAM height given
by the graduation on the ruler (i.e. decoding) there is an inherent error introduced
when the regenerated PAM sample is compared to the sent original PAM sample. This
is because of the finite number of graduations on the ruler, the binary number being
the next closest reading to the input PAM sample in the A/D process, yet the binary
number will generate a PAM sample exactly at the graduation level on the ruler during
the D/A process. The amount of difference between the sent and regenerated PAM
sample clearly varies between zero and half the distance between ruler graduations.
This is known as the quantisation error, and it is perceived by a listener as quantisation
In practical systems, the quantisation noise is kept acceptably low by using a large
number of steps, or quantisation levels, on the ruler. In addition, the quantisation
noise is kept proportional to the level of input signal by bunching the levels closer
at the lower end of the ruler and wider towards the top – following a logarithmic
progression. The A/D process, described earlier, is performed by a digital coder and
the D/A process is provided by a decoder; often the equipment needs to deal with the
Go and Return direction, and so a combined digital coder/decoder or ‘codec’ is used.
Network components
or switching
PCM words
Figure 3.17
The PCM Process
The complete A/D-and-D/A process for a telephone network, which is known as
‘pulse-code modulation’ (PCM), is shown in Fig. 3.17. This figure plots the various
representations of the signal, from analogue waveform, to PAM samples, to digital
numbers or PCM ‘words’ and back. Notice, that once converted to digital (PCM) the
signal may pass through any number of digital transmission systems and switching
exchanges across the network.
The A/D process is used for a number of applications in every day life in addition
to telephone networks. As mentioned earlier, music is digitally coded onto CDs, and
films are scanned and coded onto DVDs. Similarly, TV broadcasts are also increasingly being used in preference to analogue transmissions. Later in this book we will
consider how the trend towards the use of digital in fixed and mobile telecommunications, entertainment, information and computing creates possibilities for convergence
and what this means for the next generation of networks.
In this chapter we looked at the basic components used in the links and nodes of
telecommunications networks. Beginning with the nodes, we considered the two
types of circuit switching:
Concentrator switching, dealing with low-occupancy subscriber lines;
Route switching, interconnecting routes between concentrator units and with other
Then, we considered the alternative form of switching, generally known as packet
switching, which is now the preferred way of conveying data. The remaining nodal
functions considered include:
Control nodes;
Multiplexing (FDM, TDM and CDM);
Understanding telecommunications networks
A generalised view of transmission links was then described, noting that within the
switched and core transmission network these are 4-wire systems with Go and Return
channels, whereas in the copper access network 2-wire circuits are used for economy.
Then we considered the advantages of an IDN and how the IDN can be extended
over the access network to the subscriber using ISDN. Finally, the important A/D
conversion process was described.
REDMILL, F. J. and VALDAR, A. R.: ‘SPC Digital Telephone Exchanges’, IET
Telecommunications Series No. 21, Stevenage, 1995, Chapter 4.
ANTTALAINEN, T.: ‘Introduction to Telecommunications Network Engineering’, Artech House, Norwood, MA, 1998, Chapter 5.
REDMILL, F. J. and VALDAR, A. R.: ‘SPC Digital Telephone Exchanges’, IET
Telecommunications Series No. 21, Stevenage, 1995, Chapter 21.
Chapter 4
Transmission systems
This chapter describes the wide variety of transmission systems used in telecommunications networks today. In doing this, we will be drawing upon the concepts of
multiplexing and analogue to digital conversion introduced in the previous chapter.
Earlier chapters have also introduced the important distinction between the access
network, which serves the subscribers, and the core transmission network, which
provides links between network nodes only. In general, the transmission systems
described in this chapter may be deployed in both the access and the core transmission networks – although they are usually more appropriate to one or the other, and
this will be indicated. Chapter 2 introduced the model of the networks associated with
the PSTN (Fig. 2.11), in which the access and core transmission networks act as a
common utility for providing circuits to the various specialised networks as well as
the PSTN. However, we will leave the description of how the various transmission
systems are deployed in the access and core transmission networks until Chapter 5.
Transmission bearers
In this section we look at the general principles of transmission and the range of
practical systems used in a telecommunications network.
Transmission principles
The purpose of a transmission system is to provide a link between two distant points
or nodes. The link may be unidirectional, as in the case of radio broadcast where
transmission is only from the transmitter to the radio receivers, or bidirectional, as in
the case of a telephone connection with its two-way conversation. In telecommunication networks the links between nodes are either 2- or 4-wire circuits, as described
Understanding telecommunications networks
Modulating signal
Figure 4.1
Modulated carrier
Carrier-Wave Modulation
in Chapter 1. With 4-wire circuits, the separate ‘Go’ and ‘Return’ paths are considered as transmission channels, as explained in the generalised view of a multiplexed
transmission system in Fig. 3.14 of Chapter 3.
In general, a transmission system takes as an input the signal to be conveyed, which
may be a single channel or a multiplexed composite of channels, converts that signal
into a format suitable for the transmission medium, propagates the transmission signal
to the far end, where after conversion from the transmission signal a reproduction
of the input signal is produced (Fig. 3.13 of Chapter 3). The actual conveyance of
the channel over the transmission system is achieved by ‘modulating’ an electrical
bearer signal which travels over the transmission medium. It is the modulation or
modification of the bearer signal that constitutes the transmitted information, not the
bearer itself – indeed, the bearer signal is usually called a ‘carrier’ for this reason.
There are many forms of modulation [1]. The normal form of carrier is a sinusoidal
waveform, as introduced in Chapter 1, Box 1.1 and Fig. 1.11. Some of the common
forms of modulating sinusoidal carrier waveforms are:
Amplitude modulation (AM): in which the input signal directly varies the height
or amplitude of the carrier, shown in Fig. 4.1.
Frequency modulation (FM): in which the input signal directly varies the
frequency of the carrier, as shown in Fig. 4.1.
Phase modulation (PM): in which the input signal directly varies the phase (i.e.
positioning of the start of the wave) of the carrier.
All the three methods are used widely in radio and TV broadcasting, where a radio
carrier is modulated, and most people would recognise the commonly used terms of
‘FM’ and ‘AM’. They are also used in telecommunications transmission systems, as
described later. However, these modulation systems are analogue because the carrier
Transmission systems
is continuously moderated by the input signal and the carrier can take any value within
the working range of the system.
A digital carrier, in the form of a stream of pulses can also be modulated, taking
a fixed number of values, by an input signal. Examples of such modulation schemes
Pulse amplitude modulation (PAM): in which the heights of each carrier pulse is
moderated by the corresponding bits in the digital input signal.
Pulse position modulation (PPM): in which the position of each carrier pulse is
moderated by the corresponding bits in the digital input signal.
Pulse code modulation (PCM), described in Chapter 3, whereby a binary stream
is modified to convey the successive digital representation of a set of TDM (PAM)
speech samples.
These digitally modulated streams may themselves also be used to modulate a radio
or optical carrier (OC) on a transmission system. The input signal, in the form of
pulses, is then used to modulate the carrier in a series of step changes between a fixed
set of values. This is known as ‘keying’. Examples of such systems include: phaseshift keying (PSK) – in which the output signal steps between two or more phases
in response to the binary modulating signal; and frequency-shift keying (FSK) – in
which the output shifts between two or more frequencies in response to the binary
input. The advantage of these keying modulation systems is that the output signal can
step between more than just two values, e.g. 4, 8 or even 16 levels of phase can be
used in what is known as 4-, 8- or 16-phase PSK, respectively. In the example of the
4-phase PSK system the input modulating digital stream is taken two-bits at a time
(i.e. 00, 01, 10 or 11) to set the appropriate phase level in the output signal. Similarly,
three bits at a time (000, 001, 010, 011, 100, 101, 110 or 111) are taken for 8-phase
PSK, four bits at a time for 16-phase PSK, and so on.
All transmission media introduce some impairment to the signal being conveyed
due to the physical mechanisms involved. Examples of the most common forms of
impairment are:
Loss of power, due to the absorption, scattering and reflections of the signal within
the transmission medium, known as ‘attenuation’.
Delay to the signal, due to the propagation through the medium. This will vary
according to the characteristics of the medium (e.g. the purity of the glass in an
optical fibre) and external factors such as temperature.
Phase distortion, due to the different amount of delay introduced by the transmission medium to each of the component frequencies in a multi-frequency
transmitted signal. The resulting spread in the arrival time of the waveform is
known as ‘dispersion’.
Amplitude distortion, resulting from the different loss of power introduced by the
transmission medium for each of the component frequencies in the transmitted
Understanding telecommunications networks
Noise, the pick up of unwanted signal generated intrinsically by the medium (e.g.
thermal noise), as well as induced noise from outside sources (e.g. crosstalk from
adjacent transmission systems).
The design of transmission systems aims to minimise and compensate for the various
impairments introduced by the medium. It should be noted that the extent of the
above-mentioned impairments may change with the temperature of the transmission
medium. There are methods for compensating for phase dispersion and amplitude
distortion, known as equalisers, for example. However, these are effective only over
a defined operating range. Thus, there will be limits on the length of a transmission
system within a network if it is to operate within the accepted range of loss, delay,
noise and phase distortion. There will also be a limit to the bandwidth (range of
frequencies carried) that may be used. Normally, the usable length of a transmission
system is increased by introducing devices located appropriately along its route to
boost the signal strength, compensating for the power loss.
In the case of an analogue transmission system the signal accumulates the
impairments along the length of the medium, so the introduction of an amplifier
to compensate for power loss results in not only the wanted signal being boosted but
also that of the added noise. The ratio of the wanted signal to noise (‘signal to noise
ratio’) gets progressively worse at each point of amplification along the route of the
transmission system – so setting a limit on the acceptable number of amplifiers and
hence length of transmission link. By contrast, a digital transmission system is able
to reconstitute a noise-free signal at each point of regeneration along the route. This
is illustrated in Fig. 4.2, where the impaired signal from the line is sampled by the
digital regenerator and any signal above the threshold at the sampling instants will
result in a clean digit ‘1’ being generated and sent to line. (See also Chapter 3 and
Fig. 3.15 for a description of how this feature is exploited in an IDN.)
Fig. 4.3 illustrates the components of a generalised digital line transmission system. This is a development of the link functions shown in Fig. 3.13 of Chapter 3.
Sampling instants
Input to
Output from
Binary value:
Figure 4.2
Digital Regeneration
Transmission systems
One direction
Figure 4.3
Digital Line Transmission System Components
The input digital signal in the Go direction is shown in the left-hand side; this is
either a multiplexed assembly of many channels (SDH or PDH, see below) or one
very large capacity single channel. The signal is then converted to a format that suits
the transmission medium. This is achieved through the application of a line code to
the signal to be transmitted. Line codes are designed to convert the signal to one that
contains timing information that can be extracted by the regenerators along the line,
to ensure that they stay in synchronism with the sending end. Line coding may also
improve the efficiency of the transmission system by converting the two-level binary
signal to a multi-level signal for actual transmission over the medium. An example
of this was given with the use of four-phase PSK, earlier, where the binary signal
was converted to a four-level signal for transmission. This enables the transmission
system line rate (known as the ‘baud rate’) to be half that of the binary input signal.
In this way, a transmission system can carry higher capacities by using line codes of
increasing numbers of levels. In practice, there is a trade-off between the increased
capacity efficiency of the transmission system and the cost and complexity of the line
coding detection mechanism, which generally increases with the number of levels
used. There are many forms of line codes employed in transmission systems, each
offering different levels of efficiency against complexity and equipment cost and, of
course, suitability for the various transmission media [2,3].
Next, the line-coded signal is used to modulate the carrier for the transmission
system. In the case of an optical fibre system the signal modulates a laser sending a
corresponding stream of light pulses down the fibre (i.e. the medium) – undertaking
an electro-optical conversion. At each of the regenerators (only one is shown in
Fig. 4.3) the timing information (inserted by the line-coding process) is extracted and
used to drive the sampling gates in order to regenerate the signal, as described earlier.
At the far end the appropriate demodulation occurs, e.g. a photo detector in an
optical fibre system receives the transmitted light pulses and generates a stream of
digital electrical pulses (i.e. opto-electrical conversion). This electrical signal is then
trans-coded from the line code back to the format of the binary input.
Understanding telecommunications networks
(a) Copper pair in a multi-pair cable
Polyethylene and aluminium foil sheath
One pair with
polyethylene insulation
Optical fibre
Plastic sheath
protective cover
Central conductor
(c) Coaxial cable
(b) Optical fibre in a multi-fibre cable
Not to scale
Figure 4.4
Cable Transmission Media
The principles described above apply to all of the many forms of transmission
systems carried over the different forms of media – namely: metallic cable, optical
fibre cable and radio over free air. Given this wide range of possible transmission
options, the choice of system used in a network depends on the capacity to be carried,
the existing transmission infrastructure, the distance involved, expected capacity
growth potential and the cost of the equipment. We can now look at these aspects as
they affect a telecommunications network operator.
Transmission media
Fig. 4.4 illustrates the physical characteristics of the transmission media used in
modern networks.
Copper cable pairs
As described in Chapter 1, copper pair cable has been adopted throughout the World
as the means of providing the local loop between the subscriber and the serving
telephone exchange. But, the humble copper pair cable can also be used to carry a
variety of data high speed services, and even video, by the addition of appropriate
electronics (e.g. ADSL) at its ends, as described later in the chapter [4,5].
Copper pair cable is available as a single pair of wires, with plastic insulation,
usually deployed between a subscriber’s premises and the overhead or underground
distribution point (DP), as described in Chapter 5. However, the copper pairs are also
provided in multi-pair cable bundles of a variety of sizes (e.g. from 2 pairs up to
4,800 pairs). The thickness or gauge of the copper conductors also comes in a range
of sizes, with corresponding electrical resistance values per kilometre. Fig. 4.4(a)
Transmission systems
shows the general construction of a multi-pair cable. The individual pairs have a
thin insulating cover of polyethylene and these are grouped into quads in order to
minimise adverse electrical interference problems. The coating of the pairs is usually
colour coded to assist their identification by the technician joining the ends onto other
cables and terminals – quite a task when there are tens, hundreds or even thousands
of pairs in one sheath! The sheath is constructed from polyethylene for strength and
durability, together with aluminium foil to provide a water barrier.
Although not shown in Fig. 4.4, special forms of copper cable, arranged with one
central conductor surrounded by a copper cylindrically shaped conductor – known
as ‘transverse screen’ cable – were deployed in the United Kingdom and elsewhere
during the 1970s and 1980s to provide high-speed digital circuits in the access network
[5]. Now, optical fibre or digital transmission over standard copper pairs is used in
Optical fibre cable
Optical fibres provide a transmission medium of huge potential capacity by constraining pulses of light within a thin core strand of highly pure glass or silica. The light
stream is contained within the core by a cladding of glass which has a lower refractive index, causing the light to be totally reflected internally at the core-to-cladding
boundary [2]. The light is pulsed at the sending end by solid-state electro-optic transducers, such as light emitting diodes (LED) and lasers. Detection at the distant end
is by solid state opto-electric transducers, such as PIN and avalanche photodiodes.
The key requirement for an optical fibre transmission system is the use of a glass
(silica) fibre material which offers very low loss to the light passing through. This is
achieved through a rigorous manufacturing process in which the glass purity is strictly
controlled. In addition, advantage is taken of the transmission characteristics of the
glass which offers different attenuation to light of different frequencies, as illustrated
in Fig. 4.5 – the arrows indicate the two main windows at 1300 nm and 1500 nm for
optical fibre systems.
Optical fibre transmission has the following characteristics compared to metallic
cable transmission:
Huge potential capacity due to the inherently higher operating frequencies.
Low power loss (e.g. <0.5 dB/km), resulting in the need for few, if any, repeaters
along the route. See Box 4.1 for an explanation of the term dB used as a measure
of power levels within telecommunications networks.
The optical nature of the medium means that the transmission is unaffected by
electrical inference in the vicinity.
No escape of energy from the fibre, so there is no transmission interference
between fibres or other lines.
Physically smaller than metallic cables of comparable capacity.
Can be difficult to join optical fibres; special splicing techniques are required.
Fig. 4.4(b) shows the construction of an optical fibre within a multi-fibre cable.
The cables, which range in sizes (e.g. from 2 to 96 fibres), contain the fibres within
a plastic sheath about a central metallic wire, which provides mechanical strength
Understanding telecommunications networks
dB/Km 0.8
Wavelength (nm)
Figure 4.5
Box 4.1
Optical Fibre Operation
The Decibel Measure of Power
Power levels in telecommunications networks are measured in units known as
decibels (one tenth of a ‘Bel’). The important characteristic of dBs is that they
represent a logarithmic comparison of the measured power level against some
other power level. Thus, if the power output (P1) from a radio was turned up to
twice the level (P2), the increase in power is calculated as 10 log10 P2/P1 =
10 log10 2 = 3 dB. Since the measure is a ratio, it does not matter whether the
two powers (P1 and P2) are large or small – a doubling of power is always
3 dB and, by the same logic, a halving of power is −3 dB. The main advantage
of the use of this logarithmic measure is that all the increases and decreases
of power contributed by the elements of a circuit throughout the network are
simply added or deducted from the total to give the overall power level or
loudness rating.
However, where an absolute power level is to be measured, e.g. the output
from some electrical equipment, it is compared to a standard power level,
typically 1 mW (milliWatt). Powers are then shown with units of ‘dBm’. Thus,
a power level of 6 dBm would be equal to 4 mW, i.e. double times double
(3 dB + 3 dB) the power of 1 mW.
Transmission systems
for the otherwise very fragile bundle of optical fibres. Typical dimensions of the
optical fibres diameters are 50–200 μm core with 125–400 μm cladding for the
earlier (multi-mode) types to 5–10 μm core with 125 μm cladding for the more
recent (single mode) types. (A ‘μm’ is one millionth of a metre.) These fine glass
threads, which are about the thickness of a human hair, are wrapped in a polythene
protective cover for robustness.
Coaxial cable
As the name suggests, coaxial cable comprises a metallic central core cable encased
by a cylindrical conductor. The shielding effect of the outer conductor provides an
interference-free transport medium for high-speed electrical signals. Coaxial cable,
similar to that used for connecting aerials to a TV set, is used in the Cable TV networks to provide the video connection from the street electronics to an individual
household, as described in Chapter 2 (Fig. 2.2). Early forms of long-distance transmission networks for the PSTN employing analogue FDM systems were provided
over coaxial cables from the 1960s to the 1980s. Now, telecommunications networks
use optical fibre in preference to coaxial cables for the core transmission network.
Even though it is not now deployed in the external telecommunications networks,
coaxial cable is still used extensively for interconnecting transmission and switching
equipment within exchange and repeater station buildings. This interconnection is
usually provided across digital distribution frames (DDFs), as described in the section
on PDH (plesiochronous digital hierarchy) transmission in Section 5.3.2.
The structure of a coaxial cable is shown in Fig. 4.4(c). For physical flexibility
the shielding or screening conductor is made out of metallic braiding underneath a
plastic sheath. The screening is separated from the central core conductor by a soft
insulator so that the overall assembly of the cable is robust and flexible.
Atmosphere: radio
The medium of the atmosphere is unlike that of the confined environment of metallic
cables or optical fibres, and its unbounded nature requires the transmission systems
exploiting it to have a high degree of adaptability to the medium’s changing characteristics. Radio transmission through the air is achieved in several ways, mainly
depending on the frequency of the carrier wave used. The various types of radio paths
are illustrated in Fig. 4.6 and briefly described in the following paragraphs below [4].
Path type (a): Direct wave (also known as ‘free space wave’).
This path uses free space propagation through the air; the radio signal following
a straight line-of-sight (LOS) between antennas. The amount of transmitted power
received can be increased if the two antennas are as directional as possible. Since
the directional characteristics of antennas increase with frequency, practical systems
tend to be at microwave frequencies, that is above 1 GHz (1,000,000,000 Hz).
Path type (b): Ground reflected wave.
In addition to the direct wave there will inevitably be one or more reflected wave
paths between the sending and receiving antennas. Normally, these reflected paths
Understanding telecommunications networks
Not to scale
Figure 4.6
Different Types of Radio Paths
cause some degree of interference with the direct signal at the receiver because the
reflected signal will experience greater delay, and hence be out of phase with the
direct signal. LOS systems use a variety of methods for reducing the interference
from the reflected waves.
Path type (c) ground or surface wave.
Ground waves, which follow the surface of the Earth through the process of diffraction, are generated by electrical currents induced in the ground. Radio systems using
surface waves are able to extend beyond the LOS, following the Earth’s curvature. The
attenuation of surface waves increases with frequency, so practical systems operate
in the low frequency range (30–30 kHz) and, thus, are of low capacity.
Path type (d): skywave.
This path is created by the reflection of the radio waves off the Ionosphere, the
ionised layers belting the Earth. The waves can carry over very long distances through
multiple reflections off the Ionosphere and ground. Early long distant, i.e. intercontinental, communications was achieved using skywave radio systems operating
at HF frequencies. However, the variable nature of the Ionosphere and the use of
multi-path propagation meant that the transmission was subject to fading and daily
fluctuations. Thus, alternative transmission systems (e.g. optical fibre, microwave
radio and Earth satellite) are mainly used in preference to HF.
Path type (e): tropospheric scattering.
This path, which is created by the scattering of the radio wave by perturbations in
the atmosphere, enables communication to a receiver just over the horizon. Practical
Transmission systems
Box 4.2 The Way That Transmission Systems Use The Electromagnetic
1 mm
1 cm
10 cm
10 m
100 m
1 km
10 km
100 km
10 μm
300 THz
300 GHz
30 GHz
3 GHz
300 MHz
30 MHz
3 MHz
300 kHz
30 kHz
3 kHz
Micowave relay
radar and cellular
Free space,
Mobile radio and cellular
VHF TV and radio
Mobile radio
CB radio and
amateur radio
free space
AM broadcasting
Navigation and
transoceanic radio
Telephone and
systems operate at microwave frequencies using very high transmitting power and
large high gain receiving antennas to compensate for the very high attenuation of
tropospheric scattering paths. Such systems are used by the military to provide communication in battle conditions. BT use tropospheric systems to connect all the oil
rigs in the North Sea to mainland UK, thus extending the PSTN offshore beyond the
Box 4.2 summarises the various radio propagation techniques described earlier
and shows the radio spectrum frequency designations.
Telecommunications networks today use a mix of radio transmission systems.
Invariably, these operate at microwave frequencies in order exploit the higher bandwidth capacity and hence economy of scale that they provide. With the exception of
the tropospheric (or ‘troposcatter’) systems, these systems work on an LOS pointto-point basis to provide high capacity links for the core and access transmission
networks, or alternatively in a multipoint configuration in the access network linking
several customer locations to a central exchange.
Fig. 4.7(a) illustrates the configuration for a microwave point-to-point relay route,
providing transmission capacity between two antennas via an intermediate relay stage
(repeater or amplifier for digital or analogue transmission, respectively). The Go and
Return channels are conveyed on separate carrier frequencies. Such transmission
systems are able to provide routes across a country using many repeater antennas,
typically spaced at between 20–40 km apart.
Understanding telecommunications networks
Radio tower A
Radio tower B
Radio tower C
(a) Microwave Radio Relay
Up (Go)
Down (Return)
Down (Return)
Up (Go)
Ground station B
Ground station A
(b) Microwave Earth Satellite Relay
Figure 4.7
Not to scale
Free-Air Transmission Media
Free space: Earth satellites
By locating the microwave radio relay in a communications satellite in space orbit
around the World, the distance between a pair of antennas can be extended to intercontinental distances, as shown in Fig. 4.7(b). Clearly, the free space loss of the radio
signal is correspondingly large because of the long distances between Earth antennae
and the satellite relay – typically this so-called path loss up and down totals around
200 dB, depending of the frequency of the carrier wave and the path length, the latter
depending on the latitude of the Earth stations. Satisfactory performance requires
compensation for this loss by the use of high-gain repeaters and antennas on board
the satellite, as well as high-gain antennas on the ground. Since the gain of an antenna
increases with size, the gain of the antennas on board the satellite is limited by the
weight and diameter of the antennas that can be mounted on the satellite, so the gain
of the antennas on the ground need to be very high – hence the use of the huge satellite
dish aerials, which are a familiar sight at satellite Earth stations. Alternatively, smaller
capacity transmission links can be achieved from a central large dish antenna at one
end of the system and the use of small portable antennas at the other, e.g. smallaperture satellite systems as used by news TV broadcasters, and hand-held satellite
mobile phones used in remote areas of the World.
Multiplexed payloads
The costs of a telecommunications network are minimised by exploiting the significant economies of scale offered by the above-described transmission systems. In
Transmission systems
general, the cost per circuit carried reduces as the capacity of the transmission system
increases; thus, it pays to multiplex as many channels as possible on to each transmission link. In this section we consider the standard ways in which channels carrying
telephone calls or data communications are assembled into multiplexed composite
signals – the payloads – for conveyance over digital transmission systems, i.e. copper cable, coaxial cable, optical fibre, microwave radio or satellite, as appropriate.
(It should be noted that the multiplexed payloads first used in telecommunications
networks were analogue and based upon FDM – a technique described in Chapter 3.
However, these systems are now obsolete and so they are omitted from this book for
brevity. A description of FDM transmission payload systems may be found in [6].)
The PCM multiplexed payload: the basic building block of digital
The primary multiplexed payload created by assembling a group of PCM channels
forms the basic building block of all digital transmission and switching networks used
throughout the World. Thus, it is an important entity to understand. There have been
several versions of the PCM primary multiplex payload, but there are now just two
versions standardised by the ITU (previously known as the ‘CCITT’): the 30-channel
standard, used in Europe, Asia, and elsewhere regionally and on all international
links; and the 24-channel DS1 standard, used regionally, mainly in North and South
America. Both versions are described in the following sections. Chapter 3 introduces
the concept of TDM, A/D convertion and PCM.
30-Channel PCM multiplex
In this TDM digital standard, the primary multiplexed structure is based on a time
frame divided into 32 time slots, as shown in Fig. 4.8. The time frame length is 125 μs,
resulting from the sampling rate of 8,000 times a second – 8 kHz (the necessary rate to
sample a speech waveform with its frequency content limited to 4 kHz), as explained
in Chapter 3. Thus, each time slot in this frame occupies 1/32 of 125 μs, i.e. 3.9 μs.
The 32 time slots of the frame are permanently assigned to 30 speech channels, hence
the name of the system, and two non-speech-carrying support channels, as shown in
Fig. 4.8. The first of these support channels is carried in time slot 0 (written as ‘TS0’)
and it is often erroneously known as the ‘synch.’ or ‘synchronisation’ channel, which
is used to indicate the start of the frame, as described later. The second support channel
is carried in time slot 16 (‘TS16’) and is used to carry the call-control signalling
between the exchanges at either end of the PCM route. (Chapter 7 describes the two
sorts of signalling systems that can be carried in TS16.) Time slots 1 to 15 are used to
carry speech channels 1–15, respectively; time slots 17 to 31 are used to carry speech
channels 16–30, respectively.
The A/D conversion process within the PCM system encodes each speech sample
into an 8-bit binary word. Therefore, each of the 32 time slots in the multiplex carries
8 bits, making a total of 256 bits per frame. Since there are 8,000 frames transmitted
every second, the line rate of the 30-channel PCM multiplex is 8,000/s × 256 bits, i.e.
2,048,000 bits/s. This rate is usually designated as ‘2,048 kbit/s’ or ‘2.048 Mbit/s’ –
the latter is usually abbreviated to the more convenient and popular form of ‘2 Mbit/s’.
Understanding telecommunications networks
Speech channels
1 2 3 4 5
Time slots:
0 1 2 3 4 5
Fame alignment
signal in odd
frames (bunched)
Speech channels
26 27 28 29 30
15 16 17
27 28 29 30 31
Bits: 0 1 2 3 4 5 6 7
n n n n n n n n
Time slot :
Figure 4.8
The 30 Channel PCM Multiplex Frame Structure
Another important figure to remember with the 30-channel multiplex is the line
rate provided for each speech channel. This is given simply by the number of bits
per sample, 8, times the number of samples per second, 8,000, i.e. a line rate of
64,000 bits/s, or 64 kbit/s.
Fig. 4.8 also shows how the 8 bits of an individual channel are represented by
digital pulses every 0.488 μs (i.e. 3.9 μs/8), each pulse being 0.244 μs or 244 ns
(nano equals 1/1,000,000,000) wide.
We are now in a position to describe the function of the synch. channel in TS0. Its
role is best explained by considering the terminating end of the digital transmission
system which is receiving a stream of digital pulses arriving at the rate of 2 Mbit/s.
The timing of the pulses is extracted from the incoming stream and hence sets the
receiver sampling rate. However, this stream of pulses is meaningless unless the start
of the frame can be identified; thereafter, by counting the bits received the set of eight
bits relating to each channel can be located. This frame-start identification is indicated
by a special bit pattern, called the ‘frame alignment pattern’, which is inserted at the
sending end into the TS0 of odd frames, as shown in Fig. 4.8. At the receiving end
the first eight bits of two-frames worth of bits are examined in a digital register. If
the frame alignment pattern is not detected, the register shifts one bit and looks at the
first eight bits again. This continues until the pattern is found, the start of the frame
has then been detected and the receiver is now in ‘frame alignment’ with the sender.
The spare 8-bit capacity in the even frames of TS0 is available for special purposes,
e.g. BT used bit 5 to carry network synchronisation control signals for its UK digital
network [7].
Digital telecommunication networks have been developed and built on the basis
of the 2 Mbit/s building blocks, often referred as ‘2 Mbit/s digital blocks’. Not only
Transmission systems
are the blocks used over digital transmission networks, but they are also the entry
level into digital exchanges (local, trunk and international), as described in Chapter 6.
In addition, data can be directly inserted into a 2 Mbit/s multiplex since it is already
in digital format. Thus, the 2 Mbit/s digital block may be considered as a payload
vehicle capable of carrying 30 speech (or voice) channels or 30 data channels, each
of 64 kbit/s, over a digital transmission network.
The 2 Mbit/s block is also used by network operators as the primary rate for a
digital stream which is delivered to customer’s premises for a variety of business
services, including digital leased lines (or ‘private circuits’), connections to digital
ISDN PABXs, as well as ATM, frame relay and SMDS data services, as described
in Chapter 2. In all these cases the 2 Mbit/s block is carried through the network
and presented to the customer as an unstructured stream of 2 Mbit/s, with no implied
channel structure.
These 2 Mbit/s digital blocks, containing data or voice, may be carried directly
over a transmission link or multiplexed together with other blocks to form higher
capacity payloads for PDH or SDH (synchronous digital hierarchy) transmission
systems, as described below.
24-Channel PCM multiplex (USA DS1)
Fig. 4.9 shows the frame structure for the 24-ch. DS1 system. Although the speech
is digitally encoded into eight bits using 8 kHz sampling rate, as in the 30-ch. system
described above, there is an important difference between the two systems in the
way that it is done. This is because the spacing of the graduations on the codec ruler
(see Chapter 3) are based on the ‘A-law’ for the 30-ch. systems and the ‘Mu-law’
Speech channels
Time slots:
1 2 3 4 5
1 2 3 4 5
20 21 22 23 24
20 21 22 23 24
Frame: 125μs
Fame alignment
signal in bit 1
of odd frames
Figure 4.9
Bits: 0 1 2 3 4 5 6 7
n n n n n n n n
Time slot:
The 24-Channel PCM Multiplex Frame Structure
width = 0.324 μs
Understanding telecommunications networks
for the 24-ch. system. These two laws are different ways of spreading the quantum
steps using spacing that is roughly approximate to a logarithmic scale, ensuring a
constant signal to quantisation-error ratio over the operating range. Thus, the same
waveform would be encoded into a different set of eight bits by the two systems and
a trans-coding is required when 30-ch. PCM is connected to 24-ch. PCM links. (For
example, this trans-coding takes place at the international gateway exchanges into
the United States for calls between the United Kingdom and the United States.) All
the 24 time slots within the 125 μs time frame are normally used for speech channels
(but see below), there being no equivalence of the TS16 or TS0 of the 30-ch. system.
However, a single bit is contained at the front of the frame before TS1 which is used to
carry the frame-alignment pattern. The pattern is sent 1 bit at a time over odd frames,
i.e. dispersed over eight odd frames, rather than bunched into one frame, as in the
30-ch. system.
The frame comprises 24 8-bit time slots plus 1 bit, totalling 193 bits. The line rate
is thus 193 bits/frame, with 8,000 frames/s, giving 1,544 kbit/s – which is usually
written as ‘1.5 Mbit/s’. This rate is also known as the ‘DS1’ rate in the United States
and each of the component 64 kbit/s channels are known as ‘DS0’.
There are two ways that signalling can be conveyed over the 24-ch. system. The
earlier systems used a method known as ‘bit stealing’ in which the last bit in each
time slot of every sixth frame was used to carry signalling related to that channel.
This periodic reduction in the size of the PCM word from eight to seven bits introduced a slight, but acceptable, degradation to the quantisation noise. However, it also
prevented the time slots from being used to carry data at the full 64 kbit/s (i.e. 8 bits
at 8,000 frames/s) and the reduced rate of 56 kbit/s (i.e. 7 bits at 8,000 frames/s) was
the maximum rate that could be supported. The recently introduced alternative of
CCS avoids the need for bit stealing, since the signalling messages are carried in a
data stream carried over the bit 1 at the start of even frames, i.e. outside of the speech
channels. However, this 4 kbit/s (i.e. 1 bit every other frame) of signalling capacity
was considered too slow and now the DS1 system has been upgraded to carry CCS
at 64 kbit/s in one of the time slots, leaving 23 channels for speech.
The 1.5 Mbit/s digital block is also offered by network operators as an unstructured
primary rate of digital service delivery to customers’ premises for business services
(leased lines, ISDN, ATM, frame relay and SMDS). This 1.5 Mbit/s digital block is
frequently referred to as a ‘T1’ system in North America. (The term ‘E1’ is also used,
particularly in North America, to describe the equivalent 2 Mbit/s digital block used
in Europe and elsewhere.)
These 1.5 Mbit/s digital blocks, containing data or voice, may be carried directly
over a transmission link or multiplexed together with other blocks to form higher
capacity payloads for PDH or SDH/SONET transmission systems, as described later
in this chapter.
The time division multiplexing of digital blocks
Larger payloads may be created by the time division multiplexing of a number of
2 Mbit/s or 1.5 Mbit/s digital blocks. In Fig. 4.10 the concept is illustrated by the use
Transmission systems
2 Mbit/s frame
2 Mbit/s frame
0 1
0 1
0 1
0 1
0 1
0 1
Figure 4.10
0 1
TDM highway
0 1
Interleaving of Digital Blocks
of a rotating arm to sample each of four 2 Mbit/s digital block tributaries, each with
time slots 0–31. The 125 μs time frames for each of the tributaries is assumed to be
aligned so that they all start at the same instant. The wiper samples each tributary in
turn during the period of 125 μs; the wiper then samples the next frame’s worth of
contents of the four tributaries during the next 125 μs, and so on. Since the frame
size of the TDM highway is the same as that of the tributaries, the time spent by the
wiper on each tributary can be no more than a quarter of 125 μs. The content of the
TDM highway is therefore four times that of each single tributary, i.e. 120 traffic
channels in 128 time slots, with a line rate of four times 2 Mbit/s. At the far end of
the TDM highway the process is reversed and the appropriate contents are streamed
to the corresponding output tributary.
Clearly, for this TDM system to operate accurately the two wipers have to rotate
at the same speed with their starting points aligned so that corresponding tributaries
are being input- and output-sampled simultaneously. The speed of the distant wiper is
set by the incoming bit stream, as described earlier. The alignment is achieved using
a frame alignment pattern (actually known as a frame alignment signal – FAS) to
indicate the start of the TDM frame, in a similar way to that used to define the start
of the PCM frame, described earlier. Fig. 4.10 shows the FAS at the start of the TDM
There are two basic ways of sampling the tributaries and interleaving their contents
onto the TDM highway: ‘bit’ interleaving and ‘word’ or ‘byte’ interleaving. With the
former, one bit at a time is sampled from each tributary in turn; with the latter the full
PCM word of 8 bits (also called a ‘byte’) is sampled in one go from each tributary in
turn. Bit interleaving is used in the PDH systems, the original TDM digital payload
system, since the design is easier to construct and is readily compatible with digital-toanalogue conversion. Whilst the more recent TDM payload system of SDH (described
Understanding telecommunications networks
in Section 4.3.4), which is designed to carry both data and voice equally, and assumes
full digital networking, uses the more complex technique of byte interleaving [8].
The TDM output from the multiplexor can itself be multiplexed again with other
tributaries following either the PDH or SDH/SONET formats, and so on to create an
appropriately sized multiplexed payload, as described in the following sections.
Plesiochronous digital hierarchy system
The plesiochronous digital hierarchy (PDH) was the first internationally standardised
form of digital higher-order multiplexing and was deployed over a variety of cable and
radio systems, as well as optical fibre cable around the World. Although now being
progressively replaced by SDH systems, there is still a large amount of PDH capacity
in most incumbent operators’ transmission networks. The name ‘plesiochronous’,
meaning nearly synchronous, relates to the situation where the individual 2 Mbit/s
tributaries that are to be multiplexed over the PDH system are operating at close-butslightly-varying rates. This has several consequences. The first is that ‘stuffing’ or
‘justification’ bits need to be added to the TDM highway at each stage of multiplexing.
The concept of bit stuffing may usefully be explained by considering a cereal
manufacturer who wishes to pack up 12 packets of cornflakes into a large cardboard
box for transporting to a grocery shop. The cornflakes packets are nominally sized as
2 cm thick, which means that some boxes will be narrower (e.g. 1.98 cm), while others
might be wider (e.g. 2.04 cm). Therefore, the size of the cardboard box will need to be
greater than 12 times 2 cm, to allow for this variation in sizes. Once all 12 cornflakes
packs are placed into the box the grocer will insert pieces of cardboard as packing
to fill up any slack space so that the contents are held tightly. At the receiving shop
the cardboard packing will be discarded and the individual 12 packets of cornflakes
extracted. Different amounts of cardboard packing will be required in each subsequent
box, depending on the mix of cornflake packet sizes in each batch of 12. The boxes
containing 12 cornflakes packets can themselves be packed into larger boxes, again
using cardboard spacers to compensate for the variations in box dimensions, and this
process can continue with further stages of packing. By analogy bit stuffing in a PDH
system is applied to ensure that the TDM frame is filled tightly, irrespective of the
actual number of bits sampled from each tributary. There are several ways in which
the number and position of the stuffing bits can be indicated to the receiver of the
PDH system, so that the appropriate bits can be discarded when the tributaries are
The European standard for PDH higher-order transmission payloads is shown in
Fig. 4.11. Four stages of TDM multiplexing are defined, each being a multiple of
four. The first stage multiplexes four 2 Mbit/s digital blocks into an 8 Mbit/s stream.
The actual line rate of the output of this 2/8 muldex (the name for the multiplexor–
demultiplexor assembly, allowing for both directions of transmission) is 8,448 kbit/s.
This rate is usually referred to as 8 Mbit/s or E2 line rate. The difference of the E2
rate from four times 2,048 kbit/s results from the addition of an FAS and stuffing bits.
The subsequent stages of multiplexing similarly include frame alignment and stuffing
bits in the aggregate line rates. Table 4.1 shows all the stages of PDH multiplexing,
Transmission systems
2 Mbit/s (E1)
8 Mbit/s (E2)
Figure 4.11
34 Mbit/s (E3)
2 8/34
140 Mbit/s (E4)
2 140/
3 565
565 Mbit/s
Plesiochronous Digital Hierarchy European Standard
Table 4.1
European PDH Standard
Nominal Rate
Line Rate
No. of 64 bit/s
∗ Not standardised.
their nominal and actual line rates, the designation used for line systems at that rate
and the number of speech channels (each of 64 bit/s) that could be carried.
The North American PDH standard (Fig. 4.12) is shown in Table 4.2. In addition
to the DS designations, the line rates when made available for customers use have
the T designations; e.g. a 45 Mbit/s leased line is referred to as a ‘T3’ line system.
Although PDH digital line systems have been successfully deployed since the
1970s around the World, they have proved to be inflexible and cumbersome to manage in large-scale transmission networks due to the so-called multiplexor mountain
problem. Fig. 4.13 shows that 42 multiplexors are required to provide a fully equipped
140 Mbit/s PDH multiplexed system (for clarity, the line transmission equipment is
not shown). In order to extract a 2 Mbit/s digital block (e.g. a 30-ch. module) from the
140 Mbit/s system each stage of de-multiplexing is necessary in order to extract the
stuffing bits from the 140, 34 and 8 Mbit/s frames, respectively. (This is analogous
to extracting and discarding the cardboard packing sheets from each nested grocery
box until the individual box of cornflakes is obtained.)
Not only does the multiplexor mountain of a PDH network incur a large number of
multiplexors, with the consequent cost and potential fault liability, but also each item
Understanding telecommunications networks
1.5 Mbit/s (DS1)
6 Mbit/s (DS2)
1.5 Mbit/s (DS1)
45 Mbit/s (DS3)
4 M34
Figure 4.12
274 Mbit/s (DS4)
Plesiochronous Digital Hierarchy North American Standard
Table 4.2
North American PDH Standard
Nominal Rate
Line Rate
No. of 64 kbit/s
(or 56 kbit/s)
DS1, T1
DS2, T2
DS3, T3
DS4, T4
of equipment needs to be connected appropriately. Such connections are usually made
manually at the time of setting up a transmission route, using coaxial jumper cables
across a distribution frame in the core transmission network station in an exchange
or a stand-alone building, as described in Chapter 5.
SONET and synchronous digital hierarchy system
The PDH standard described earlier was defined at the start of the introduction of
digital transmission networks during the 1970s. However, by the end of the 1980s most
of the analogue transmission and switching equipment had been replaced by PDH
digital and there was an increasing proportion of optical fibre in use. A new standard
of digital transmission was defined, originally by ANSI in the United States, which
was designed to overcome the multiplexor mountain and lack of flexibility of PDH and
take advantage of the predominance of digital optical fibre working within the network
and the advances made in very large scale integrated circuits (VLSI) technology.
Transmission systems
2 Mbit/s 8 Mbit/s
34 Mbit/s
34 Mbit/s
8 Mbit/s
2 Mbit/s
140 Mbit/s
Figure 4.13
A 140 Mbit/s PDH Line Multiplexed System
Critically, the system is based upon optical rather than electrical interfaces between
transmission and switching equipment in the network. The system is called SONET,
derived from ‘synchronous optical network’. Later the ITU-T defined a universal
standard based upon SONET and extended it to cover the European conditions called
‘synchronous digital hierarchy’. The main features of SONET and SDH are:
Direct access to tributaries through the add/drop multiplexing function (unlike
the PDH multiplexor mountain);
Standard optical interfaces to optical fibres carrying the aggregate signal;
Compatible to both European and North American standards;
Reduced operational costs compared to PDH because of the elimination of manual
jumpering between the multiplexors.
Network management features;
End-to-end performance monitoring capability.
The overall effect of the above features is that SONET and SDH networks use less
equipment than the PDH equivalent, with the consequent improvement in operational
costs and quality due to the resulting lower fault rate. However, the biggest advantage
is that the network management features allows the SONET–SDH transmission to
be deployed as part of a coherent managed transmission network, as described in
Chapter 5.
As the names SONET and SDH suggest, the system is based upon multiplexing
synchronous tributaries, i.e. they are all operating exactly at one of a fixed set of
standard bit rates [9]. This eliminates the need for the use of variable amounts of
bit stuffing bits and the tributaries are time-division multiplexed directly by byte
interleaving onto an aggregate highway. The location of a particular synchronous
tributary within the frame can easily be found by counting the appropriate number
of bytes. The system can also multiplex PDH tributaries without the need for bit
Understanding telecommunications networks
9 rows
270 × N columns
125 μs frame
64 kbit/s
STM-N = Synchronous transport module, N defined for
1, 4, 16 (8 & 12 for United States also)
Gross bit rate = N × 155,520 kbit/s
Figure 4.14
Basic STM-N Frame Structure
Table 4.3
Standard Optical Interfaces for SONET and SDH
Data Rate
Optical Carrier Level
Synchronous Transport
Module (STM-N)
51.84 (52)
155.52 (155)
622.08 (622)
1,244.16 (1.2 Gbit/s)
2,488.32 (2.5 Gbit/s)
9,953.28 (10 Gbit/s)
stuffing by an ingenious use of pointers, which indicate where in the frame the start
of a particular PDH tributary is to be found.
The tributaries are packed into a basic container, which is known as the synchronous transport module (STM) in SDH or the optical carrier (OC) in SONET.
Fig. 4.14 illustrates the STM frame structure [10]. Although based on the 125 μs
frame, the module actually stretches over nine such frames, creating a matrix of 270
columns and nine rows. Each element in the matrix is 64 kbit/s, i.e. the capacity of a
single voice channel in the basic 2 Mbit/s block, and this is the smallest tributary that
can be multiplexed into and extracted from the basic module. The overall rate of the
basic STM is 155 Mbit/s. However, higher speed STMs are created by increasing the
number of rows that are contained within the 125 μs frame, each being a multiple ‘N’
of the basic STM – designated ‘STM-N’. The standard rates for the modules OC-N
and STM-N for SONET and SDH, respectively, are shown in Table 4.3.
Transmission systems
Lower-order path layer
Tributary unit pointer:
Higher-order path layer overhead:
Path name, path status, path content,
block error detection, and far-end block
error status.
High speed
Figure 4.15
SDH Packing Mechanism
When packing the tributaries into the modules (i.e. OCs or STMs) other channels
in the form of management information as well as the pointers, described earlier,
are also included. This packing process takes place in two stages of multiplexing,
as shown in Fig. 4.15. In the first stage tributaries are placed into a lower-order
container which, together with an ‘overhead’ containing transmission-link network
management information, is inserted into a virtual container (VC). Several of these
VCs are then collected together to form a transmission unit (TU) using pointers to
indicate their location. Several TUs form a group (TUG), which is the output of the
first stage of multiplexing. In the second stage of multiplexing the TUGs are packed
into higher-order containers, which together with an overhead carrying the network
management information for the higher-order path – potentially comprising several
links – is inserted into the higher-order VC. Finally, several higher-order VCs may be
grouped into an administrative unit (AU) which forms the entire payload of the STM1, again using pointers to indicate the location of the constituent higher-order VCs.
There are a variety of ways that the tributaries may be multiplexed up to the
STM-N level, depending on the size of the tributaries and the configuration of the
transmission links [10], and these are illustrated in Fig. 4.16.
Whilst this multiplexing structure for SDH and SONET might appear complicated, the basic concept used is relatively straight forward. It can be simply explained
by considering the synchronous transmission modules (STMs) as containers on large
transport lorries, which are trundling in a continuous stream down the (SDH/SONET)
highways between warehouses (transmission network nodes). Fig. 4.17 shows this
simplified view of the concept with a lorry containing a single container (STM), in
Understanding telecommunications networks
VC-4 C-4
139,264 kbit/s
VC-3 C-3
44,736 kbit/s
34,368 kbit/s
VC-3 C-3
Higherorder VCs
C = Container
VC = Virtual container
TU = Transmission unit
TUG = TU group
AU = Administrative unit
AUG = AU group
STM = Synchronous transport module
Figure 4.16
VC-2 C-2
6,312 kbit/s
VC-12 C-12
2,048 kbit/s
VC-11 C-11
1,544 kbit/s
Lowerorder VCs
SDH Multiplex Structure
Administrative unit (AU)
Transport unit group (TUG)
Transmission unit (TU)
Figure 4.17
Virtual container
Transmission systems
Optical Interfaces:
155 Mbit/s
622 Mbit/s
1.2 Gbit/s
2.5 Gbit/s
10. Gbit/s
Optical Interfaces:
155 Mbit/s
622 Mbit/s
1.2 Gbit/s
2.5 Gbit/s
10. Gbit/s
Optical fibre
Optical fibre
Electrical interfaces:
155 or 274 Mbit/s
34 or 45 Mbit/s
1.5 or 2 Mbit/s
64 kbit/s
Figure 4.18
A Synchronous Add–Drop Multiplexor
which two different freight companies each have a share – thus forming two administrative units (AU) – separated by a net curtain inside the container. Within each AU are
a number of large crates (transport unit groups) each containing several smaller crates
(transport units). It is these transport-unit crates that contain the various-sized parcels
for delivery (i.e. virtual containers). The location of the parcels within the crates
is indicated by information on written sheets (i.e. ‘dockets’ or ‘inventory sheets’) ,
corresponding to the pointers in the SDH multiplexing system. STMs-4, -16, and -48
are analogous to proportionately larger containers on the lorries or possibly lorries
towing a series of container trailers.
As described earlier, one of the big advantages of equipment based on SONET
or SDH is the ability to inject and extract a tributary from a multiplexed aggregate
bearer using just one component. The functions of such a component, known as an
‘add-drop multiplexor’ (ADM) is shown in Fig. 4.18, which lists the potential optical
interfaces on the input and output ports to the optical fibre bearers, and the electrical
interfaces to the set of tributaries.
The network aspects of a SONET or SDH transmission network are described in
Chapter 5.
The range of transmission systems
In this section we consider the range of transmission systems currently used in
telecommunications networks. Each of these systems is carried over an appropriate
medium with the required capacity provided by the use of a direct or multiplexed payload, as described in Sections 4.2 and 4.3. The range of transmission systems is wide.
Understanding telecommunications networks
Network operators need to choose which systems to deploy based on the costs, the
extent of existing infrastructure (i.e. whether cables or ducts are already in place), the
nature and capacity of transmission channels required, as well as future growth expectations etc., as described in Chapter 5. In addition, of course, the technologies available
for transmission systems are continually being improved in terms of performance and
costs, which can further complicate the choices for the network operator.
Metallic-line systems
Originally all transmission systems used in telecommunications networks were based
on metallic – predominantly copper (but also aluminium) – lines, carried as single
pairs or bundled into cables, slung overhead between poles. These pairs carried just
one telephone circuit as payload, either from the subscriber’s premises (i.e. in the
access network) or between exchanges providing ‘trunk’ and ‘junction’ routes. Today,
single circuit payload copper pairs still constitute the majority of the subscriber lines
in all of the PSTNs in the World. In addition, copper line systems – usually overhead –
are still used in many networks, particularly in rural areas, between local exchanges
and between a local exchange and its trunk exchange.
However, there has been a progressive use of, first analogue then digital, multiplexed payloads added to the single copper pair subscriber lines to provide extra
capacity for operational or customer-service use. An example of the operational use
was the first generation of so called pair gain systems, in which a two-channel analogue FDM system enabled two subscriber’s lines to be provided over a single pair
between the adjacent premises and the serving local exchange [11]. This method of
saving on the cost of copper pairs, which provides standard telephone service to the
subscribers, is still used today by many network operators – although, the more recent
deployments use digital pair-gain systems, based on DSL technology, described in
Section 4.4.2.
The provision of multiplexed payloads over copper lines for customer-service
reasons has until recently be directed at business subscribers, providing private circuit
and multiple channels access from their premises to the local exchange. Initially, FDM
systems giving up to 12-channels (48 kHz) capacity were carried over the copper
pairs. Since the 1980s, digital payloads (1.5 Mbit/s T1 in the United States and
2 Mbit/s E1 in Europe) have been deployed over copper pairs in the access network.
Interference between pairs carrying such payloads in a single cable is minimised by
using separate pairs for Go and Return direction. A further measure is the assembly
of the pairs into ‘balanced quads’ within the cable sheath [4]; such grouping reduces
the mutual interference between pairs carrying the high bit rate signals. However, the
distances the (quad) copper pairs can carry 2 Mbit/s payloads is restricted to just a few
kilometres and it is possible to use only a few pairs in each cable with such signals
in order to control the level of interference. Thus, more robust copper systems (with
longer reach and complete fills) such as transverse screen and co-axial cable are used
to carry payloads up to 140 Mbit/s between the business subscriber premises and the
serving exchange. To cope with long distances the transmission signal is boosted by
digital repeaters located periodically along the length of the cable, housed in boxes
Transmission systems
submerged in the street, i.e. footway boxes, or mounted on the poles supporting the
overhead or ‘aerial’ cable [12].
An extreme version of long-distance metallic cable transmission is that of submarine cable systems, used throughout the World to cross rivers, seas and oceans. The
first generation of submarine cables used analogue FDM payloads using different sets
of frequencies for the Go and Return direction of transmission over the single cable.
These specialised systems are designed to withstand the rigours of being deployed
under water over rough terrain (e.g. the sea bed) and generally being inaccessible
for repairs. Apart from (shark resistant) armour plating on the cables the system’s
sub-sea analogue amplifiers use high levels of component redundancy to achieve the
necessary reliability, since repairing amplifiers located on the ocean bed is a costly
and difficult task! Apart from providing a low loss transmission path for the payload,
the submarine metallic cable also needs to carry electrical power from its two end
stations (‘cable landing points’) to all the amplifiers along its length [13,14]. Finally,
of course, submarine transmission systems require specialist cable ships to lay the
cables initially and later to recover and repair installed cables.
Digital subscriber line transmission systems
Recent advances in integrated circuit technology, particularly digital signal processors
(DSPs), have enabled high-speed digital payloads to be carried over standard existing
subscriber-line copper pairs originally provided for telephony only. There is actually
a family of such digital subscriber line (DSL) systems, known generically as ‘xDSL’,
as shown in Fig. 4.19. Each of the systems has the same basic architecture, i.e.
digital send and receive transmission equipment located at each end of the copper
local loop. However, there are important differences within the range. There are two
main requirements for an xDSL system. The first requirement is to cope with the
generic digital subscriber loop
high bit-rate DSL
asymmetric DSL
basic rate
Figure 4.19
The DSL Family
very high speed DSL
symmetric DSL
Understanding telecommunications networks
impairments introduced by the copper local network. Actually, the copper line is able
to carry signals far in excess of those used for simple telephony – which are either
constrained to 4 kHz by filters or, if unfiltered, extend to only 15 kHz – typically
up to some 20 MHz over short distances. However, the attenuation introduced by
the copper line rapidly increases with the higher frequencies. In addition, the use of
high frequency signals over the copper access network causes interference between
adjacent pairs within cables. The name for this interference is ‘crosstalk’, because
of the way that speech from one pair is picked up and can be clearly heard by the
listener on the other pair. The interference can occur at the near end to the transmitted
signal – ‘near-end crosstalk’ (NEXT) or at the far end – ‘far-end crosstalk’ (FEXT).
There are several techniques used by the DSL systems to cope with the attenuation,
as well as the FEXT and NEXT introduced by the copper lines.
The second requirement is for both Go and Return signals to be carried by the
digital line system over the copper pair, i.e. ‘duplex’ working. Even though the
copper pair can support both directions of transmission simultaneously, as described
in Chapter 1, the xDSL electronic equipment at either end is inherently unidirectional
because of its dependence on semiconductor technology. Generally, xDSL systems
use a form of multiplexing to separate the Go and Return signals, e.g. TDM (use
of ‘ping pong’ alternate Go and Return bursts); FDM (use of separate frequency
bands for the two directions) or SDM (use of separate Go and Return pairs). Other
techniques, such as echo cancellation, are also used.
The first xDSL systems were associated with providing so-called basic rate ISDN
over copper. These systems extend the IDN to the subscriber’s premises allowing
digital data and telephony calls over the PSTN, as introduced in Chapter 3 and further
described in Chapter 6. However, the xDSL family is now considered to refer mainly
to those systems that enable ‘broadband’ access over the local loop. Four examples
of such systems are described below [3,15,16].
ADSL (asymmetrical digital subscriber line) system
This system was designed originally to provide ‘video on demand’ service over a
copper pair, but is now widely deployed over many PSTNs in the World to provide
broadband services (mainly high-speed Internet access) to consumers and small businesses, where the deployment of optical fibre to the premises is not warranted. ADSL
uses a filter at the subscriber’s premises and at the exchange end of the local loop
to extract the standard (4 kHz) telephony signal from the composite broadband signal; thus the telephony service can be used simultaneously and separately from the
broadband service. The latter comprises a medium-speed upstream data channel and
a high-speed downstream channel – hence the term ‘asymmetrical DSL’.
The system copes with crosstalk and other forms of interference by using a technique known as discrete multi-tone (DMT) modulation, or alternatively known as
orthogonal frequency division multiplexing (OFDM) [17]. This approach conveys
the data signal over a large set of individual carriers, each set at a precise frequency in
the available range – for ADSL the frequency separation is 4.3 kHz. At the receiving
end of the link the composite waveform is processed so that each modulated carrier
Power level
Transmission systems
High-pass filter
138 150
Frequency in kHz
Telephony (bothway)
Figure 4.20
Spectral Allocation for ADSL
is sampled at the appropriate frequency. The use of OFDM enables the maximum
possible rate of data throughput despite the limitations of the line due to crosstalk
and interference across the whole range of frequencies used. This results in ADSL
systems having variable data rates which continuously differ in real time from the
nominal advertised rate, as conditions on the line vary.
Duplex working, i.e. transmission in both directions over the single copper pair,
is achieved using separate bands of frequencies for up and down stream, a technique
known as frequency division duplex (FDD). Fig. 4.20 shows the spectral allocation of
ADSL in the United Kingdom [15], indicating the frequency spread of the low-band
splitter filter and the high-band data extraction filter. With perfect conditions, i.e. in
the absence of interference, data rates of up to some 2 Mbit/s can be carried over
about 5 km of copper line, with up to 8 Mbit/s over lengths of about 2.7 km. These
figures will improve with progressive upgrades to the ADSL equipment, but they are
in any case highly dependent on the quality of the local loop line plant within the
PSTN. More details, including a block-schematic diagram showing the architecture
of an ADSL line, is given in Chapter 5 and Fig. 5.4.
VDSL (very-high bit rate digital subscriber line) system
VDSL is a derivative of ADSL which provides significantly higher data rates by
restricting the use of copper line used to just a few hundreds of metres. This requires
the deployment of optical fibre from the exchange building to the VDSL equipment
housed in street cabinets – typically physically adjacent to the primary connection
points (PCPs) of the access network (see Chapter 5). This hybrid optical fibre–copper
system provides data rates up to some 52 Mbit/s over 300 m of copper pair or some
12.96 Mbit/s over 1.350 m of copper line [16,18].
Understanding telecommunications networks
HDSL (high bit rate digital subscriber line) system
HDSL provides high-speed data transmission symmetrically over two or three pairs
of copper cable. Rates of 2 Mbit/s symmetrical can be carried over three pairs up to
distances of 4.6 km [3]. Network operators typically use HDSL to provide the local
ends of digital private circuits (‘leased lines’) for business customers where optical
fibre local ends cannot be justified, or in advance of later deployment of optical fibre.
SDSL (symmetric digital subscriber line) system
SDSL is a derivative of ADSL, but with equal (symmetric) transmission capacity
upstream and downstream. This means that the NEXT interference dominates at each
end and consequently the achievable speeds are much lower than for an ADSL system
over a single pair in the same network.
Point-to-point optical fibre
Optical fibre systems offer the network operator the highest bandwidths of all transmission systems – potentially up into the Tbit/s (1,000,000,000,000’s bits/s) range.
They were initially introduced as replacements for the coaxial metallic cable systems carrying PDH payloads in the trunk network, but they are now used extensively
throughout the national Trunk and International networks and increasingly in the
access networks, carrying SDH/SONET payloads.
The basic architecture of an optical fibre point-to-point digital transmission system is as shown in the generic picture of Fig. 4.3. The carrier modulation system uses
a LED or laser diode to pulse a single colour light beam at the send end and Avalanche
photo diodes or PIN diodes are used to detect the light pulses at the receive end, as
described in Section 4.2.2. The operating characteristics of the send and receive transducers and the windows in the attenuation of the single mode optical fibre now used
in transmission systems – as shown in Fig. 4.5 – give a wide range of potential wavelengths (colours) of light, centred on 850, 1300 and 1500 nm. ( Note: a nanometre, nm,
is one billionth of a metre.) Special line codes, e.g. ‘code-mark inversion’ are used in
conjunction with the payload coding (e.g. scrambling) to ensure that there is sufficient
number of on–off transitions in the pulsed light sent down the optical fibre for successful timing recovery by the regenerators along the line [14]. Actually, the very-low
rate of attenuation of modern optical fibre systems means that the regenerators can
be spaced at least 50 km apart – a figure that is continually improving with equipment development. This means that on many optical fibre transmission links within
medium-or-small-sized countries, such as the United Kingdom, no regenerators are
required along the route, saving equipment, accommodation and operational costs.
Normally, separate optical fibres within a single cable are used to provide the
Go and Return directions of transmission. The alternative configuration, i.e. the two
directions of transmission carried as separate frequency bands (FDM) over a single
optical fibre, incurs extra electronic equipment costs and is only employed where the
number of optical fibres in a single cable are unduly restricted.
Since the mid 1980s submarine transmission systems have used optical fibre
in preference to digital coaxial metallic cables [19]. There is now a vast network of
Transmission systems
optical fibre submarine cable systems throughout the World. The potential capacity of
each new system laid continues to increase each year. For example, the trans-Atlantic
telecommunications (‘TAT’) optical fibre cable system laid in 2001, known as ‘TAT14’, is capable of carrying 1,024 STM-1’s across the ocean (compared to TAT-1 which
opened in 1956 with 36 telephone circuits). The system achieves this huge capacity by
carrying payloads of STM-64 (10 Gbit/s) on each of 16 different wavelengths of light
over the fibre pair using dense wave-division multiplexing (DWDM), as described
in Section 4.4.4. It is configured as a pair of optical fibres (Go and Return) looping
from Manasquan in New Jersey, United States to Denmark, then Holland, France and
the south western tip of the United Kingdom (Bude-Haven) and back to Tuckerton in
NJ, USA, totalling some 15,800 km.
Dense wave-division multiplex system
Network operators have progressively introduced to their transmission networks digital line systems of increasing speed, in order to get improvements in the cost per
channel carried. Of course, the full economy of scale on transmission systems is only
achieved if the transmission systems are fully occupied by traffic. Since the early
1980s the demand for transmission capacity has generally grown in line with the
development of ever increasing line speeds. For example, between 1980 and 1990
the development of digital transmission systems over optical fibre has increased from
2 Mbit/s to 2.5 Gbit/s. However, this progression is based on increasing the speed of
optical transmitters and receivers and the repeater systems, all operating on a single
wavelength of light over the optical fibre. Since the mid 1990s development work
has focused on the use of several different wavelengths of light, i.e. colours, as a
way of increasing capacity on line systems beyond the 10 Gbit/s achieved on one
wavelength in 1995. These systems, which are referred to as ‘dense wave-division
multiplexed’ (DWDM), require very precise optical filters to separate the component colours. Fig. 4.21 presents a historical view of the development of transmission
capacity up to the current maximum of 10 Gbit/s over a single wavelength of light
over an optical fibre, and the trend for increasingly higher numbers of wavelengths
within DWDM systems.
Fig. 4.22 illustrates the general arrangement for a DWDM system with n wavelengths. At the sending end, n lasers are required, each generating a different colour
(i.e. wavelength) of light. The optical fibre carries the composite DWDM signal containing the n wavelengths. At the receiving end of the system a filter is required to
separate the n wavelengths and these are passed to n individual photodiodes to detect
the set of sent signals. For the 16-wavelength system described for TAT-14 above the
potential capacity is 16 × 10 Gbit/s, i.e. 160 Gbit/s, which is the equivalent capacity
of some 2.4 million voice channels!
Passive optical fibre network
So far, we have considered only point-to-point optical fibre transmission systems.
Whilst this configuration is appropriate for delivering large capacities between network nodes, or to those business customer sites requiring many broadband and
Understanding telecommunications networks
Stage 2
128 64 Capacity log scale
32 4
Figure 4.21
16 Stage 1
Transmission Capacity Per Fibre
10 Gbit/s max
Figure 4.22
1 fibre (per direction)
at 10n Gbit/s line rate
Dense Wave-Division Multiplexing (DWDM)
telephony circuits, it is not economical for delivering small payloads. In particular, network operators have been seeking a cost-efficient method of deploying optical
fibre systems in the access network to serve residential and small business customer
sites, as described in Chapter 5. There are two main cost factors that need to be
addressed: the cost of deploying the many thousands of optical fibres from a local
exchange to each subscriber premise, and the terminating equipment at each end
needed to extract/insert the signal to/from each fibre. Passive optical network (PON)
systems have been developed to address these factors.
The basic premise of the PON system is that the cost of fibre and electronics is
minimised by using a tree and branch configuration, whereby a single fibre from the
local exchange splits several times in succession in order to serve many subscriber
premises. The splitting is achieved by fusing bundles of optical fibres so that the
light energy splits equally into each tributary. Since this is done along the fibre route
without the use of electronics – it is known as a ‘passive’ system. The tree and branch
Transmission systems
Optical fibre
Optical fibre
Splitter 1
Optical fibre
3 Optical fibre
Splitter 8
Figure 4.23
Optical fibre
Splitter 8
configuration also minimises the cost of electronics since only one device per fibre
spine (or ‘tree trunk’) is needed at the exchange end. Fig. 4.23 shows the basic concept
of a PON system. The optical fibre splitters are typically housed in footway boxes
beneath the street. In the example shown, the first splitter is 4-way, and the second
set of splitters is 8-way; thus, 32 subscriber premises are served by a single optical
fibre spine from the exchange.
At the exchange end optical line termination (OLT) equipment associated with
each optical fibre spine assembles the GO transmission for the 32 channels into a
time-division multiplexed signal that is broadcast over the PON fibre tree structure.
Thus, at each subscriber premises a terminal unit, usually called an optical network
termination (ONT), extracts the appropriate time slot from the received TDM broadcast signal. For the Return direction of transmission, each ONT is allocated a fixed
time slot during which a burst of the subscriber’s digital signal is sent. In effect, this
is a dispersed version of a TDM multiplexed system, similar to that described in
Chapter 3, but with each of the 32 tributaries located in different subscriber premises.
An important variant on this configuration, is the amalgamation of a set of subscriber
ONTs in a single unit, known as an optical network unit (ONU), which is located at
a business customer’s premises where it serves the many terminations via internal
wiring. With the business configuration there is either one or no stages of splitting
in the network, with the necessary TDM channel extraction/injection performed by
the ONU.
PONs were first developed in the BT Labs and trialled in 1987, with the first
application designed for telephony subscribers – known as telephony over passive
optical network (TPON). Since then there have been a series of PON developments
Understanding telecommunications networks
defined by the full service access network (FSAN) consortium and the ITU-T; more
recently in the United States the IEEE 802.3ah standard has been defined. There are
several types of PON, as follows [20,21].
APON: ATM over PON is a broadband version using ATM (asynchronous transfer
mode, see Chapter 8) over the PON system to provide broadband service. The third
generation provides a total capacity of 622 Mbit/s symmetrical for the maximum
32-split PON system.
BPON: Broadband PON is an expanded version of APON designed to carry video.
Use is made of DWDM to add a separate wavelength for the video services above the
non-video traffic on the PON. For example, a BPON system may carry voice, data
and bi-directional video for 32 customers over a single fibre spine. Two wavelengths
are used to carry voice and data at 622 Mbit/s downstream and 155 Mbit/s upstream,
and a third wavelength is used to carry video up to 1 Gbit/s bidirectional.
EPON: Ethernet PON offers the LAN packet-based protocol of Ethernet to
interface to the data networks on business customers’ premises.
GPON: Gigabit/s PON is the latest high-speed PON design, offering up to
2.5 Gbit/s symmetrically.
Line-of-sight microwave radio systems
The basic configuration of a line-of-sight (LOS) microwave radio system is shown
in Fig. 4.7(a). The key components are narrow-beam antennas, which can operate as
precision point-to-point LOS systems, allowing re-use of the same carrier frequencies
around the country without mutual interference. Frequency bands allocated by the
regulator and the ITU-R for this use by network operators reside at 4, 6 and 11 GHz –
these are mainly used on systems providing trunk links between exchanges. In addition
frequency bands around 19, 23 and 29 GHz are also available, which are mainly used
for shorter distance trunk links and subscriber access [22] and higher-frequency bands
are periodically being made available. In general, the size of the required antenna
reduces with the carrier frequency, but the allowable spacing between antennas also
reduces and the radio transmission becomes more susceptible to water absorption,
i.e. rain and clouds.
LOS microwave radio systems were initially introduced into operators’ networks
in the 1960s using analogue transmission, as an economical solution for the conveyance of TV programmes between the various studios, TV switching centres and
the TV broadcast mast antennas around the country. Later, FDM payloads of up to
3,600 telephony channels were carried using FM over these radio systems between
trunk exchanges.
Since the 1980s the microwave radio systems have been designed for digital
transmission, initially carrying the standard PDH payloads of 34 and 140 Mbit/s (and
multiples of 45 Mbit/s in the United States). With the advent of SDH/SONET, the
microwave systems carry the standard 155 Mbit/s STM-1 digital block.
An alternative configuration for LOS microwave radio systems is point-tomultipoint, in which an omni-directional central antenna radiates uniformly to any
Transmission systems
antenna/receiver in the vicinity. Such systems can be attractive in providing subscriber access links. A number of high-speed or ‘broadband’ multipoint microwave
radio systems have been developed to support the delivery of data services to business
customer premises. One such example from the United States is the local multipoint
distribution service (LMDS), which operates in the 28 GHz range and offers capacities
up to 155 Mbit/s per customer (although this has not been commercially successful).
The deployment of point-to-point and multipoint microwave radio systems is covered
further in Chapter 5.
Earth satellite systems
The established system of telecommunication satellites uses a configuration of three
satellites located equally spaced on an equatorial plane at a distance of 36,000 km
above the Earth. At this height the satellites spin around the Earth at exactly the
same rate as the rotational speed of the Earth, and they therefore appear stationary
to antennas on the Earth’s surface [23,24]. Whilst this positioning enables all areas
(except the far North and South) of the World to communicate with at least one
of the satellites, the high orbit does incur a transmission propagation time for the
microwave radio signal of some 250 ms, even travelling at the speed of light. This
means a round trip delay of about half a second. Such a delay is noticeable to two
people speaking over a satellite link, resulting in some difficulty in maintaining a
natural conversation. Consequently, telephone call connections are never provided
over more than one satellite hop; the remaining distance is always provided over
terrestrial links with their low delay.
However, the delay issues with geostationary satellites is not a serious problem for
one-way communication, such as broadcast TV (e.g. Sky ) or data communication.
More recently, two new satellite configurations have been introduced: medium Earth
orbit (MEO) and low Earth orbit (LEO) systems, both of which incur less transmission
distance and hence delay [24]. MEO and LEO systems comprise a constellation of
satellites set in an elliptical orbit around the World, designed to give coverage only to
specific areas on the Earth’s surface. Since the satellites are not in a geostationary orbit
they will appear to move continuously across the sky; thus the Earth station antennas
need to be able to track and move to receive the satellite’s signal. Once a particular
satellite has moved to the limit of visibility the Earth station antenna needs to switch
to a following satellite which should have just come into view. Typically each satellite
in a LEO system is visible for about 10 min, orbiting about every 120 min at some
780 km above the Earth. These LEO systems tend to be of low capacity but can be
tailored to give good coverage of north and south regions which are difficult to access
from geostationary satellites, and they have the major advantage of incurring only
some 10 ms propagation delay. Alternatively, MEO systems use orbits around 10,000
km and incur some 70–80 ms propagation delay.
Unlike the three satellites necessary for full coverage with a geo-stationary system,
MEO systems require some 12 satellites, while LEO systems may require up to 200
satellites. The complexity of the antenna steering control, the overall satellite system
control and the large number of satellites required means that LEO and MEO systems
Understanding telecommunications networks
are not currently economical for large-scale commercial application – although, they
are used for specialist applications, such as communications in the remote north of
Russia and Canada.
Wireless LANs
There is a family of short-distance microwave radio systems designed for conveyance
of data services between a user’s terminal (e.g. a laptop computer), which is normally
static, and some serving node in the network. Since the configuration is really an
extension of the LAN concept, originally used in offices but now also deployed in
residences, the systems are known as ‘wireless LANs’. In the United States these systems have been specified within the ‘IEEE 802.11’ range, a nomenclature which has
now been widely adopted worldwide. They are also popularly referred to as ‘WiFi’.
These systems use techniques, such as OFDM described earlier for ADSL, to cope
with the multi-path distortions to the radio signal (see Chapter 9). The transmission
network aspects of wireless LANs and their data network aspects are covered in
Chapters 5 and 8, respectively.
Wireless MANs
A new generation of LOS microwave access systems has been defined which provides
broadband data links between LANs or between a LAN and a central server, creating
so-called wireless metropolitan area networks (MANs). The system, defined by the
IEEE 802.16 standard, is popularly known as ‘WiMax’ [25,26]. It uses 256 or 2048carrier frequency OFDM systems to provide maximum data throughput over the
multi-path conditions experienced over radio links. However, the standard allows
many variations in the way that manufacturer’s equipment is designed, thus giving a
range of performance and cost options for users. Although WiMax is specified for use
in the 10–66 GHz range of the radio spectrum, initial interest is in the unlicensed band
of 5.725–5.850 GHz, and in the licensed bands of 2.5–2.69 GHz and 3.4–5.80 GHz.
The data throughputs possible range from about 3–7 Mbit/s over a 5 MHz wide radio
band; however, this capacity is shared among all the users in the WiMax cell. The
use of WiMax as a form of broadband access is considered in Chapter 5.
In this chapter we have looked at the wide ranging subject of transmission systems
as used in telecommunications. In Section 4.2 we began with the basis mechanism of
transmission and then looked at the physical transmission media: copper cable, coaxial
cable, optical fibre and radio. Section 4.3 addressed the various ways of building a
digital multiplexed payload for carrying over the transmission media, covering:
PCM and the basic 2 or 1.5 Mbit/s digital block – the basic building block of
digital networks;
PDH, the original digital transmission multiplex payload system;
Transmission systems
SONET and SDH, the synchronous digital payload system that facilitates a managed transmission network, as well as eliminating the multiplexor mountain
problems of PDH.
The wide range of transmission systems currently in use by network operators were
then briefly reviewed in Section 4.4, including:
Metallic cable;
The family of xDSL (i.e. ISDN, ADSL,VDSL, HDSL and SDSL) systems;
Optical fibre systems, primarily point-to-point;
PON systems, for cost-effective access applications;
DWDM, which by using many colours of light is providing ever increasing
capacity over optical fibre;
Microwave radio systems;
Earth satellite systems (geostationary, LEO and MEO);
Wireless LANs and wireless MANs.
The deployment of digital multiplexed payloads over radio, metallic line or optical
fibre systems in the access, core and international transmission networks is covered
in the next chapter.
FLOOD, J. E.: ‘Transmission Principles’. Chapter 2 of ‘Transmission Systems’,
edited by FLOOD, J. E. and COCHRANE, P., IET Telecommunications Series
No. 27, Stevenage, 1995.
DORWARD, R. M.: ‘Digital Transmission Principles’. Ibid., Chapter 7.
ANTTALAINEN, T.: ‘Introduction To Telecommunications Network Engineering’, Artech House, Norwood, MA, 1998, Chapter 4.
FLOOD, J. E.: ‘Transmission Media’. Chapter 3 of ‘Transmission Systems’,
edited by FLOOD, J. E. and COCHRANE, P., IET Telecommunications Series
No. 27, Stevenage, 1995.
DUFOUR, I. G.: ‘Local Lines – The Way Ahead’, British Telecommunications
Engineering, Vol. 4, Part 1, April 1985, pp. 47–51.
KINGDOM, D. J.: ‘Frequency-Division Multiplexing’. Chapter 5 of ‘Transmission Systems’, edited by FLOOD, J. E. & COCHRANE, P., IET Telecommunications Series No. 27, Stevenage, 1995.
REDMILL, F. J. and VALDAR, A. R.: ‘SPC Digital Telephone Exchanges’, IET
Telecommunications Series No. 21, Stevenage, 1995, Chapter 10.
FERGUSON, S. P.: ‘Plesiochronous Higher-Order Digital Multiplexing’. Chapter 8 of ‘Transmission Systems’, edited by FLOOD, J. E. & COCHRANE, P.,
IET Telecommunications Series No. 27, Stevenage, 1995.
SEXTON, M. J. and FERGUSON, S. P.: ‘Synchronous Higher-Order Digital
Multiplexing’. Ibid., Chapter 9.
WRIGHT, T. C.: ‘SDH Multiplexing Concepts and Methods’, British Telecommunications Engineering, Vol. 10, Part 2, July 1991, pp. 108–115.
Understanding telecommunications networks
TAYLOR, G.: ‘Electronics in the Local Network’, Chapter 13 of ‘Local Communications’ edited by GRIFFITHS, J. M., IET Telecommunications Series No.10,
Stevenage, 1983.
FLOOD, J. E.: ‘Transmission Systems’. Chapter 4 of ‘Telecommunications Networks’, Second edition, edited by FLOOD, J. E., IET Telecommunications Series
No. 36, Stevenage, 1997.
HOWARD, P. J. and CATCHPOLE, R. J.: ‘Line Systems’. Chapter 10 of ‘Transmission Systems’ by FLOOD, J. E. & COCHRANE, P., IET Telecommunications
Series No. 27. Stevenage, 1995.
PETTERSON, O. and TEW, A. J.: ‘Components for Submerged Repeaters
and Their Qualification’, British Telecommunications Engineering, Vol. 5, July
1986, pp. 93–96.
KERRISON, A.: ‘A Solution for Broadband’. Chapter 5 of ‘Local Access
Networks’, edited by FRANCE. P., IET Telecommunications Series No. 47,
Stevenage, 2004.
ABRAHAM, S.: ‘Technology for the Fixed Access Network’. British Telecommunications Engineering, Vol. 17, Part 2, 1998, pp. 147–150.
ENGELS, M.: ‘Wireless OFDM Systems: How To Make Them Work?’, Kluwer
Academic Publishers, Norwell, MA, 2002, Chapter 3.
CLARKE, D.: ‘VDSL – The Story So Far’, The Journal of the Institute of British
Telecommunications Engineers, Vol. 2, Part 1 January–March 2001, pp. 21–27.
HORNE, J. M.: ‘Optical-Fibre Submarine Cable Systems – The Way Forward’,
Ibid., pp. 77–79.
JAMES, K. and FISHER, S.: ‘Developments in Optical Access Networks’.
Chapter 9 of ‘Local Access Networks’, edited by FRANCE, P., IET Telecommunications Series No. 47, Stevenage, 2004.
LOW, R.: ‘What’s Next After DSL – Passive Optical Networking?’, The Journal
of The Communications Network, Vol. 4, Part 1, January–March 2005, pp. 49–52.
de BELIN, M. J.: ‘Microwave Radio Links’. Chapter 12 of ‘Transmission Systems’ by FLOOD, J. E. & COCHRANE, P., IET Telecommunications Series
No. 27. Stevenage, 1995.
NOURI, M.: ‘Satellite Communication’. Chapter 13 of ‘Transmission Systems’,
edited by FLOOD, J. E. & COCHRANE, P., IET Telecommunications Series
No. 27, Stevenage, 1995.
SCHILLER, J.: ‘Mobile Communications’, Second edition, Addison-Wesley,
Harlow, 2003, Chapter 5.
Standard 802.16: A Technical Overview of The WirelessMAN™ Air Interface
For Broadband Wireless Access’, IEEE Communications Magazine, June 2002,
pp. 98–107.
GHOSH, A., WOLTER, D. R., ANDREWS J. G. and CHEN, R.: ‘Broadband Wireless Access With WiMax/802.16: Current Performance Benchmarks
and Future Potential’, IEEE Communications Magazine, February 2005,
pp. 129–136.
Chapter 5
Transmission networks
Chapter 4 introduces the concept of modulating the information (voice, data, video,
etc.) onto a variety of transmission systems. We now need to consider how these
systems are used in telecommunication networks. As discussed in Chapter 1, all
telecommunication networks, and indeed most other sorts of networks, comprise an
access portion which connects the population of users to a serving network node, and
a core portion which interconnects the set of serving network nodes. These portions
are transmission networks in their own right, containing links and nodes, and are
usually known as theAccess Network and the Core Transmission Network (or similar),
respectively. (Of course, the core portion also contains switches or exchanges, data
routers, control nodes, etc., which are contained in separate core switching, data
routeing, control networks, etc.) The Access Network may be provided over fixed
transmission links (using wires, optical fibre cable or radio systems) or over mobile
radio links (using a mobile handset), while the core is always provided over a fixed
link. This chapter describes the structure and characteristics of the fixed Access and
Core Transmission Networks; Chapter 9 covers the mobile access.
5.2 Access networks
5.2.1 Scene setting
The role of the Access Network is to provide a link between each user (referred to
as ‘subscriber’) and their serving node in the network. The PSTN of each country
provides the basis for the access network, with the deployment of copper cable to
(practically) all parts of the country to provide basic telephony service for residential
and business subscribers. However, overlaying the copper or ‘local loop’ network
of the PSTN is a variety of transmission systems – transverse-screen copper, optical
fibre, terrestrial microwave radio or satellite microwave radio – provided specifically
100 Understanding telecommunications networks
to deliver many types of services in addition to telephony to business subscribers.
Where possible, the near-ubiquitous physical infrastructure of the local loop, i.e. the
duct ways, poles, and joint boxes, carrying the cables, is used to carry these overlaid
transmission systems.
It should be noted that the so-called local-loop network, which comprises the
copper pairs connecting subscribers to the telephone exchange, is also used by several
non-telephony services, such as analogue and digital private circuits (leased lines),
alarm circuits, telex and, of course, broadband (e.g. using ADSL). Indeed, there are
a few local exchanges in central London where copper pairs used for private circuits
outnumber the telephony pairs terminating at the exchange.
5.2.2 The copper (‘local loop’) access network
The structure of a copper access network is best described as a set of trees and their
branches radiating from the serving exchange to all the subscribers in the catchment
area. Fig. 5.1 shows a stylised view of the physical architecture of a typical copper
network. At the exchange, all of the copper pairs are connected to a large metal distribution frame, located on the ground floor. This frame, known as the main distribution
frame (MDF), is the demarcation point in the exchange between external cabling on
its input and internal cabling on its output. Jumper wires are used to link each external pair, connected to a particular subscriber, to the internal wires to the appropriate
point (SCP)
n ca
frame (MDF)
cross connection
point (PCP)
Figure 5.1
The Physical Architecture of A Copper Access Network
Transmission networks 101
termination equipment dedicated to that subscriber (i.e. the ‘line card’as described
in Chapter 6) on the local switching unit in the exchange building. It is also at the
MDF that jumpers are used to link the external pairs to internal wiring associated
with private circuit equipment, broadband terminating equipment (as described later
in this chapter), etc. In general, the MDF is the physical point of presence in the
exchange where an exclusive electrical contact can be made to a particular copper
pair for operational activity, such as testing or application of special tones, etc. Apart
from its dominance functionally as the exchange end of the copper access network,
the MDF is physically an imposing sight for a visitor to an exchange building, being
a metal structure some 2.5 m high and arranged in rows of 10–15 m long, often
occupying most of the ground floor and, exceptionally, covering several floors.
Whilst considering the exchange end of the copper access network, it is convenient
to describe the other major physical entity usually on the ground floor of a serving
(i.e. local) exchange – namely, the battery. As described in Chapter 1, a voltage of
50 V DC (with slight variation between countries) is applied to each telephony copper
pair in order to power the telephone instrument. Actually, most PSTN operators are
required to provide sufficient power for several instruments on any one subscriber’s
line. The limit is set by the power required to ring a standard bell at the instrument –
defined as ringing equivalent number ‘REN’. In the United Kingdom the REN is set
at four, which allows any number of phones to be connected to the one line, provided
the combined REN values do not exceed four.
The exchange battery is also an impressive sight within a local exchange building.
It can best be described as similar to a large car battery, about 1 m high, occupying a
large-sized room. The 50 V DC power from the centralised battery is carried to the
switching equipment over aluminium or copper bus bars suspended from the ceiling
and walls, measuring about 20 cm wide and 1.3 cm thick. The exchange battery is
charged from the mains electricity on a trickle basis. During normal operation the
power to the exchange is provided by the electrical main supply with the battery
connected in parallel (with its voltage so called ‘floating’) across the charge input,
creating a stabilising effect on the 50 V DC supplied to the telephones. In the event
of a mains failure, the battery can provide power to the subscribers’ telephones for
several hours. A local generator may be used in the event of long mains power outages.
This ability to enable the telephones to work during a mains electrical power outage
is viewed seriously by the regulatory authorities because it allows subscribers to call
the emergency services at all times.
The progressive introduction of electronic equipment within exchange buildings
since the mid1980s, serving not only digital telephony switching but also data transmission and routeing, has increased the need for much lower voltages than the standard
50 V DC. This has been met by the use of distributed power equipment racks located in
adjacent rooms to the relevant equipment within the exchange building. These racks
are directly fed by the standard AC electrical mains (e.g. 240 VAC) and include power
rectifiers to convert to the required DC voltage for the served electronic equipment,
e.g. 6 V DC, as well as compact stand by batteries. The power is distributed over the
metal bus bars described above. Most exchange buildings tend to have a mix of the
central and distributed systems [1,2].
102 Understanding telecommunications networks
Immediately outside of the serving (local) exchange, all the cables from the MDF
are hubbed at an overhead or, more usually, underground assembly (known as the
‘exchange manhole’ in the United Kingdom). Here the copper cables are carried
through ducts laid under the streets radiating from the exchange building. As Fig. 5.1
shows, those subscribers living close to the exchange are served directly from one
of these cables. For the more distant subscribers, the exchange cables terminate on
frames within street cabinets (known as PCPs) where distribution cables radiate down
individual streets within the PCP catchment area. For the most distant subscribers,
or for the more rural areas, a further stage of connection at secondary connection
points, SCP (also called ‘pillars’ in the United Kingdom), is used. Jumpers are run
across the frames in PCPs and SCPs to create a continuous circuit (loop) between the
exchange-side cable and the distribution-side cables – known as ‘E-side’and ‘D-side’,
The final link (or ‘final drop’) to the subscriber’s premises is made either via a
drop wire from a pole in the street (referred to as ‘overhead distribution’) or via a
lead-in pair from the cable running underground along the street. For the overhead
case, the cable from the PCP or SCP runs underground to the base of the pole and
then terminates at the top of the pole in a distribution box containing, typically, a 10or 20-pair terminal block. Each drop line is then slung from a secure point on the
outside of the subscriber’s premises. With underground distribution, the lead-ins are
run to joints made, along the cable, in footway boxes below street level, as it passes
the subscriber’s premises. In general, business premises and all premises in urban
areas are served by underground distribution, with the suburban residential premises
generally served by overhead distribution. It is interesting to note that in the United
Kingdom about half of the residential subscriber final drops are provided overhead.
The arrangement of PCPs and SCPs described earlier allows a flexible development of a tree and branch structure, which can minimise the total cost of the copper
cables to serve an exchange catchment area. First, the quantity of pairs deployed is
minimised by taking fewer pairs to the E-side of the PCP/SCP than on the D-side, a
jumper-link between the two sides being made only when a new subscriber is added to
the exchange. Second, the cost is minimised by using as thin a gauge of copper wire as
possible. The limit to the size of gauge is set by the required electrical characteristics
of an exchange line, since the thinner the gauge, the greater the attenuation of the
local loop and hence the greater the reduction in loudness of the telephone connection.
(For example, in the United Kingdom, the attenuation limit is set at 15 dB, measured
at a frequency of 1,600 Hz.) The maximum allowable local-loop resistance is set by
the need for a sufficient current to operate the signalling relays in the exchange line
card. Typical limits for local loop resistance are 2 k. In practice, a combination of
different gauges is used, with narrow gauge serving the shorter subscriber lines and
a combination of narrow, medium and thicker gauges serving the longer subscriber
lines. Fig. 5.2 illustrates a typical arrangement of the gauges of the copper pairs and
the size of the cables in the access network.
In rural areas, particularly where the terrain is rocky or mountainous, the E-side
cables are carried overhead, slung from poles running along the road or cross-country,
as appropriate. Special rugged cables are used in order to cope with the exposure to
Transmission networks 103
100pr; d
20pr; b
200pr; d
800pr; d
1pr drop
; b DP
PCP2 200pr; b
600pr; b
400pr; b
200pr; b
to other DPs
200pr; b
DP 20p
Primary connection point (cabinet)
Distribution point
a: 0.63 mm copper cable
b: 0.50 mm copper cable
c: 0.40 mm copper cable
d: 0.32 mm copper cable
Figure 5.2
10pr; b
100pr; c
r; b
r; a
r; a
20pr; a
Typical Local Loop Make-Ups
wind encountered in such situations and the weight of the cable over long catenary
lengths between poles. Very-long overhead cables also require amplification or reduction in loss using loading coils at points along the route to ensure that appropriate
electrical characteristics are achieved.
5.2.3 The optical fibre access network
Optical fibre cables are also carried in the underground duct of the access networks.
However, there are several important differences between the copper and optical fibre
networks. First, the optical fibre is deployed to serve the larger businesses – where the
link takes a large capacity directly to the terminating point in the business premises.
Alternatively, the fibre is deployed as a hybrid optical-copper delivery method for
residential and small business subscribers.
Direct fibre delivery
The optical fibre taken directly to the subscriber’s premises terminates on a bank
of multiplexing equipment which extracts the various channels and services. These
may be private circuits, data circuits, digital multiplexed speech circuits, analogue
individual voice circuits, etc. Most business subscriber premises contain an equipment room for the terminating equipment (e.g. multiplexors) of the network operator,
together with a smaller version of an MDF for copper wires and a DDF for jumpering
104 Understanding telecommunications networks
the digital channels to the various internal communication equipment, e.g. PABX,
ATM switches, IP routers, frame relay switches, Ethernet LANs.
In the case of PDH transmission systems, a typical arrangement is an 8 Mbit/s
digital block carried over the optical fibre from the serving exchange to the subscriber
premises. This is demultiplexed to four 2 Mbit/s digital blocks in the subscriber’s
equipment room, giving, say, one 2 Mbit/s ISDN link from the PSTN switch to the
PABX and three 2 MBit/s private circuits (or two 2 Mbit/s private circuits and one
spare slot for future expansion).
Now the preferred high-capacity delivery system within the access network for
businesses is SDH. Typically, this is at the STM-1 (155 Mbit/s) level, delivered over
the fibre from the exchange with add–drop multiplexors in the subscriber’s equipment
room extracting the required tributaries.
The optical fibre network is structured as a tree and branch network, similar to that
of the copper access network. Large capacity optical fibre cable (up to 96 fibre pairs
per cable) radiate from the local exchange, with smaller cables branching out at joints
in boxes at appropriate points – with the final drops to customers’ premises using
typically four-fibre (Go and Return with spare pair) optical cables. The optical fibre
network then extends from the serving local exchange, the parent trunk exchange
and perhaps to other trunk exchanges and neighbouring exchanges, as appropriate –
thence joining to the core transmission network.
Passive optical network delivery
Passive optical networks provide an alternative to the use of a single optical fibre
per customer’s premises, and hence they potentially offer network operators an economical solution to avoiding further deployment of copper cables. The basic system
is described in Chapter 4. In practice the PON systems are installed in parts of the
access network where there is growth in demand for standard telephone, ISDN service or medium speed private circuit services. The optical splitter assemblies are
usually housed in footway boxes, since no power supply is required en route. As
with the direct optical fibre systems, the active line terminating equipment (LTE) is
housed in the customer’s premises. The derived tributaries are connected at small
distribution frame to the internal wiring to the terminal devices within the customer’s
Hybrid fibre-copper delivery
Here, the optical fibre terminates on street-located cabinets, usually adjacent to the
PCPs, where individual channels are derived from the high capacity on the fibre by
multiplexing equipment. The channels are carried by copper cable to the PCP, where
they are jumpered to the appropriate copper final drops to the subscriber’s premises.
Such arrangements are used by PON systems, in which the street-based multiplexors extract 48 subscribers’ circuits (Go and Return Channels), which are delivered
over copper pairs either as underground or overhead final drops. This use of PON,
as opposed to taking the PON fibres direct to each served premises described earlier,
is primarily to provide relief for the existing copper access network by delivering
Transmission networks 105
capacity for new telephony subscribers at the outer edge of the exchange catchment
area without requiring new copper cables in the E-side.
5.2.4 Radio access network
Fixed line-of-sight microwave radio
Radio links are also an important part of many operators’ access networks. Such
links are provided on a permanent or semi-permanent basis with one end terminating
on a fixed radio receiver attached to the subscriber’s premises and the other end
termination at the serving exchange building. This use of radio is therefore part of
the fixed network and is different to the way radio is deployed in a mobile network,
as described in Chapter 9.
Multipoint radio systems, based on an omni-directional antenna, typically located
on the serving-exchange roof, are mainly used to provide permanent access links to
subscribers wanting single-line telephony. The choice of whether to use multipoint
radio is based on the degree of geographical remoteness of the subscriber and practical
problems with using the cheaper copper-pair option, e.g. the serving of subscribers
in the mountains or across rivers. Consequently, multipoint radio has been deployed
extensively since the 1970s in such areas as the Highlands and Islands of Scotland.
Many of these earlier systems use analogue transmission. However, more recent
deployments use digital transmission, offering a better performance quality (e.g.
clarity and lack of noise) for the user.
Digital microwave point-to-point systems using two dish antennae, one on the roof
of the subscriber’s premises (usually an office block) and the other on the roof of the
exchange, provide high capacity access links. A typical example is the use of a 19 GHz
digital radio system to provide for a bi-directional 2 Mbit/s digital private circuit from
an exchange to the business customer’s premises (i.e. the ‘local end’). Another use
of microwave radio is that of providing high capacity distribution of satellite links,
where the provision of individual satellite terminals on customers’ premises is not
possible. Here, the 29 GHz transmitter is located on the roof of the serving exchange
building along with a satellite receive antenna. The video transmission received from
the satellite is then relayed via the 29 GHz radio system to antennas on several
customers’ buildings on a point-to-point or multipoint basis [3].
Unfortunately, the performance of radio links, particularly microwave radio, is
impaired by atmospheric conditions, e.g. rain, as well as reflections from buildings.
(The biggest reflection problems are caused by the tall cranes on building sites which
are constantly swinging around!) Many operators, therefore, install microwave radio
links initially to serve a business customer, taking advantage of the speed of provision
of radio compared to running a new optical fibre link. Later, once the optical fibre route
is completed, the service is swapped to the optical-fibre system and the microwave
radio link recovered, ready for deployment elsewhere.
The key aspect for a network operator deploying a microwave radio system is the
need for so-called radio-frequency planning. Each network operator has a portion of
the spectrum allocated by the national regulator for their use of specific types of radio
systems. The operator, therefore, has to ensure that their proposed deployment of a
106 Understanding telecommunications networks
Local exchange
Area 2
Server (authentication)
Area 1
Figure 5.3
Wireless LAN Access Networks
radio system will not interfere with any existing radio systems in the vicinity or other
operators or future potential radio users in the area. They must also consider the profile
of the landscape and potential points of interference or areas of severe attenuation so
that the radio path provides the required level of transmission performance. Planning
tools, usually computer-based, are used to determine the appropriate power levels
and frequencies within the allocated range to be used within the proposed fixed radio
Wireless LANs
In contrast to the fixed microwave radio systems above, wireless local area networks
(‘Wireless LANs’) use frequencies in the unregulated part of the spectrum. This means
that non-radio-expert service providers, as well as network operators, can rapidly
deploy WLAN hot spots around the country. The low power and short distance of the
wireless LAN coverage means that there is little chance of interference with other
radio users. Typically, WLAN are provided in airport lounges, coffee shops, railway
stations, etc., enabling customers to gain broadband data access from portable data
terminals, such as computer laptops. This form of radio access network gives the user
the ability to move within the catchment area of the antenna, typically 10–20 m; the
user may also log on to other wireless locations around the country – giving a form
of so-called nomadic mobility.
Fig. 5.3 illustrates the arrangement [4] for a public service. The WLAN antenna
is centrally located at the serving location, e.g. a coffee shop, giving a catchment area
determined by the power of the transmission and the characteristics of the building,
since this determines the absorption of the radio signal. The antenna may be connected
by a private circuit over the access network, typically using optical fibre to the serving
local exchange building and then to a server on the service provider premises, as
shown in Fig. 5.3. Alternatively, this backhaul link is carried on ADSL over copper.
Transmission networks 107
The service provides authentication and logging-on functions. Access to the Internet
is then provided over another private circuit, ADSL or a data service to the serving ISP.
As described in Chapter 4, the second generation of wireless MAN, standardised as
IEEE 802.16 and popularly known as ‘WiMax’, offers the possibility of ubiquitous
broadband data access. It conveys a broadband link from an exchange to an external
antenna on a customer’s building and thence to a LAN or other in-building network.
Both unlicensed and licensed parts of the frequency spectrum may be used, giving a
trade-off between freedom of use and minimal planning but with greater chances of
interference versus more predictable performance, but with a greater degree of planning required, respectively. Although, the WiMax standard is still relatively immature,
there are hopes that the use of this radio technology will prove useful for providing
broadband data coverage to many customers who cannot be reached economically by
copper cable and ADSL or optical fibre.
5.2.5 Broadband access
So far in this chapter we have considered the delivery of telephone service over the
copper cables and the delivery of higher-capacity services to business customers using
optical fibre or radio in the access network. However, since the late 1990s telecommunications network operators and Cable TV operators have been delivering high-speed
data services, primarily fast Internet access, to residential and small business customers – the so-called ‘broadband’ service. Cable TV operators provide broadband
service to their subscribers by adding cable modems to carry the fast Internet-access
data over a spare TV channel, as described in Chapter 2 (Fig 2.9).
Telecommunications network operators have deployed ADSL electronic equipment at each end of copper lines to convey the fast Internet access data channel in
addition to the telephone service. ADSL is now the main method that incumbent network operators use to provide broadband service to the residential and small business
market. This new use of the copper local network originally deployed just for the provision of telephony (i.e. so-called narrowband service) represents a significant new
lease of life for the copper cables. Since it is only the incumbent operators that have
the ubiquitous copper network, there is usually a regulatory requirement for this asset
to be made available to alternate network operators so that they can deploy their own
ADSL equipment over the copper pairs, through a process known as ‘unbundling’,
as described in the following section.
ADSL and the unbundling of the local loop
Chapter 4 introduces ADSL equipment as one of a family of digital transmission
systems designed to operate over copper pairs in the access network, and Chapter 2
(Fig. 2.8) describes the overall routeing of the Internet access over ADSL. Now we
will take a more detailed view of the arrangement within the copper access network,
as shown in Fig. 5.4. In order for a network operator to provide broadband service to
a particular subscriber the ADSL terminating equipment, i.e. the DSLAM, needs to
108 Understanding telecommunications networks
256 kbit/s
(500 kbit/s, 1 Mbit/s,
2 Mbit/s, 2 Mbit/s,
8 Mbit/s, etc.)
Figure 5.4
To data
Basic ADSL Architecture
have been installed at the serving local exchange. In practice, the functions of splitter and DSLAM multiplexor are incorporated within equipment capable of serving
several hundred subscribers. (Chapter 8 considers the data transport capabilities of
ADSL systems.) When a subscriber requires broadband service an ADSL splitter and
termination is installed at the subscriber’s premises, and a jumper connection is made
at the exchange MDF to link the copper pair to a port on the DSLAM equipment.
The need for the network operator to send a technician to the subscriber’s premises
is avoided by the self installation of small filters at each of the subscriber’s telephone
In many countries, particularly those of Europe and North America, the Regulator
requires the incumbent PSTN operator to make parts of the Access Network available
to other PNOs so that they can provide ADSL-based broadband service to their own
customers. The term ‘unbundling’ is used to describe this opening up of the access
network at various points. In theory, the local loop can be unbundled at any point
between the subscriber premises and the MDF, for example, by allowing the PNO
to connect to the subscriber’s pair at a street cabinet or primary connect point. The
latter, which is known as ‘sub-loop unbundling’ is applicable to VDSL where the
PNO optical fibre terminates at a multiplexor in the street and uses the copper loop
for just the final drop. Full loop unbundling is, however, applicable to ADSL delivery
(described later). There is a final unbundling option, in which the incumbent network
operator shares the copper pair with a PNO; the incumbent providing telephony (i.e.
‘plain old telephone service’ or ‘POTS’) over the bottom 4 kHz of bandwidth (also
known as ‘baseband’), while the PNO uses the higher frequencies for the broadband
Transmission networks 109
service. Needless to say, the line-sharing unbundled option requires both operators to
agree careful operational procedures for the maintenance of the copper local loop, as
well as the provision and maintenance of the telephony and broadband services [5].
However, the most common approach is the use of full unbundling, in which the
demarcation between the two operators is at the MDF in the serving local exchange.
Here, a pair of copper jumper wires is extended from the subscriber’s line termination on the MDF to a ‘hand-over distribution frame’ (HDF) associated with the
PNO’s DSLAM equipment, which is typically located in a segregated area of the
exchange. The incumbent network operator allocates a protected and segregated area
of the exchange to various PNOs providing unbundled services. The latter pay rental
charges to the network operator for both the subscriber pairs and the floor space in
the exchange, use of power, standby arrangements, etc. The alternative arrangement
is for the PNO to extend a tie cable from the incumbent’s MDF to their remotely
located HDF and DSLAM housed in their own accommodation.
The performance of ADSL is progressively being upgraded via advances in the
underlying electronics (i.e. DMT or OFDM) technology. However, it is important to
appreciate that the rate at which data can be carried over a copper pair is decreased
by the presence of:
(i) crosstalk due to interference from signals (e.g. ADSL or HDSL) on other pairs
in the same cable and
(ii) impulsive noise outside the control of the network operator (e.g. due to electrical
induction from power cables or from other operators’ cables in the vicinity).
Both effects increase with distance and with the proportion of pairs in a cable carrying
DSL systems (known as ‘the fill’).
The first generation ADSL is capable of about 8 Mbit/s downstream over pairs up
to some 2.5 km in the absence of interference, but nearer 1.5 km on cables with realistic
amounts of interference from other pairs in the same cable. The second generation
equipment – ‘ADSL2’ – provides an increased maximum rate of about 12 Mbit/s
(due to the improved frame structure). However, the increased rates will be available
only on short lines, and just a modest improvement over first generation ADSL will
be achieved over longer lines. A further development, ‘ADSL2+’, increases the rate
still further to about 24 Mbit/s through the use of a wider spectrum over the line –
2.2 MHz instead of the 1.1 MHz used by ADSL and ADSL2. But again, due to the
rapid attenuation of the higher frequencies with distance, in practice the higher data
rates will be restricted to the shorter lines. Thus, overall ADSL2 and ADSL2+ will
provide higher downstream data rates on only a relatively small proportion of lines.
With the move to so-called next generation network (NGN), operators will have
the option of incorporating the ADSL DSLAM equipment into multi-purpose terminating equipment located in street electronics or exchange buildings. For BT’s NGN
this new equipment is known as full service access nodes (FSAN) in which a range
of subscriber access lines are terminated, including ADSL, standard POTS and ISDN
lines, as described in Chapter 11.
110 Understanding telecommunications networks
Microwave radio
Copper pair
Optical fibre
Coaxial cable
Microwave radio
(POTS only)
ADSL = Asymmetric DSL
VDSL = Very high speed DSL
FTTK = Fibre to the kerb
HFC = Hybrid fibre-coax
PON = Passive optical network
POTS = Plain old telephone
FTTH = Fibre to the home
FTTO = Fibre to the office
PCP = Primary connection
LE = Local exchange
Mux= Multiplexor
PSTN = Public-switched
telephone network
Figure 5.5
Cable TV
Street Mux
Direct fibre
Broadband Options for the Access Network
Broadband options for the Access Network
There are a wide range of transmission systems technologies that a telecommunications network operator or Cable TV network operator can use to provide broadband
access to customers, as illustrated in Fig. 5.5. As described earlier in this chapter, this
range covers:
Copper pair cable
• ADSL2+
Hybrid optical fibre and copper pair cable
Hybrid optical fibre and coaxial cable (HFC)
Cable modem
Optical fibre
Direct fibre (FTTH, FTTO)
Transmission networks 111
Microwave radio (generically known as broadband radio access, BRA)
WiFi (IEEE 802.11)
WiMax (IEEE 802.16)
LOS microwave radio
Point-to-point and point-to-multipoint radio
All of the above systems have differing characteristics in terms of bandwidth delivered, technology used, operational and distance constraints, and – of course – cost.
In general, the unit cost of transmission systems deployed in an access network
depends on the density of subscribers served. This is because the cost of the electronics located in an exchange, and street cabinet, has to be apportioned across the
number of subscribers served within the equipment’s catchment area. Against the
cost dimension, the network operator needs to offset the potential revenues from
the services delivered. A typical choice of access systems for the various subscriber
densities and market segments (indicative of the expected revenues) is illustrated in
Table 5.1. In practice, the type of systems adopted will differ depending on the many
operational and operational factors relating to the network.
Table 5.1
Segmentation of Broadband Access Systems
Customer Density
SME2 business
Direct optical
fibre, PON
Large business
Direct optical
fibre, PON
Very small business
radio (e.g.
Coroprate business
ISP & service
Direct optical
fibre, PON
1 WiMax may be directly deployed in this sector for some customers or may be combined with optical
2 SME: small-medium enterprise.
ADSL: Asymmetrical digital subscriber line; HFC: Hybrid fibre-coax; LMDS: Local microwave
radio distribution service; PON: Passive optical network; SDSL: Symmetrical digital subscriber line;
VDSL: Very-high speed digital subscriber line.
112 Understanding telecommunications networks
The copper pair cable is still the main basis of all the World’s PSTNs, even
though the technology and design dates from the early 1900s. Although optimised
for voice, the copper network has proved remarkably versatile in its ability to carry
ever greater bandwidths using the addition of electronic systems, such as ADSL, and
so continuing its utility in today’s ‘Internet age’. Nevertheless, the high levels of
costly manual interventions involved in the provision of service, rearrangements and
maintenance of the copper access network has been a driver for network operators
to seek replacements with lower operational expenditure (‘opex’). Unfortunately, all
the potential alternatives, e.g. direct optical fibre, fibre to the kerb (FTTK), PON,
multipoint radio, etc., have proved to be uneconomical as wholesale replacements of
the copper network. This is due to the difficulty in achieving opex savings in practice
with such conversions and the absence of sufficiently large extra revenues from the
new services that could be offered from the use of the new broader bandwidth systems.
The latter is often referred to as the lack of a ‘killer application’, i.e. a profitable
application or service that can be supported only by the new technology.
The continual downward pressure on prices, and in particular the relatively low
value people place on extra bandwidth, may prolong the difficulty in finding the business justification for the wide-scale replacement of copper. Furthermore, the growth
in the unbundling of the copper network, which creates a dependency by an increasing
number of alternative network operators on copper, is likely to exacerbate the position. This problem has been tackled in some countries by the national governments
assisting the move to a politically attractive broadband-enabled access network by
subsidising the incumbent operators in their large scale deployment of optical fibre
as a replacement for copper pairs. In the absence of such subsidies or a suitable killer
application, the access networks are likely to continue to have a large component
of copper pairs, together with a mix of optical fibre and radio systems for the near
5.2.6 Planning and operational issues
Planning and operating the Access Network represents the major annual cost of
any PSTN operator. The copper access network, in particular, involves substantial
amounts of manual intervention in the provision of service to new customers, maintenance and the installation of new or replacement capacity. Consequently, by far the
largest number of an operator’s technicians work in the access network, incurring the
highest contribution to the current account cost.
Network operators constantly endeavour to reduce the amount of manpower
involved in the routine actions which are performed many thousands of times a day,
such as the provision of a new telephone line, or the repair of a fault on the line.
Such reductions are enabled by improvements to the processes being followed (i.e.
‘process engineering’) and by the use of computer support systems to provide line
plant inventory, track fault histories, etc. Computer support systems can also help the
despatch supervisor in the deployment of technicians to an area, taking into account
the location of the various jobs that need to be done that day, so minimising the
amount of travelling time spent going to subscriber premises.
Transmission networks 113
Provision of a new subscriber’s line
In a copper access network, the provision of a new subscriber’s line involves a visit
to the subscriber’s premises and relevant parts of the external network, as well as
actions within the exchange building, the service-management centre (SMC), and
within the administration offices of the network operator. Briefly, the actions may be
summarised as follows:
• Order taken by sales administrator (over the telephone, by e-mail or fax), who
is able to check availability of line plant to the street address and allocate a telephone number. Both actions are usually supported by a computer order-handling
• The order is received from the sales administrator by the technician dispatcher,
who schedules an external line plant technician job. In parallel, a job is set up for
the internal exchange work and for the SMC.
• At the SMC a technician at a control console enables the allocated line card in
the local concentrator switch of the serving exchange (see Chapter 6) against the
given telephone number. Dial tone is then detectable on the wires connecting the
line card to the MDF.
• A technician adds a jumper wire from the exchange side of the MDF to the
allocated pair in the external cable to the serving street cabinet (i.e. the PCP).
• The external technician then visits the street cabinet and installs a jumper wire
to link the scheduled E-side (the line going to the exchange) and D-side (the
line going to the subscriber) copper pairs. If there is more than one flexibility
point involved (e.g. an SCP) then this needs to be jumpered also. The technician
should be able to detect 50 V DC and the dial tone on the completed loop to the
exchange. The technician makes the final drop connection by either running an
overhead drop wire or an underground lead in to the subscriber’s premises, as
described earlier. The exchange line is terminated in the subscriber’s premises
at an appropriate block terminal – the NTTP (see Box 5.1). This is the point
from where the subscriber is able to run internal wiring to their own equipment,
such as cordless telephone systems and answer machines. At this stage of the
process, the subscriber should then be able to verify that telephone service has been
• Successful completion of line provision is then noted in the administrative system
so that directory information is recorded and the billing process initiated.
A corresponding process is required for the installation of ADSL to this basic copper
pair, as mentioned above. In addition, there are processes for the installation of all the
broadband access systems listed in Table 5.1. For business premises with sufficient
demand for telephony, corporate voice networks, private circuits, VPN, centrex and
data services (as described in Chapter 2) the network operator will deploy SDH transmission payload over direct optical fibre. Chapter 11 looks at the architectural aspects
of the management of operations and maintenance within a telecommunications
114 Understanding telecommunications networks
Box 5.1
Network Termination
The key feature of an Access Network is the user-to-network interface presented
at the subscriber’s premises. The actual interface depends on the telecommunications service being provided by the network operator and the type of user’s
terminal. Because of the influence on terminals and service – and hence the competitive opportunities and constraints – this interface is defined by the national
telecommunications regulator. In the United Kingdom, a network termination
and test point (NTTP) is identified for all operators [9], usually referred to as
the NTP.
As Fig. 5.6(a) shows, the NTP is actually a reference point located inside the
network terminating equipment (NTE) on the user’s premises. The NTE terminates the line transmission system and may also include a safety barrier between
the subscriber’s equipment – which may be powered by mains electricity – and
the access line. This provides protection for the operator’s technicians working
on the copper lines as well as protecting the user’s from any dangerous induced
voltages on the copper line.
The NTE is deemed to be the edge of the network, and hence the point
where the operator’s network service ends, as shown in Fig. 5.6(b). However,
the user’s application and the actual service seen by the users really extend
between the ends of the apparatus or terminals at both ends. Very often, users
will attach several terminals directly and indirectly to the NTE.
The NTP is also a major test point for maintenance activities, such as fault
location and diagnosis. Operators arrange for a loop to be placed across the
line at the NTP to check continuity within the access link, so that a fault can be
located to the network side or to the user’s terminal side. Often, this looping
within the NTE can be initiated under remote control from a maintenance centre
in the network.
Spare plant
Most network operators aim to be able to provide a new subscriber line within a
matter of a few days after the order is received. Clearly, for this to be achieved there
needs to be a stock of spare pairs throughout the copper access network sufficient to
meet the demand for new service. If there are no spare pairs in the serving D-side
cable or E-side cable connecting the PCP to the exchange, then either new capacity
needs to be added to the network or some alternative means found to provide service. The installation of new capacity will incur an appreciable delay (e.g. weeks or
months) before service can be provided to the subscriber. Alternative means of providing service include the use of other cables off another PCP – sometimes possible
in dense urban areas, or the use of radio links. Sometimes, the use of a pair-gain
system (as described in Chapter 4), using electronics to derive two local loops off
a single pair, can enable service to be readily given to a new subscriber using the
existing pair of a neighbouring subscriber. However, in general, alternative means of
Transmission networks 115
Safety barrier
(a) The NTE functions
(b) The extent of network service
Network service
User’s application
Figure 5.6
Network Termination
providing a subscriber line are more expensive than providing service from pairs in
There is an economic trade-off between having sufficient spare copper pairs in
the ground in order to meet demand for new service directly from stock on the one
hand and the sunk cost of the unused asset of the idle capacity. This asset management
problem for a network operator is similar to that faced by a grocery shop: what is
the optimum number of tins of baked beans that needs to be on the shelf so that
all expected customer demand can be met – not having tins available means loss of
business, having too many ties up capital unnecessarily? For most PSTN operators the
inability to provide a new telephone line within a reasonable time will either cause loss
of business as the potential customers take their service instead from a competitor, or
the intervention of the national regulator. Operators determine an internal target figure
for meeting demand for telephone service from stock, based on their assessment of
this economic trade-off of stock size, the cost of alternative methods of provision and
loss of business, typically set at around 95 per cent.
Copper access network planners therefore use the 95 per cent (say) figure with their
forecasts of future demand for new telephone lines to dimension the cable network.
Finally, it is worth emphasising that the highly granular nature of the Access Network
means that the planning decisions are critical – capacity is needed down the actual
streets to serve potential subscribers; having plenty of spare capacity in cables running
down other streets will not enable this demand to be met.
116 Understanding telecommunications networks
(a) No aggregation
(b) With aggregation
AC + AB + AD + AE + AF
CA + BA + DA + EA + FA
BC + BA + BD + BE + BF
CB + AB + DB + EB + FB
AD + AE + AF +BD + BE
+ BF +CD + CE + CF + DA
+ DB +DC + EA + EB +EC
+ FA + FB + FC
30 Traffic routes carried
on 15 small transmission
EA + EB + EC + ED + EF
AE + BE + CE + DE + FE
FA + FB + FC + FD + FE
AF + BF + CF + DF + EF
30 Traffic routes on 5 large transmission systems
Figure 5.7
Economies of Transmission Aggregation [Ward]
Core Transmission Networks
5.3.1 Scene setting
Unlike the Access Network, which may link some tens of thousands of subscribers to a
single serving local exchange, the Core Transmission Network provides links between
relatively small numbers of network nodes, typically spread across the whole country.
These transmission nodes are points in the national network where bundles of circuits,
serving telephony, private circuits, data services, etc., are extracted from or entered
onto the required transmission links. Economies of scale are achieved by multiplexing
onto as few, but large, transmission systems as possible – since the unit costs of transmission systems decreases with system size. Thus, in many cases it is economical
to unpack and re-pack transmission links at intermediate points within a network in
order to achieve optimum loading of transmission links. This principle is illustrated in
Fig. 5.7, where 30 unidirectional traffic routes require 15 small transmission systems
when carried directly (Fig. 5.7(a)), whereas the same set of traffic routes could be carried over just five large transmission systems using aggregation (Fig. 5.7(b)). Whilst
this leads to economies of scale in the transmission cost, it does also incur the need
for packing and unpacking at the C and D nodes. This activity involves demultiplexing and multiplexing together with a manual or automatic means of jumpering (i.e.
cross-connection) between them – a process known as ‘transmission-flexibility’. The
nodes at which this flexibility is provided are known generally as ‘core transmission
stations’ (CTS), and as ‘transmission repeater stations’ (TRS) in the United Kingdom.
CTS are usually located at each of the trunk and international exchanges and
many of the larger local exchanges. However, in addition there may be some CTSs
located in buildings which contain only transmission equipment. Fig. 5.8 illustrates
the concept with an example of five CTSs (A to E), which are supporting the core
transmission between a set of trunk telephone switching units, data nodes and private
Transmission networks 117
Exchange A
Exchange B
Optical fibre cable
Exchange D
Figure 5.8
= Transmission flexibility at CTS
PC = Private circuit node
DU = Data node
TSU = Trunk telephone switching unit
Exchange E
Transmission Network Configuration
circuit nodes. Exchange building A contains the CTS-A which serve the co-located
trunk switching unit, data unit and private circuit node, and a similar situation exists at
exchange buildings B, D and E. However, the CTS providing transmission flexibility
for core transmission links between A, B, D and E is not located in an exchange
building, but rather at a transmission-only building at an appropriate transmission
network nodal point.
5.3.2 PDH network
Chapter 4 and Fig. 4.13 introduced the concept of the ‘multiplexor mountain’
associated with Core Transmission Networks using PDH equipment. The practical
arrangement at a CTS is shown in Fig. 5.9, in which a 2 Mbit/s block is extracted
from the incoming 140 Mbit/s PDH transmission system on the left and inserted into
the outgoing 140 Mbit/s PDH transmission system on the right (and vice versa for the
return direction of transmission). The CTS is composed of racks of transmission terminal equipment (not shown in Fig. 5.9 for clarity) and three stages of multiplexors:
140/34, 34/8 and 8/2 – each connected to a common DDF, or to one of several DDFs.
In this example, the incoming 140 Mbit/s system terminates on 140/34 Mbit/s mux.
No. 1, from which a jumper wire is run from output No.1 to the input of 34/8 Mbit/s
mux. No. 1, and so on via the central DDF to the output multiplexor mountain through
to 140/34 Mbit/s mux. No. 11.
118 Understanding telecommunications networks
DDFs may also
be located here
2 Mbit/s
140 Mbit/s
systems 34 Mbit/s
8 Mbit/s
Figure 5.9
2 Mbit/s
140 Mbit/s
2 Mbit/s
DDFs may
also be
PDH Core Transmission Network Station
Fig. 5.9 also shows the extraction at the DDF of 2 Mbit/s blocks from the input
140 Mbit/s transmission system (via 34/8 Mbit/s muxs No. 1 and No. 4, and 8/2 Mbit/s
muxs No. 1 and No. 13) to the co-sited telephone switching unit.
Transmission flexibility is also possible at any of the intermediate transmission
rates of 34 and 8 Mbit/s through appropriate jumpering at the DDFs. Such a facility
would be used, for example, to route a 34 Mbit/s digital private circuit through
the CTS.
The jumper wires on the DDFs are coaxial type cables, similar to those used
to connect the domestic TV to its aerial. Each of the jumperings have to be made
manually when the routes are set up. Not only does this represent a current account
cost for the network operator, but the periodic intervention on the DDFs to add or
remove jumper wires also introduces disturbance to the established jumpers, often
resulting in damage and faults. Electronic replacements for the jumper wires and
DDFs, using digital cross-connect equipment, known as ‘DXC’ in the USA, are used
by some network operators. However, for cost reasons most network operators tend to
use DXC to provide the transmission flexibility only for the high value international
transmission systems at the larger international gateway CTSs.
5.3.3 SDH network
An SDH CTS uses a combination of add–drop multiplexor (ADMs) and DXC equipment (see Chapter 4) to provide the necessary transmission flexibility, as shown in
Transmission networks 119
2 Mbit/s
140 Mbit/s
155 Mbit/s
2 Mbit/s
140 Mbit/s
155 Mbit/s
Figure 5.10
SDH Core Transmission Network Station – DXCs
Fig. 5.10. This automated system contrasts with the more-complex and manually
intensive arrangement shown for PDH shown in Fig. 5.9. The configuration of the
ADM and the DXC are managed through a computer-based controller, which may
be co-sited or remotely located. As described in Chapter 4, use of SDH equipment
enables a fully managed transmission network to be created. The SDH configuration
controller allows the network operator to manage the configuration of the CTS flexibility points, through planning and assignment processes, as well as re-configurations
in real time to compensate for transmission link breakdowns. SDH also allows the
end-to-end performance of an SDH routeing through several CTSs to be monitored
automatically – a particularly attractive feature for digital private circuit customers.
An SDH CTS comprises a DXC on which the high speed transmission links terminate. For the example of a 2 Mbit/s block extraction from the incoming 155 Mbit/s
link the DXC needs to be able to identify and manipulate the appropriate 2 Mbit/s
tributary from the incoming SDH link and pass it to the outgoing link. This requires
the DXC to have a 2 Mbit/s ‘granularity’, i.e. to be able to cross-connect at the
2 Mbit/s level. Typically, a DXC will need a range of granularities to cope with the
PDH rates still used in the network (i.e. 2, 8 and 34 Mbit/s), as well as the SDH line
rates (i.e. 155, 622 Mbits/s and 2.5 Gbit/s) to enable a practical set of transmission
flexibility levels. This is illustrated in Fig. 5.10, in which the DXC is divided into
a higher-order DXC switch-block handling the SDH transmission rates and a lowerorder DXC switch-block handling the 2 Mbit/s and other PDH rates. Any legacy
PDH transmission systems still in the network terminate on special PDH terminal
interfaces to the SDH DXC system, as shown in Fig. 5.10 [6].
120 Understanding telecommunications networks
2 Mbit/s
155 Mbit/s
155 Mbit/s
Figure 5.11
SDH Core Transmission Network Station – ADMs
At smaller transmission nodes ADMs only are used to extract digital transmission
blocks (at the PDH rates of 2, 8 and 34 Mbit/s) for the co-sited telephone switching
unit, private circuit, and data units within the exchange building associated with the
CTS, as shown in Fig. 5.11.
In order to maximise their management capability, SDH networks are usually
structured in a set of hierarchical levels. Fig. 5.12 illustrates a typical SDH network
structure using DXCs and ADMs. A useful analogy which can help in visualising
this structure is to consider it as an upside-down set of trees, with the trunks at the
top level of the national SDH network and the succession of branches reaching down
into the lower transmission levels. (Unfortunately, the tier designations increase the
lower the level!) At the top level (Tier 1) of the national network is a mesh of highcapacity SDH transmission links between flexibility nodes (CTS) of DXCs. This
forms the inner portion, or backbone of the Core Transmission Network and links
the major trunk exchanges, as well as private circuit and data nodes. Hanging off this
level at Tier 2 is a set of SDH rings linking ADMs within a region of the country,
serving smaller trunk exchanges, local exchanges and other nodes. Tier 2 nodes serve
the access network, with links to large business customers (‘Access SDH’), small
local exchanges or other customer service aggregation points. Above the Tier 1 of
the national network is the international portion of the Core Transmission Network,
the Tier 0, which links to the network’s transmission gateways to the transmission
networks of other countries – via submarine cable landing stations, microwave radio
stations or satellite Earth stations [7].
Of course, the technology used in both the Access and Core Transmission Networks continues to develop in terms of cost per circuit or unit bandwidth carried and
Transmission networks 121
Tier 0
Earth station
Tier 1
= Digital crossconnect (DXC)
= Add–drop mux (ADM)
Tier 2
Tier 3
Figure 5.12
SDH Transmission Network Structure
management and service features. Consequently new types of transmission systems
will be introduced periodically to the network. For example, DWDM, described in
Chapter 4, has been added to the existing SDH over optical fibre major transmission
links within the core network to increase the payload capacity. Whilst adding DWDM
on a link by link basis gives proportionate capacity increases, the widespread availability of multiple wavelengths over the optical fibre network offers the attractive
possibility of managing transmission routeings across the network on a wavelength.
Wavelengths can be added and extracted to/from optical fibre networks using wavelength add–drop multiplexors (WADM) and interconnected by optical cross connects
(OXCs) to create so-called λ-routeing (‘lambda routeing’), thus creating a so-called
managed optical platform, in which a connection-pattern of different wavelengths of
light support the SDH payloads carried over the optical fibre network [8]. This may
give rise to new network structures and architectures, as discussed in the section on
NGN in Chapter 11.
5.3.4 Transmission network resilience
In practice, from time to time there will be events that cause a transmission link
to fail. Typical examples of such events include the physical severing of the link
resulting from storm damage of an overhead cable, or a digger slicing through an
underground optical fibre, or alternatively the terminal equipment in the CTS might
become faulty. There are other occasions where planned work by the operator, e.g.
adding new capacity or re-arranging the use of transmission links, will necessitate
an arranged outage on one or several links. Whatever the cause, the effect is the
withdrawal of connectivity across the Core Transmission Network, affecting perhaps
122 Understanding telecommunications networks
several thousand channels of capacity between two nodes. The ability of the network
to cope with this disruption is often referred to as ‘resilience’.
The inherent resilience of the network can be enhanced by remedial action within
the Core Transmission Network, as described in this section, and at the switched
traffic level, as described in Chapter 6 – while the overall contribution to network
performance and quality of service (QOS) as perceived by the customers is considered
in Chapter 11. All methods of increasing resilience rely on the addition of some extra
network capacity. Thus, a resilient network is one in which its optimum structure and
capacity is augmented by some degree of redundancy. The appropriate extent of this
over provision is, of course, determined by a cost-benefit balance set by the value of
the traffic being carried over the network which would otherwise be lost.
Fig. 5.13 illustrates the various levels of resilience possible in a telecommunications network. At the top is the switched network, where a set of calls (i.e. ‘traffic’)
between switching units at exchanges A and B could automatically be routed via the
switching unit at Exchange C in the event of a failure on traffic route A–B. This alternative routeing of the calls between exchanges A and B would incur throwing extra
load on the traffic routes A–C and C–B, as well as the cost of switching the extra calls
at C. Although this use of alternative routeing gives a good increase to the overall
network resilience, its use is limited by the ability of the alternative routes, A–C,
C–B, which are already carrying their normal load of calls, to handle the offloaded
traffic from A–B without incurring unacceptable overload and congestion. Automatic
alternative routeing (AAR) is considered further in Chapter 6. The remainder of this
section will consider the resilience measures that can be taken in the four levels of
the Core Transmission Network below the traffic switched level shown in Fig. 5.13.
Switched-traffic level
(re-routeings of traffic
paths between exchanges)
Transmission-circuit level
(re-routeings between DXC)
Figure 5.13
Levels of Network Resilience
Transmission-path level
(e.g. diversity, rings)
Transmission-media level
(sparing & serviceprotection networks)
Physical-media level
(cables, radio links)
Transmission networks 123
The various measures described below fall into two camps: those that enable
instantaneous full replacement of the lost capacity, so ensuring service with little or
no perceived interruption, and those that mitigate the loss to an acceptable degraded
level of service. At the transmission-circuit level of the Core Transmission Network
(Fig. 5.13) the traffic route between exchanges at A and B can be provided by a direct
a–b transmission circuit route or, in the event failure, routed via a transit CTS d.
Each transmission route can itself be protected at the transmission path level by the
use of physical diversity or the use of rings. Each of the diverse routes, or links of
the rings, can be further protected at the transmission-media level by the automatic
switching to standby capacity using sparing or service-protection networks. Finally,
at the physical media level the resilience can be enhanced by the use of physically
protected line plant.
Transmission-circuit level switching at digital cross connection units
Apart from facilitating the deployment of capacity, the use of DXCs at the top tier of an
SDH core transmission network (Fig. 5.12) enables the circuits to be re-routed in the
event of link failures. This form of resilience relies on appropriate spare link capacity
being available, as well as the ability to detect link failures and decide on the suitable
re-routeings. The latter may be handled manually or through a fully automated control
system, which is usually complex and expensive. Obviously, this form of resilience
is available only at the Tier 1 of the core transmission hierarchy and it is, therefore,
limited to protecting just the links (which also tend to be large capacity) between the
major nodes.
Transmission path level: diversity
This simple, but highly effective, method of increasing resilience is through the
spreading of the link capacity across two, three or sometimes four parallel separate
(b) Physical diversity
of three
(a) Physical diversity
of two
Figure 5.14
Transmission-Path Resilience: Diversity
124 Understanding telecommunications networks
transmission routeings, as shown in Fig. 5.14. As an example, we will consider a
typical three-way diverse routeing of, say, 90 circuits capacity on the traffic route
between exchanges A and B (Fig. 5.14(b) refers). This means that 30 circuits are
carried over each of the three parallel paths (‘engineering routeings’) – No. 1 (A–B
direct), No. 2 (A–z–B), and No.3 (A–x–y–B). Each of these ‘routeings’ is chosen to
give as much physical diversity as possible, i.e. through the use of separate cables
and CTSs following different paths (roads and towns) between exchanges A and B.
In the event of failure of any one of the diverse routeings only one third, i.e. 30
circuits, of the traffic route A–B capacity would be lost. If telephony service was
being carried, this loss would be perceived by the users as an interruption to the
calls in progress over the affected link and a general increase in the chance of call
congestion on the whole traffic route due to its effective reduction to 60 circuits. This
is significantly more acceptable than the total failure of all calls between exchanges A
and B that would otherwise have occurred (in the absence of resilience at the switched
traffic level). The cost to the operator of increasing resilience by diversity results from
the use of three smaller sets of transmission links, with their inherently higher unit
costs, rather than a single large more cost-efficient link.
Transmission path level: SDH and SONET rings
SDH and SONET transmission systems have the valuable facility of a supervisory
channel which enables them to work in a managed transmission network, as described
in Chapter 4. In particular, resilience at the transmission-path level can be provided by
the use of spare capacity between ADM when associated with SDH or SONET carried
over optical fibre deployed in a ring structure. These systems are referred to as selfhealing rings (SHR). There are several types of SHR used in current networks, e.g:
Multiplex section – shared protection ring (MS-SPRING)
Multiplex section – dedicated protection ring (MS-DPRING)
These may be based on two or four optical fibre arrangements.
Fig. 5.15(a) shows a simple example of a two-fibre MS-SPRING comprising six
ADMs, labelled i to n. The spare capacity within the optical fibres joining the ADMs
is shown as dotted lines, and the working capacity between the various ADMs is
shown as full lines. Attention is drawn to circuit 1, entering at ADM j, routeing via
ADMs k and l, and leaving at ADM m; and circuit 2 entering at ADM k and leaving
at ADM l. (Note, only one fibre carrying one direction of transmission is shown for
simplicity; a corresponding fibre is required for the return direction.) In the event
of a break in link k−l, ADMs k and l initiate a switch of circuit 1 onto the spare
capacity back to ADM j, and hence via i, n, m, and switch at l and back out at m –
thus restoring the j−m connectivity, as shown in Fig. 5.15(b). Similarly, the circuit 2
connectivity is restored by switching within ADM k onto spare capacity which routes
via ADMs j, i, n, m and switches out at ADM l.
Transmission-media level: use of spare links and service protection network
An individual transmission link can be protected by running a duplicate standby link
in parallel. There are three main configurations for the use of spare links, as shown in
Transmission networks 125
Ring with vacant
standby paths
Break between k and l;
Circuits 1 and 2 routed over
standby paths around RING
Figure 5.15
Transmission-Path Resilience: SPRINGS
Fig. 5.16. The simplest but most robust and expensive arrangement is the so-called hot
standby, or 1 + 1 Protection, in which the live traffic is run over both the worker link
and standby duplicate link simultaneously (Fig. 5.16(a)). Both line system terminal
equipments (LTE) continuously monitor the signal from both links, and in the event
of failure of the worker the output from the standby link takes precedence.
A more-complex, but generally cheaper arrangement is the 1 : 1 protection system
shown in Fig. 5.16(b) – one of several automatic protection systems (APS). Here, the
standby link is normally idle and the traffic is switched at the LTEs from the worker
link in the event of its failure. This has the advantage of allowing the standby link to
be used by the network operator to carry other, but lower priority, traffic while it is
in standby mode, on the basis that it will be lost whenever the standby link needs to
take over from the worker.
Finally, Fig. 5.16(c) shows the more usual ‘1 : N Protection’ scheme for transmission networks, in which ‘N’ worker links are protected by a single standby link.
Whilst this is cheapest of all three schemes in terms of the least level of link capacity
redundancy, it does suffer from the limitation of being only able to handle the traffic
from one failed worker link at a time. Again, the standby link can be used to carry
low priority traffic when in standby mode.
An extension of this concept of the use of spare links is used in service protection
networks (SPNs) – in which a full network of standby links extending across the
country is deployed. In the event of failure, transmission capacity is switched over
to the SPN, but unlike the single link 1 : N spare arrangement, several links across
126 Understanding telecommunications networks
(a) 1 + 1 Protection
(b) 1 : 1 Protection (APS)
N workers
(c) 1 : N Protection (APS)
Figure 5.16
Transmission-Media Resilience: Use of Spare Links
the SPN may be used to carry the offloaded traffic. An SPN therefore needs not only
a set of spare transmission links but also an automatic control and switch system to
detect and respond to link failures rapidly. The expense and complexity of an SPN
means that they are used to protect only the major transmission links between the
most important nodes in a network.
Physical media level
At the individual physical media level, resilience can be increased by the use of
extra strengthening in the construction of the optical fibre cables, copper cables,
etc. Two typical examples of such measures are the use of armour-plated cable in
submarine cables – providing protection against shark attack and fishing trawlers –
and strengthened aerial cable for use in exposed, windy terrain, such as hills and
This chapter considered how the various transmission systems described in Chapter 4
are used to construct transmission networks. In Section 5.2 we addressed the Access
Network, covering the ubiquitous copper local loop network, as well as the use of
optical fibre and microwave radio. We noted that a key aspect is the ability of the
access network to carry increasingly higher capacities to the residential and small
business users – the so-called broadband access – as well as for business customers.
The importance of the planning, operational and regulatory issues for the access
network were also described.
Transmission networks 127
In Section 5.3 we considered the structure of the Core Transmission Network and
the role of the CTSs at the transmission nodes, looking in particular at:
• PDH networks
• SDH networks
Finally, the various ways in which the resilience of the Core Transmission Network
can be increased were identified.
A simple example of dimensioning a Core Transmission Network in association
with the switched telephony network is given in Chapter 6. Subsequent chapters
describe intelligence, signalling, data and mobile networks.
PARSONS, J. R.: ‘The Development of Power and Building Engineering Services in Telecommunications’, British Telecommunications Engineering, Vol.
14, April 1995. pp. 71–80.
PARSONS, J., AKERLUND, J., RIDDLEBERGER, C. et al.: ‘Powering the
Internet. Data Communications Equipment in Telecommunications Facilities:
The need for a DC Powering Option’, British Telecommunications Engineering,
Vol. 17, 1998, pp. 185–197.
SCOTT, R. P. and COOKE, P. J.: ‘A 29 GHz Radio System for Video
Transmission’, British Telecommunications Engineering, Vol. 10, July 1991,
pp. 141–146.
BEGLEY, M. P. and SAGO, A: ‘Wireless LANS’. Chapter 12 of ‘Local Access
Network Technologies’, edited by FRANCE, P., IET Telecommunications Series
No. 47, Stevenage, 2004.
BUCKLEY, J.: ‘Telecommunications Regulation’, IET Telecommunications
Series No. 50, Stevenage, 2003, Chapter 8.
BALCER, W. R.: ‘Equipment for SDH Networks’, British Telecommunications
Engineering, Vol. 10, July 1991, pp. 126–130.
GALLAGHER, R. M.: ‘Managing SDH Network Flexibility’, British Telecommunications Engineering, Vol. 10, July 1991, pp. 131–134.
HAWKER, I.: ‘Evolution of the BT UK Core Transport Network’ British
Telecommunications Engineering, Vol. 17, January 1999, pp. 298–300.
VALDAR, A. R.: ‘Provision of Telecommunication Service in a Competitive and
Deregulated Environment’, British Telecommunications Engineering, Vol. 8,
Part 3, October 1989, pp. 150–155.
Chapter 6
Circuit-switching systems and networks
Having considered in Chapters 1 and 2 how a call is conveyed across one or more
networks, this chapter is all about how the exchanges – or more precisely, the switching units within the exchanges – in a PSTN actually work. Section 6.2 considers the
basic components of circuit-switching systems, often referred to as ‘Voice switches’.
For clarity, this chapter describes the arrangement for the fixed-network exchanges
(also called ‘wireline exchanges’); consideration is given to the additional elements
that are included in a mobile network exchange in Chapter 9.
In Chapter 5 we looked at how a core transmission network carries a required
capacity between two network nodes. But, how does a network operator determine
what capacity is required to be carried? For example, are 90 circuits between Leicester
and Coventry sufficient to ensure that the expected volume of calls between these
exchanges can be carried satisfactorily, or could the operator save money by providing
just 60 circuits? Section 6.3 considers how the size of a traffic route between two
exchanges is determined, and how the flow of traffic through a network of many
exchanges is routed so as to give maximum utilisation of the capacity while achieving
the appropriate level of resilience (at the switching level). Finally, we consider a
simple example of how a network is dimensioned at the switching level, and how this
impacts the dimensioning of the Core Transmission Network.
Circuit-switching systems
6.2.1 Introduction
Circuit-switching systems form the nodal functions of the PSTN and they are thus
optimised for voice, although they do successfully carry non-voice (i.e. data) traffic
also. There are two basic types of switching unit: those that terminate and switch
subscribers’ lines – ‘local exchanges’ (or ‘Class 5 exchanges’ as they are known in the
130 Understanding telecommunications networks
Box 6.1
Exchange Designations in North America
The terminology used in the United States and Canada for the exchange units in
the PSTN differ from those used in the United Kingdom. They are summarised
Equivalent UK Designation
Regional Centre Trunk Exchange
Sectional Centre Trunk Exchange
Primary Centre Trunk Exchange
Toll Centre
Junction Tandem, Wide Area Tandem
End Office
Local Exchange
The ten Class 1 centres in the United States and two in Canada are fully interconnected (i.e. top of the routeing hierarchy). There are some 70 Class 2, 230
Class 3, 1,300 Class 4 and around 19,000 Class 5 exchanges in North America.
United States and Canada), and those that terminate and switch trunk and international
lines only, with no subscribers lines. Examples of the latter include ‘trunk exchanges’
also known as ‘main switching units’, ‘junction tandems’ and ‘toll switches’, and
‘Classes 1, 2, 3 or 4’ in North America; as well as the international exchanges.
(Box 6.1 describes the North American designations for exchanges in a PSTN.) As
is explained in the following section, it is the switching of subscribers’ lines that
incurs the major complexity and cost, and so we will start by considering the local
exchange units and later note the cut-down version that constitutes the trunk, junction
and international exchanges.
6.2.2 Subscriber switching (local) units
A simplified generalised local switching unit is shown schematically in Fig. 6.1.
This representation combines the functional diagrams of Fig. 1.6 of Chapter 1 with
Figs 3.2(a) and (b) of Chapter 3. At the left side of the unit subscriber’s lines terminate
on line cards forming the front-end of a subscribers’ concentrator switch. (These line
cards incorporate the off-hook detector, ring-current feed and power feed shown in
Fig. 1.6.) A call initiated by subscriber A is connected across the concentrator to
one of the chosen free outlets on the internal link to the route switch. If the call is
to another subscriber on that exchange – known as an ‘own-exchange call’ – it is
connected through the route switch to a chosen free outlet to the called subscriber’s
(i.e. B’s) concentrator switch. The call can be completed by connecting through this
concentrator switch to the line card of subscriber B, and hence to the terminating
local line. This path, shown as a-b-c-g-h in Fig. 6.1, is established exclusively for
this call until subscriber A or B clears down – a characteristic of circuit switching.
The signalling and control functions necessary to set up the calls across the switches
Circuit-switching systems and networks 131
Concentrator switch
acting in
Signalling and control
Local exchange switching unit
= Subscriber’s line card
Figure 6.1
= Control link
Simplified Generalised Local Exchange
are considered in Chapter 7. There are two key points concerning Fig. 6.1 to note at
this stage.
i) Although the call is initiated by A and so the traffic flow is deemed to go from A to
B via the three switches along the connection path described above, transmission
paths within the switches are of course required to flow from A to B (‘Go’ direction) and from B to A (‘Return’ direction) to enable a conversation, corresponding
to the 4-wire circuit described in Chapter 3 (Fig. 3.14).
ii) With the direction of traffic flow shown for path (a) subscriber A’s concentrator
switch is concentrating calls from many lines onto fewer outlets (a concentration
ratio of 10 : 1 being typical). The path (c) through the route switch is not concentrated, since the latter has an equal number of inlets and outlets; the switch is
therefore performing the function of interconnecting subscriber A’s concentrator
switch to subscriber B’s concentrator switch. However, on the exit side of the
exchange the traffic flow over path (g) is passing through subscriber B’s concentrator switch over path (h) from concentrated side to non-concentrated side,
since it has, say, ten times more outlets than inlets. The call path is therefore
experiencing ‘expansion’ through this switch.
Fig. 6.1 also shows that external (traffic) routes from other exchanges terminate
on the route switch. This enables calls from these exchanges to be connected via the
route switch to the input to subscriber B via its concentrator switch (albeit, acting in
expansion mode), as shown by traffic path i-e-g-h.
Calls from subscriber A that need to be routed via other exchanges are also
switched across the route switch (‘Go’ and ‘Return’) to the appropriate outgoing
traffic route, shown as path a-b-d-j in Fig. 6.1.
132 Understanding telecommunications networks
Finally, the switching unit could act as a tandem by providing connectivity (‘Go’
and ‘Return’) between traffic routes terminating on the route switch, as shown by
path i-f-j in Fig. 6.1.
We now need to redraw the simplified configuration of Fig. 6.1 to reflect the fact
that subscribers A and B may actually be on the same concentrator. This leads to the
concept of folding the switch configuration so that all subscribers’ lines are on the
left-hand side and all external traffic routes are on the right-hand side of the Route
switch. The connections through the route switch thus may need to loop back, as
shown in Fig. 6.2 for path a-b-c-g for calls between subscribers A and B on two cosited concentrator switches, and path a-b-k-l-m for calls between subscribers A and C
on the same concentrator switch. We now have a generalised circuit-switching unit,
comprising several concentrator switch units and a route switch. Thus, the common
but somewhat trivial labelling of a telephone exchange as a ‘switch’ hides the fact that
in practice it will contain several switches. Indeed, as we shall see in Section 6.2.3,
the concentrator and route switches themselves contain several stages of switches.
Confusion can be avoided by using the term ‘switch-blocks’ to denote an assembly
of switching stages forming the role of concentration or route switching.
It is a simple matter now to consider the generalised block schematic diagram
for a trunk, JT or international exchange, since these comprise a single large route
switch-block only – since no subscriber lines terminate on these exchanges. We will
look at these in a little more detail in the description of digital switching later in this
Signalling and control
Local exchange switching unit
= Subscriber’s Line card
Figure 6.2
Generalised Local Exchange (Folded)
Circuit-switching systems and networks 133
6.2.3 Digital telephone switching systems
Chapter 3 introduced the concept of converting analogue speech into a digital format
for the purposes of switching and transmission across the PSTN. This conversion
gives capital and current account (operational) cost savings for the PSTN, as well
as giving the telephone users a significantly improved quality of the sound. In fact,
the vast majority of the telephone exchanges in the World, both fixed and mobile,
are digital. They are also computer controlled – using a technique known as ‘storedprogram control’ (SPC) [1]. In this section we will look at how a call connection
is established across a digital switch-block within a fixed network exchange; we
consider the differences involved in a mobile-switching system in Chapter 9, after we
examine the data-network-based alternatives to switching voice calls in Chapter 8.
The signalling and control functions for all types of switching systems are considered
in Chapter 7.
However, before we start to examine the process of digital switching it may be
helpful to review the two standard forms of digital multiplexed systems. Chapter 3
introduced the concept of slicing a single channel of analogue speech into a series of
samples taken 8,000 times a second, each of which is converted to an 8-bit binary
number. This stream of 64,000 bits/s (64 kbit/s) is then multiplexed with the 64 kbit/s
streams of other speech channels. The process is formally known as pulse-code modulation (PCM). There are two standard formats for PCM systems: the 32-channel
international standard used throughout Europe and many parts of the World as well as
on all international links, and the 24-channel North American standard, as shown in
detail in Figs 4.8 and 4.9 of Chapter 4, respectively, and summarised for this chapter
in Fig. 6.3. Both systems contain their channels within a frame size of 125 μs (i.e.
Speech channels
Time slots:
1 2
Speech channels
26 27 28 29 30
3 4
15 16 17
27 28 29 30 31
Frame: 125 μs
(a) 30-Channel International Standard
Speech channels
Frame alignment
Time slots:
20 21 22 23 24
3 4 5
20 21 22 23 24
2 3 4
1 2
Frame: 125 μs
(b) 24-Channel North American Standard
Figure 6.3
Summary of PCM Frame Formats
134 Understanding telecommunications networks
125 millionths of a second), being the reciprocal of the universally agreed sampling
rate of 8,000 times a second. However, there are differences in the algorithm used by
the systems to convert the speech samples to digital format (the encoding laws). In
addition, there are the following differences in frame format which are important to
consider in our treatment of digital switching.
• The 30-channel PCM system actually has its frame divided equally into 32 time
slots of 3.9 μs duration, each containing 8 bits, making a total frame size of
256 bits. Time slot 0, written ‘TS0’ is used only for PCM system management – primarily to carry the frame-alignment signal, which is used by the
receiving terminal to determine the start of the frame. Time slots 1–15 carry
speech channels 1–15, respectively, and time slots 17–31 carry speech channels
16–30, respectively. Time slot 16 is normally not used for speech, instead it
is used as a bearer for signalling between exchange-control systems – mainly,
the international standard signalling system 7 (SS7), as described in Chapter 7.
The total rate of the 30-channel PCM system is 2,048 kbit/s, usually written as
‘2 Mbit/s’.
• The 24-channel PCM system has its frame divided into 24 time slots of 8-bits
each, and a single bit at the front used to carry the FAS, making a total frame size
of 193 bits. If SS7 signalling needs to be carried this is placed in one of the time
slots, typically TS24, which means that such PCM systems can support only 23
speech channels. The total rate of the 24-channel PCM system is 1,544 kbit/s,
usually written as ‘1.5 Mbit/s’.
For simplicity, the rest of this chapter will consider digital switching using just the
example of the 30-channel PCM format.
Fig. 6.4 illustrates a simple block schematic diagram of the main functional entities in a DLE; this can be seen as a development of the generalised local exchange
of Fig. 6.2. Each subscriber copper line is terminated on the MDF (see also Fig. 5.1
Subscriber concentrator unit
line card
MDF Subscriber’s
Route switch unit
Exchange-control system
Figure 6.4
SPC Digital Exchange: Functional Overview
Circuit-switching systems and networks 135
in Chapter 5) from where an internal copper pair cable is run within the exchange
building to the appropriate subscriber line card on the concentrator switch-block.
As described later, the line card performs all the necessary subscriber-line termination functions as well as converting the speech from the analogue signal on the
subscriber’s line to a 64 kbit/s digital stream. The outputs from 30 of these line cards
are multiplexed together to form a standard 30-channel PCM frame, with a line rate
of 2 Mbit/s. Each multiplexor is connected at the 2 Mbit/s level to the input of the
concentrator switch-block. From the other side of this switch-block, 2 Mbit/s links
are taken to the co-located route switch-block for connection to the 2 Mbit/s digital
trunk links to other exchanges. As described in Chapter 5, the 2 Mbit/s output from
the switching units are multiplexed together with other links onto either PDH or SDH
higher-speed transmission links at the serving core transmission station (CTS), which
is usually in the same exchange building as the switch-blocks. The route switch-block
is also connected with 2 Mbit/s links to the exchange’s SS7 signalling equipment, for
signalling to the control systems of distant exchanges.
As also shown in Fig. 6.4, each subscriber’s line card is monitored by a controller
which is used by the exchange-control system to manage the setting up and clearing
of calls, while the signalling from subscribers’ telephones is received by the MF
equipment; both systems are described in Chapter 7. The tone equipment supplies
supervisory tones, e.g. ‘network busy’, as well as recorded announcements, such as
‘Sorry we are unable to connect your call’.
Before we start to look inside the subscriber concentrator switch-block and route
switch-block, some attention should be paid to the importance of the subscriber
line card. This piece of equipment, physically about the size of a man’s wallet, has
to perform all the functions of terminating a copper line on the switching system.
First and foremost of these functions is the provision of an interface between the
high voltages and currents present on the external copper pair and the relatively
low voltages of the semiconductor electronic equipment of the digital switch-blocks.
Thus, the line card includes equipment that cannot be constructed using semiconductor
technology – this makes it relatively expensive. The key point is that a line card must
be provided for each subscriber’s line, irrespective of the number of calls made. In
fact, the capital cost of all the line cards on a typical local exchange is about 70% of
the total exchange cost!
It is important to note that all costs of a PSTN are allocated on a per-line basis
for the network between the local-loop termination at the subscriber’s premises and
the line card, while all network costs from the output of the line card and through all
the switching and core transmission networks are allocated on a per-call basis. The
UK regulator identifies the former as the ‘Line Business’ and the latter as the ‘Calls
Business’. The way that these different sets of costs are recovered by the telephone
price structure is briefly addressed in Chapter 10.
In fact, the issues associated with the subscriber’s line card are really fundamental to telecommunications networks; all forms of user terminations on any type of
network (data, voice, fixed, mobile) incur the cost and complexity which must be
borne on a per-user access basis. This principle will also be addressed in Chapters 8
and 9, looking at data and mobile networks, respectively.
136 Understanding telecommunications networks
The set of functions required by a copper-line termination and provided by a
subscriber’s line card are best remembered by applying the acronym ‘BORSCHT’,
as follows:
Battery: application of the 50 V DC power supply to the subscriber’s line (as
described in Chapter 1).
Overload: protection of the delicate semiconductor equipment of the digital
switch-blocks from any induced voltages on the external copper line – it is
particularly important to provide protection against lightning.
Ringing current: provision of an interrupted electrical signal, with an appropriate
cadence, of about 75 V DC and 200 mA is required to ring a subscribers’ set of
telephone instruments.
Supervision: that is, the detection of the subscriber going off-hook to signify call
initiation, and detection of off-hook conditions in the presence of ringing current
to signify call answer.
Codec: the A/D conversion in the Go direction and the digital-to-analogue conversion in the Return direction – A/D coding and decoding (hence the term
Hybrid: 2-to-4 wire conversion (as described in Chapter 1) from the local loop to
the Go and Return format of the switches and core transmission network.
Test: the application of electrical continuity testing at the end point of the local
loop. This is usually applied automatically under remote control from an exchange
maintenance console.
Fig. 6.5 illustrates the items on a typical subscriber’s line card and shows how these
correspond to the BORSCHT functions listed above [2].
64 kbit/s
Figure 6.5
Line unit
The Components of a Subscriber Line Card
and trip
Line card
Circuit-switching systems and networks 137
TS3 TS20
Figure 6.6
Digital Switch-Blocks
Now to consider how the digital switch-blocks work. Fig. 6.6 presents a simple
example of a digital switch-block (either concentrator or route) with five 2 Mbit/s
PCM highways at the inlets and five 2 Mbit/s PCM highways at the outlets. It is
necessary for all the PCM systems to be synchronised and aligned with each other so
that the timing of all their time slots coincide. Thus, for example, all the PCM systems
pass the 8 bits relating to their time slot 1’s simultaneously. Two call connections are
due to be set up. First a call arrives on the channel in time slot 3 (TS3) of highway
C destined for highway Y, and a connection is set up across the switch-block to exit
onto TS3 of highway Y. The stream of 8 bits now flows every time TS3 occurs, i.e.
every 125 μs, from inlet C to outlet Y across the switch-block. However, a problem
occurs if another call were to arrive in the channel occupying TS3 of highway A also
destined for highway Y, since the TS3 on the outlet Y is already occupied with the call
from highway C. The only solution is to transfer the call carried on TS3 of highway
A to a vacant time slot channel on the required outlet highway Y, e.g. TS20, as shown
in Fig. 6.6.
The conclusion to the above reasoning is that two forms of switching are required
in a practical digital switch-block. The first is the switching of a time slot’s contents
from an input 2 Mbit/s highway to the same number (i.e. coincident) time slot on an
outgoing highway, a function known as ‘space switching’. The second is the switching
of the contents of a time slot of an inlet highway to a different value (i.e. noncoincident) time slot on the outlet highway, a function known as ‘time switching’. All
digital concentrator and route switch-blocks comprise both space (written as ‘S’) and
time (written as ‘T’) switches. Typical configurations for a concentrator switch-block
are T-S and T-S-T and for a route switch-block are T-S-T and T-S-S-T.
Digital time switching
The process of time switching is achieved by the use of computer storage to delay
the contents of an incoming time slot until the moment when the required output time
slot arrives. A simple example is given in Fig. 6.7, in which we assume the input
138 Understanding telecommunications networks
TS: 1 2 3 4
TS: 1 2 3 4
Read address
for CM
Read address for SM
Figure 6.7
TS1 – TS4
TS2 – TS3
TS3 – TS1
TS4 – TS2
The Principle of Time Switching Using Storage
highway has a digital stream with just four time slots (as opposed to some 512 time
slots in practical switches). The required connection pattern, as determined by the
exchange-control system is as shown, i.e. input TS1 to output TS4, and similarly TS2
to TS3, TS3 to TS1, TS4 to TS2. This connectivity is achieved by using a simple
electronic counter to set the write address for the contents of each time slot, and
the ‘Connection Memory’ to hold the read address. Thus, the contents of incoming
channel during TS1 are ‘written’ (i.e. transferred) into location 1 of a computer store,
known as the ‘Speech Memory’; the contents of the channel during TS2 are written
into location 2 and so on. The speech memory has as many locations, or cells, as
there are time slots on the inlet highway. After the contents of the channels in the four
time slots have been passed sequentially into speech memory the counter continues
with another round, and the next set of speech samples in the time slots are written
into the respective cells in the speech memory. This continues for the lifetime of the
switch-block (assuming no system failures)!
The connection memory also has four locations, the contents of which are ‘read’
(i.e. transferred out) under the control of the counter (cycling 1–4). The contents of
each location in the connection memory indicates the read address of the cell in the
speech memory that must be read during that time slot. The exchange-control system
places the appropriate read addresses in the respective locations of the connection
memory to achieve the desired output sequence from the time switch. The shifts in
time for the four channels arriving in time slots 1–4 is illustrated in Fig. 6.7. The
connection pattern will continue cyclically 8,000/s until there is a need to cease one
Circuit-switching systems and networks 139
1 2 3
1 2 3
Row 100
Row 011
Figure 6.8
Connection pattern
A/TS1 – E/TS1
B/TS1 – F/TS1
A/TS2 – F/TS2
B/TS2 – E/TS2
B/TS3 – H/TS3
Digital Space Switching
of the calls or establish a new one, when the connection memory contents will be
altered by the exchange-control system appropriately.
Digital space switching
Digital space switching is achieved through the time-division multiplexing of a simple
matrix array, with the functional equivalent of digital logic AND gates at the crosspoints. As usual, the basic requirement is for all the highways and the matrix array to
be synchronised and aligned, so that all corresponding time slots precisely coincide.
The operation of the space switch relies on the scheduled closure of the specific crosspoints during the period of each time slot so that the contents can be shifted between
input and output highways. Fig. 6.8 illustrates a simple example, again with just four
time slots on the input and output highways. Each column of the matrix is controlled
by a connection memory (CM), the contents of which contain the address of the crosspoint that should be closed during the time that it is being read. Thus, each of the
connection memories is read simultaneously during each time slot period under the
read address control of a counter, cycling 1–4. When TS1 appears the address read
from the first location of CM-E is 100; this appears on the control bus for the column,
which is decoded by the cross-point on row 100 causing it to close; the contents of TS1
on input highway A are then transferred to TS1 on output highway E. At the same time,
the address 011 is read from the CM-F onto the control bus causing the cross-point on
row 011 to close, allowing the contents of TS1 of input highway B to be transferred
to TS1 of output highway F. During the instance of TS2 the second locations of all
the CMs are read and the addresses of the appropriate cross-points used to establish
140 Understanding telecommunications networks
the required connections across the matrix: contents of TS2 on highway A to TS2
on highway F; similarly TS2/B to TS2/E. During TS3 only one cross-point opens
giving TS3/B to TS3/H connectivity, and there are no connections required during
TS4. This connection pattern across the matrix continues 8,000 times a second until
the contents of one of the connection memories is altered by the exchange-control
system to enable another call to be established or an existing one to clear down, and
the new pattern then continues.
Digital switch-blocks
Practical digital switch-blocks use space and time switches constructed from highspeed semiconductor technology using a large degree of integrated circuitry. Maximum efficiency is gained by further time-division multiplexing the 2 Mbit/s highways
terminating on the switch-block up to typically some 512 time slots per 125 μs frame
(i.e. still with a new frame every 8,000 times per second). An example of a typical
T-S-T switch-block is shown in Fig. 6.9. Here, a call connection is to be established
between TS10 of the highway A2 entering input time switch A2 and TS45 of output
highway C1 leaving output time switch C1. The exchange-control system determines
that the next freely available pathway through the space switch is via TS124. Thus,
time switch A2 is set up to shift the contents of TS10 to TS124 on the highway b2. The
space switch provides connectivity during TS124 between highway b2 and highway
B1 and into the output time switch C1. The latter performs time switching between
TS124 and TS45 onto its outgoing highway C2. This pattern of connectivity continues
until the contents of the CMs are changed [3].
Space switch
Figure 6.9
T-S-T Switch-Block
Circuit-switching systems and networks 141
6.2.4 PBX
The concept of private automatic branch exchanges (PABX or PBX) is introduced in
Chapter 2 (Box 2.1 refers). These privately owned exchange systems are essentially
the same as the public network equivalents, using digital switch-blocks, as described
in Section 6.1.3. However, there are many different designs of digital PBX resulting
from the wide range of sizes from just a few extensions (i.e. internal telephone lines)
up to tens of thousand of extensions, depending on the extent of the customer’s
business. Generally, the PBX are connected to the serving public local exchange over
one or more 2 Mbit/s (or 1.5 Mbit/s in the American countries) digital links. Small
PBX systems are based on T-S switch-blocks and the larger systems based on T-S-T
6.2.5 Digital exchange structures
We are now in a position to consider the key structural features of digital exchanges.
Fig. 6.10(a) shows a block schematic diagram of a DLE serving both subscriber
copper lines and PBXs. The line card is provided by the concentrator switch-block
in the local exchange in the case of subscribers served by copper local lines, and by
the PBX switch-block for all its extensions. The 2 Mbit/s digital links (carried over
optical fibre or high-speed transmission over copper, e.g. HDSL) from PBXs are
terminated on a digital line termination unit (LTU) at the exchange building, close
to the MDF, from where a digital link terminates on a special PABX line card on the
concentrator switch-block. This special line card simply needs to provide a timing
interface between the digital line and the switch-block; it does not need to perform all
the BORSCHT functions, since these are provided by the line card in the concentrator
switch-block of the PBX. A special signalling link between the control systems of
the PBX and the local exchange is carried over the 2 Mbit/s digital link (in TS16),
enabling the PBX to establish and receive calls to and from the PSTN, as described
in Chapter 7.
Digital public exchanges are designed to meet a high level of system reliability
to ensure continuity of telephone service to the users – typical objectives being no
breaks in service more often than once every 20 years! This means that all of the key
components of the exchange contain redundancy and employ automatic cut-over to
standby units in the event of failure. In particular, the concentrator and digital route
switch-blocks are fully duplicated, operating on a worker-standby arrangement [3]
(see also Chapter 5).
Another element in the reliability of the exchange system is that of synchronisation. As will be apparent from the description of digital switching, it is necessary for
all the digital line systems terminating on the exchange switch-blocks, as well as the
switch-blocks themselves to be operating at exactly the same speed, i.e. the number
of bits/s. This means that in all exchanges in the PSTN either all of the digital clocks
generating the bit streams need to be precisely accurate and stable or some mechanism is required to synchronise the speed of all the clocks. The actual arrangements
for synchronising may differ among national networks, but they usually involve the
142 Understanding telecommunications networks
Processor exchange
Exchange-control system
Subscriber’s line card
Remote concentrator unit
Processor exchange
Exchange-control system
Subscriber’s line card
Figure 6.10
(a) Local Exchange Structure-1, (b) Local Exchange Structure with
Remote Concentrator Units
dissemination of the timing from a central atomic clock through a mesh of synchronisation links to control equipment in each exchange [4,5]. More recently, network
synchronisation systems have been introduced that use reception of GPS (Global
Positioning Satellite) timing pulses at each network node. It should be noted that
network synchronisation is used by all digital components in the various networks
(not just the PSTN) of all network operators in a country. Usually, the incumbent
network operator provides the definitive timing source which is taken by all the other
network operators (fixed voice, data and mobile) as well as subscribers operating
private networks and PBXs.
Circuit-switching systems and networks 143
In practice, digital concentrator switch-blocks are designed to terminate some
2,000 subscriber lines. Therefore, in an exchange configuration as shown in
Fig. 6.10(a) there may be several concentrator switch-blocks co-located within the
exchange building, each connected to the single large route switch-block. Typical
capacities of such exchanges range from about 5,000 to 50,000 lines. When serving
a catchment area of fewer than about 4,000 lines a DLE based on the configuration of Fig. 6.10(a) becomes uneconomical, with unacceptably high per-line cost due
to the dominance of the fixed costs of the exchange-control system. There are two
main solutions to this problem: the use of special cost-reduced very small-exchange
systems or the use of remote concentrator units (RCUs), as described below.
Cost-reduced small exchange systems
These very small systems, which typically range in capacity from some 10 to 600
lines, have a cut-down exchange-control system providing a reduced set of service
features for the users and a severe restriction in their traffic carrying capacity. These
very small systems tend to be deployed in remote rural areas, e.g. BT has deployed
some 500 units (designated ‘UXD5’) in the Highlands and Islands of Scotland, parts
of Wales and Northern Ireland [6].
Remote concentrator units
An economical solution for serving some 100 to about 4,000 lines is achieved by
locating just a subscriber concentration switch-block at the focal point of the copper network within a small town, and linking this to the route switch-block and the
exchange-control system at a distant parent local exchange. The latter is often called
the ‘Processor Exchange’, because that is where the large exchange-control system,
comprising a multi-processor cluster, is located. With this configuration, the fixed cost
of the exchange-control system is spread over the subscriber lines connected to all the
co-located and remote concentrator switch-blocks served. The RCU has a small local
control system to support the concentrator switch-block and line card functions, but
this has to be supported by the main control system at the parent exchange, as shown in
Fig. 6.10(b). In fact, the RCU is dependent on the parent for its operation. All calls on
the RCU, including those to subscribers on the same concentrator switch-block, must
be switched to the route switch-block at the parent exchange, from where it routes
to the destination concentrator switch-block – this may be at the parent exchange,
another RCU off that exchange, the originating RCU, or taken via a trunk route to
another exchange to complete the call. In the case of an own RCU call, the to-and-fro
routeing via the parent exchange is known as ‘tromboning’ for obvious reasons.
If the digital link between the RCU and the parent exchange is cut for any reason
the RCU is isolated. In such conditions a restricted form of own-RCU calls can
usually be provided by the local controller, but without many of the facilities including
charging being available. Normally, several parallel PCM digital links are provided
from the RCUs to their parent exchanges, with as much spatial separation in the line
plant routeing as possible (see Section 5.3.4 on resilience in Chapter 5), to minimise
the risk of RCU isolation. In the UK, the RCU-to-parent exchange distances are
144 Understanding telecommunications networks
typically some 5–20 miles, although in some other countries far greater distances are
6.2.6 ISDN exchanges
Following the successful introduction of digital switching for voice in the 1980s, the
next logical step of evolution was that of ISDN. The concept of the ISDN is that the
network of digital telephony exchanges is enhanced to allow the switching of data
calls. This requires the following additions to the digital PSTN:
1. The introduction of a digital transmission system onto the subscriber line to
extend the digital path from the exchange to the subscriber’s data terminals at
their premises. (The ISDN digital line system is covered briefly in Chapter 4.)
2. The provision of enhanced signalling between subscriber and the exchange (see
Chapter 7).
3. The provision of enhanced signalling between exchanges (see Chapter 7).
4. The provision of extra control software to enable the more-complex form of data
calls to be handled (see Chapter 7).
Two forms of internationally standardised ISDN are available:
Basic rate: Two 64 kbit/s channels (called the ‘B’ channels), together with a
16 kbit/s signalling channel (‘D’ channel). This is designated “2B + D”. The user
rate is 144 kbit/s (i.e. 2 × 64 kbit/s + 16 kbit/s) and is normally carried over a single
copper pair.
Primary rate: 30 × 64 kbit/s (B) channels, together with a 64 kbit/s signalling
channel (D). This is designated ‘30B + D’. The user rate is 2 Mbit/s and this is carried
over an optical fibre, microwave radio link or possibly HDSL over three parallel
copper pairs.
In practice, most PSTN exchanges today are combined ISDN and telephony systems. When a subscriber wishes to have basic rate ISDN service the existing copper
line is shifted from the analogue line card to an ISDN line card on the same concentrator switch-block (requiring a re-jumpering at the MDF). A network terminating unit
(NTU) is installed at the subscriber’s premises to terminate the digital line system
and present a set of sockets for the subscriber’s data and telephony terminals. There
is usually a terminal adaptor associated with the NTU to provide a conversion of the
analogue telephone to digital so that it can use the ISDN line. In the case of primary
rate ISDN service the termination at the subscriber concentrator is at the 2 Mbit/s
level, requiring an appropriate digital line termination – as provided for digital PBX.
Fig. 6.11 shows the arrangement for ISDN, with an indication of how the BORSCHT
functions are spread between the ISDN line card and the NTU on the subscriber’s
premises, instead of being solely in the line card as for analogue subscriber lines.
Non-data calls from ISDN subscribers are routed through the switch network as for
telephony with each B channel being treated separately. Data calls are also normally
handled on a single B-channel basis, with the required extra features indicated through
the call set-up across the network through the ISDN signalling system. ISDN lines are
Circuit-switching systems and networks 145
Home or
small business
=2Mbit/s line card
Figure 6.11
MDF ISDN subscriber’s line card
Basic rate ISDN
144 kbit/s
PBX subscriber’s
line card
Exchange-control system
Local exchange
NTU = network terminating unit LTU = Line terminating unit
ISDN Lines on a Digital Local Exchange
given standard telephony numbers and are charged at the same rate as telephony. In
addition, some network operators provide the ability for basic rate subscribers to use
the two B channels as a composite 128 kbit/s channel; the ISDN exchange then has to
route two parallel calls through the network simultaneously to the same destination
ISDN line. (However, this 128 kbit/s service has not proved to be a commercial
Network dimensioning
6.3.1 The concept of switched traffic
Most people are comfortable with the concept of traffic, say, as a flow of vehicles
down a motorway or aircraft taking off and landing on an airport runway. Likewise,
telephone traffic is the set of calls being carried over the line between a user and the
serving exchange or over the link between two telephone exchanges (i.e. a ‘traffic
route’). Just as there are vehicles of various lengths on the motorway, and planes
requiring long and short take off runs, telephone calls also vary in duration. The need
is to be able to dimension the number of channels or circuits on the traffic route
between two exchanges (equivalent to the number of lanes on the motorway) so that
at the expected level of calls during the busiest time of day, there is a sufficiently
low probability that any new calls arriving will encounter a lack of free capacity on
the route, i.e. congestion, – similar to all lanes on the motorway being so tightly
packed that no more cars can join it. With the existing conventional circuit-switched
telephone exchanges, any calls meeting congestion will be ‘lost’, with the user experiencing equipment-busy tone leading to the call attempt being abandoned. However,
there are switching systems in which calls are delayed rather than lost when meeting
146 Understanding telecommunications networks
congestion – so-called delay systems. This alternative form of telephone switching
using packet technology is considered in Chapter 8.
Network operators use dimensioning techniques to determine the size of traffic
routes between exchanges – the number of circuits required to avoid congestion – as
well as the capacity of the exchange equipment that handles the calls. Both of these
dimensioning activities involves an implicit economic trade-off between the cost
penalty of over-providing exchange switching capacity or circuits on a traffic route on
the one hand and the impairment to the quality of service (QOS) offered to the user due
to under-provision on the other. This is a common dilemma for any service industry.
We have already used the analogy of the motorway to describe the dimensioning
of the traffic routes. Another analogy may help in considering the dimensioning issues
for the exchange equipment. For example, consider the staffing levels required in a
public post office in which several attendants serve simultaneously at the counter.
By arranging for a single queuing system, arriving customers can be directed to
the next available attendant. Obviously, the serving time for all the customers will
vary depending on the nature of their transactions and so there will be the chance
that all attendants are busy when the next customer arrives in the post office. This
chance can be reduced by increasing the number of attendants at the counter. The
dimensioning challenge for the post office management is to determine how many
attendants should be provided: trading their cost against the disadvantages of having
dissatisfied customers having to wait to be served. For the post office situation, a
short wait in the queue is usually deemed acceptable, whereas the loss-system nature
of a telephone exchange means that a lack of available servers will cause a call to fail
(‘be lost’), with a perceived degradation of QOS for the customer and a potential loss
of revenue if the repeat attempt is made via another network operator.
6.3.2 Call distribution
For a network operator to dimension a telephone exchange a forward view of the likely
amount of calls on that exchange is required. Whilst the instantaneous demand from
individual telephone users will vary according to circumstances, the total population
of lines on a telephone exchange will tend to follow a pattern of peaks and troughs
of numbers of calls from day to day which is fairly stable, and over a course of time
predictable growth (or decline) forecasts of traffic can be made. A typical daily pattern
of telephone demand on a local exchange is shown in Fig. 6.12, arranged in hourly
segments. This shows that during the typical weekday the demand for calls increases
from a low level at night from around 8 am to a peak around 11 am, followed by a
slight drop in early afternoon and a second lower peak in mid afternoon – driven by
activity at the work place. There is then a domestically driven evening peak around
8 pm; more recently with the advent of people using the PSTN to dial up the Internet
recreationally from home after work, this peak can in fact be greater than the midafternoon peak. A different pattern applies at weekends and public holidays with the
majority of calls being generated by residences rather than from the work place. There
may also be a distinctive seasonal pattern, for example, with much higher levels of
demand in the summer months at exchanges serving seaside resorts.
Circuit-switching systems and networks 147
‘Busy Hour’
Network capacity
and National
Figure 6.12
4 am
Time of day
Typical Telephony Traffic Profile
6.3.3 Traffic flow
As Fig. 6.12 shows, the required exchange capacity is set by the one-hour period
of highest demand, traditionally known as the ‘Busy Hour’. The network operator
needs to ensure that there is sufficient capacity in the switch-blocks to cope with the
current demand, with an appropriate allowance for exceptional increases. In addition,
the operator arranges for the exchange capacity to be augmented periodically to cope
with the predicted growth in demand over the following 12–18 months. This futureproofing also ensures that there is spare capacity in the exchange to deal with any
unusual daily fluctuations. The extent of this spare capacity is obviously an economic
trade-off, as described earlier.
So, now to look at the process of dimensioning an exchange. Fig. 6.13 illustrates
the concept of traffic occupancy, taking the example of a system of four channels –
these might be channels within a switch, as described earlier, or channels in an
outgoing route to another exchange. The incidents of calls being carried over each
channel is plotted over time, indicating the different durations or holding times of
the calls. At any one time the number of simultaneous calls on the system varies
from zero to three, as shown in the lower plot in Fig. 6.13. The summation of all
the channel occupancies over time constitutes the telephone traffic intensity, which is
measured in units called ‘Erlangs’ (named after A. K. Erlang, the Danish pioneer of
traffic theory) [4]. An Erlang is defined as the average number of simultaneous calls
on a system during a one-hour period. (An alternative unit is used in North America
which measures traffic intensity in units of ‘hundred call seconds per hour’ – CCS,
i.e. equal to 1/36th of an Erlang.) In general, the traffic intensity of flow A, measured
148 Understanding telecommunications networks
Traffic flow
Figure 6.13
Traffic and Occupancy
in Erlangs, is given by:
A = Ch
where C is the average number of calls during the hour and h is the average holding
time of those calls. By definition, a traffic flow of one Erlang is equivalent to a circuit
being occupied for a complete hour.
As described in the introduction above, there will be occasions when a system
serving telephone traffic will be fully occupied and so temporarily unable to service
any more calls – a state known as call congestion. Any calls arriving at that time will
be lost (the subscriber will have to abandon the call). The probability of this occurring
is defined as the ‘Grade of Service’ (GOS), represented by the symbol ‘B’ in traffic
dimensioning formulas. Telephone exchanges are typically dimensioned to a GOS of
one lost call in 200, i.e. during the busy hour there is a 0.5 per cent chance that a call
attempt will meet congestion (outside the busy periods the chances of congestion are
negligible). It should be emphasised that even this low probability of call failure will
only be experienced when the capacity of the exchange is just at the design level; as
explained earlier, usually spare capacity has been provided in anticipation of growth
and so the actual GOS will at all other times be even better than the design limit.
The name of Erlang is also associated with a formula for determining the number
of circuits required to carry a specified level of telephone call intensity at a specified
GOS – the famous ‘Erlang’s Lost Call’ or ‘Erlang’s B’ Formula [7,8]. Table 6.1 below
shows the variation in the traffic that can be carried on a given number of circuits
at various GOSs, calculated using Erlang’s B formula. There are several points to
note. The first is that the amount of traffic that can be carried by a set of circuits
increases as the GOS decreases, showing the trade-off between capacity carried and
the chance of congestion or loss of traffic. The second point is that the larger the
group of circuits the proportionately higher the amount of traffic carried by each
circuit, i.e. the circuit loading. For example, with a GOS of 1-in-100 a group of 20
circuits carries 12.8 Erlangs (making 0.64 E/circuit) compared to a group of a 100
circuits which carries 84.1 Erlangs (making 0.81 E/circuit), an improvement of circuit
loading from 64 to 81 per cent, respectively. It is apparent from this table that the
Circuit-switching systems and networks 149
Table 6.1
Loading of Circuits with GOS
No of Circuits
Erlang Capacity with GOS
traffic characteristics of a group of circuits or servers are non-linear, with improved
efficiency with size. In general, therefore, network economies can be achieved by
carrying and switching telephone traffic over fewer but larger exchanges with the
fewest and largest size routes between exchanges as possible. This process was already
identified in Section 6.2.2 when the role of concentrator switch-blocks was described
as the traffic concentration of the lightly loaded subscriber lines (at, say, 0.05 E/circuit)
onto the smaller set of more highly loaded outlet trunks (at 0.5 E/circuit), assuming
a 10 : 1 concentration ratio.
6.3.4 Traffic routeing
We are now in a position to consider how traffic is handled by the exchange. First,
of course, the subscribers on the exchange will ‘originate’ traffic, i.e. initiate calls,
and they will ‘terminate’ traffic, i.e. receive calls. This is shown in Fig. 6.14 with
the direction of the arrows indicating the traffic flow (the conveyance of speech is
of course in both directions). Thus, the subscribers originate a total of 1,500 Erlangs
(1.5 kE), of which 1.2 kE is destined for subscribers on the same exchange (known
as ‘own exchange’ traffic) and 0.3 kE is outgoing trunk traffic; the subscribers also
terminate some 1.4 kE of traffic of which 0.2 kE is from other exchanges. Although
in practice the traffic distribution from the various exchanges differs, typically some
Figure 6.14
Exchange Dimensioning
1.5 kE
0.3 kE
1.2 kE
1.4 kE
0.2 kE
150 Understanding telecommunications networks
80 per cent of originated traffic is own exchange, and as this example shows the
originating and terminating traffic is not usually symmetrical. In this simple example
the total switched traffic is 1.2 + 0.3 + 0.2 = 1.7 kE. The 40,000 subscribers originate
1.5 kE and terminate 1.4 kE in the busy hour, giving an average both-way calling rate
of 0.0725 E per line. Typical both-way calling rates are around 0.06 E per line for
residential subscribers and 0.18 E per line for business subscribers [9].
Fig. 6.14 shows the full exchange configuration for the example explained earlier.
The route switch-block needs to be dimensioned to carry the 1.7 kE of total switched
traffic as well as provide terminations for all the trunk routes (total traffic of 0.3 kE
outgoing and 0.2 kE incoming). Since a subscriber concentrator unit can typically
terminate a maximum of 2,000 lines, some 23 or 24 units (depending on the fill
factor) would be required to support the 40,000 subscribers. These concentrators will
be either co-located with the route switch-block in the exchange or located at several
remote small exchange buildings, depending on the disposition of the subscriber lines
in the exchange catchment area.
Chapter 1 described a typical PSTN and introduced Fig. 1.9 as a simplified diagram
of BT’s telephone network, which is reproduced as Fig. 6.15 for convenience. A PSTN
is arranged as a hierarchy of exchanges, with local exchanges connected to a parent
trunk exchange, and in the case of the UK each trunk exchange is fully interconnected.
This enables calls from subscribers on any exchange to be routed to subscribers on
any other exchange via a maximum of four inter-exchange links. Calls between two
local exchanges on the same parent trunk unit would normally be tandem switched at
that trunk unit. However, where the level of traffic between two exchanges becomes
BT’s International
BT’s specialised
Other network
Optional route
Central LE
Very small
Figure 6.15
BT’s Public Switched Telephone Network
Circuit-switching systems and networks 151
sufficiently high it becomes economical to establish a direct route. Such routes are
referred to as ‘optional’ or ‘auxiliary’, whereas the hierarchical routes (e.g. local
exchange to parent trunk exchange) are mandatory.
An example of an optional route is given in Fig. 6.16, in which there is 25 E
of traffic between Exchange A and the parent trunk exchange B, and 30 E between
Exchanges B and C (assume all traffic quantities are total bothway), and the planning
question is whether the 5 E of direct traffic between A and C should go via the transit
at B or directly. This is known as the direct versus tandem (d/t) routeing decision. The
cost comparison comprises: the cost of the transmission of 5 E of traffic between A
and B (in practice this may actually be carried over line plant A to B and B to C rather
than over a new A to C link) versus the cost of 5 E extra on to the existing 25 E route
A–B, plus the marginal cost of switching 5 E extra at B, plus the cost of 5 E extra
on the existing 30 E route B–C. However, because of the non-linear characteristic
of traffic loading with route size, the additional 5E on link A–B incurs just six extra
circuits, and on link B–C just five extra circuits compared to the need for 12 circuits to
carry the new 5 E route. The cost of a direct route would not be warranted in this case.
Thus, if the d/t > 1 it is cheaper to route via the tandem; with the d/t < 1 the direct
route is justified. The d/t ratio does not normally approach 1 until the traffic between
two exchanges is around 10 E to 12 E, depending on the actual cost of switching and
transmission equipment deployed in the network.
Direct Cost: Trans cost (5E){ [A–B] + [B–C]}
= Trans cost (12 circuits)[A–B] + Trans cost (12 circuits)[B–C]
Transit cost: Trans cost (30E–25E)[A–B] + Switching cost (5E) + Trans cost (35E–30E)[B–C]
= Trans cost (6 circuits)[A–B] + Switching cost (5E) + Trans cost (5 circuits)[B–C]
Figure 6.16
Example of Direct Versus Tandem Routeing Comparison
152 Understanding telecommunications networks
Figure 6.17
A to
Automatic Alternate Routeing (AAR)
In the case of the BT PSTN shown in Fig. 6.15, the potential for optional routes is
shown with a dotted line. JTs are used to carry traffic between the local exchanges in
large metropolitan areas such as London and Birmingham; direct optional routes are
only used where the level of co-terminal traffic is above the d/t threshold. Actually,
the addition of these optional routes improves the resilience of the switched traffic
network by imposing a mesh structure on top of the basic hierarchical star structure.
Furthermore, the appropriate use of AAR enhances both the resilience and the
overall cost efficiency of the switched traffic network. This facility uses the ability of
the processor control system of a digital exchange to select a number of alternative
routeings to a destination exchange on a call-by-call basis. Fig. 6.17 shows a part of a
network of trunk exchanges A, B, C, D and E, in which the first choice and alternative
routeings from A are shown in the table. For example, the first choice routeing of A to
B is over A–B, but in the event of this route being full (congested) the second choice
is to route A–C–B (i.e. tandem switching at C). Similarly, the choices of alternative
routeing are shown for the other routes from A, with the case of A–E having a third
choice routeing via A–C–D–E, i.e. incurring two sets of tandem switching. Clearly,
this process introduces increased resilience by virtue of the quick fall back from
first choice to second or third choice routeings, without the users being aware of the
changes. However, it is also used to decrease overall costs by under-dimensioning
all but the final choice of route. Thus, in the example of Fig. 6.17 only route A–C is
fully provided (according to Erlang’s B formula), all other routes are dimensioned to
a lower level, known as ‘high usage’ – the occasional peaks of traffic experiencing
congestion on these routes being switched to the alternative fully-provided route.
Typically, optional routes are dimensioned on a high-usage basis, while mandatory
routes are fully provided.
6.3.5 Exchange capacity planning
A network operator’s planners have the task of ensuring that there is sufficient equipment in each of the exchanges to meet the traffic demand on a daily basis. To this end
measurements of the traffic demand are made periodically at each exchange. Forecasts of future demand are made, based usually on historical data, and the planners are
then able to determine when each exchange will ‘exhaust’ – that is when the forecast
Circuit-switching systems and networks 153
Capacity increments
Average spare capacity
and take up
Average capacity
Figure 6.18
Exchange Capacity Design Periods
future demand equals the installed capacity. The aim is for new capacity to be added
to an exchange sufficiently in advance of exhaustion to avoid any congestion on the
one hand and not to have unnecessary amounts of spare equipment sitting idly in
the exchange on the other. Fig. 6.18 illustrates this exchange capacity management
process. The demand curve is shown, being the actual measured demand up to current date and forecast forward in time. Exchange equipment is supplied in appropriate
capacity increments, in terms of extra switch-block (concentrator and route) capacity,
trunk LTUs, signalling system units and exchange-control system units. Each increment will give capacity sufficient to meet growth in demand for a period of several
months – this is known as the ‘design period’. There will also be constraints on the
capacity increments due to module sizes of the equipment.
Design periods are an economic trade-off between the cost of planning and installation versus the burden of undue spare capacity. Typical design periods for exchanges
are around 18 months to 2 years. Fig. 6.18 also shows the plot of average spare capacity; again, this needs to be kept at a level which meets the economic trade-offs, as well
as giving a useful buffer for unforeseen sudden increases in demand. It is interesting to note that the alternative approach of ‘just-in-time’ provision would require the
demand curve to be tracked closely by the installed capacity resulting in only minimal
amounts of spare capacity, but this requires a rapid delivery and installation process,
which is generally impractical for the telecommunication network environment.
In Section 6.2 we considered how an exchange (or more precisely the switching
equipment in the exchange building) switches telephone calls. First the generic local
154 Understanding telecommunications networks
exchange structure is described, comprising switch-blocks to concentrate the traffic
from subscriber lines (subscriber concentrator units) and a remote or co-located route
switch-block. The mechanism of digital time and space switching – the heart of
modern digital exchanges – was explained. We introduced the acronym BORSCHT to
help identify all the functions necessary to support a subscriber’s line at an exchange.
Section 6.3 then looked at the network aspects of switching: the concept of traffic
flow, how it should be routed and the economic and resilience aspects of switched
network structures. Finally, the planning of switch capacity was briefly examined.
A simple example of the stages of network planning of a PSTN, bringing together
all the aspects covered in this chapter and the core transmission network planning
covered in the previous chapter was given in Box 6.2.
Box 6.2 A Simple Example of the Stages of Planning a PSTN
As a simple example of the stages of planning a PSTN we will consider the
small hypothetical island with population distribution as shown in Fig. 6.19.
Stage 1 is the determination of the number of exchanges to serve the island
and their catchment areas, also shown in Fig. 6.19.
Stage 2 (Fig. 6.20) is the identification of the exchange locations, which
need to be as close as possible to the centre of gravity of these catchment areas
so as to serve all potential customers with minimum total subscriber-line costs.
Stage 3 is the estimation of the traffic flows between each of the five
exchanges, in both directions, as set out in the matrix. In practice these flows
would be determined from measurement of existing calling distributions, or in
the absence of data use of the gravitational formula (traffic between A and B
= [population of A× population of B]/the square of the distance apart) may be
appropriate (Fig. 6.21 refers).
Stage 4 is the application of the direct versus tandem calculation to all the
traffic flows, resulting in the ‘d’ or ‘t’ decision, as shown in the matrix of
Fig. 6.22. The final pattern of traffic routes is then a star structure centred on
exchange C, with direct traffic routes (in either direction) between exchanges A
and B.
Stage 5 is the planning of the transmission network, which in this simple
example is a cable route following the single main road on the island. Transmission cable networks usually follow the road infrastructure for practical reasons
(Fig. 6.23).
Stage 6 is then the important process of mapping the traffic network to the
physical cable infrastructure – the latter are known as the ‘engineering routes’.
The resulting mapping is shown in Fig. 6.24.
Finally in stage 7 the dimensions of the cable, engineering routes, can be
determined by summing all the traffic routes over each transmission link, using
the matrix of stage 6 (Fig. 6.25).
Circuit-switching systems and networks 155
Figure 6.19
Planning Example; Stage 1: Customer Distribution [Ward]
Figure 6.20
Planning Example; Stage 2: Exchange Location [Ward]
156 Understanding telecommunications networks
Traffic streams
n(n−1) = 20 traffic streams
Traffic Matrix
Figure 6.21
Planning Example; Stage 3: Traffic Distribution [Ward]
Traffic routeing matrix
Traffic routes
Figure 6.22
Planning Example; Stage 4: Traffic Routeing [Ward]
Circuit-switching systems and networks 157
Figure 6.23
Planning Example; Stage 5: Physical Cable Routes [Ward]
Enginering routeing
of traffic routes
Traffic routes
Cable routes
Figure 6.24
Planning Example; Stage 6: Traffic Route To Engineering Route Marix
158 Understanding telecommunications networks
Required transmission link capacities
Link A–B = Traffic Ccts. {A–B/B–A) + (A–C/C–A)}
Link B–C = Traffic Ccts. {(A–C/C–A) + (B–C/C–B)}
Traffic routes
Link C–D = Traffic Ccts. {(C–D/D–C) + (C–E/E–C)}
Link D–E = Traffic Ccts. (C–E/E–C)
Ccts. = circuits
Cable routes
Figure 6.25
Planning Example; Stage 7: Transmission (Engineering Route) Dimensioning [Ward]
REDMILL, F. J. and VALDAR, A. R.: ‘SPC Digital Telephone Exchanges’, IET
Telecommunications Series No. 21, Stevenage, 1995, Chapter 11.
Ibid., Chapter 7.
Ibid., Chapter 6.
Ibid., Chapter 10.
BREGNI, S.: ‘Synchronization of Digital Telecommunications Networks’, John
Wiley & Sons Ltd., Chichester, 2002, Chapter 4.
FITTER, M. A.: ‘The Introduction of UXD5 Small Digital Local Exchanges’,
British Telecommunications Engineering, Vol. 4, Part 1, April 1985, pp. 27–29.
FLOOD, J. E.: ‘Telecommunications Switching, Traffic and Networks’, Pearson
Education, Prentice Hall, Harlow, 1999, Chapter 4.
BEAR, D.: ‘Principles of Telecommunications Traffic Engineering’, IET
Telecommunications Series No. 2, John Wiley & Sons Ltd., Chichester, 1980,
Chapter 7.
SMITH, S. F.: ‘Telephony and Telegraphy A – An Introduction To Telephone
Exchange and Telegraph Instruments and Exchanges’, Oxford University Press,
London, 1969, Chapter 7.
Chapter 7
Signalling and control
The concept of a telephone call being routed through a succession of exchanges was
first introduced in Chapter 1, extended to the interconnection of several different
operators’ networks in Chapter 2 and considered from the switching systems’ viewpoint in Chapter 6. We are now in a position to examine how the control of such call
routeings across the networks (PSTN and others) is achieved. To this end, this chapter
first considers the mechanism of signalling between nodes across the networks, and
then the principal features of call switching control are introduced.
7.2.1 An overview of signalling
Signalling may be defined generically as the extension of control information from
one network node or user to another over one or more links. It is employed in two
distinct domains: between the user and the serving node (subscriber concentrator
switch) – known as ‘user’ or ‘subscriber’ signalling, and between network nodes
(e.g. exchanges) – known as ‘inter-exchange’ or ‘inter-nodal’ signalling. The latter
may extend across network boundaries, perhaps concatenating several operators’
networks in the case of an international call. Fig. 7.1 shows the domain of subscriber
and inter-exchange signalling.
Network operators create a clear demarcation within the exchanges between subscriber and inter-nodal signalling in order to ensure network integrity. It is important
that the control that an individual user can have on the network – seizing of links,
routeing of calls, etc. – is carefully constrained to avoid a rogue subscriber deliberately or inadvertently monopolising key parts of the network. (This is a feature that
engineers are now trying to introduce into the Internet!)
160 Understanding telecommunications networks
Figure 7.1
The Domains of Signalling
Signalling in either domain has two purposes. The first is that of ‘supervisory’
or ‘line’ information, necessary to manage the link resources for the call. A simple
example is the ‘off-hook’ signal (a loop established in the absence of ringing current)
which signifies that the subscriber wishes to initiate a call. A corresponding example
for inter-exchange signalling is the indication to the distant exchange that a circuit
has been seized by the calling-end exchange as part of the call set-up sequence, as
described later. The second form of information sent over signalling links is that of
‘address’, e.g. the number of the called subscriber.
Over the local copper line, subscriber supervisory signalling relies on the presence
or absence of loop conditions. Since the earliest days of automatic exchanges, rotating
dials on the telephone injected a series of breaks in the loop at the rate of 10 per second
to convey the address information (e.g. three breaks indicated digit 3) – a system
known as ‘loop disconnect’ (LD) signalling. The resulting electrical waveform on the
line during signalling is a train of pulses indicated by each turn of the dial (referred to
as ‘10 PPS (pulses per second) signalling’. Most dial telephones have been replaced by
the tone signalling system, as described below, but the subscriber even today is said to
‘dial’ a number (even if most users have never experienced a dial!). The modern tone
telephone has a set of 12 push buttons used to indicate the address signalling, each of
which selects the appropriate pair of tones from six frequencies – ‘Multi-frequency
(MF) signalling’ – giving rise to the characteristic musical feedback to the user. The
signals are detected at the far end by a set of six filters, each tuned to one of the
frequencies. In addition to the digits 0–9 the MF phones provide the ‘star’ and ‘hash’
(termed ‘pound’ in the United States) keys for extra address information. The latter
facility is useful in allowing users to send additional information after the call has
Signalling and control 161
(a) Channel-associated signalling
(b) Common-channel signalling
Speech circuit
Signalling circuit
Figure 7.2
Signalling Relationship to Voice Channel
been established, e.g. the sending of PIN codes to authorise entrance to conference
A further important distinction with signalling is the relationship between the
signalling channels and the corresponding speech channels. The simplest arrangement
is ‘channel associated’, in which the signalling and speech share the same transmission
path on a permanent and exclusive basis, as shown in Fig. 7.2(a). The standard
telephone subscriber local loop uses channel-associated signalling (CAS) with both
the line and address signalling carried over the subscriber’s pair. Whilst channelassociated systems benefit from having a simple relationship between signalling and
the speech channels, they do suffer from the restriction of sharing the transmission
path, e.g. being unable to signal during the conversation phase. Also, because of the
one-to-one relationship with the number of speech channels, the cost of the signalling
channels needs to be kept low.
The alternative arrangement is to have a single common separate channel which
carries the signalling for a set of speech circuits, as shown in Fig. 7.2(b). This single channel can carry the signalling at any time, including during the conversation
phase, and it can be feature rich since its costs are being shared across several speech
channels. Typically, channel-associated working is applied within the Access Network for customers served by copper cable; whereas, CCS systems are used for the
more-complex needs of high speed and data services, such as ISDN and PBXs [1].
Historically, a wide range of channel-associated systems were developed for internodal signalling, each appropriate to the transmission media used at the time, e.g.
‘DC’, ‘AC’ and ‘MF’ signalling systems [2]. However, today, nearly all inter-nodal
162 Understanding telecommunications networks
signalling is common-channel, using the ITU-T (previously known as CCITT) Common Channel Signalling System Number 7 – chosen because of the wide range of
messages possible and the high speed of working, as described in the following
section. (See Appendix 1 for information on ITU-T and CCITT.)
7.2.2 Applications of common-channel signalling systems
Common-channel signalling systems provide data messaging between the computers
of distant exchange-control systems. For this reason these systems are also called
‘inter-processor signalling systems’ and, as we shall see in Section 7.2.3, the set of
data messages carried is intimately linked to the call set-up process. CCS is also used
between PBXs and their serving public local exchange, as well as on ISDN local
links (e.g. ITU system Q931). A special form of CCS is used between PBXs that
form a private or corporate network, e.g. DPNSS (digital private network signalling
system), which enables calls to be established using private numbering schemes
and special charging arrangements. However, in addition to these various call set-up
signalling applications the data message characteristics of CCS also enable exchangecontrol systems to make enquiry transactions with remotely located data bases to
help handle complex calls. This extra application of CCS is referred to as ‘non-circuit
related’signalling because, unlike the application described above, there are no speech
circuits between the exchange and the control data base. Fig. 7.3 illustrates all these
different applications of CCS. We will now concentrate on the widely deployed CCS
SS7 Circuit related
SS7 Non-circuit related
Q931 (DS1)
Figure 7.3
Applications of CCS
Signalling and control 163
standard ITU SS7 system, which is used for both circuit- and non-circuit-related
7.2.3 ITU common-channel signalling system no. 7 (CCSS7, SS7 or C7)
This universally applied international standard CCS system is called variously:
common-channel signalling system No. 7 (CCSS7), signalling system 7 (SS7),
ITU-T 7, CCITT No. 7 or C7. For convenience, in this book we use the term ‘SS7’.
It is important to appreciate that the SS7 signalling links between the various
nodes (as shown in Fig. 7.3) really constitute a separate telecommunications network
overlaying the PSTN, and the other specialised networks that it serves. Whilst the
signalling nodes are physically part of the switching systems, as described in Chapter 6, and the signalling links are carried over the Core Transmission Network, as
described in Chapter 5, they are functionally part of a stand-alone network, which is
managed and dimensioned separately. Each node is identified by an address, known
as the signalling point code. We may therefore view the basic function of SS7 as
the reliable delivery of signalling messages between the various SS7 nodes. These
messages can be routed directly between nodes or sent via one or more intermediate
nodes, similar to the direct or tandem routeing of telephone calls. Indeed, there are
special SS7 nodes which act as tandem points for SS7 messages, known as signal
transfer points (STPs), which will be considered later.
Before we look at the system itself, it is instructive to consider a simple example
of the flow of SS7 messages involved in setting up a telephone call. Fig. 7.4 sets out
a simple scenario with a call from subscriber A on local exchange A, which is set
Subscriber A
Subscriber B
Initial address message
Address complete
Initial address message
Address complete
Clear forward
Release guard
Figure 7.4
Clear forward
Release guard
A Typical Telephone Call Sequence
164 Understanding telecommunications networks
up via an intermediate trunk exchange to the terminating local exchange B serving
called subscriber B; time progresses vertically down the figure.
(i) On receipt of the dialled digits from subscriber A the control system of
exchange A determines that the call needs to be sent to the trunk exchange for
completion and so selects a free channel (time slot), say channel 17, on the
traffic route to the trunk exchange.
(ii) Exchange A then sends the first of the SS7 messages to the trunk exchange.
This message indicates that a call is to be set-up for channel 17 and gives the
phone number or ‘address’ of the called subscriber B. For obvious reasons,
this is known as the ‘initial address message’ (IAM). The message is sent via
the SS7 system using the relevant source and destination point codes of the
nodes at exchange A and the trunk exchange, respectively.
(iii) The control system of the trunk exchange then determines that the call is
to one of its local exchanges and seeks a free channel on the traffic route,
say channel 25, to exchange B. The control system then requests an IAM
to be sent to exchange B, which contains the same destination address as
the message from exchange A, but with the related channel number of 25
(rather than 17). The message is sent via the SS7 system using the source and
destination addresses for the nodes at the trunk exchange and exchange B,
(iv) On receiving this message the control system at exchange B then checks
whether the line of subscriber B is free. If it is free, ringing current is sent to
the line. A reply SS7 message is then initiated by exchange B, which indicates
that a correct telephone number has been received and that ringing current is
being applied to subscriber B’s line. This ‘address complete’ message is sent
from exchange B to the trunk exchange, referring to channel 25 and using the
appropriate point codes.
(v) The control system at the trunk exchange recognises that the message from
exchange B relates to a particular call in progress, and sends an ‘address
complete’ message to exchange A.
(vi) On receipt of this response, the control system of exchange A sets up the
‘switch through’, i.e. a continuous speech path is now established through the
switch-blocks from the line of subscriber A to channel 17 on the outgoing traffic route to the trunk exchange. A continuous path then exists from subscriber
A’s telephone all the way through the network to exchange B. The ringing tone
generated by exchange B can now be heard by the calling subscriber A. (In
general, the ringing tone heard by the caller originates from the destination
local exchange, which is why the style of ringing tones may vary when calls
are made to different networks.)
(vii) As soon as the control system at exchange B detects that subscriber B has
answered the phone by going ‘off-hook’, ringing current is ceased and the
switch-through exchange B’s switch-blocks is established between channel
25 from the trunk exchange and subscriber B’s line. The speech path has thus
been extended to subscriber B.
Signalling and control 165
(viii) A return ‘answer’ message initiated by exchange B, indicating successful call
set-up, is sent to the trunk exchange, which in turn sends a similar message
to exchange A. The call now moves to the conversation phase and the control
system at exchange A starts to log the call details in a ‘call record’ – part of
local computer storage within the exchange-control system. (Call records are
periodically dumped off the storage at each local exchange and transferred to
off-line billing centres for processing.)
(ix) The call–clear-down sequence begins as soon as the caller replaces their handset. (If the called subscriber clears first the call connection clear-down will start
once the caller clears or after a few minutes time out, which ever is the sooner.)
A ‘clear-forward’ message is sent from exchange A to the trunk exchange,
which in turn sends a clear-forward message to exchange B. The speech path
is then cleared down across the switch-blocks of all three exchanges. The
charging for the call ceases and the call record is completed by exchange A’s
control system.
(x) Finally, return ‘release guard’ messages are sent from exchange B and the
trunk exchange to free up the speech channels 25 and 17 respectively on the
traffic routes.
In this typical call scenario of Fig. 7.4, ten SS7 messages are sent between the three
exchanges. Obviously, if there are more exchanges involved in a call the number
of SS7 messages required increases. If any of the messages are corrupted during
transmission or lost, a re-transmission of the message is requested and made. Morecomplex calls, e.g. from mobile exchanges (ME) , also incur more messages and from
a wider repertoire than for a simple telephony call. We can now appreciate that the
SS7 system comprises a mechanism ensuring the safe delivery of the messages, and
several portfolios of messages, each relating to different types of calls.
Fig 7.5 gives a simplified representation of the delivery mechanism and the messages. First, we will consider the delivery mechanism. The confusingly termed ‘signal
unit’ (SU) is a structured data packet, i.e. an assembly of bits (1’s and 0’s) which acts
as a conveyor of the signalling messages. A so-called flag, which is a special binary
pattern (i.e. 01111110), indicates the front end of the SU and there is another flag at
the end. Between the flags within the SU there are specified areas, or fields, allocated
to specific functions, although for simplicity only the main fields are shown. The first
is the check sum, a 16-bit binary number which results from the multiplication of the
contents of the SU by a special error-correcting polynomial binary code. Examination of this check sum at the distant end of the SS7 signalling link enables corruption
or errors to be detected. The receiving end sender/receiver may then either directly
correct the contents of the received SU or, if it is too severely corrupted, request a
re-transmission of the SU. Next is the signalling information field – i.e. the actual
message to be sent. This field may vary in length up to a maximum of 272 bytes,
depending on the nature of the signalling information being sent. The next field is
used to indicate which of the portfolio of messages is being used. Since the size
may vary according to the amount of signalling information being sent, the next field
indicates the length of the SU. The final two fields are used to carry the forward and
166 Understanding telecommunications networks
Point codes
and forwards
Sequence No.
Signalling information field
Signal unit
Figure 7.5
SS7 Message and Signal Unit Formats
backward sequence numbers. These are used to count and identify the sequence of
SUs sent in each direction so that re-transmissions of specific SUs can be requested
and later placed in the correct sequence upon receipt.
Each type of message will have a different format depending on the nature of the
signalling information to be sent [3]. However, as an example we will look at the
format of the IAM, described earlier, as shown in Fig. 7.5. The first two fields contain
the destination and source points codes, each of which is the 14-bit address of the
relevant SS7 node. The next field indicates the circuit number (or channel) to which
the message relates, e.g. 17 or 25 in the example explained earlier. Header codes in
the next field indicate the type of message, in this case an IAM, so that the content
can be appropriately interpreted. The final field contains the called and, if required,
the calling telephone numbers.
We are now in a position to consider Fig. 7.6, a block-schematic diagram showing
the inter-working of the SS7 system and the exchange systems (covered in Chapter 6).
This shows trunk exchanges A and B connected by a traffic route comprising two
2 Mbit/s digital streams, each with 30 speech channels. The SS7 link terminates in
either exchange on a sender/receiver (S/R) system linked to the exchange-control
systems and the digital route switch-blocks. Exchange A’s control system initiates
a message, say an IAM, by sending an instruction to the SS7 signalling-control
subsystem through the control link, passing all relevant information, including dialled
number, circuit or channel identity and the point code of the destination SS7 node (i.e.
the SS7 S/R at exchange B). The formatted message is then passed to the signallingtermination subsystem, where it joins a queue with other messages waiting to be sent.
This subsystem then creates a signalling unit to carry the message by the addition of
the SU sequence numbers, length indicator, flags, etc., and the generated check sum
from the error-control subsystem. As soon as a vacant slot appears on the signalling
link, the SU is transmitted via the time slot 16 (TS16) in one of the 30-channel
Signalling and control 167
Exchange B
Exchange A
30 Speech channels
30 Speech channels with
SS7 in TS16
Transfer of signalling units
Figure 7.6
Digital Exchanges and the SS7 System
2 Mbit/s systems, as shown in Fig. 7.6. The transmission rate of SS7 signalling over
the network is therefore 64 kbit/s.
At exchange B the TS16 contents from the 2 Mbit/s link are permanently switched
through to the SS7 signalling-termination subsystem. On receipt of the SU the
error-control subsystem checks for errors and corrects or seeks a re-transmission
if necessary. The SU is then decomposed and the message part transferred via a
queue to the signalling-control subsystem for decoding. The relevant information is
then passed to exchange B’s control system. A corresponding procedure occurs for
the SS7 messages in the reverse (backward) direction. The call set-up and clear-down
then progresses following the sequence described in the above example.
In theory, the SS7 system has sufficient capacity to handle the signalling for
some 4,096 traffic channels – based on the circuit identity field of 12 bits! However,
network operators view a dependency of so many calls on a single signalling link as far
too vulnerable, and significantly lower capacities are used in practice. For example,
the typical configuration for a remote concentrator is for the mandatory route to its
parent exchange to be made of up to eight 2 Mbit/s (30-channel) systems, with their
physical routeing spread over at least two separate cable routes following divergent
paths (a diversity factor of two). The SS7 signalling link is carried over TS16s of two
of the separated 2 Mbit/s streams, the TS16s of the remaining six 2 Mbit/s streams
kept empty. The signalling load is shared across the two TS16 paths, with the ability
to send all signalling units over either one in the event of a cable failure.
The resilience of the SS7 signalling network is further enhanced by automatically
sending signal units over alternative routes in the event of congestion or failure on the
168 Understanding telecommunications networks
(a) Fully associated
(b) Quasi-associated
= SS7 signalling link
= Speech traffic link
Primary Choice:
A–B, B–C, C–D
(c) Alternative routeing and non-associated working
Figure 7.7
Modes of Association
first choice SS7 route, similar in concept to the automatic alternate routeing (AAR) for
telephone traffic described in Chapter 6. In this way the routeing of the signalling links
can become separated from the traffic routes that they serve, as illustrated in Fig. 7.7.
The intermediate SS7 nodes may be located at telephone exchanges, or provided as
stand-alone STPs providing a tandem capacity exclusively for SS7 messages. The
use of STPs can in general provide economies in terms of number of SS7 signalling
routes and, as we shall see later in this section, more efficient routeing of signalling
messages to network data bases.
SS7 has a flexible architecture which can be enhanced by the addition of new
portfolios of messages as new types of service are introduced by network operators.
Fig. 7.8 presents a simplified architectural view, which is based on a common message
transfer part (MTP), i.e. generating and transferring the signal units, and a series of
user or application parts, i.e. the message portfolios. The set of messages required
for telephony call control is contained in the telephone user part (TUP), as used
in the example earlier. The messages for the more-complex call control for ISDN
are contained in the ISDN user part (ISUP). (Other user parts have been specified
in anticipation of new services, e.g. data (DUP) and broadband ISDN (B-ISUP)
which have yet to be adopted.) There are also a set of message portfolios in socalled application parts, namely, intelligent networks (INAP), mobile (MAP) and
operations, maintenance and administration (OMAP), which have to be supported by
what might be called the non-circuit related routeing part, since these are not directly
associated with traffic circuits. This part sets up a logical routeing for the messages to
follow through the SS7 network and supports the message flow, taking account of the
requirements of the application or service involved. Details of this part, which contains
Signalling and control 169
Operations &
Transaction capability application part (TCAP)
Intermediate service part (ISP)
Signalling connection control part (SCCP)
Non-circuit related
routeing part
Message transfer part (MTP)
Figure 7.8
Simplified Architectural View of SS7
the TCAP (transaction capabilities application part), ISP (intermediate service part)
and SCCP (signalling connection control path), can be found in References 1 and 4.
A typical example of the use of the INAP set of SS7 messages is the enquiry by
an exchange of a remote data base in order to complete a complex call. The call is
initially routed from the local exchange to a trunk exchange using SS7 TUP or ISUP
set of messages. Here, the exchange-control system determines that a translation of
the dialled digits to an appropriate destination number is required from a remote
data base – part of an intelligent network (IN) – in order to complete the call. First,
an INAP message seeking a translation of the dialled number is sent over the SS7
signalling links, directly or via several SS7 links, using the non-circuit related routeing
part to steer the enquiry to the remote data base. The response to this inquiry, advising
the new destination number, is then sent as an INAP message back over the SS7
network routeing to the trunk exchange control. Now the call can be completed
through the PSTN using the SS7 TUP or ISUP messages in the normal way. This noncircuit related form of SS7 signalling (using INAP messages) is now widely deployed
in support of the advanced voice services provided by the various intelligent networks,
as described in Section 7.3.2.
7.2.4 ITU H323 and session initiation protocol
Although SS7 is widely deployed in nearly all PSTN and intelligent networks around
the World, there is a new breed of signalling systems being developed to support the
use of IP networks to provide voice calls (‘voice over IP’) – an essential part of the
so-called NGNs. We will briefly consider two of these new signalling systems. The
170 Understanding telecommunications networks
first, H323, is the signalling system specified by the ITU primarily for the setting up
of calls via multimedia networks, e.g. a video conference call. This system uses the
transport capability of the IP packet network to convey the messages (equivalent to
the message transmission part of SS7) and a series of message sets to set up variable
bandwidth and other features. Standard telephone numbering is used to identify subscribers [5]. It also uses the ITU-T Q931 protocol, shown in Fig. 7.3, to manage the
telephony aspects.
More recently a simpler system has been specified by the IETF (Internet Engineering Task Force) specifically for the voice-over-IP networks. (For more information on
the IETF see Appendix 1.) The session initiation protocol (SIP) uses URL addresses
(i.e. like an e-mail address) rather than telephone numbers to represent the users and
the protocol is based on a series of text-based messages. Like H323, SIP uses the
transport capabilities of the IP network to convey the signalling messages between
nodes. At its simplest level, a session is set up by the interchange between two computers (with telephones attached); an interchange of messages ensures compatibility
between the two terminals and acceptance of the ‘call’. The SIP signalling process
is simply a series of requests and acknowledgements between software agents –
known as ‘user agents’ – acting on behalf of a user or inanimate computer equipment
(‘automatons’) [6].
A simple example is given in Fig. 7.9(a), which shows the SIP message flows
involved in establishing and clearing down a multimedia call, say voice and text,
SIP user agent B
SIP user agent A
Media session
(a) SIP message sequence
Call ID
(b) SIP message format
Figure 7.9
The SIP Concept
Caller ID
Caller ID
proxy ID
(e.g. invite)
Signalling and control 171
between users associated with computer terminals A and B. The user agent for A
(resident in the software of A’s computer) generates an ‘invite’ message which is
conveyed to B over the appropriate data network – usually an IP network, as described
in Chapter 8. User agent B responds positively with an ‘OK’ message, to which user
agent A responds with an ‘acknowledge’ message. The media session is thus initiated;
the session itself comprises a flow of data (IP) packets between the A and B computers.
This session is ended by either party sending a ‘Bye’ SIP message, which is responded
to by the other party sending an ‘OK’ message, as shown.
The SIP messages comprise two parts: a header, containing the logic of the message and a body containing the identities of the parties (actually, the user agents)
involved in the signalling, as shown generically in Fig. 7.9(b). Special SIP names are
given to identify the user agents (although for convenience these are made as similar
to e-mail names as possible) – the ‘caller ID’ and ‘called ID’. There is also space in
the body for the ID of any intermediate nodes involved in routeing or interrogating
the SIP message to be included in the body of the SIP message (shown as ‘via proxy’
in Fig. 7.9(b)). Since the message will be one of several relating to a call, a ‘call ID’ is
included in the body of the message, together with a sequence number which is used
should a re-transmission of the whole message be requested due to faulty receipt of
the original.
SIP is a simple signalling system which depends on IP and several supporting
protocols for its transmission. However, it is versatile and readily extendible to incorporate new service features, covering voice, data, video, etc. In particular, it has
been adopted as part of the future 3G all-IP mobile network specification (UMTS
Release 5), as described in Chapter 9. It is also a strong contender for call-control
signalling in future IP-based voice networks and NGNs, as discussed in more detail
in the Chapters 8 and 11, respectively.
Call control
7.3.1 Exchange-control systems
Having seen the close association of subscriber and inter-nodal signalling with the setup of a telephone call, it is now appropriate for us to consider the call-control system
within an exchange. The block-schematic diagrams of a local exchange shown in
Chapter 6 (Figs 6.4, 6.10 and 6.11), show a single entity labelled exchange-control
system. In fact, this system typically comprises multi-processor clusters and a set of
regional processors, as shown in Fig. 7.10. The supervisory line conditions detected by
the subscriber line cards (e.g. ‘off-hook’ and loop-disconnect dial pulses) are gathered
by a line controller, which periodically scans all the line cards on the concentrator
switch-block. The outputs of the line controllers associated with each of the co-located
subscriber concentrators are gathered by a regional processor which formats the offand on-hook indications and any dialled digits into a series of computer compatible
messages for analysis by the cluster controller. This regional processor also assembles
the dialled digits by the tone phones as detected by the MF sender/receiver. All of
these messages from the regional processor indicate the appropriate subscriber’s line
172 Understanding telecommunications networks
card Concentrator
Local or remote
OA&M console
S/R Switch-block
Line controller
Regional processor
control line signals
assembles digits into messages
Cluster controller
Cluster 1
Figure 7.10
Cluster 2
Exchange-Control System
number so that the messages can be associated with the relevant call set-up or cleardown currently being handled by the processor cluster. Similarly, messages to and
from the SS7 sender/receiver are passed by the cluster controller to the cluster dealing
with the call.
The clusters comprise several processors (central processing units, CPUs) and
associated memory, containing a copy of the call-control program and relevant
exchange connection data. The cluster controller shares the call-control work across
all the clusters. Call control is achieved by following a logical sequence that leads
to appropriate actions within the switch-blocks, based on the digits received from
the calling subscriber, or distant exchange in the case of an incoming call. When the
controlling cluster determines that a connection is required across one of the switchblocks a message is sent via the cluster controller to the appropriate switch-block
control – in the form of a request to ‘connect a particular inlet time slot to a particular
outlet time slot’. The switch-block control is then able to set up the connection by
examining the current occupancy details as held in a local store (known as ‘map in
memory’). A routeing table held in the cluster memories holds all the first choice
and subsequent routeing details for all traffic routes from the exchange. These cluster
memories also save all the call records for later off-line processing of the subscriber
bills, as well as recording traffic-usage statistics for use by the network operators to
dimension the exchange, as described in Chapter 6. The control system time shares
its attention between all the active calls by processing the set-ups and clear-downs in
Signalling and control 173
stages, progressing each call to the next stage and then turning its attention to the other
calls during the relatively long times of waiting for digits to arrive or connections to
be set up [7].
Actually, the exchange-control systems provide extremely powerful and reliable
real-time processing. Typically, they are capable of handling up to 1,000,000 busy
hour call attempts (known as ‘BHCA’), operating on a truly continuous basis (‘24 ×
7 × 365’). The call-control and maintenance programs for these systems are complex
and lengthy. To make the point, engineers have estimated that a printed version of
the source program of a typical digital SPC exchange would fill about 10 miles of
computer print-out paper! (Indeed, some particularly feature-rich switching systems
are known as ‘12-mile exchanges’.) The software is updated periodically by the
switching system manufacturer to incorporate new features for customer services
and to correct any bugs on earlier versions. Network operators typically change the
software on each of their processor and trunk exchanges once every 12–18 months –
amazingly, without any break in service! This feat is achieved through the partitioning
of the processor clusters into working and off-line status; changing the software on the
latter; and then swapping the status so that the software on the remaining processor
clusters can be changed. These change-overs are usually performed during the quiet
periods of exchange usage, e.g. in the early hours and at week ends. Obviously, this
standard of reliability is significantly greater than that generally experienced on office
and home computers.
A set of operations, administration and maintenance (OA&M) consoles are provided locally within the exchange building and remotely at one or more OA&M
centres, supported by a man-machine interface (MMI) system attached to the cluster
controller, as shown in Fig. 7.10. These consoles are used by the operator’s technicians
and administrators to, variously:
set up each new subscriber’s line (‘service provision’),
amend the facilities settings for individual subscribers,
support fault clearance of a subscriber’s line,
initiate traffic monitoring of the exchange,
change the routeing table for the exchange,
manage changes to the control software,
change the equipment status and configuration during software upgrades and
installation of new capacity,
manage the recording and off-loading of billing information.
The key challenge for network operators in managing the exchange call-control systems is the assembly of the correct data and software versions for each exchange – for
use in initiating or upgrading exchanges. This is usually referred to as ‘data build’.
The set of software versions deployed at each exchange-control system in the network
needs to be recorded at the central operations centre, so that upgrades and software
patching can be managed accurately. As far as the data is concerned, each control
system needs to be kept up to date on the latest routeing details, including any changes
of codes, subscriber number blocks or parenting arrangements of local exchanges,
174 Understanding telecommunications networks
etc. A further major area of changes of routeing details is incurred by the addition or
withdrawal of other network operators and changes to the routeings to their POI.
7.3.2 Intelligent network (IN)
In the consideration of inter-nodal signalling above we recognised that call services
range from simple telephony to the more complex calls requiring some form of
translation of the dialled number, or other additional processing.
Whilst the software in the exchange-control system is designed to provide the
basic telephony service, it may not be able to provide the additional capabilities
necessary for the service features needed by a network operator. One option is for a
network operator to commission the switching system manufacturer to develop the
additional program code to provide the extra features, for inclusion in the next upgrade
to the exchange-control software. However, this may be expensive due to the need
to include the upgrade on all exchanges installations despite the initial uncertainty of
the amount of take-up or usage that customers will make of the new service. There
will be other situations where the complexity of the new features cannot practically
be provided on standard exchange-control systems. The rise in the need for ever more
service features led to the development of the IN concept during the 1980s and 1990s.
The IN concept is based on the managed access from the PSTN exchanges to
a centralised data base where the complex and expensive control software and any
associated service data is concentrated. The standard architecture of an IN is shown in
Fig. 7.11, in which the IN is shown as an overlay to the PSTN. When a call is received
by the PSTN that is recognised from the number dialled as an ‘IN call’, it is routed
Open network provision
(ONP) interface
switching point
Figure 7.11
The Intelligent Network Concept
Signalling and control 175
to an IN gateway exchange – known as a ‘service switching point’ (SSP). Normally,
the major trunk exchanges in the PSTN take on this additional role of being an IN
SSP. The call is held at the SSP, with the normal call-control sequence interrupted (or
‘triggered’) so that an interrogation can be made of the remote IN service control point
(SCP) and its IN data base (INDB). Typically, there are two or three SCPs located
within the network serving the whole country through several hundred SSPs. The
interrogation of the SCP is achieved by using non-circuit related signalling over the
SS7 network, perhaps using STPs to route the messages, as described earlier in this
chapter. On receiving the response to the interrogation, the SSP is able to complete
the call using the new destination number and perhaps other control information, as
advised by the SCP.
The other key element of the IN architecture shown in Fig. 7.11 is the so-called
service creation environment (SCE), a development facility associated with the IN
where new services can be constructed using software-based features as basic building
blocks. This facility can act as the network operator’s test bed for new services or
extra features on existing services, and can also tailor specific services to the particular
needs of large business customers. Finally, the standard architecture identifies opennetwork provision (ONP) interfaces at key points, with a view to allowing different
manufacturer’s equipment to be used throughout the IN. For example, the interface at
point A is designed to allow different manufacturer’s exchanges to operate as an SSP
working to a standard SCP. Similarly, the interface B is designed to allow different
manufacturer’s SCEs to work to one or more different manufacturers’ SCPs [8,9].
Actually, there is no such thing as an ‘IN service’, since all services could be
realised either by using resident exchange-control systems, perhaps with additional
control hardware, or by using an intelligent network solution. The choice is essentially
one of economics. In general, exchange-control-based solutions are used for the
following categories of services:
(i) Where service features are required on every call, e.g. basic call set-ups with
standard charging arrangements,
(ii) Mass usage calls, e.g. ‘ring back when free’,
(iii) Low-value calls (i.e. very low charges or free), e.g. ‘1471’ in the UK.
However, the following categories of services are appropriate to an IN-based
(i) Conditional number translation, e.g. ‘Freefone’ calls, where the destination number depends on one or more conditions (time of day, geographical
area of the caller, instantaneous loading on call centre attendants, staffing
arrangements at call centres, etc.)
(ii) Complex charging arrangements, e.g. ‘Freefone’ calls, where the charges are
fully or partly paid by the recipient or other third party rather than wholly by
the caller.
(iii) Customer-specific management information, e.g. where usage and distribution
information on calls from their clients to a business customer are assembled
and notified automatically to the customer.
(iv) Corporate numbering schemes, e.g. for corporate VPNs.
176 Understanding telecommunications networks
(v) Multi-operator or complex number translation, e.g. number portability of
geographical or non-geographical telephone numbers.
(vi) Mobile calls, e.g. the determination of the location of telephone handsets in
real time.
(vii) Calls that require a network-wide view in order to determine how they need
to be routed, e.g. distributing calls to a dispersed set of manual call centres
around the country depending on the queues at each of the centres.
As a general rule, new services are developed and provided on an IN initially. Within
a few years, the volume of the customer usage and take-up may increase to such a
point that it becomes cheaper to arrange for the necessary enhancements to the control
software and hardware at the exchanges, so that the service can be off-loaded from
the IN and handled instead by the PSTN. Thus, a network operator will use both the
PSTN and the IN to provide a range of services for both residential and business
Fig. 7.12 illustrates a practical arrangement of a set of intelligent networks, as
might be used by an incumbent national network operator in conjunction with a
PSTN. In this example, the trunk exchanges of the PSTN act as the SSPs for all
three IN networks. The non-circuit related SS7 signalling messages from these SSPs
are routed via signalling transfer points to the appropriate IN SCPs according to
the nature of the service. The main IN SCP provides several large volume services,
each based on a transaction server and associated data base. Resilience is assured
Service creation
Small volume
services IN
Service 3
New services IN
(small volume)
Service 2
Service 1
multi-purpose IN
STP = Signal transfer point
Figure 7.12
A Typical Set of INs
= Server
= Non-circuit related
SS7 signalling
Signalling and control 177
through the triplication of the SCP, each unit being located at separate buildings
across the country. Signalling traffic to these SCPs is spread evenly, although the three
SCPs are periodically kept in step so that in the event of any single SCP failure the
other two can share the work load. Relatively small volume but specialised business
services are provided by a dedicated IN SCP (an example is customised automatic call
distribution – ACD). There may also be a new-services IN, comprising a small-scale
development facility with a powerful SCE facility. New services are launched on this
SCP, where they remain until they grow in volume to a level where it is economical
to transfer them to the large-volume IN. Alternatively, the services might instead be
transferred (i.e. migrated) to the exchange-control systems on dedicated specialised
exchanges or local and trunk exchanges within the PSTN.
7.3.3 Future network intelligence
Despite the successful implementations of INs in most countries during the 1990s
enabling the provision of a range of advanced voice call services, as described above,
not all of the attributes originally hoped for were achieved. In particular, the implementation of truly open interfaces (ONP) has proved difficult to achieve. Also, the
extent of service creation has been limited to just those functions associated with callcontrol software – hence omitting the non-exchange aspects of service management.
However, the biggest change in the way that intelligence is introduced to modern
networks is commercial. The IN architecture is centred on a single network operator
having the control of a service creation and its later execution, a construct that does
not fit well with the growing multi-operator open environment that is being progressively introduced in many countries since the year 2000. There are now several ways
in which intelligence based on a computer server can be made available in a network
to provide the extra service capability – the IN approach may now be seen as just one
One of the significant changes in the way that intelligence is managed in the
network meeting the new commercial position is defined in the approach of the
Parlay consortium [10]. Fig. 7.13 gives a simplified view of the Parlay architecture
as associated with a fixed network operator. The key component is a Parlay gateway,
provided by the operator, which offers a set of software functions that can be used by
one or more service providers to create a service that can run on the network. The set
of software functions is made available through an application programming interface
(API). The Parlay consortium has defined the API in such a way that service providers
can develop services on a variety of different networks: PSTN, IN, mobile, IP data
networks, etc., in a commercially attractive way within the confines of maintaining
the integrity and resilience of the networks used.
In this chapter we saw the close association of signalling with the control of call
connections across the network. First, the different forms of subscriber signalling
178 Understanding telecommunications networks
Service provider
(Parlay client)
Data link (e.g. IP)
Data link (e.g.IP)
Figure 7.13
The Parlay Concept
were identified, ranging from the old fashioned use of a dial, the ubiquitous tone
dialling system and the more advanced forms of CCS from ISDN terminals and PBX.
We then focused on the inter-nodal signalling, specifically the widely deployed ITU
CCS system No. 7. After considering the message flow involved in a simple telephony
call across three exchanges, we examined the use of SS7 to convey data transactions,
typically enquiries of a data base – the non-circuit related form of signalling. Finally,
the new signalling systems, SIP, being introduced with the voice-over-IP networks
were discussed.
In Section 7.3, we considered the control aspects, initially looking at the exchangecontrol system in a telephone exchange. We then looked at how more-complex calls,
which require number translation or more call processing than normal telephony, can
be handled by an IN. The interrogation messages to and from the SCPs of the IN
use non-circuit related signalling carried over the SS7 network, including the use of
STPs. This chapter concluded with a brief examination of the role of APIs for service
providers, as defined by the Parlay consortium.
MANTERFIELD, R. J.: ‘Common-Channel Signalling’, IET Telecommunications Series No. 26, Stevenage, 1991, Chapter 1.
2 WELSH, S.: ‘Signalling in Telecommunications Networks’, IET Telecommunications Series No. 6, Stevenage, 1981, Chapters 3–8.
3 MANTERFIELD, R. J.: ‘Telecommunications Signalling’, IET Telecommunications Series No. 43, Stevenage, 1999, Chapter 6.
4 Ibid., Chapter 7.
Signalling and control 179
5 TANENBAUM, A. S.: ‘Computer Networks’, Fourth edition, Prentice Hall,
Amsterdam, the Netherlands, 2003, Chapter 6.
6 WISELY, D. R. and SWALE, R. P.: ‘SIP – The Session Initiation Protocol’.
Chapter 11 of ‘Voice Over IP: Systems and Solutions’, edited by SWALE, R.
P., BT Exact Communications Technology Series 3, IET Publishers, Stevenage,
7 REDMILL, F. J. and VALDAR, A. R. ‘SPC Digital Telephone Exchanges’, IET
Telecommunications Series No. 21, Stevenage, 1990, Chapter 15.
8 TURNER, G. D.: ‘Service Creation’, Chapter 8 of ‘Network Intelligence’ edited
by DUFOUR, I. G. BT Telecommunications Series, Chapman & Hall, London,
9 ANDERSON, J.: ‘Intelligent Networks: Principles and Applications’, IET
Telecommunications Series No. 46, Stevenage, 2002, Chapter 2.
10 Ibid., Chapter 6.
Chapter 8
Data (packet) switching and routeing
Previous chapters have concentrated on the communication of voice, primarily in the
form of the telephony service as carried over the PSTN. We are now in a position
to consider the various non-voice services, generally referred to as data, and the
networks that carry them. At the end of this chapter we will investigate the way that
these data networks can also convey voice, and that leads us to consider the new
generation of communication networks.
8.2 The nature of data
Data is intrinsically different to voice. By definition, data is information which generally originates in the form of a digital representation – normally binary 1s and 0s
and it therefore does not need to be converted to digital within the network, unlike
voice which originates from the telephone microphone as an analogue signal. Data
originates from terminals, such as temperature measuring devices, or from a laptop or
computer, consoles, control monitors and ATMs (i.e. ‘automatic teller machines’, not
the data service described later in this chapter!) in the wall of a shop or bank. Applications generating data from these devices include the sending of e-mails, transfer
of information between files and other forms of enquiry, remote control of machinery and other forms of surveillance, etc. A data service is one that conveys data
between the terminals and other devices. There are several key differences between the
characteristics of voice and data and hence the different requirements for successful
communications, as summarised below.
(i) Data applications may be tolerant or intolerant of delay on a communications
network, known as ‘latency’ tolerance or intolerance. (Unlike voice, which is
always latency intolerant.)
182 Understanding telecommunications networks
(ii) Data may be generated as a continuous stream or in bursts. (Digitised voice is
continuous stream.)
(iii) Data may flow at any speed, ranging from the extremely high billions of bits
per second to the very low few bits per second. (Digitised voice is carried at 64
(iv) Data is usually highly susceptible to error, every bit is precious. (Digitised
voice is fairly robust to errors and even the loss or corruption of many bits is
not perceived by the listener.)
(v) There is a wide range of ways that the data can be organised, i.e. many protocols
are used. (Voice encoding is highly standardised.)
(vi) The data may be encrypted (as is digitised voice over mobile networks).
Circuit-switched telephone connections through a PSTN can be used to carry
data services, indeed the earliest form of switched data service in the 1960s and
1970s was known as ‘datel’ – data over telephone. However, the nature of circuit
switching, with its continuous connection held for the duration of the call, does not
provide the most economical communications medium for the majority of the types
of data. Also, the relatively slow speed of call set-up can be a problem for short
bursts of data. The specific needs of data services have, therefore, been addressed by
communications networks based on packet rather than circuit switching. Not only does
packet switching give better network utilisation and economics than circuit switching,
but it also copes well with the speed variation and other special characteristics of data.
The remainder of this chapter now considers packet switching for data. Later we will
examine the treatment of voice as just another form of data service and look at the
switching of voice over packets rather than circuits.
A brief note on terminology: The technical literature associated with data networks, describes the size of the chunk of data being considered either as a number of
‘bits’ (i.e. 1s and 0s) or in units of eight bits known as ‘bytes’. Similarly, the speed of
data transmission is described as either ‘bits/s’ or ‘bytes/s’ – obviously, the two differ
by a factor of 8. (See also p. 325 for a summary of binary multipliers.)
Packet switching
A data packet comprises a string of bits with the front end formed into a header,
giving the destination address, and the remainder of the packet forming the payload
carrying the actual data for transmission. The stream of data is thus deposited into
one or more packets, each with an appropriate address, and the packets sent down a
transmission bearer interleaved with packets from other sources. Fig. 8.1 contrasts the
packet approach with the continuous path of a circuit-switched connection. A useful
analogy for packet switching is that of a postal service. Users put their letter into
an envelope onto which they write the destination address. Similarly, the data is
placed into the payload of the packet and the destination address inserted into the
header. Like the postal service, other information may be included in the header,
e.g. originating (‘from’) address, date of despatch and a reference number. Different
capacity envelopes are used in the postal case to handle the range of lengths and size of
Data (packet) switching and routeing 183
Continuous circuit connection
Payload (contents)
Figure 8.1
Packet and Circuit Switching
letters – and a similar arrangement exists for the data packet services, as we shall see
throughout this chapter. Finally, the equivalent of packet switching is the role of the
sorting office, i.e. the reading of the envelopes and the depositing into the appropriate
outgoing bags destined to be conveyed to the next stage of sorting. Different sizes of
envelopes are accommodated by the postal bags and all the envelopes are intermingled
or interleaved within the bags as they are transported by van to the next sorting office.
Just like the postal service, data packets can get misrouted or lost – once this is
detected the recipients then ask the sender for a replacement packet.
There are two fundamental ways that packets can be switched through a network,
following either ‘connection-orientated’ or ‘connectionless’ modes of operation.
8.3.1 Connection-orientated packet mode
Fig. 8.2 illustrates the connection-orientated mode. This works on the principle of
establishing a so-called virtual path (VP) for the packets to follow. The three phases
involved – virtual path set-up, data transfer and then virtual path clear-down – are
analogous to the set-up, conversation and clear-down phases of a circuit-switched
call. However, unlike the latter no actual path is established for the packets. Rather,
a relationship is set up between all the packet switches involved in the virtual path
across the network. In the example of Fig. 8.2, during the set-up phase it is determined
that all packets sent between terminals x and y will have a virtual path identity label
43. Each of the designated packet exchanges, in this case A, D, and E, have the
appropriate outgoing route identified in their routeing tables against VP 43. All packets
flowing from either terminal x or y with the label 43 in their header will follow this
predetermined routeing for as long as the virtual path exists. (This is analogous to
the sorter at the originating office scribbling ‘43’ on the envelopes and all subsequent
sorters merely using this number, rather than reading the full address and looking up
the appropriate way to forward the letter, and all subsequent letters would follow the
same routeing across the country.)
The paths may be set up for permanent operation – known as ‘permanent
virtual circuit’ (PVC) working; or they may be set up whenever a data session is
184 Understanding telecommunications networks
Circuit 43
Circuit 43
Figure 8.2
Box 8.1
Connection-Orientated Packet Switching
X25 Data Service
The first connection-orientated public packet data network service was based
on the CCITT (now ITU) X25 international standard as implemented in many
countries during the 1970s. It was introduced at a time when public networks
were still based on analogue transmission, the noisiness of which tended to
create errors on the X25 data stream, requiring high levels of correction.
The X25 packet comprises a 3-byte (24 bits) header containing the virtual
circuit number and packet acknowledgement and request sequence numbers,
followed by a payload for the carried data. The maximum size of the packet
is 128 bytes. Non-X25 terminals need to be connected to converters (packet
assembler/disassembler, PAD) before being linked to the serving X25 packet
switch in the network, using a 48 or 56 kbit/s modem driven digital stream over
analogue and digital private circuits or dial up over the PSTN. Although still
used in many parts of the World today, X25 has generally been superseded by
frame relay service [6,10].
required – a service known as ‘switched virtual circuit’ (SVC). The earliest form of
public packet data service, known as CCITT X.25 or just ‘X25’, used the PVC form
of connection orientated operation (see Box 8.1). Other notable examples of this form
of packet switching are frame relay (see Box 8.2) and ATM, which is described in
more detail later.
Data (packet) switching and routeing 185
Box 8.2
Frame Relay
Frame relay packet networks are designed to exploit the low noise capability of the digital transmission networks which are now prevalent around the
World. The connection-orientated system, therefore, has very little of the error
correction overhead of X25 and so is able to operate at much higher speeds,
from 56 kbit/s up to some 45 Mbit/s. A wide range of data stream types can
be carried in the variable payload of the frame relay packet, up to a maximum of 4 kbytes. The main application of frame relay network services is to
provide interconnection between LANs on corporate sites, research stations
and university campuses. The link from the customer sites to the frame relay
packet switches is usually provided over private circuits, typically at 2 Mbit/s
(or 1.5 Mbit/s in North America). The large variation permitted in the packet
length and the consequent range in packet transport delays means that there is
only a small amount of performance manageability possible with frame relay
Figure 8.3
Connectionless Packet Switching
8.3.2 Connectionless packet mode
In contrast, the connectionless mode of packet switching treats each packet in a
stream completely independently within the network, operating what is known as
a ‘datagram’ service. Fig. 8.3 shows a network made up of packet switches, which
are normally referred to as ‘routers’ in this mode of connectionless working, and a
terminal x transferring a stream of packets to terminal y. The routers, A to E, handle
186 Understanding telecommunications networks
Box 8.3
Switched Multimegabit Data Service (SMDS)
This connectionless service is primarily directed at linking LANs, i.e. working
as MAN and WAN (metropolitan and wide-area local area networks, respectively). It is a relatively simple system with no provision for error correction
and with no single standard version, so inter-working between different manufacturers’ equipment is difficult. The user data is encapsulated into a packet
with some 40 bytes of header containing the destination and source addresses
(at layer 3). This packet is then chopped into 53-byte packets or frames for
transmission at layer 2. User speeds range from 56 kbit/s up to 45 Mbit/s.
Although deployment of SMDS is widespread, particularly within Europe, it
has generally been overshadowed by the use of frame relay, particularly in
North America [21].
packets from many sources, with the received packets queuing on their inlet routes.
As the first packet from terminal x is received at router A, the address contents of the
header is examined and the outgoing route, say to router D chosen from a look-up
schedule, i.e. the routeing table. In this way, the packet progresses through to router
E. The second and third packet may follow the same routeing, each router examining
the full address in the header and making an independent decision on how to route
the packet onwards. As the loading on the network changes with time due to the
generation of packet flows from other users, the queues on the routers will fill and
temporarily become unavailable causing different choices of routeing for subsequent
packets from terminal x. Thus, the fourth packet might be routed A–C–D–E, and the
fifth packet routed A–B–D–E. This independent handling of each packet means that
the stream of packets between x and y could potentially all follow different routeings
and consequently arrive at various times and out of sequence.
The most famous example of connectionless packet working is that of the Internet
Protocol (IP), as used in the (public) Internet and corporate private intranets. The other
notable example of connectionless working is switched multimegabit data service
(SMDS), see Box 8.3.
8.3.3 Comparison of packet switching modes
It is interesting to compare the merits of the two modes of packet switching. The
connection-oriented approach provides a high degree of capacity and performance
management of the services running over its switches. This is because the control
mechanism will allow new VPs to be established only if there is sufficient capacity
through the nominated switches. Connection-orientated packet switching may thus
be characterised as incorporating a fairly elaborate control mechanism – seen by
some data network designers as a disadvantage – on the one hand, while on the other
hand it has the distinct advantage of ensuring predictable managed performance.
By comparison, the connectionless mode has relatively simple control, which is
robust and able to cope with contention and failure of routers in the network, but
Data (packet) switching and routeing 187
it does not provide any performance prediction or management. Not surprisingly,
network operators try to get the best of both worlds by using connectionless packet
systems running over connection-orientated networks! Another way of looking at this
relationship between the two modes is to refer to the OSI reference model (see Box 8.4)
which places the connectionless packet switching at layer 3 and the connectionorientated mode at layer 2.
Box 8.4
Open System Interconnect Reference Model
The open systems interconnection (OSI) model, which was devised by the International Standards Organisation (ISO) in 1983, is the widely accepted way of
delineating the functions that need to be performed by equipment involved in
the communication of data [10,16]. Like many reference models it is based
on the concept of layers, each of which performs functions or services for the
layer above and is in turn supported by services provided by the layer below. By
clearly defining the functions of the layers, many different manufacturers and
programmers are able to develop terminal equipment, communications equipment and software applications that successfully work together. Interactions
between functions in the same layer at either end of a communications link are
governed by appropriate sets of rules, the so-called protocols.
Fig 8.4 illustrates how the OSI model’s seven layers apply to the communication between a data terminal (the client) and the server computer. Working
from the bottom up, the functions of each of the layers are summarised below:
Layer 1: physical: This layer is concerned with the actual transmission of the
binary signal over the wires. Layer 1 specifications therefore cover the functions
of each pin in the plug and sockets between terminals and communications
equipment, and the shape and size of the electrical waveform in representing
binary 1s and 0s. Actually, the physical transmission media, i.e. the optical or
metallic cables or radio links, are deemed to reside in a hypothetical Layer 0
outside of the OSI model.
Layer 2: data link: The main task of this layer is the safe and reliable transmission of the actual data messages or stream. It therefore concerns the formatting
of the data into frames, packets or cells, which are manageable groups of bits
which can be identified at the distant end and an acknowledgement returned.
The Layer 2 specifications also include error detection and flow control.
Layer 3: network: It is at this level that the routeing over one or more links
through switches or routers is managed. The specification includes the naming
and addressing of network nodes.
Layer 4: transport: This layer provides the management of the data communications on an end-to-end basis, masking any transmission actions at the
network layer below. Thus, the transport layer functions include the allocation
of the source data into numbered packets, sequencing received packets into the
correct order, packet-flow management and re-transmission requests.
188 Understanding telecommunications networks
Layer 5: session: This layer allows two distant computers to run a dialogue for
the duration of the ‘session’. The specification includes a protocol for ensuring
proper hand-over between machines, avoidance of the same function being
tackled on both machines simultaneously (known as synchronisation), etc.
Layer 6: Presentation: It is at this layer that the actual content of the information being communicated is formatted in a mutually understandable format
for the two ends. The specification covers the syntax, semantics and coding of
the content.
Layer 7: application: The actual application being executed over the data
link is defined at this top layer. Examples include the protocols for sending
e-mails, accessing a web page over the Internet, data file transfer, etc.
Application protocol
7. Application
7. Application
Presentation protocol
6. Presentation
6. Presentation
Session protocol
5. Session
5. Session
Transport protocol
4. Transport
4. Transport
3. Network
3. Network
3. Network
2. Data link
2. Data link
2. Data link
1. Physical
1. Physical
1. Physical
Host A
Host B
Figure 8.4
The OSI Reference Model
8.4 Asynchronous transfer mode
First, let’s deal with the curious name of ‘asynchronous transfer mode’ (ATM). It was
coined in the 1980s by the engineering-standards fraternity, who wanted to differentiate the emerging system from the majority of existing data systems which were
synchronous, i.e. all actions initiated under the control of pulses from a digital clock,
and the sending of idle bits in the absence of data. ATM was developed by the telecommunications industry, led by the network operators, as the universal switching system
capable of handling all existing and expected future services, ranging from voice to
Data (packet) switching and routeing 189
User services
ATM cell stream
Figure 8.5
The Multi-Service Capability of ATM
video and all sorts of data: delay tolerant and intolerant, high and low speed. In fact,
ATM was deemed to be the specification for broadband ISDN, the then-hoped-for
future all-purpose service-switching mechanism for public networks, and the natural
successor to the ‘narrowband’ ISDN recently introduced to most PSTNs (as described
in Chapter 6).
The principle of the ATM all-service capability is that an adaptation system allocates the various inputs to a set of fixed-sized 53-byte packets, known as ‘ATM cells’,
as shown in Fig. 8.5. As soon as each cell is filled it joins the queue for despatch;
slow speed inputs are carried in relatively few cells per second, while higher speed
inputs are carried at a proportionately higher cell rate. The user–network interface
is an ATM multiplexor which performs the adaptation function, feeding the set of
user data streams onto the ATM highway – this act of interleaving the set of ATM
cells onto the highway is known as ‘statistical multiplexing’. These ATM highways
are carried through the Access and Core Transmission Networks to the serving ATM
switch. Here, the cells are switched onto the VPs carried over the routes to other
ATM switches, as described in Fig. 8.2 above. Actually, the VPs are capable of being
divided into virtual channels (VC). Thus, for example a VP between two separated
office buildings could contain several VCs providing connectivity between individual
rooms in the two buildings, serving the two sets of accounts departments, financial
management, human resources, etc. An ATM switch therefore requires two switching
stages, one for the VPs and a further stage for the VCs, as shown in Fig. 8.6.
We can now look at the adaptation process more closely. The wide range of users’
data services are divided into five categories for adaptation into payload of the ATM
cells. These categories are defined as ‘ATM adaptation layers (AAL) 1 to 5’, as shown
in the table of Fig. 8.7, which are used by the ATM system to create four classes of
service (A to D) for the user. The great strength of ATM is that the chosen class of
service (CoS) is applied by the ATM switches to all cells passing through the network.
190 Understanding telecommunications networks
VC switch
VP switch
Figure 8.6
VCs, VPs and Switches
Service class
Timing relation
Bit rate
Connection orientated
AAL protocol
AAL3/4 AAL3/4
47 bytes
Figure 8.7
1 byte
identifier identifier
5 bytes of Header
ATM Service Classes, AAL and Cell Structure
This ensures that priority is given at the switches to cells carrying delay-intolerant
services, for example. At the distant receiving end the ATM multiplexor extracts the
data from the cells according to the choice of AAL and reconstitutes the service at
the output tributary port to the user. Briefly, the service classes are:
Class A: Constant bit rate (CBR) service, in which the rate is guaranteed to be kept
stable; e.g. ordinary video or voice.
Data (packet) switching and routeing 191
Class B: Variable bit rate (VBR) service with timing, meaning that although the rate
may vary the ATM service will immediately despatch the cell; e.g. VBR video or
Class C: VBR connection-orientated service, in which delay can be tolerated, e.g.
ordinary data transmission.
Class D: VBR connectionless service, this has the lowest priority within the ATM
network, e.g. data transmission of short messages (datagram). Network operators
use ATM Class D to backhaul the ADSL (‘broadband’) service traffic from the
local exchange to the data network.
The CoS is indicated to the ATM switches during the VP set-up. The switch-control
system will not allow a new VP to be established unless there is sufficient capacity
available to guarantee that the requested CoS can be assured. Once the VP is established the switches respond only to the VP and VC 12-bit identifiers in the short
header of the ATM cell, as shown in Fig. 8.7. The ATM cell header also contains an
8-bit error detection pattern, an indication of whether the cell payload is users’ data
or messages relating to the VP set-up phase, and an indication whether cells can be
jettisoned in the event of switch congestion. The number schemes for ATM networks,
used during the set-up phase of the VP, are considered in Chapter 10.
Before we leave our consideration of ATM it is useful to consider how it is used in
the network operator’s and private networks (see Fig. 8.8). The key point to note is that
an operator’s ATM network is used to provide ATM service to its business customers,
offering an ATM network–user interface (AAL-based) – so-called ‘native ATM’, as
well as using theATM infrastructure to support (at OSI layer 2) the conveyance of other
services provided by the network operator. Examples of the latter include IP service,
Figure 8.8
ATM access
Applications of ATM
ATM Core
Providing service for ATM customers
Support of other networks (IP)
192 Understanding telecommunications networks
video and TV links, broadband access and voice. This use of the ATM infrastructure
enables network operators to manage the transmission capacity within the Access and
Core Transmission Networks through the setting up of VPs between nodes so that the
various services carried receive the appropriate QoS. It also represents an economical
way of partitioning the transmission capacity of the networks. We will look at ‘IP over
ATM’ later in this chapter. Business customers also use ATM within their corporate
networks to provide managed data links between their office, research and factory
Internet protocol
The ‘Internet protocol’, as its name suggests, is the basic protocol used on the Internet
(at layer 3 of the OSI model). However, the tremendously widespread adoption of this
connectionless packet standard is also due to its applicability to so many situations. IP
forms a simple means of transferring data among users on different types of computer
and devices; it forms the common interface to many applications run on computers
and, last but not least, it provides the connectivity infrastructure for corporate data
networks (i.e. intranets). As we shall discuss later in this and subsequent chapters,
IP-based technology is capable of forming the next generation of telecommunications
networks. The operation of the IP service consists of the transfer of users’ data by IP
packets, each of which is managed individually and passed progressively from router
to router, using an IP addressing scheme, until the destination terminal or computer
host is reached. There are currently two versions of IP being used in networks: Version 4 (IPv4) is the original and still most widely used, while Version 6 (IPv6) is
being introduced as a replacement in order to take advantage of its greater addressing
capacity and improved facilities [1]. The addressing aspects of both IP versions are
considered in Chapter 10.
The IPv4 packet structure comprises a fixed header of 20 bytes and a variable
payload of a maximum of 64 kbytes (although in practice this is normally limited to
1.5 kbytes to avoid congestion within the network), as shown in Fig. 8.9. The header
contains the 32-bit source and destination addresses, as well as an indication of the
length of the overall IP packet. Other fields in the header provide a checksum (for
the header contents only), the version of IP being used and a simple type of service
indicator. An intriguing field called ‘time to live’ is used to set the maximum number
of hops between routers before the packet is discarded; this is necessary to prevent the
clogging of the routers by unsuccessfully delivered packets continuously circulating
around the IP network.
As with any communications network, there are two basic nodal functions in
public and private IP networks: the edge routers supporting a number of access lines
from the data users, and a set of core routers (usually of high capacity) providing the
interconnectivity between edge routers and any other IP networks. The role of an IP
router is shown in Fig. 8.10. Packets arriving at the input ports of the router are queued,
normally on a first-in first-out basis, and the header of each is examined in turn by
the control system. A first choice outgoing route to the next router is obtained from a
Data (packet) switching and routeing 193
20–60 bytes
Payload (max 64 kbytes)
Time to live
1 1
check sum
2 1 1
Fragment offset
Length in bytes
Figure 8.9
and header
IPv4 Packet Format
Routeing table
Destination Next router
IP packet stream
Figure 8.10
IP Packet Routeing
simple look-up in the routeing table, using the appropriate part of the destination IP
address. The packets then join queues associated with the output ports.
There are a number of ways in which the routeing table is generated and updated.
Routers are able to discover the status of all the other routers in their network, i.e.
whether they are experiencing long queues, are out of service due to faults, and
whether new routers have come on stream. Discovery protocols control the investigative stream of IP packets sent periodically by routers to each other (e.g. ‘open
shortest path first’, OSPF). In this way a router-control system can build up a picture
of the best way to route to particular destination routers by taking the least number
194 Understanding telecommunications networks
Autonomous system no.3
Autonomous system no.1
Boundary routers
Inter-AS network using BGP
Autonomous system no.2
Autonomous system no.4
Boundary routers
Figure 8.11
Autonomous IP Systems
of hops or travelling over the least distance – the latter being determined by timing
the round-trip of an exploratory packet. Routeing tables are automatically modified
several times a day, reacting to changing circumstances in the IP network [2].
The terminals and hosts using an IP network may be part of a localised data
network, i.e. a sub-network, within the workplace or even a residence (LANs are
discussed later). There is therefore a need for an incoming IP packet addressed to the
sub-network to be associated with the appropriate terminal or host machine. This is
done by the use of one of several available protocols, e.g. reverse address resolution
protocol (RARP) and dynamic host configuration protocol (DHCP) [3,4].
A network of IP routers owned by a particular organisation is usually formed into
a group known as an ‘autonomous system’ (AS) for the purposes of management. So,
for example, an AS owned by a corporation may have a gateway to other IP networks,
but need to limit the extent of other network users gaining access to the corporate
network. Therefore, there are two algorithms required to determine routeings: one
for routeing within the AS, known as interior gateway protocol (IGP), and one for
routeing between one AS and another, known as exterior gateway protocol (EGP).
Fig. 8.11 illustrates the relationships between IP autonomous domains.
In its raw state IP does not actually provide a useful data conveyance service. Some
additional functionality is needed at the transport layer (i.e. layer 4 in the OSI 7-layer
reference model) to manage the flow of IP packets. The two main contenders for
this role are the transmission control protocol (TCP) and the user datagram protocol
(UDP). Fig. 8.12 illustrates the role of TCP, which acts as a connection-orientated
protocol between the sender and receiver end points (terminals or host machines) at
layer 4. None of the routers in the IP network are aware of the TCP activity. The
Data (packet) switching and routeing 195
TCP header
IP header
IP packet
Flow control
Routers examine
only IP packet
Figure 8.12
IP header
IP packet
IP packet
TCP header
in sequence,
of delayed or
errored packets
Transmission Control Protocol (TCP)
main function of TCP is the numbering of IP packets, ‘datagrams’, at the sending
end and the examination of these numbers at the receiving end – and requesting the
re-transmission of any packets not arriving within a specified limit. This number
sequence is also used by the TCP facility at the receiving end to assemble received
packets into the correct order. Finally, TCP provides a monitoring of the rate of flow
of packets getting through the IP network and consequently adjusts the launch rate of
IP packets. The TCP envelopes the source data by a 20-bit header and the composite
is then carried by the IP packet as a payload.
UDP operates in a connectionless mode, as a datagram, providing some similar
facilities to TCP but with the important difference that no re-transmissions of lost
or errored IP packets is requested. UDP therefore does not incur the delays due to
re-transmissions, making it suitable for real-time applications such as VOIP.
It is instructive at this stage to try to picture all enveloping of packets within
packets that occurs in an IP network. The example given in Fig. 8.13 is for a request
to see a web page using hyper-text transfer protocol (HTTP) – this is the actual data
packet coming from the application. This user data is then encapsulated in the TCP
packet, which in turn is deposited into the payload of the IP packet. This may not be
a one-to-one mapping since the TCP packet could be spread over several IP packets.
In this example we assume that the IP is carried over ATM, as described later in this
chapter. Because of the small fixed size of the ATM cells, IP packets are typically
chopped into several chunks of 53 bytes for transmission over an ATM virtual path.
This stream of ATM cells may then be deposited into a virtual container within an
196 Understanding telecommunications networks
HTTP Request
TCP HTTP Request
Application layer
HTTP Request
Transport layer
TCP HTTP Request
Network layer
Link layer
Physical layer
Figure 8.13
Message Transmission Using Layers
SDH frame for transmission over the core transmission network! At the distant end,
in our example the web page host machine, the reverse process takes place with each
of the packet/cells being extracted at each level, as shown in Fig. 8.13.
One way of trying to make sense of all the many protocols involved in IP networks
is the use of an architectural diagram, as shown in Fig. 8.14. This shows all the
appropriate protocols mapped against the seven layers of the OSI reference model.
At the physical layer are all the various transmission systems in the access and core
transmission networks. At layer 2, there is an array of data link systems, including
ATM, frame relay and Ethernet LAN. The PSTN is also shown as a possible link
mechanism because in the case of dial-up access to the Internet the switch path
through the PSTN acts as a simple data link between the caller’s computer and the
ISP. (It should be noted that the PSTN is, of course, operating at layer 3 while the call
is being set up.) IP resides at layer 3. At layer 4 the two transport protocols, TCP and
UDP, support the latency tolerant and latency intolerant applications, respectively, as
indicated in this architectural diagram of Fig. 8.14. There is a vast array of protocols
covering the many possible applications that may be run over an IP network – Internet
or intranet – or an IP-based computer system, only a selection of which are shown.
8.6 The Internet
Chapter 2 introduced the concept of the Internet, describing it as a constellation of
(IP) networks. There is a vast range of these IP networks, some of which form an
entire AS, others may be small parts of an AS. Unlike the public telecommunications
Data (packet) switching and routeing 197
L7: Application
L6: Presentation
L5: Session
L4: Transport
L3: Network
L2: Link
e.g: ATM, PSTN, frame relay, point-to-point protocol (PPP), ethernet
L1: Physical
Domain Name Server
Hyper Text Transfer Protocol
File Transfer Protocol
Internet Protocol
Transmission Control Protocol
Virtual Terminal
eXternal Data Representation
Figure 8.14
Post Office Protocol
Real-time Transport Protocol
Simple Mail Transfer Protocol
Simple Network Management Protocol
User Datagram Protocol
Network File System
Open Systems Interconnection
The Internet and OSI Protocol Reference Models
networks, which are designed to international standards recommended by the formalised International Telecommunications Union (ITU), the design of the Internet is
based on a consortium of users, academics and data equipment manufacturers known
as the IETF. The latter have set recommendations for numbering, routeing, and all the
various protocols at OSI reference model layers 2–7 – although the IETF considers
layers 5–7 as a single entity for convenience.
It should be appreciated that the Internet has gone through rapid development:
originally a US defence department project (ARPANET) conceived in the mid 1960s,
going international with the first non-US node in University College London in 1973,
with the addition of TCP in 1983 and becoming a widespread information-sharing
network, mainly used by universities and research institutes around the World. However, it was the development of the HTTP, which allowed information in many forms
to be easily shared, that led to the introduction of the now famous ‘World Wide Web’
(WWW) in 1991 [5]. Undoubtedly, it was the utility of the ‘Web’ in allowing access
to information – text, pictures, sound and moving image (video) – resident on distant
computer storage facilities that has made the Internet the phenomenon it is today.
Without the Web it seems improbable that the original academic and specialist Internet would have been adopted by the general populous as both a work and domestic
facility. In fact, it was the sheer size of the problem of being aware of and gaining
access to the enormous wealth of information stored on servers at libraries, business
and research organisations, government agencies, all manner of commercial enterprises, etc., that led to the introduction of search engines. These have not only eased
the problem of finding information but have also provided commercial opportunities
for companies providing an entry into the Internet/www, the so-called portals. Of
198 Understanding telecommunications networks
course, the biggest change to daily life has been caused by the wide-scale adoption
of electronic mail (e-mail), beginning in the mid 1990s, and now forming one of the
major applications run over the Internet.
At the turn of the century use of the Internet took another leap forward as several
mechanisms for ensuring security of credit card information were introduced. Customers were then able to make purchases over the Web: undertaking all the stages of
perusing the catalogues, placing an order, paying and tracking the progress of delivery
of the goods. This, together with other developments, such as electronic signatures
and public key encryption, has given rise to the derivative phenomenon of electronic
commerce, better known as ‘e-commerce’.
It must be emphasised that all the developments associated with IP, both technical
and commercial, have two distinct separate, but related, areas of application: the
public Internet and the private intranets. The same technology is used in both areas, but
the control of who uses the facility and the commercial implications are significantly
different. In the case of the Internet there is freedom of access to all the subscribers
of ISPs and the loading on the routers and links can vary enormously, giving rise
to variable and unpredictable performance (download times, etc.). However, the
biggest problem with the open-to-all policy of the Internet is its vulnerability to
malicious users sending e-mails with attachments that inject corrupting software into
the recipient computers, i.e. a virus attack. When directed at major ISPs this can cause
the complete locking-up of the routers causing a wide-spread failure in the Internet.
On the other hand, the intranets are self-contained IP networks, using all the IETF
standards for numbering, etc., in which all the users are members of a commercial or
government organisation located at their offices, factories, laboratories or academic
institutes. The performance of an intranet can be controlled by the organisation’s
communications department, since the network can be appropriately dimensioned for
the well-defined population of users. Organisations create a protected environment
for the intranet allowing entry or ‘ingress’ from the Internet only via a firewall – the
latter being a server that critically examines every IP packet to apply an admission
policy, rejecting those packets that do not conform.
As with all networks, the nodal and end points within the Internet (or intranet) need
to be identified by an address or number in order to route the IP packets accordingly.
The identification of either an e-mail subscriber or a web page is made using a humanfriendly name (e.g. ‘name@company.co.uk’). This name needs to be translated into a
destination address that can be placed into the header of the IP packets. As a simplified
example of this process we will consider the routeing of an e-mail across the Internet
between two corporate intranets, as shown in Fig. 8.15. The request to send an e-mail
is detected by the server in the local network 1 (an intranet) and a brief exchange of IP
packets takes place with the serving ISP, which requests the address of the e-mail. The
ISP router then needs to pass an enquiry IP packet to the nearest data base of names
to seek the destination address. This data base is known as a ‘domain name server’
(DNS) and there is a DNS protocol (shown in Fig. 8.14) used for this interrogation.
Once the originating computer has received a packet containing the Internet address
for the e-mail recipient it can then use that address in all the IP packets conveying
the e-mail contents. Each of these packets is then routed through the Internet from
Data (packet) switching and routeing 199
Local DNS
Local Net 1
Local Net 2
Figure 8.15
Example of Internet Routeing
ISP-1’s router to the destination ISP router, and hence the recipient’s computer on
local network 2. (The role of DNS and IP numbering and addressing is considered
more fully in Chapter 10.)
8.7 Voice-over-IP
The multi-service capability of packet networks has been exploited for the conveyance
of voice for many years. Some frame relay corporate networks were carrying internal
voice calls (VOFR) as early as the 1980s [6]. Such deployments were small scale
and specialised with the poor quality being acceptable for corporate private networks
because of the cost savings. More recently, voice traffic has been handled by ATM
switches (VOATM), achieving much higher levels of quality, but still applied to
special situations. Then in the early 2000s BT replaced many of their trunk exchanges
with a hybrid system using circuit switching and ATM to create a ‘next-generation
switch’ (NGS) [7]. However, we will now concentrate on VOIP because this has now
become the generally accepted successor to the circuit switching of voice.
There is actually a range of scenarios in which voice can be carried over an IP
network, as shown in Fig. 8.16. The earliest application was the use of a telephone
handset, earphones and microphone plugged into a computer connected via the Internet to a colleague using a similar arrangement – ‘Internet Telephony’, shown as A–B
in Fig. 8.16. This arrangement requires both parties to be on line to the Internet simultaneously and for an e-mail pathway to be established in order for a conversation to
200 Understanding telecommunications networks
An Intranet
Figure 8.16
The Internet
Voice-Over-IP Scenarios
occur. The quality is highly dependent on the instantaneous loading of the Internet
and can vary from appalling to adequate. However, recent developments have led to a
new generation of A–B VOIP applications (e.g. ‘Skype’) which have much improved
quality because of the use of broadband access, and are a widely popular form of
cheap (or even free) public telephony, as described later. A corresponding arrangement applies to the corporate network situation (‘intranet telephony’), in which the
computer-handset terminals are connected via the company intranet –shown as C–D
in Fig. 8.16. This application has proved to be increasingly acceptable because the
quality can be managed by appropriate dimensioning of the intranet, and there is
opportunity for the voice and data infrastructure within office blocks to be combined,
resulting in only one set of internal office wiring and the elimination of the (circuitswitched) PBX [8]. Calls can also pass between the users on the intranet and the
Internet, e.g. C–B in Fig. 8.16, with the limitation that the IP packets will have to be
vetted by the corporate firewall.
However, the biggest opportunity for VOIP is the substitution of some or all of
the PSTN by an IP alternative. This is achieved by the use of a gateway between the
PSTN and the Internet, known as a PIG (PSTN-to-Internet gateway). Referring to
Fig. 8.16, calls from subscriber E are switched in the PSTN as far as the PIG 1, where
the call then progresses over the Internet to the exit PIG 2 where it terminates in the
PSTN and subscriber F’s line. There are several issues to bring out here.
(i) The choice for use of the Internet for some of the call is made by subscriber E
dialling to an ISP and subscribing to a VOIP service.
(ii) The Internet routeing could bypass the long distance part of a PSTN call routeing, or even one or more international routes around the World if subscriber F
Data (packet) switching and routeing 201
Figure 8.17
[SS7 in IP]
e.g. SIP,
IP packets
IP Network
PSTN-to-IP Gateway (PIG)
is in a foreign country. Since no charge is made for the use of the Internet,
subscriber E can enjoy a considerable saving in call charges.
(iii) The quality of the overall call connection will be dictated by the Internet portion,
whose performance depends very much on its instantaneous loading.
(iv) The scenario could be extended to allow calls to be routed via the Internet to
an intranet, enabling E–C communication, for example.
An alternative scenario for E–F communication is for one or more of the exchanges in
the PSTN to be replaced by VOIP equipment, with an intranet rather than the Internet
used to provide the IP networking between PIGs. In this situation, unlike the above
examples, the subscribers may not be aware that their call is being handled fully
or partially by VOIP equipment, with the tariff and quality appearing as a standard
telephony offering.
Fig. 8.17 illustrates how a call is set up between two PIGs following this alternative
scenario. There are two links between the ‘egress’ (i.e. exit gateway) exchange and
associated PIG: a signalling link carrying SS7/ISUP (see Chapter 7) which terminates
on the SS7 signalling gateway; and the 30 channels of PCM speech which terminate
on the media gateway (MG). Now we can consider how a connectionless IP network is
manipulated, or as the Internet fraternity say ‘spoofed’, into establishing a telephone
call connection.
(i) The call set-up is initiated by an IAM on the SS7 signalling link to the signalling
gateway (GC) in PIG 1. This generates a request message from the SG to the
media gateway controller (MGC).
(ii) The MGC now has to examine the PSTN number of the called subscriber and
determine the IP address of the appropriate egress PIG to serve this number.
202 Understanding telecommunications networks
This would normally be done on the basis of the trunk or area code for a
national call, and the country code for an international call. A collated data
base would normally hold this information.
The MGC now passes the IP address of the destination PIG to the MG,
indicating the channel identity of the calling speech circuit.
The MGC sends a message in an IP packet to the egress IP’s MGC requesting
the call set and indicating the destination PSTN number.
The egress MGC also instructs the SG to send an SS7 IAM to the associated
PSTN exchange. A sequence of IAMs now progresses through the PSTN to set
up the call. Once the ‘address complete’ SS7 message is received (indicating
calling subscriber is receiving ringing tone) the SG advises its MGC of the
call status.
The egress MGC now sends an IP packet to the ingress MGC to advise that
the speech reverse path should now be established over the intranet.
The egress MG now extracts each PCM byte from the called channel (the
speech is already in digital format) and deposits them into the payload of successive IP packets, setting the ingress PIG address as the packets’ destinations.
A speech path now exists over the IP network (and the PSTNs) between called
and calling subscriber and the latter can hear the ringing tone generated by
the destination PSTN exchange.
Once the answer SS7 message is received by the egress SG the speech path
from the ingress MG to the egress MG can be established – each PCM byte
being deposited in the payload of subsequent IP packets addressed to the
egress MG.
Any SS7 messages that need to be passed between the PSTN without interrogation by the PIGs are passed in IP packets directly between the ingress and
egress SGs.
Fig. 8.17 also shows some of the protocols that are used between PIGs. The speech is
carried in IP packets associated with UDP at the layer 4 and real-time protocol (RTP)
at layer 5 to minimise transmission delay and provide as high a quality of perceived
performance as possible. Control messages between the MGs are sent using SIP or
H323 (see Chapter 7) or with the new MGC-to-MGC protocol. MGC to MG messages
may use Megaco (media gateway control) or MG control protocol (MGCP) [9].
Consideration of the elements of Fig. 8.17 will reveal that the switching of voice
calls over an intranet or the Internet– using MGs, SGs and controllers – is achieved
without any switch-blocks! In effect, the IP network is performing the switch-block
function. The essence of the call is really created by the software on the MGCs. For
this reason, MGCs are often referred to as ‘soft switches’.
8.8 VOIP over broadband
Probably the most profound development affecting the adoption of VOIP has been the
introduction of broadband access over ADSL for the residential and small business
customers. This has given a cheap facility for carrying data from homes and offices,
Data (packet) switching and routeing 203
which not only provides wide bandwidth (typically up to 2 Mbit/s downstream and
256 kbit/s upstream) but also an ‘always on’ capability. Since VOIP enables voice to
be treated as data over an IP network, the broadband access facility can also be used
for voice calls. This gives rise to so-called ‘voice-over-IP-over-broadband’. There are
several companies providing public telephony services using VOIP over broadband
technology, each using their own proprietary designs – so full inter-working between
the systems is difficult. Examples of VOIP companies include ‘Vonage’ and ‘Skype’.
Overall, the VOIP-over-broadband services perform well, often with exceptionally good sound quality, as a result of using high-speed codecs which convert the voice
to IP packets without needing the normal 4 kHz frequency filtering and the wider
bandwidths of broadband compared to the ‘narrow band’ (i.e. 4 kHz and 64 kbit/s)
PSTN. This high sound quality is, of course, subject to the loading on the Internet
which if high can lead to loss of packets and hence degradation to the sound quality.
There are also limitations to such services due to the fact that they are essentially data
rather than switched-voice based. For example, calls are possible over most of the
systems only when both subscribers have their computers on line (i.e. connected to
the Internet), so calls have to be on the basis of pre-booking a time or waiting until
the other party comes on line. Nonetheless, since the broadband access to the ISP and
hence the Internet usage is usually on a subscription and not usage basis, the voice
‘calls’ using VOIP over broadband are cheap or at no extra cost. For many people,
such services are used in preference to the PSTN-based telephony.
Fig. 8.18 illustrates the principle of VOIP-over-broadband telephony. First we
consider the case where both subscribers, A and B, have their computers connected
Data network
Data network
Figure 8.18
Network operator serving A
Network operator serving B
IP packet stream
Authentication, authorisation & accounting
VOIP Over Broadband Services
204 Understanding telecommunications networks
to a broadband Internet access services, using ADSL as described in Chapters 4
and 5. Thus, there is a data path established by the serving network operators from
subscribers A and B to their respective ISPs (described in the next section). In the
simplest form of service, the subscribers’ terminals comprise computers with telephone headsets. Voice communication is then essentially provided by a stream of
VOIP packets between the computers of subscribers A and B, flowing over the path
set up by the broadband connection at both ends and the connection to the Internet
provided by the two ISPs, as shown in Fig. 8.18. Two actions are required before this
flow of IP packets can be initiated: first, both subscribers must have logged on to the
VOIP service when they went ‘on line’ (i.e. switched on their computer and logged
on to the ISP) and second, the calling subscriber needs to be made aware that the
called subscriber is currently on line. The latter may be achieved using the ‘presence’
indications available from instant messaging facility, i.e. the fact that subscriber B
is on line is indicated to all subscribers of the VOIP service, or at least those that
have subscriber B in their personal VOIP-service address book. Setting up the call
between A and B is then a matter of setting the appropriate IP addresses in the IP
packet headers and a simple interchange of ‘ring’ and ‘answer’ message packets. This
direct data communication between A and B is often referred to as ‘peer-to-peer’
working, since there is no intermediate server involved in setting up the path. (Note:
not all VOIP-over-broadband services work on a peer-to-peer basis.)
The initial registration of subscribers to the service is managed by an authentication, authorisation and accounting (AAA) system managed by the VOIP service
provider. This system monitors the status of each subscriber (i.e. whether they are
currently online and logged on), as well as providing accounting information for
billing (if appropriate). The presence mechanism may be managed centrally by the
AAA, or on a peer-to-peer basis. Fig. 8.18 shows the IP packet paths for subscribers
A and B logging on to the AAA system.
In the case where a call is to be made to a subscriber on another network, e.g.
the circuit-switched PSTNs or mobile networks, the IP packet stream needs to be
routed to a gateway (i.e. a PIG), as shown in Fig. 8.18. Such ‘break out’ calls are
usually charged for by the VOIP service provider to compensate for the interconnect
charges they incur from the PSTN, etc., operators. Similarly, calls from the PSTN or
other networks are routed to the VOIP service provider’s gateway so that an IP stream
can be established to the called subscriber, using the principles described earlier for
the PIG.
Network aspects: IP over ATM
Public networks need to provide connectivity for a wide range of services – switched
voice and data, and un-switched leased lines. In practice, this means that capacity
within the access and core transmission networks has to be partitioned so each service
gets the bandwidth and throughput that it needs. Network operators therefore use a
combination of ATM and IP, as well as SDH (and perhaps PDH), to carry the range
of data services and, increasingly, voice services as well.
Data (packet) switching and routeing 205
Point of
140 Mbit/s (VC4)
ADSL splitter
Figure 8.19
IP Over ATM Using ADSL Access
In the access network, ADSL is used over a single copper pair to provide a bothway broadband data circuit from subscribers to the local exchange, as described in
Chapter 5 and shown in Fig. 5.4. The network configuration to support ADSL lines
is shown in Fig. 8.19. Network economies are achieved by introducing contention
at the aggregation point, as provided by the DSLAM. Typical contention ratios are
20 : 1, 25 : 1 and 50 : 1 (compared with the 10 : 1 traffic concentration ratio at the
subscriber concentrator switch block in a DLE). This contention is provided by an
ATM switch within the DSLAM, in which the maximum cell arrival rate at the inputs
from all the subscriber lines, is proportionately greater than the cell capacity on the
output highway. Cells from the subscribers are queued, and if necessary discarded,
in order to match the output cell rate. In practice, the subscriber lines are relatively
lightly loaded with only a small proportion active at any one time so the contention at
the DSLAM, causing a drop in throughput, is not normally perceived by subscribers.
The ATM path is established from the ADSL termination on the subscriber’s premises
where ATM adaptation is performed through to the DSLAM input. The ATM cells
are then switched on to VPs from the DSLAM carried over SDH transmission links
to the first ATM switch. Here the network operator is able to intercept the cell flow
to monitor and authorise the subscriber’s use of the broadband service. Cells are then
switched onto VPs directly to one of the many ISPs or via other ATM switches to the
serving ISP. The IP packet flow is, thus, between the computer on the subscriber’s
premises – over the ATM paths provided by the ADSL, through the DSLAM, the ATM
switches – and the ISP. The ATM VPs marshal the flow of IP packets and establish
manageable permanent paths from the many hundreds of ADSL aggregation points
to the, typically, several hundreds of ISPs serving a country.
206 Understanding telecommunications networks
MSP Core
IP trunks
Leased ATM
High capacity Optical fibre
ATM access
Figure 8.20
over ADSL
MSP Edge
ATM trunks
Multi-Service Platform (MSP)
IP over ATM is also used as the core infrastructure of a so called multi-service
platform (MSP), as implemented by many operators to provide their range of different
data services economically. The concept is shown in Fig. 8.20. The objective of an
MSP is to use the multi-service capability of an IP-over-ATM network as the core
supporting all the services, with equipment at the edges of the network to provide the
individual services to the subscribers – presenting the appropriate protocols, interfaces
and data rates. In this way, the aggregated traffic from the many access lines is carried
over the MSP core. Any temporary or long-term variation in demand on an individual
service (above or below forecast) is absorbed by the aggregated core capacity. A
further advantage for a network operator is that the MSP should also be able to
provide complete cross-country connectivity for any new low-volume data service at
marginal cost.
Multi-protocol label switching
Interestingly, the merits of ATM in providing a connection-orientated layer 2 to support the connectionless IP packets at layer 3 have been recognised by the Internet
fraternity (IETF, etc.) in so far as they have invented a more IP-friendly substitute
called multi-protocol label switching (MPLS). Essentially, MPLS uses labels to establish a VP for the IP packets to follow. Unlike ATM, however, MPLS does not chop
up the IP packets into 48-byte chunks (hence the epithet of ‘Data shredder from hell’
used by IP advocates to describe ATM), rather it envelops the complete IP packet.
The MPLS header, which is attached to the front of each IP packet, comprises the
Data (packet) switching and routeing 207
Label Address
Figure 8.21
Label Address
Example of MPLS Forwarding of Packets
20-bit label, a 3-bit COS indicator, a 1-bit bottom-of-stack indicator and an 8-bit
time-to-live indicator (a total of 4 bytes).
An example of the use of the MPLS forwarding of IP packets is given in Fig. 8.21.
The IP routers enhanced with the MPLS capability form domains within an IP network.
The routers at the edge of the domains are known as label edge routers (LER), while
the routers within the domain are called label switch routers (LSR). The required VPs,
known as ‘switched label paths’(SLP), are established using a set-up procedure similar
to that used with ATM. Fig. 8.21 shows the label switch routeing tables associated
with three of the MPLS-enabled routers. There are two SLPs associated with the
LER-1: packets destined for IP address are switched to outgoing port 1
and assigned a label of 20; IP packets destined for are also switched to
outgoing port 1 but assigned a label of 10. All IP packets have the appropriate MPLS
header (with label) added. At the second router, which is an LSR, all MPLS encased
packets with label 20 are given a new MPLS header with a label 4 and sent on output
port 0. Similarly, packets with label 10 are re-labelled 16 and sent on output port 1.
In this way the various streams of IP packets are quickly routed through the network
following the pre-determined VP (SLP).
MPLS offers the network operator many advantages over simple IP also called
‘native IP’, namely:
(i) Quality of service (QOS)
• Segregation of QOS classes can be applied by the routers in the network by
reference to the service classification in the MPLS headers.
208 Understanding telecommunications networks
Company A
Company B
Company C
Company B
Company A
Company A
Company C
Figure 8.22
MPLS-switched label paths
Company B
• The SLP set-up is made with reference to the required QOS.
• Prioritisation is applied at each MPLS router based on class-of-service
• Selective resilience can be applied in an IP network, through identifying
the COS of each packet.
(ii) Virtual private networks (VPN)
• Isolation of closed-user groups within a public IP network is provided by
the SLPs, so enabling the establishment of a series of VPNs to be maintained. Fig. 8.22 shows an MPLS-enhanced IP network supporting VPNs
for companies A, B and C. These are usually referred to as ‘IPVPNs’.
• Accountability is possible, since the IP packets relating to individual VPNs
are segregated and easily monitored.
(iii) Faster routeing
• The use of simple look-up tables based on the reading of labels in the MPLS
header is a faster process than having to open up the IP packet and examine
the full IP address. This is because label switching can be performed using
hardware, whereas normal IP routeing involves the slower processing of
software to deconstruct and decode the IP address. Indeed, it was through
the development of higher-speed IP routers that the concept of MPLS
Data (packet) switching and routeing 209
(iv) Traffic engineering
• The use of explicit routes, i.e. SLP, for the conveyance of IP packets means
that the capacity of the network can be dimensioned and used predictably.
• Resources (i.e. capacity in the routers) can be reserved using the SLP set-up
• Rules, such as constraint-based routeing, can be applied to the establishment
of SLPs.
• Overall, traffic engineering similar to that available with circuit-switched
networks is possible.
Network operators offer MPLS as a premium service to data customers, who in turn
use it to create quality managed corporate networks (intranets, LANS, etc.). MPLS
is also used by network operators within the public network to provide IPVPNs for
corporate customers (as described earlier). Finally, the quality assurance of MPLS
makes it an essential component for VOIP networks. As Chapter 11 discusses, IP with
MPLS forms the basis of NGNs, whose VOIP capabilities enable them to replace the
existing PSTNs.
Local area networks
Rather than considering the external or public network aspects of data transport, we
now address the way that data is carried between the terminals on people’s desks and
the computers, printers, servers, etc., in the workplace (e.g. office, campus, research
centre and factory). The term ‘local area network’, LAN, is used to describe the data
networks which provide this facility. It is useful to consider briefly how LANs work
because this helps to understand how customers generate the need for data services
from network operators. Also, the Ethernet LAN specification is used as standard
interface between user’s data terminals and a network operator’s data service.
The principle of a LAN is that it uses a common highway or ‘bus’ (short for
omnibus – meaning ‘for all’) to link all the devices. As with any network system,
LANs need an addressing scheme to identify each termination on the bus, and a
mechanism to allow the terminations to use the transmission capabilities of the bus
one at a time. The LAN may be physically constructed using twisted pair cable, coaxial
cable or optical fibre cable and configured as a single bus or as a set of buses in a tree,
ring or star topology. In practice, clusters of terminals are connected in star formation
onto a hub device where they terminate on a common bus, as shown in Fig. 8.23. In
the case of many terminals additional hubs may be added and interconnected via a
LAN switch.
Several standard types of LAN have been developed and deployed since the first
was introduced in 1976, namely, Ethernet, Token Ring and Token Bus [10]. By far
the most popular and successful standard system is that of the Ethernet family (IEEE
802.3, see Box 8.5), the focus of the rest of this section. The distinguishing feature of
an Ethernet system is that the mechanism for managing the joint use of the bus is based
on collision control. This relies on each termination wishing to communicate sending
a burst of data and its destination address within an Ethernet ‘frame’ onto the LAN.
210 Understanding telecommunications networks
2 Mbit/s
2 Mbit/s
10 Mbit/s
2 Mbit/s
10 Mbit/s
10 Mbit/s
25 Mbit/s
2 Mbit/s
Figure 8.23
Box 8.5
10 Mbit/s
LAN Concept
IEEE 802 Range of Standards
The US-based Institution of Electrical and Electronic Engineers (IEEE) is the
largest professional group in the World. Amongst its many activities are those
of standards group 802, addressing a range of LANs and associated systems.
Many people are now aware of particular IEEE 802 standards, e.g. ‘802.3’ and
‘802.11’, but there is a long list of outputs from the group, as shown below.
802.1 Overview and architecture of LANs
802.2 Logical link control
802.3 Ethernet LAN
802.4 Token Bus LAN
802.5 Token Ring LAN
802.6 DQDB (Dual queue dual bus, an early form of MAN)
802.7 Technical advisory group on broadband technologies
802.8 Technical advisory group on optical fibre technologies
802.9 Isochronous LANs
802.10 Virtual LANs and security
802.11 Wireless LANs
802.12 Demand priority (proprietry LAN)
802.13 Spare
802.14 Cable modems (now defunct)
802.15 Personal area networks, e.g. Bluetooth
802.16 Broadband wireless (wireless MAN)
802.17 Resilient packet ring
Data (packet) switching and routeing 211
0–1,500 bytes
8 bytes
Start of frame
Figure 8.24
0–46 bytes
6 bytes 6 bytes 2 bytes
4 bytes
Data being
Padding bits
Length of frame
Ethernet LAN Frame Format
If any other terminals simultaneously send data the resulting ‘noise’ (mutilated signals) caused by the collision of two sets of data frames is detected by both terminations,
which stop immediately. These terminations each wait for a short random period and
then retry – normally, one of the terminals is first and the other waits for the bus to
clear. This mechanism, known as ‘carrier-sense multiple-access/collision detection
(CSMA/CD)’, is analogous to a conference telephone call without a chairman, when
two or more participants inadvertently speak at the same time, each stops and tries
again later. Actually, CSMA/CD is one of several protocols – known as MAC, part of
OSI layer 2 – used in data transport systems to control the access of several terminals
to a single resource.
Fig. 8.24 shows the format of the Ethernet frame, mentioned earlier, which is really
a form of data packet. An 8-byte binary number acts as the so-called preamble to allow
the recipient terminals to lock into the actual speed of transmission of the received
frame. The frame proper starts with the 6-byte destination and source addresses and
these are followed by the field for the actual data, which has a maximum capacity of
1,500 bytes. There is also a small field to indicate the length or type of the frame and
a 46-byte padding field used to ensure that the minimum-allowable 64-byte length of
the frame is achieved. Errors in the received frame are detected through examining
the 4-byte checksum. More detail on Ethernet LANs is available in the Reference
section [10–14].
There are currently three speeds of Ethernet LAN: 10 Mbit/s, 100 Mbit/s (‘Classical Ethernet’) and 1 Gbit/s (‘Gibabit Ethernet’). The LAN provides a managed data
path between terminals and servers, allowing terminal to terminal, terminal to server
and server to server communication – effectively at layer 2 of the OSI reference model.
IP packets then flow over the LAN path at layer 3. The Ethernet LANs usually have
a gateway giving access to links to other LANs or to one of the public data networks.
The transmission capabilities of 1 Gbit/s over optical fibre cable make it a candidate
for the Access Network, linking a LAN site, say at an office, to the first network node,
thus extending Ethernet beyond the customer’s premises. New generations of PON
use the Ethernet interfaces [15].
A corporate network comprising several interconnected LANs at various sites
within an area of some 3–30 miles is known as a ‘metropolitan-area network’, MAN.
212 Understanding telecommunications networks
The interconnection between LANs, often called the ‘sub network’ may be realised
using leased lines or a public data service, e.g. ATM, frame relay or SMDS, which is
provided by the network operator. A further concept is that of the ‘wide-area network’,
WAN, which links widely separated LANs and MANs (typically over distances greater
than 30 miles) [10,16,17].
8.12 Wireless LANs
In addition to the wired form of LANs, there is a family of wireless-based systems,
known as Wireless LANs, ‘WLANs’, ‘Wireless Ethernet’ or colloquially ‘WiFi’.
These LANs serve a group of terminals, typically laptops or PCs, each with a network interface card (NIC) transmitting over a wireless link to a centrally located hub,
the ‘Access point’ (AP) as shown in Fig. 8.25. A special media-access control (MAC)
protocol is used to cope with the special requirements of wireless links, namely:
‘carrier-sense multiple-access/collision avoidance’ (CSMA CA) which operates similarly to the LAN CSMA CD except that the terminals carefully listen for existing
transmissions before transmitting to avoid collisions. Since all the terminals share
the wireless capacity on an AP, larger capacities are achieved through having two or
three APs, each operating on a different wireless carrier frequency. The AP are linked
by cable to an Ethernet switch. The NICs are able to determine the appropriate AP to
join by comparing the signal strengths of the various carrier frequencies within the
WLAN. A collocated server is connected by cable to the APs. All communication is
between the NICs and the AP, since the terminals are not able to link directly.
WLANs were originally designed for use within the office environment, giving
users a degree of freedom to connect to the office computing facilities without using
Figure 8.25
Wireless Ethernet LAN (‘WiFi’)
To exit
Data (packet) switching and routeing 213
cables. However, the progressive reduction in the costs of the equipment and the fact
that the system operates in uncontrolled wireless spectrum (i.e. no fees or licence)
has opened up the WLAN applications far beyond the office environment. A widely
growing application is the use of WLANs in so-called hot spots, e.g. hotels, railway
stations, airport departure lounges and coffee shops around town. Here, people can
communicate to their corporate LAN or the Internet from their laptop computers, if
equipped with an appropriate NIC, within the catchment area of an AP. This provides
what is, in effect, a form of nomadic communication mobility, whereby someone on
the move can communicate within the vicinity of any hot spot, but not when moving
between them. In addition, WLANs are now being installed in homes, linking PCs,
laptops, printers, routers, etc., meeting personal and home-working needs.
There are several versions of WLAN, all in the IEEE 802 range (see Box 8.5).
Systems adhering to 802.11b were the first to be deployed, operating with wireless
carriers at 2.4 GHz, offering gross data rates of speeds up to 11 Mbit/s (giving actual
user rates of around 5 Mbit/s) at a range of up to 150 m. The more recent 802.11a
operates at higher carrier frequencies (at 5 GHz) and consequently shorter ranges of
50 m, but offers gross data rates up to 54 Mbit/s. There is also an 802.11g standard
which operates in the frequency range of 802.11b, but with higher user rates.
The data frame structure for the 802.11 range of WLANs is shown in Fig. 8.26.
As will be noticed, the 802.11 frame is similar to non-wireless LANs except for
the additional ‘Address 3’ and ‘4’ fields which enable NICs to communicate with
terminals or hosts on other wireless cells, and the use of duration (transmission time
of the frame burst) rather than length of frame to indicate the variable amount of data
being carried in the frame [11,18].
0–2,312 bytes
7 or 16 bytes
6 bytes
Signal rate
30 bytes
Source and
4 bytes
Data being
Check sum
Indicates the speed of transmission
(1, 2, 5.5, or 11 Mbit/s)
and the length of the frame
Includes provision
for two extra addresses
Figure 8.26
A Simplified Illustration of the Frame Format of IEEE 8-2.11 Wireless
214 Understanding telecommunications networks
To complete the picture of wireless data transmission systems, we should note the
emergence of the broadband wireless link defined in 802.16, also known as ‘WiMax’
[19]. This is really a high-capacity fixed link between a stationary user and a stationary
server; many operators consider it an alternative to the provision of ADSL over copper
cable to provide broadband to residential and small business users (see Chapter 4). We
also note the short distance (up to 30 m) wireless system, known as ‘Bluetooth’, which
is defined in IEEE 802.15 standard. Bluetooth provides a convenient replacement
for wires between equipment around the office or home, e.g. the cordless headset.
It is also widely used in mobile phones and personal digital assistants (PDA) for
device-to-device communications within buildings.
In this chapter we looked at the variety of packet technologies used to support the wide
range of data services used by both business and residential customers. After noting
the differences between voice and data services, we then discovered that actually
several of the packet systems are equally capable of providing voice switching. We
looked at ATM in detail, noting that the virtual paths established enabled the QOS
offered to customers to be predicted and assured. IP was then examined to see why it
has proved to be ubiquitous and so popular as a network service and interface to user
applications run on computers and terminals. Consideration of the Internet provided
an insight into the applicability of IP and, with the advent of the World Wide Web,
the massive impact on every day life at work and home.
An important next step was the examination of how IP can be used to switch voice
and hence become the contender for replacing the circuit-switched PSTN. Importantly, we considered the joint working of ATM and other layer 2 technologies with
IP. In particular, we looked at the new service offerings and network traffic engineering made possible by MPLS. The latter, with VOIP, forms the basis of the so-called
NGNs, as discussed in Chapter 11.
Finally, we considered the role of LANs, particularly Ethernets, in serving the
communication between terminals and servers within an office environment and generating the need for public data services as provided by network operators. We noted
that a recent extension of the LAN concept are the wireless LAN systems.
DAVIDSON, R.: ‘IPv6: An Opinion on Status, Impact and Strategic Response’,
The Journal of the Communications Network, Vol. 1, Part 1, April–June 2002,
pp. 64–68.
2 CHAO, H. J., LAM, C. H. and OKI, E.: ‘Broadband Packet Switching Technologies – A Practical Guide to ATM Switches and IP routers’, John Wiley &
Sons, New York, 2001, Chapter 13.
3 TATENBAUM, A. S.: ‘Computer Networks’, Fourth edition, Prentice Hall,
Amsterdam, the Netherlands, 2003, Chapter 5.
Data (packet) switching and routeing 215
LOSHIN, P.: ‘TCP/IP Clearly Explained’ Second edition, Academic Press, A P
Professional, San Diego, CA, 1997, Chapter 20.
TATENBAUM, A. S.: ‘Computer Networks’, Fourth edition, Prentice Hall,
Amsterdam, the Netherlands, 2003, Chapter 7.
GOLENIEWSKI, L.: ‘Telecommunications Essentials’, Addison-Wesley,
Boston, MA, 2003, Chapter 7.
NEWMAN, R., ROBINSON, G. and TURNER, N.: ‘Evolution of the BT
PSTN Trunk Network using Next-Generation Switches’, The Journal of the
Communications Network, Vol. 1, Part 1, April–June 2002, pp. 69–74.
CATPOLE, A. B. and MIDDLETON, C. J.: ‘IP Telephony Solutions for the Customer Premises’. Chapter 4 of ‘Voice over IP: Systems and Solutions’, edited by
SWALE, R., BT Exact Telecommunications Technology Series 3, IET, Stevenage,
ROSEN, B.: ‘VOIP Gateways and the MegaCo Architecture’. Chapter 7 of
‘Voice over IP: Systems and Solutions’, edited by SWALE, R. BT Exact
Telecommunications Technology Series 3, IET, Stevenage, 2001.
TATENBAUM, A. S.: ‘Computer Networks’, Fourth edition, Prentice Hall,
Amsterdam, the Neherlands, 2003, Chapter 1.
Ibid., Chapter 4.
ROSS, J.: ‘Telecommunication Technologies: Voice, Data & Fiber-Optic
Applications’, Prompt Publications, Indianapolis, IN, 2001, Chapter 4.
GOLENIEWSKI, L.: ‘Telecommunications Essentials’, Addison-Wesley,
Boston, MA, 2003, Chapter 8.
DENNIS, A.: ‘Networking in the Internet Age’, John Wiley & Sons, New York,
2002, Chapter 4.
JAMES, K. and FISHER, S.: ‘Developments in Optical Access Networks’,
Chapter 9 of ‘Local Access Network Technologies’, edited by FRANCE, P.,
IET Telecommunications Series No. 47, Stevenage, 2004
DENNIS, A.: ‘Networking in the Internet Age’, John Wiley & Sons, New York,
2002, Chapter 1.
ROSS, J.: ‘Telecommunication Technologies: Voice, Data & Fiber-Optic
Applications’, Prompt Publications, Indianapolis, IN, 2001, Chapter 1.
DENNIS, A.: ‘Networking in the Internet Age’, John Wiley & Sons, New York,
2002, Chapter 8.
PATACHIA-SULTANOIU, C.: ‘Deploying Wimax Certified Broadband Wireless Access Systems’, The Journal of The Communications Network, Vol. 3, Part
3, July–September 2004, pp. 105–116.
SHEPARD, S.: ‘Telecom Crash Course’, McGraw-Hill, New York, 2002,
Chapter 6.
DENNIS, A.: ‘Networking in the Internet Age’, John Wiley & Sons, New York,
2002, Chapter 6.
Chapter 9
Mobile switching systems and networks
Mobile phones are now such an everyday feature of life throughout the World that in
total they outnumber the population of fixed telephone lines. For many, the mobile
phone is the preferred means of communication. This chapter explains how a mobile
phone system works and the relationship between mobile and fixed networks. We
begin by considering the nature of a mobile system. After a simple review of how a
telephone call can be carried over a radio link, we examine the various cellular mobile
network systems, and conclude by looking at how the mobile and fixed networks in
the future may begin to merge.
Characteristics of mobile networks
Most people would characterise a mobile phone as something that uniquely relies
on the use of radio. Indeed, mobile phones do use radio – but, not all systems using
radio are actually mobile. An example of non-mobile usage of radio is the point-topoint radio links between exchanges and users’ premises, as described in Chapter 5.
There are, in fact, several characteristics in addition to the use of radio that together
distinguish mobile from fixed telecommunications systems. The key characteristics
are described below.
9.2.1 Tetherless link
A so-called tetherless link is used between the mobile handset and the exchange
to enable the required freedom of movement. Whilst in theory there are several
technologies that could provide a tetherless link, including infra-red, laser light, etc.,
invariably a radio link is used. In effect, this radio link performs the role of the local
loop in a fixed network.
218 Understanding telecommunications networks
9.2.2 Need for handset identification
Taking a fanciful view, one could imagine that if a technician standing at the termination of the local loop cable at the entry to the fixed exchange (i.e. at the MDF) took
hold of a copper pair and gave it a great jerk, at some distance away the telephone on
the subscriber’s desk would move! Whilst clearly not true, this story does illustrate
the point that the identity of any telephone on a fixed-line network is determinable
from its point of entry to the exchange. By contrast, mobile networks do not have
this fixed physical relationship to enable identification of the (telephone) handset. All
mobile networks therefore need a mechanism to identify its subscribers’ handsets.
9.2.3 Need to track the location of users
Users of mobile phones may move extensively across the domain of the serving
network, or indeed move into the catchment area of a foreign mobile network – the
latter action known as ‘roaming’. A key attribute of a mobile network is therefore a
mechanism to determine the instantaneous location of its users’ handsets, and to track
this location during a call. This is known as ‘mobility management’.
9.2.4 Need for a complex handset
The handset for a mobile network not only acts as a telephone terminal, but it also
provides the unique mobile functions, including those that are undertaken in the line
card of a fixed telephone exchange, as well as the transmission and reception of
radio, power management and the provision of handset identification. Consequently,
the mobile handset is significantly more-complex and expensive than a fixed-line
telephone. It is interesting to note that the ubiquitous GSM handset contains processing power equivalent to a Pentium microprocessor! Yet this complex device is now
viewed by many users as a fashion accessory – and the success of mobile service
offerings are highly dependent on the attractiveness, i.e. the look and feel, of the
handset. Indeed, many handsets now contain a variety of other functions, such as
cameras, MP3 players and PDAs. [For convenience, we will use the term ‘handset’ to
mean a mobile terminal (MT) whether it is hand-held or incorporated in some other
device, such as a hands-free unit in a car or a plug-in card to a computer laptop. There
are several terms used for a handset within the industry, including ‘mobile station’
(MS) and ‘user equipment’ (UE).]
9.2.5 Use of complex commercial model
Due to both the nature of their services and the more competitive commercial
and regulatory environments pertaining at the time when they were introduced,
mobile networks tend to operate within a more-complex commercial model than
the longer-established fixed networks. This means that there are usually several different companies – or ‘players’ – involved in the provision of mobile services, each
undertaking one of the following roles:
Mobile service providers (MSPs), who sell to the user, have the billing relationship
and provide the overall service management;
Mobile switching systems and networks 219
Mobile network operators (MNOs), who build and operate the mobile network;
Mobile virtual network operators (MVNOs), who appear to the users to be a full
MNO and service provider with a distinctive brand – although they actually do
not own any network, rather leasing capacity from an MNO instead.
The resulting commercial relationships (who serves whom and how the money flows)
can become complex, particularly when there may be several MNVOs and MSPs,
all dependent on a single MNO. Of course, when considering data and web access
services via a mobile handset this complexity increases greatly with the addition of
ISPs and information providers.
9.2.6 Need for specialised service support
The more-complex nature of the services available from a mobile network, e.g. roaming abroad, data calls and the separation of the service provider role from the network
operator role – requires specialised service support to handle customers’ enquiries,
ordering, fault handling and billing.
9.2.7 A simple generic model of a mobile system
We can now consider Fig. 9.1, a simple picture of a generic mobile network based
on the characteristics described earlier. The heart of the mobile system, indeed some
would say the crown jewels, is the set of radio frequencies allocated exclusively to the
operator: the radio system. A mechanism is required to control the access of the many
handsets to the constrained set of frequencies in the radio system, a process known as
multiple access, as described later. (In fact, this control of access to the radio resource
Service support
Figure 9.1
A Simplified Generic Mobile System
220 Understanding telecommunications networks
is an example of the generic function of ‘medium-access control’ (MAC), used in
networks whenever a transmission medium is shared across many users.) Fig. 9.1
shows the functions of mobility management, and identification (and authentication)
at the exchange, together with the switching and routeing to other mobile or fixed
networks. Finally, as described earlier, our generic picture of a mobile system needs to
show the service support function (including network operations and maintenance).
We will look at all these functions later in this chapter. But first, a simple consideration
of how radio works is appropriate.
How does radio work?
The concept of the use of radio to provide transmission paths for telecommunications is introduced simply in Chapter 4 (Figs 4.6 and 4.7). Radio paths of vibrating
electromagnetic energy (waves) are set up between transmitting and receiving antennae. Different forms of radio propagation are possible depending on the frequency of
the electromagnetic vibrations, which can range from about 3,000 Hz (pronounced
Hertz, the unit for vibrations per second) to around 300,000,000,000 Hz (300 GHz,
pronounced ‘gigahertz’) – see Chapter 4 and Box 4.2. For the purposes of considering
mobile systems we will concentrate on radio frequencies in the microwave range, i.e.
300 MHz (300 million Hz, pronounced ‘megahertz’) to some 3 GHz, in which the
radio waves follow a LOS propagation method coupled with reflection and refraction.
Radio systems use an electromagnetic signal in the shape of a sine wave, as shown
in Fig. 1.11 of Chapter 1. The most efficient form of antenna of such sine waves is
a metallic pole, known as a ‘dipole’, whose length is half the wavelength of the
sine wave. In the case of a mobile network using frequencies at, say, 2 GHz, the
wavelength (λ) is given by dividing the speed of the radio wave (which is the same as
that of light), 300,000,000 m/s, by the frequency (2,000,000,000 Hz) equalling 15 cm.
So the efficient antenna size is λ/2, i.e. 7.5 cm long. However, for convenience and
compactness the less-efficient but smaller antenna size of λ/4 (i.e. 3.75 cm), or the
even less-efficient λ/8 (i.e. 1.88 cm), is used in mobile phone handsets and car
mounted antennae, as appropriate.
Fig. 9.2 illustrates the principle of radio propagation [1]. The electrical signal is
passed through the dipole antenna causing an electromagnetic field to exist around
the length of the dipole. It radiates the radio wave strongly equally in all directions at
right angles to its length, but there is no radiation at the two ends of the dipole. The
radio wave travels through the air just like a light wave – in fact, of course, light is
also an electromagnetic wave, but vibrating several thousand times per second faster
than the radio wave. A similar dipole placed some distance away will be ‘excited’ by
the propagated radio wave causing a corresponding but much attenuated electrical
signal to be generated across the dipole and passed to the receiving equipment, where
it is amplified or regenerated.
In the ideal conditions for radio propagation, i.e. free space, without any air or
physical matter in the way, the radio signal will experience attenuation (i.e. a reduction
in magnitude) at the rate of the square of the distance travelled – known as ‘free space
Mobile switching systems and networks 221
Figure 9.2
Radio Propagation
loss’. This loss is due to the transmitted energy having to be spread across an everincreasing size of wave front as it radiates out from the antenna (similar to the surface
of an expanding spherical balloon).
However, in practice mobile networks are operating in cities, towns and countryside where there are a range of buildings, metal bridges, railways, vehicles, hills,
trees, rivers, etc., all of which cause additional loss to radio transmissions. The various
causes of the additional losses to free space loss are as follows:
Blocking by large buildings, walls, etc. The loss becomes progressively worse at
higher radio frequencies.
Reflection by obstacles such as buildings or large metallic building cranes, the
strength of the reflected signal depends on the degree of absorption by the obstacle.
Refraction and scattering by roofs, trees and mountain tops, etc.
Multi-Path effects caused by the radio wave reaching the destination via many
paths, each experiencing different degrees of reflection and absorption and different path lengths, as shown in Fig. 9.3. The net effect is that multiple versions
of the transmitted signal are received staggered in time and at differing strengths.
If severe this can cause the received radio transmission to be too mutilated to be
reliably recovered.
Rain absorption and scattering causing attenuation that varies with the weather –
the heavy density of water droplets in a fog can have a pronounced effect on the
radio signal strength. The effect worsens with increasing frequency of the radio
222 Understanding telecommunications networks
Figure 9.3
Multi-path Propagation
Other propagation effects experienced by radio waves in a mobile network include:
Doppler shift in the frequency of the radio signal due to the receiver moving relative to the transmitter, the effect worsening as the speed increases. (A commonly
experienced example with sound waves is when the pitch of a train’s whistle
changes as it rushes towards then away from someone standing at a station.) The
distortion to the received frequency caused by the Doppler shift is clearly an
issue for mobile users in a moving car or train; consequently, the faster a mobile
user travels relative to the transmitter the poorer the quality of the sound (or data
throughput) they receive, unless the effect is mitigated by corrective processing
of the received signal.
Electrical noise caused by interference from other radio sources, such as other
mobile phones, public radio communication systems, discharges from faulty or
unsuppressed electrical systems, etc.
Designers of communication systems using radio therefore have to cope with the
loss in signal strength (attenuation) due to free space loss, absorption, scattering,
etc., and the spreading and distortion to the signal shape due to multi-paths and
Doppler shifting. This diminished and distorted signal then has to be extracted from a
background of radio noise also picked up by the receiving antenna. The development
of mobile telecommunications systems has been characterised by a succession of
improvements to the way that the radio paths give ever greater usable capacities by
ingenious ways of injecting and then recovering voice and data waveforms, despite
the attenuation and distortion to the radio carrier.
So far we have just considered the propagation of a single sine wave along the
radio path. However, the communication channel needs to be superimposed onto
this basic sine wave, the latter then becomes the ‘carrier’ wave. That is to say,
Mobile switching systems and networks 223
the communications channel waveform ‘modulates’ the carrier wave, also shown
in Fig. 9.2 and as described in Chapter 4. Of the many modulation techniques available, digital mobile phone systems such as GSM mainly use a form of frequency
shift keying (FSK), in which the digital signal to be transmitted switches the carrier
between two frequencies [1]. The more recent so-called third generation mobile systems use the alternative technique of PM, in which the encoded (spread spectrum)
digital signal causes the carrier’s waveform to jump ahead of its sinusoidal sequence in
discrete steps. (The now-obsolescent analogue mobile systems use FM, in which the
frequency of the carrier is varied in sympathy with the analogue speech waveform.)
As described above, the radio-carrier frequencies available to mobile operators
are a valuable and strictly limited resource, which are allocated by national governments either on a quota basis or through auctioning. The range of frequencies
made available to the mobile operators must comply with the overall management of
the radio spectrum in that country, which in turn conforms to internationally agreed
allocations administered by Wireless Regulation Congress, WRC, (part of the ITU).
Mobile networks are allocated bands of radio frequencies variously at around 450,
900, 1,800, 1,900 and 2,000 MHz (i.e. 2 GHz). The exact frequencies allocated vary
between countries according to local circumstances and depending on the needs of
other radio users, such as cordless telephone systems (e.g. DECT system). In addition,
radio users such as Wireless LANs use unregulated parts of the spectrum, in which
there are no specific allocations by governments.
A large part of the operational management of a mobile network is the allocation
of the set of frequencies owned by an operator to the various parts of its network –
a function known as ‘radio planning’. This needs to take account of the natural terrain
of the country or area served by the network, sources of interference, etc., as well
as the expected distribution of demand for calls and data services. There is also a
need for all the mobile operators in a country to co-ordinate their radio planning
activities to ensure that there is minimal radio interference between their respective
networks. (This is necessary, even though all operators are allocated separate ranges
of frequencies, because all radio systems are capable of generating diminished but
detectable radiation in adjacent frequency bands as a result of faulty equipment,
abnormal weather conditions, etc.) We will now look at how a mobile operator makes
maximum use of its allotted frequency bands.
Cellular networks
Although small capacity radio–telephone car systems were in use as early as the 1950s,
large-scale mobile networks really started during the 1970s with the introduction of
cellular wireless technology, based on concepts developed in Bell Laboratories [1,2].
The initial cellular mobile systems were based on analogue transmission: AMPS in
the United States, NMT in the Nordic countries, and TACS in the United Kingdom.
During the 1990s the so-called second generation of cellular systems based on digital
transmission were introduced: a wide range of systems in the United States (e.g.
D-AMPS, TDMA, CDPD, CDMAone), PDC in Japan and GSM initially in Europe
224 Understanding telecommunications networks
and subsequently most of the World. In most countries the analogue mobile systems
have now been replaced by second-generation (2G) digital systems. This move was
driven by the improved user experience – better quality transmission, privacy from
eavesdropping through the inherent encryption used in digital systems – and for the
operators an improvement in the use of spectrum. As we shall see in Section 9.7, an
interim generation of mobile, the so-called Generation 2 12 , providing small capacity
packet switching for data services, has been installed alongside existing 2G systems
during the late 1990s. The full move to packet-based mobile cellular systems, the
much-vaunted third generation or ‘3G’, began with small-scale deployments in the
early 2000s. However, it is expected that the huge existing capacity of 2G will remain
alongside the 2 12 G and 3G cellular mobile networks for several years.
Fig. 9.4(a) illustrates the basic concept of a cellular mobile radio network. The
radio network is configured as a honeycomb of cells (shown as hexagons for convenience although in practice they have a variety of shapes), each with a radio mast at its
centre. Separate small bands of radio frequencies are allocated to each cell. However,
by careful management of the strength of the transmitted power from each radio mast
to a level that is just sufficient to serve the area of its cell, the power received in the
next-but-one cell will be so low as to appear as indecipherable noise to any mobile
handsets. This means that the same set of frequencies can be used in non-adjacent
cells right across the network. It is through this frequency re-use that mobile operators
have been able to construct the widespread high capacity networks throughout the
World using relatively small amounts of the radio spectrum.
(a) Full frequency re-use in a 7-cell structure
(b) Micro and pico-cells
Figure 9.4
Cellular Structure of a Mobile Radio Network
Mobile switching systems and networks 225
Several cellular frequency patterns are possible, giving 4, 7, 11 or 21 repeat
patterns. Fig. 9.4(a) shows a 7-cell repeat pattern, the most commonly adopted by
MNOs. The required size of the cells depends on the size of the frequency band
available and the forecast number of simultaneous calls from within that cell during
the busy period. Therefore, cell sizes will vary, typically, from around 1.5 km radius
in a town centre to some 15 km radius in a rural area. However, in areas of heavy
concentrations of demand for calls (i.e. ‘hot spots’), e.g. in a shopping mall, small
(350 m radius) or very small (50 m radius) cells, known as ‘micro-cells’or ‘pico-cells’,
respectively, are used, as shown in Fig. 9.4(b).
For a cellular system to work there are several essential functions required,
(i) A power management system controlling the strength of transmitted signal from
the antenna serving the cell, and the transmitted power of the handsets in the
(ii) An automated hand-over system to enable handsets to move between cells
during a call.
(iii) The ability to locate and deliver calls to and from any handset in any cell.
(iv) A common control-channel available continuously in all cells to all handsets.
We can now consider the basic architecture of a cellular network, as shown in Fig. 9.5.
Each of the cells contains a radio mast with transmitting and receiving radio equipment
and a local control system operating at the allocated send and receive set of frequencies
for that cell; this assembly is known as a ‘base station’. These BSs are connected by
a fixed terrestrial transmission link to a central BSC using SDH or SONET over
optical fibre cable or a microwave radio system (as described in Chapter 4). The
Figure 9.5
Architecture of a Cellular Network
226 Understanding telecommunications networks
BSC manages the temporary allocation of radio channels to handsets on demand in
each of its dependent cells through the use of a control channel with each of the BSs.
(The various ways that the radio capacity in each cell is made available to the mobile
handsets is described in Section 9.4.)
The send and receive speech paths from the handsets are then extended from the
BSC over a terrestrial transmission system to a centrally located MSC. Typically, an
MSC would have several BSCs in its catchment areas, each of which has a few up to
a 100 or so dependent cells (and hence BSs). A full national mobile network might
comprise some 20–100 MSCs, depending on the size of the country.
Comparing the simplified generic view of a mobile system of Fig. 9.1 with the
architecture of a cellular network in Fig. 9.5 we can see that the access control and radio
system are provided by the BSs and BSCs. The mobility management, identification
and authentication of the handsets is provided by systems within the MSC, as is
the control, switching and routeing of calls. The MSC acts as a combined local and
trunk exchange within the mobile network: switching calls between handsets in its
catchment area, and routeing calls to other MSCs in the operator’s mobile network
or to the PSTN and other mobile networks, as appropriate.
From a traffic switching point of view, looking at Fig. 9.6 we can see that the
equivalent of the concentration function provided by the subscriber concentrator
switch-block in a fixed network is undertaken by the BS serving the cell. Here, radio
channels are allocated to handsets only on demand when a call is in progress – this
constitutes traffic concentration between the many handsets and the small number of
radio channels serving the cell. Clearly, as in any traffic-loss system (see Chapter 6),
in the event of all radio channels being occupied in a cell no further handsets will be
unable to initiate or receive a call until an existing call is cleared.
We can take this consideration of the architecture of a cellular network a stage further by identifying the location of the functions necessary for supporting telephones.
In Chapter 6 (Fig. 6.5) these functions are captured by the acronym of ‘BORSCHT’.
In a fixed network, with the normal analogue local loop (copper pair), the BORSCHT
(No hybrid or test)
Call concentration
Figure 9.6
BORSCHT Functions in a Mobile System
Mobile switching systems and networks 227
functions are all provided by the subscriber line card in the concentrator switch-block
of the local exchange. However, as Fig. 9.6 shows, in a cellular network these functions are spread between the mobile handset and the MSC. In the handset there is a
battery (B), a ringing tone device (R) powered by the battery and part of the supervision function (S), and also the codec (C) providing theA/D (and vice versa) conversion
of the voice. No hybrid (H) is required since the communication from the handset is
on a completely ‘four wire’ basis, with separate send and receive channels throughout the mobile system. At the BS, BSC or MSC there is no provision of any per-user
equipment, so unlike a fixed-network local exchange there is no equivalent of the test
(T) or electrical overload protection (O) functions.
We can conclude, therefore, that the functions undertaken by equipment that represents some 70 per cent of the cost of a fixed-network local exchange (see Chapter 6)
are provided by the user’s handset in a mobile network. This means that the mobile
handset is not only much more expensive than a fixed-line telephone instrument but
also the technical specification is more-complex and exacting since the handset must
perform many of the critical network functions. As we shall see in Section 9.6, the
mobile handset also contains extremely complex and powerful digital-signal processor microchips, which manage the interaction with the BS and the control of its radio
transmission power. It is worth noting that the handsets are sold in a competitive
consumer market and they are produced by a variety of companies, many of whom
are not telecommunications network equipment manufacturers. Often, the relatively
high cost of the mobile handset is hidden from the users by the service providers
devising tariffs that subsidise the handset cost by increased call charges spread over
a period of 6, 12 or even 18 months.
9.5 Access mechanisms in cellular networks
As mentioned in the generic description of a mobile system, a mechanism is required
to enable the handsets to gain access to the radio frequencies for the duration of a
call. The mechanism must allocate this capacity in both send and receive directions –
known as ‘duplex’ working. There are several ways in which the band of frequencies
allocated to a cell can be made available on demand to the handsets currently in
the cell. These so-called multiple-access schemes are based on the three types of
multiplexing regimes introduced in Chapter 3: FDM, TDM and CDM. The choice
of which multiple access scheme to deploy depends on their technical characteristics
and different efficiencies in the use of the spectrum, the maturity of the appropriate
technical standards at the decision time and, to some extent, commercial and political
(i) Frequency-division multiple access (FDMA): in which the allocated frequency
band is divided into a set of frequency blocks, as in an FDM system. When
required, handsets are then allocated to a free send-frequency block from the
FDM stack on a pre-determined or random basis. It is convenient to have
the receive-frequency block determined by a simple separation, i.e. the sendand receive-frequency blocks are paired; this is known as frequency-division
228 Understanding telecommunications networks
960 MHz
200 kHz
935.2 MHz
TDMA frame
915 MHz
890.2 MHz
5 Send
time slots
5 Receive
time slots
Figure 9.7
Multiple Access and Duplex Arrangments for Mobile Networks
duplex (FDD). An example is shown in Fig. 9.7, in which 124 frequency blocks
of 200 kHz are stacked as a send group (occupying 925.2 MHz to 960 MHz)
and 124 frequency blocks are stacked as a receive group (occupying 890.2 MHz
to 915 MHz) [3].
(ii) Time-division multiple access (TDMA): in which one half of the allowance of
frequencies for that cell are shared across a set of time slots, as in a TDM
system (Fig. 9.7). When required a time slot is allocated to a handset for the
send direction. The other half of the allocated frequencies are used in a TDMA
system for the receive direction, i.e. using frequency division to give duplex
working (FDD).
(iii) Code-division multiple access (CDMA): in which one half of the allowance
of frequencies for that cell are continuously made available to a set of send
channels, each separated by the application of a different digital code, as in
a CDM system. Again, the second half of the frequencies may be used for a
CDMA system in the receive direction (FDD working). When required a send
and a receive channel are allocated to a handset.
In addition to the frequency division method of duplex working, mobile systems also
use the alternative time-division duplex working (TDD). Here the send direction digits
are sent in a burst followed by a burst in the receive direction, and so on alternately,
giving rise to its popular name of ‘ping pong’ working.
9.6 The GSM system
The term ‘GSM’, which for many people is now a household word, was originally
a designation for a standards definition group within CEPT (see Appendix 1) – the
‘Groupe Speciale de Mobile’ – set up in the 1980s to specify a common second
generation mobile system for Europe. The key attribute sought was the ability for
mobile users to roam freely between European countries, enjoying a constant level
Mobile switching systems and networks 229
of service and with charges recoverable by the host network operator from the parent
operator. In fact, since its launch in the early 1990s the GSM standard has been a
spectacular success, not only across Europe but also across all of the World continents.
Even in the United States, where the wide-spread deployment of several incompatible
proprietary second-generation-mobile systems resulted in an inability to roam nationwide or even regionally, GSM has been adopted belatedly as the only universal
system. Consequently, the term ‘GSM’ has been re-designated as ‘Global System for
Mobile’. In view of its importance as a second-generation system we will consider
its architecture and means of working; it also gives a lead into a brief description of
2 12 G and 3G systems later in this chapter.
The original GSM networks launched in Europe operated at 900 MHz, then a
second wave of GSM networks using the 1.8 GHz (1,800 MHz) band, called ‘PCN’
(personal communication networks) or ‘GSM1800’ was introduced. The handsets
contain the radio transmitters/receivers (known as ‘transceivers’) capable of operating
at 900 MHz and 1.8 GHz – ‘dual band’ – so that users can move freely between both
types of networks. Generally, in order to be universal, handsets also have the capability
of operating at the 1.9 GHz band, as used by GSM [or ‘personal communication
services’ (PCS)] operators in the United States; these are referred to as ‘tri-band’.
The main features of GSM are as follows:
• Automatic hand-over between cells. This enables a user to continue a call uninterrupted as they pass through a succession of cells, e.g. when the user is travelling
by car. The call will, however, drop out at the boundary of two cells if there are
no spare channels for the call to be handed over to in the new cell.
• Automatic roaming between operators’GSM networks in different countries. This
requires a commercial agreement to exist between the parent and host operators so
that charges can be recovered between them. When a mobile user enters a country
where there are several operators, all of whom have commercial agreements with
the parent operator, the choice of host allocated to the user is either on a random
or sequential basis.
• In order to minimise the amount of radio spectrum used per channel in the GSM
cells, normally speech is encoded at 13 kbit/s, i.e. much lower than the standard
64 kbit/s used in the PSTN (Chapter 3). Whilst this does give a noticeably degraded
but acceptable sound quality for the users under normal conditions, such low-bit
rate coding does incur perceptible delay – due to the time taken by the codec to
process the speech signal – and a vulnerability to adverse conditions, when the
performance deteriorates sharply.
• Data calls can be established giving 9.6 kbit/s data rate (e.g. dial up to the Internet).
• Short message service (SMS) enables users to ‘text’, using standard telephone
keys, handsets on any GSM network with whom a commercial agreement has been
set up by the parent operator. As described later, this low-speed message service,
originally specified for engineering supervisory use, is carried in the signalling
links between GSM exchanges. SMS was initially an unexpected commercial
success with users, but now forms a major source of extra revenue for mobile
230 Understanding telecommunications networks
Base station subsystem
Network subsystem
Figure 9.8
VLR = visitors location register
HLR = home location register
EIR = equipment identity register
AuC = authentication centre
SMSC = short message service centre
EC = echo canceller
TC = transcoder
IWF = inter-working functions
Architecture of GSM Network
9.6.1 GSM system description
We will now work through a simple description of a GSM system by considering the
block-schematic diagram of Fig. 9.8, which presents the standard GSM architecture.
Beginning from the left-hand side, the user’s handsets, known as ‘Mobile Stations’
in the GSM standard literature, are linked via the radio channels to the cell basetransceiver (i.e. transmitter–receiver) station – the BTS.Astandard interface is defined
between handsets and the BTS, so that any type of handset adhering to this so-called
Air Interface may be used on the GSM network.
As described earlier, the BTSs are linked by a fixed transmission system to the
parent BSC. However, at the entrance to the BSC, the speech channels are converted
(i.e. trans-coded) from the low-bit rate of 13 kbit/s, used over the radio links, to the
high-quality standard 64 kbit/s rate which is used throughout the rest of the non-radio
parts of the GSM network and, of course, any fixed networks that might be included
in the call.
Each of the BSCs are then linked by a fixed terrestrial transmission system (optical
fibre or micro-wave radio, as appropriate) to its parent MSC. (It should be noted that
the terrestrial fixed links between the BTS, BSC and the MSCs may be provided by
the mobile operator or carried over leased lines provided by a fixed network operator.)
As Fig. 9.8 shows, the ‘A interface’ is defined between BSCs and their MSC – again,
to enable a mobile operator to use compliant BSC and MSC equipment from different
The MSC itself is essentially a digital switching unit, as described in Chapter 6.
Thus, the MSC comprises typically a T-S-T digital switch-block, through which a
Mobile switching systems and networks 231
series of circuit-switched 64 kbit/s connections are established for the mobile calls.
Control is provided by a multi-processor system with some degree of dispersed
regional controllers using micro-processors on the individual modules, as described in
Chapter 7. As with the fixed-network exchanges, signalling in the MSC is provided
by SS7 systems, but using the mobile message set of the mobile application part
(MAP), rather than the fixed-network message set of ISUP or TUP (see Chapter 7).
The mobility management functions are provided by adjunct systems associated with
the MSC, as follows.
Home location register
The home location register (HLR) is the heart of the mobility management of the
GSM network. This is the central data base containing the definitive information on
all the subscribers registered on the GSM network. All the MSCs have direct links to
the single, centrally located HLR. The HLR is supported by an authentication centre
(AUC) which holds the details of the registered SIM (subscriber information module)
cards allocated to users. The SIM is a card measuring some 1.0 × 2.5 cm which, when
stimulated, generates a unique code which can be compared with the details held in
the AUC. The SIM card is fitted into the handset, which then becomes part of the
GSM network; users may change to another handset by swapping the SIM card over.
Thus, the SIM in the handset provides the MSC with an authoritative indication of
the identity of the user, just as the point of entry of the local loop does in a fixed
network – as described at the beginning of this chapter.
Visitor location register
Each MSC has a visitor location register (VLR), which is used to record the information relating to any handsets currently in the MSC catchment area. These handsets are
either registered on the HLR of the GSM network operator, or they are visitors from
other GSM networks and so their registration is with their parent operator’s HLR.
Interworking functions (IWF) and echo cancellers (EC)
One or more of the MSCs in a GSM network also acts as a gateway for calls to and
from the PSTN. The main interworking function that needs to be provided here is
the conversion between the mobile (MAP) and the fixed set (ISUP) of SS7 signalling
messages, as described in Chapter 7. There is also a need for echo-cancellation
equipment at this gateway to compensate for the relatively long delays introduced
by the 13 kbit/s low-bit rate encoding of the mobile handsets. The echo cancellation
equipment normally copes with all but very long delays – the latter causing difficulties
for telephone users on that call.
Equipment identity register
The equipment identity register (EIR) is a data base held by a GSM operator containing
the status of all the mobile handsets registered with that network. The serial numbers
232 Understanding telecommunications networks
of the handsets are placed in one of three categories: white for authorised, grey for
needed to be kept under surveillance and black for unauthorised. In this way, stolen
or faulty handsets are prevented from using the GSM network.
Short message services centre
The short message services centre (SMSC) co-located with the MSC provides a
message-switching function for SMS messages. This is essentially a store-andforward system associated with the SS7 signalling of the GSM network. The SMS
messages from a handset, although addressed using the recipient GSM mobile telephone number, are not in fact switched as telephone calls within the MSC. Instead,
the SMS messages are passed by the BSC to the SMS centre (SMSC) directly or
via a semi-permanent path through the MSC switch-block. At the SMSC, the SMS
messages are stored and examined one at a time; they are then sent to the destination
handset on the GSM network, or another operator’s GSM network, as appropriate, carried as messages in the SS7 network between MSCs. Clearly, before the text messages
can be despatched, the location of the destination handset needs to be determined –
this is provided by the location management system as used for telephone calls, and
described in Section 9.6.2.
Operation and network management centre
The operations and maintenance subsystem contains the computer-based support systems used to manage the GSM service. These systems provide the following functions
for the MSC:
• Data build for routeing table;
• Management of software loads onto the exchange-control system;
• Alarm management;
• Usage traffic statistics;
• Management of billing data.
In addition, the subsystem includes the AUC and the EIR.
9.6.2 Location management in a GSM system
The key attribute of a mobile system is, of course, the ability to determine the location
of all active handsets on the network. GSM, like all second generation systems, relies
on the use of a common control radio channel permanently available for all active
handsets to access. The location management process is initiated as soon as a handset
is switched on (i.e. ‘made active’) and continues until it is switched off. (Although
the common radio control channel is permanently available, send and receive radio
channel pairs are allocated only when a call is to be made, as described earlier in this
Mobile switching systems and networks 233
Handset initiated within the same MSC area
Whenever a handset is switched on it sends an initial signal message over the alwaysavailable control channel. The serving BSC then forwards the message together with
an indication of the identity of the cell-transceiver currently serving that the handset –
in effect giving the location of the user, to within the cell or sub-cell area – to the
VLR at the serving MSC. This message is simply the subscriber identity number, as
given by the SIM card. The record in the VLR indicates that this terminal is already
registered as being in this cell, BSC and MSC areas. Thus, all that is required is for the
status of the handset to be recorded as ‘active’. The VLR then sends a signal over the
control channel to the handset indicating it is now an active participant on the GSM
network, capable of making and receiving calls and SMS messages. This indication
is in the form of the name of the GSM service provider (which, of course, may be
different to the network operator in the case of a MVNO) appearing on the screen of
the handset.
Location update with handset in a new MSC area
If the handset is switched on in a different MSC area to when last active, the initiating
signal will be passed over the control channel to a new VLR. On receipt of this signal
the VLR determines that this is a ‘visiting’handset and commences the location-update
procedure, which is illustrated in Fig. 9.9. This begins with an enquiry message
Area of
Figure 9.9
Location Update in GSM within Same Network
Area of
234 Understanding telecommunications networks
comprising the SIM identity sent by the new VLR, i.e. VLR(B), to the HLR for
the GSM network – carried over SS7 signalling links between the new MSC, i.e.
MSC(B), and the HLR (shown as No. 2 in Fig. 9.9)
The HLR first determines the status of the subscriber’s registration, e.g. whether
they are authorised to make calls and whether any special service features apply. The
HLR then sends a message (No. 3) to VLR(A), i.e. the one on which the terminal was
last registered, seeking any update status information and advising the VLR that the
terminal has moved to VLR(B). The VLR(A) sends a signal message (No. 4) back
to the HLR advising of updated information, and that the temporary storage of this
terminal’s details have been cleared.
The HLR now sends a final message (No. 5) to VLR(B) detailing the relevant
status and service features information relating to the terminal. An active participant
signal is then sent by VLR(B) to the terminal.
Location update with handset in the serving area of another GSM network
An important variant on the location-update procedure described above is where the
handset is moved into the serving area of another GSM network operator, often this
is in another country. In principle, the procedure is the same as above, except that
VLR(B) will not have links from its MSC to the parent HLR since the latter is part of
another GSM network, The VLR(B) to HLR signalling (messages No. 2 and No. 5)
must therefore be carried between the two GSM networks (‘host’ and ‘home’) over
the SS7 signalling network as shown in Fig. 9.10. Normally this capability is provided
by the PSTN, or two PSTNs with a section of international network in the case of host
and home networks being in different countries. Exceptionally, certain GSM networks
may have direct traffic routes between them which carry these SS7 messages.
9.6.3 Mobile call in a GSM network
Once the location-update procedure is completed the serving VLR has all the necessary location and status information and the handset is ready to make or receive
calls. After keying the digits the caller (A) presses the send button on the terminal,
which initiates a call-request message to the serving MSC containing the destination
number. The MSC then handles the call in a similar way to a fixed-network telephone
exchange, with the dialled digits examined by the exchange-control system. If the
called number is part of the numbering range of A’s GSM network, then the MSC
interrogates the HLR to determine the current location of the called handset. The HLR
indicates the identity of the destination VLR currently supervising the called handset
so that the originating MSC can switch the call directly to the destination MSC/VLR.
The called handset is then set to ring by a signal message from the destination MSC;
but unlike a fixed network telephone, the ringing is generated by the handset itself.
(This gives the user the ability to personalise their ringing sound by either selecting
from a pre-set range built into the handset or downloading prepared tones from an
information provider.)
Mobile switching systems and networks 235
Area of
Figure 9.10
Area of
Location Update in GSM on Other Network
Fig. 9.11 illustrates the slightly more-complex example of a call between two
registered users of a GSM mobile network (home), but handset B is temporarily
located in the area of another GSM network (the host). As a result of the locationupdate procedure described in Section 9.6.2, the home HLR contains the identity of
the host VLR currently supervising handset B. Thus, on receiving the dialled digits
the home MSC has the message interchange with the HLR (No. 2 and No. 3) described
earlier, but now the HLR needs to communicate with the host VLR on the distant
network in order to complete the call. This is normally achieved by signalling over
SS7 PSTN network between the two GSM networks (No. 4 and No. 5). Once the host
VLR has indicated that the terminal B is active and free (message No.5) the call can
be completed by establishing a standard circuit switched connection over the PSTN,
including an international section if appropriate. This PSTN call connection is routed
using a telephone number temporarily applied to that call from a pool allocated to the
host MSC for this purpose. On receipt of this call from the PSTN, the host MSC is
able to initiate ringing of mobile B (No. 7).
9.6.4 Cell hand-over and power management
As described at the beginning of this chapter, it is necessary for the power of the radio
transmitter (at the BS) in a cell to be adjusted to a level that enables good reception
by the handset up to the boundary of the cell; then the level should rapidly deteriorate
at greater distances. Similarly, the transmitted power from the handset needs to be
236 Understanding telecommunications networks
Area of
Figure 9.11
Area of
Mobile Call in GSM System
set at just sufficient level to be adequately received at the BS’ transceiver. A GSM
network’s cellular structure critically depends on efficient power management.
Since the handset is likely to be on the move for much of the time, the GSM
power management system needs to check and set continually the radio transmission
power level of active handsets. The variation in required power depends not only
on the distance from the cell’s base station (BS), but also on the local reflective and
absorptive environment – which can change in only a few metres of handset travel.
Typically, power management control signals are sent to active terminals every five
minutes or so.
The GSM handset not only monitors the signal strength from its BS but also from
the BSs of the adjacent cells, and sends back an indication of the signal strengths to the
respective base stations (BTSs). Thus, in the case of a moving handset approaching
a cell boundary, once the radio channel received by the adjacent cell becomes the
stronger a hand-over is initiated. Where the two cells BSs are controlled by the
same controller (BSC), it can manage the hand-over. This involves the send and
receive speech paths from the MSC to be switched at the BSC from the link to Base
Station 1 (BS1) to BS2. The uninterrupted call then continues via the radio send and
receive channel combinations allocated to BS2. Clearly, in the case where all of the
send/receive channels are currently occupied in cell 2, the call cannot be sustained
and it has to be abandoned – ‘call drop out’.
Fig. 9.12 illustrates the GSM hand-over arrangements in the situation where the
adjacent cells are dependent on separate BSCs, in which case the common-parent
Mobile switching systems and networks 237
Figure 9.12
GSM Hand-over
MSC controls hand-over; and where the BSCs are on separate MSCs, in which case
the two MSCs work together to manage the hand-over.
9.6.5 GSM frame structure
Finally, in considering the GSM system we will briefly look at the way that the
radio capacity is allocated to users making calls – namely, the GSM frame structure.
Actually, GSM uses a combination of the FDMA and TDMA access mechanisms
described earlier in this chapter. The arrangement is illustrated in Fig. 9.13. First,
the spectrum owned by the network operator is split into a send and receive set
of frequencies (known as ‘up link’ and ‘down link’), each of which is frequencydivision multiplexed into a group of GSM frame channels of 200 kHz width. (In
the example of Fig. 9.13, the 124 send frame channels occupy 935–960 MHz and
124 receive frame channels occupying 890–915 MHz.) Each of the GSM channel
frames is then divided into eight time slots using TDM. Each call is allocated a send
and receive time slot from the send and receive channel frames. We may, therefore,
describe the multiple access arrangement for a GSM system, first, as a time-division
multiple access (TDMA) of a 200 kHz channel, which itself is in frequency-division
multiple access (FDMA) to the available spectrum. Finally, the up link and down
link directions are separated by frequency bands, creating frequency division duplex
(FDD). The scheme may therefore be described as ‘TDMA/FDMA/FDD’. The time
slots derived in this way carry not only the speech, or ‘traffic channels’, but also
various types of signalling and control channels.
Each GSM time slot, whether carrying traffic or control channels, occupies
546.5 μs (about 12 millisecond) and carries 114 bits of encoded speech or control
information. In addition, there are tail bits which indicate the start and finish of the
time slot, together with 26 so-called ‘training’ bits, which are used to compensate
for radio transmission impairment. Since, as we discussed in Chapter 3, speech with
238 Understanding telecommunications networks
935-960 MHz
124 channels (200 kHz)
890-915 MHz
124 channels (200 kHz)
Higher GSM frame
GSM time slot (normal burst)
User data
3 bits
57 bits
S Training
User data
57 bits
26 bits
Tail space
546.5 μs
Figure 9.13
The GSM Frame Structure
4 kHz bandwidth must be sampled at 8,000 times per second, i.e. once every 125 μs,
the TDMA time slot actually contains a ‘burst’ of the equivalent of four consecutive
speech samples. For system-management purposes, the TDMA 8-time slot frames are
grouped into 26 or 51 to form ‘multiframes’, which in turn are grouped into 50s to
form ‘superframes’, 2048 of which fit into a ‘hyperframe’. These frames are similar
in concept to the TDM frames used in SDH and PDH transmission systems described
in Chapter 4. More details of the GSM system are available in Mouly and Pautet’s
authoritative book [4].
So, to summarise, the voice of the caller on a GSM network is carried over the radio
portion as a series of burst of bits in a time slot, which is part of a TDMA frame, which
itself occupies part of the frequency range used in the cell. This combined FDMA
and TDMA arrangement may seem unnecessarily complicated, but it does provide an
efficient use of the precious radio spectrum. In addition, the TDMA system is flexible
so that the time slots can be used in other ways to create new higher-speed services,
e.g. GPRS, as described in the following section.
General packet radio service
In the beginning of this chapter the move to providing packet-based mobile systems
was described. The general packet radio services (GPRS) is the interim packet-based
Mobile switching systems and networks 239
system – the so-called 2 12 generation (2.5G) – which has been added as an overlay
to many of the existing GSM networks. This has given mobile operators and users
the opportunity to develop and use packet-based services at moderately increased
speeds in advance of the widespread deployment of the more complex full-facility
higher-speed 3G networks.
Generally, the GPRS mobile terminals (e.g. handsets) enable the users to make
normal GSM-like voice calls as well as packet-based data sessions to the Internet
or private corporate networks (intranets). However, in addition a dedicated range
of data-only GPRS terminals are available which plug directly into a PC or laptop
computer, thus making the computers act as the mobile terminal. The design philosophy of GPRS is that it should be added to an existing GSM network, exploiting as
much of its installed capability as possible. Thus, GPRS handset/terminals use the
GSM (TDMA/FDMA/FDD) radio access system. At the BSC voice calls from GPRS
handsets are passed to the GSM MSC for circuit switching and they are carried as
standard calls within the GSM network, as described earlier.
When a packet session is initiated by the GPRS terminal the (GPRS-enhanced)
BSC passes the ‘call’ to the serving GPRS support node (SGSN). There is another
important difference with the packet ‘call’, and that is the ability for it to use more than
one pair of send and receive time slots in the radio access link. In fact up to eight time
slots can be concatenated in order to give the user higher data rates. The allocation
of the additional time slots to a data session is made on a real-time basis, tracking
the instantaneous spare capacity on the radio multiple access system managed by the
BSC. Priority is given to the demand for time slots by voice calls. Thus, the speed
of packets throughput on a particular GPRS packet session will vary continuously
between one and eight (maximum) depending on the number of spare time slots
available at the BSC. This concept is shown in Fig. 9.14.
We can now consider the specific GPRS network which handles the data packets,
shown in Fig. 9.15. The main component in the network is the SGSN which provides
the packet switching capability. AGPRS network comprises a number of these SGSNs,
typically co-located with the GSM MSCs, each of which serves a catchment area of
several BSCs. The link between each BSC and the SGSN is provided by a frame relay
packet system (see Chapter 8). Thus, virtual paths are established (at layer 2 of the
OSI 7-layer model) for each of the packet sessions from the GPRS terminals to the
parent SGSN; IP packets to and from the terminals then flow (at level 3 of the OSI
model) over the frame relay virtual paths to an IP router in the SGSN [5].
The GPRS network of SGSNs is linked by an IP network, usually by core transmission using ATM over SDH transmission on optical fibre or microwave point-to-point
radio, as described in Chapter 8. One or more of the SGSNs also acts as a gateway
from the GPRS network to all the external networks – these are designated gateway
GPRS support nodes (GGSN). In practice, a large number of ports are established at
the GGSN with virtual path links provided by network operators to a range of ISPs,
corporate LANs, and the public data networks, as shown in Fig. 9.15. The GSM
network’s VLR-MSC and the HLR provide the mobility management for the GPRS
overlay network, as described earlier in this chapter.
240 Understanding telecommunications networks
on BSC
Available for packet
Total time
voice call
Voice calls
Time of day
Figure 9.14
Allocation of Spare Capacity for Pocket Switching in GPRS
SME users,
Figure 9.15
GPRS Outline
Mobile switching systems and networks 241
Box 9.1
Other Cellular Systems
HSCSD. The basic GSM network can be upgraded to give higher-speed data
transmission using the high-speed circuit-switched data (HSCSD) system. This
uses addition software in the handset/mobile terminal and MSC that enables up
to four of the TDMA traffic channels to be concatenated to give data throughput
speeds up to 43.2 kbit/s, compared to the 9.6 kbit/s data call available over
standard GSM [5,11].
EDGE. The enhanced data rates for global evolution (EDGE) system, which
is essentially an enhanced version of GPRS, is viewed as a step towards full
3G by some and an early 3G system by others – as shown in Fig 9.16. It offers
data rates up to 384 kbit/s through the aggregation of TDMA time slots and
higher modulation levels [11].
DECT. The digital enhanced cordless telecommunications (DECT) system is
a European-wide standard cellular system used for cordless telephones (in the
office and home) as well as campuses, factories, railway stations, etc. DECT
is designed primarily for indoor use within cells of up to 300 m, but typically
around 30 m radius, and with high density of users [5].
TETRA. The European standard for terrestrial trunk radio (TETRA) system
offers business communities (e.g. taxi services, vehicle fleets, police service)
shared access to a set of radio frequencies for voice and data services. The
system has a similar architecture to GSM [5].
The introduction of a packet-based mobile system, such as GPRS, raises the need
for a different form of charging for data service compared to the circuit-switched voice
calls on GSM (mobile) and the PSTN (fixed). The main difference is in the fact that
a packet system enables an always-on status for the GPRS terminal while it is active.
Bearing in mind that mobile terminals are typically kept active almost continuously,
packets can flow from the terminal via GPRS to an ISP for Internet access at any
time. The charging is therefore based on a flat rate, irrespective of packet activity, or
alternatively the charging is based on packet or data throughput. An example of the
latter is the charging for megabytes of data carried per month.
There are several other digital cellular systems currently being deployed –
a selection of which are briefly reviewed in Box 9.1.
9.8 Third generation (3G) mobile systems
Whilst the 2G and 2 12 G mobile systems offer good quality voice plus some degree of
data switching, the third generation (3G) mobile system has been designed specifically
to support a full range of multimedia services. This encompasses good quality voice,
packet access to the Internet and corporate LANs, video, interactive games, etc.,
with higher data rates. As described earlier in this chapter, the deployment of mobile
networks throughout the World has followed various progressions between 2G to
242 Understanding telecommunications networks
Figure 9.16
Single carrier
Standards Evolution from 2G to 3G
2 12 G systems and now 3G systems. The variety of standards is indicated in Fig. 9.16.
Thus, the generally used term 3G actually covers three different standards, all offering
similar facilities, but with little compatibility between them. The European mobile
operators favour the universal mobile telecommunications system (UMTS) which,
as its name suggests, was seen as the natural successor to GSM and with similar
expectations about worldwide adoption. The other two major standards are CDMA
200 mainly for the United States, and TDSCDMA in China.
9.8.1 Universal mobile telecommunications system
UMTS is a unified system which supports multimedia services, global roaming,
service portability and Internet style addressing (in addition to telephone numbering).
It supports a range of new multimedia terminals with high-speed user rates depending
on the extent of the user’s movements: up to 2 Mbit/s for stationary terminals; up to
384 kbit/s for pedestrian movement and up to 144 kbit/s for vehicular movement. The
principal feature of UMTS is that it relies on a new radio access system, known as the
UMTS terrestrial radio access network (UTRAN), which provides the higher-speed
multimedia capability. The modulation and access mechanism is based on a wideband
form of code-division multiple access (‘W-CDMA’) using either FDD or TDD duplex
working [6].
The UTRAN uses the set of frequencies allocated to 3G services, which in many
countries were auctioned by national governments to the mobile operators. W-CDMA
Mobile switching systems and networks 243
User Air
Circuit-switched interface
ISDN, etc.
Packet-switched interface
Figure 9.17
UMTS System Architecture
gives an efficient exploitation of available spectrum, supporting a variety of stationary
and moving terminals, requiring differing data rates. It has the unusual characteristic
of so-called cell breathing: whereby, the transmission power from the base station
is varied to cope with the instantaneous density of users, so causing the extent of
the cell to increase or decrease. This is done to ensure that all terminals within the
cell receive an adequate transmission quality (e.g. as measured by the signal-tointerference noise ratio) at all times. (As described in Chapter 3, all CDMA handsets
within a cell involved in a call/session use the same set of frequencies continuously,
with individual codes enabling each handset-to-BS link to be separated.)
The initial form of the architecture for UMTS is shown in Fig. 9.17. The UTRAN
system comprises Node Bs, which are similar to the BSs of a GSM network, and
radio network controllers (RNC), which provide similar functions to the BSC of GSM
networks. Like GPRS, the two types of calls are separated at the interface of the radio
access (UTRAN) to the communication network (CN) where two paths of switching
are provided. Voice traffic is extended by the RNC to the circuit-switched MSC,
similar to the GSM MSC. Voice calls interconnect with the PSTN or other circuitswitched mobile networks via one of the MSCs, acting as the gateway (GMSC).
Packet-based traffic is sent by the RNC to the IP router-based serving support
node (SGSN). IP packets from the users’ terminals are therefore carried over the
radio access system (UTRAN), over the ATM-based virtual paths from the RNC and
through the IP routers in the SGSN. Egress to the Internet, ISPs, corporate LANs,
etc., is provided via the gateway SGSN (SSSN), as described earlier for GPRS.
244 Understanding telecommunications networks
So, to summarise, the initial architecture for the European 3G system, UMTS, is
based on a new radio access network (UTRAN), but with the mobile management and
voice switching handled by the GSM standard MSC and the HLR/VLR system; while
the data traffic is handled by the IP packet-based GPRS serving node. Voice traffic
interconnects with other mobile and fixed networks via standard circuit-switched
exchange interfaces, using SS7. Data traffic interconnects with external data networks
and ISPs via an IP interface.
However, the 3G systems are being progressively developed towards a data-only
architecture. It is expected that later versions of UMTS will replace the MSC with a
voice-over-IP (VOIP) system, so that voice as well as data will be handled by a single
packet system. Thus, future 3G networks will eventually interconnect their voice and
data traffic with external networks via an IP interface only.
9.8.2 Network planning considerations
The deployment (or ‘rollout’) of GSM or other 2G mobile networks has been undertaken over a period of 10–15 years, with the result that for many countries the coverage
of the cellular network is sufficient to serve 90–99 per cent of the population –
although, many sparsely populated areas are still not covered, The additional rollout
of 2 12 G system, e.g. GPRS, is potentially as widespread as the 2G network, since the
same radio access network is used, although the coverage will be dependent on the
extent of the BSC upgrades and the deployment of new packet-switching equipment.
In the case where an active GPRS terminal is moved out of the GPRS coverage area,
the call will either drop out or revert to a simple telephony call if it is within the GSM
area of the mobile operator.
However, the cost of acquiring the spectrum then building 3G mobile networks,
with the new radio access network (e.g. UTRAN) will result in their coverage areas
being smaller than that of the 2G networks for many years. 3G mobile operators will
therefore have to arrange for voice calls to revert to 2G (e.g. GSM) when terminals
leave a 3G area, and similarly for packet sessions to revert to GPRS, as shown in
Fig. 9.18. The 3G operator will therefore have to make suitable commercial arrangements with 2G and 2 12 G operators to ensure continuity of service for their customers.
Clearly, the 3G handsets will need to be multi-standard – working at the 3G and 2 12 G
frequencies, and capable of using modulation and multiple access mechanisms for
the two sets of networks. The techniques involved in planning the cell structure for
3G networks is described in Laiho and Wacker [7].
9.9 The wireless scene
It is useful to consider the services provided by mobile networks and their evolution
in the context of all the uses of wireless in communications. Fig. 9.19 gives a graphical indication of the range of applicability of the current wireless systems using
the parameters of user communication rate (in Mbit/s) and the degree of mobility
Mobile switching systems and networks 245
3G Coverage
GSM coverage
GPRS coverage
Service drops
back to GPRS
Service drops
back to GSM
Figure 9.18
Mobile Network Coverage
Wireless wide area networks (WLAN)
Figure 9.19
Wireless local area
networks (WLAN)
The Wireless Scene
Bluetooth, offering speeds up to some 0.5 Mbit/s, is designed for data or
voice transmission over a short range (maximum 10 m) within buildings.
246 Understanding telecommunications networks
A typical application is the provision of hands-free wireless telephone headsets,
or connection between two mobile terminals over a short distance.
DECT (digital enhanced cordless telecommunications), the European standard
for cordless telephones, provides a digital wireless link within buildings, with a
range up to about 100 m.
GSM and GPRS provide full mobility in and outdoors, even for fast moving
UMTS offers full mobility indoors and higher rates – up to 2 Mbit/s, depending
on the speed of movement of the terminal.
Wireless LANs (IEEE 802.11), whilst primarily offering local or hot spot data
communication, also provide freedom of movement within some 100 m radius
(see Chapter 8). The ability to communicate, say, from a laptop to an ISP and the
Internet, from one hot spot area then from another, but not when moving between,
is known as ‘nomadic working’.
As Fig. 9.19 illustrates, the areas of application of the wireless communication overlap in both the degree of mobility and the user speeds offered. This means that there
are often alternative systems available to meet people’s communication needs provided by competing suppliers/operators. Furthermore, there is rapid development of
systems offering ever increasing speeds and usable range. Standardisation work is
currently harnessing the various wireless technologies associated with nomadic data
applications with a view to defining the so-called fourth generation (4G) of mobile
systems [8].
Fixed–mobile convergence (FMC)
Because of their differences in nature, mobile networks tend to be managed separately
from fixed networks. This separation is due to several factors. First, most national
regulators apply different rules to fixed and mobile network operators, and they may
be separately licensed. The technical standards for many aspects of mobile networks
are different to those applied in fixed networks. Also, the quality of performance –
and, indeed users’ expectations – differs for the two types of networks, as does the
commercial and service management. However, for most users it is the terminals that
most obviously differentiate their perception of the fixed and mobile networks. This
differentiation, of course, is reinforced by different tariffs, telephone numbers and
service features.
However, this separation is not static. The progressive improvements in mobile
handsets and the declining price of voice calls have driven many users to prefer the
use of mobile networks even for calls inside buildings, where fixed telephones are
readily available. This has given rise to the so-called mobile substitution of fixed
calls. Furthermore, the advent of wireless LAN hot spots (see Chapters 4 and 8) with
their freedom of movement by broadband users within the coverage areas gives rise
to the semi-mobile or nomadic services, mentioned in the context of the wireless
scene. Thus, the demarcation between fixed and mobile is in many ways shifting,
Mobile switching systems and networks 247
with many overlaps between the two domains. One manifestation of this overlapping
is the development of fixed-mobile convergence (FMC) services.
FMC is a generic term that describes services which span both fixed and mobile
networks. They use a single dual purpose handset/terminal, which is linked to either
the fixed network – say, at home and at the office – or a mobile network when out on
the move away from home and office. Calls started when the terminal is linked to the
fixed network will continue as the terminal seamless shifts to the mobile network [9].
There is a variety of tariff arrangements that can apply. Other forms of FMC cover
data and multimedia services. It is important to note that with FMC it is the service
as seen by the user that is ‘converged’ – the fixed and mobile networks supporting
the service remain separate.
Another related movement is fixed-mobile network integration (FMI) through
the use of common technical standards, and the deployment of completely integrated
equipment, where appropriate. An example is the use the of the technical standard for
the IP and multimedia service control systems of UMTS in the 3GPP specification
being used also for the fixed NGN [10], as described in Chapter 11. This FMI is
being pursued to give technical and operational advantages, and may also be used to
provide some FMC services.
In this chapter we looked at the basic structure of cellular mobile networks and noted
their evolution through first, second, 2 12 and third generations. As part of this review
the basic principles of radio transmission, as used in the access portion of cellular
networks, were described. The architecture and method of working of the ubiquitous
GSM system was examined in some detail, covering the location-management function, hand-over between cells, and the use of standard circuit-switching within the
MSC. We then considered the introduction of packet data service, initially through
2 12 G systems, such as GPRS, grafted onto the 2G networks, and more recently through
the rollout of the 3G networks. Finally, the use of mobile systems was put in the context of the total use of radio systems in providing different degrees of mobility, reach
and user speed – leaving an intriguing question around the nature of the 4th generation
mobile systems, as well as the future of fixed-mobile convegence.
SCHILLER, J. H.: ‘Mobile Communications’, Addison-Wesley, Harlow, 2000,
Chapter 2.
2 ANTTALAINEN, T.: ‘Introduction to Telecommunications Network Engineering’, Artech House, Norwood, MA, 1999, Chapter 5.
3 SCHILLER, J. H.: ‘Mobile Communications’, Addison-Wesley, Harlow, 2000,
Chapter 3.
4 MOULY M. and PAUTET, M. B.: ‘The GSM System for Mobile Communications’, Telecom. Publishing, Paris, 1992.
248 Understanding telecommunications networks
SCHILLER, J. H.: ‘Mobile Communications’, Addison-Wesley, Harlow, 2000,
Chapter 4.
WALKE, B., SEIDENBERG, P. and ALTHOFF, M. P.: ‘UMTS – The Fundamentals’, John Wiley & Sons, Chichester, 2003, Chapter 1.
LAIHO, J. and WACKER, A.: ‘Radio Network Planning Process and Methods
for W-CDMA’. Chapter 6 of ‘Advances in UMTS Technology’, edited by BIC,
J. C. & BONEK, E., Hermes Penton Ltd., London, 2002.
BRISO, C., ALONSO, J. and BAYONA, R.: ‘4G Networks: Global Vision,
Reference Architecture and Applications’, The Journal of the Communications
Network, Vol. 3, Part 3, July–September 2004, pp. 89–92.
CLARK, R. and ABERNETHY, T.: ‘Future Voice Services and CPE Trends’,
The Journal of the Communications Network, Vol. 4, Part 2, April–June 2005,
pp. 25–29.
CUEVAS, M.: ‘The Role of Standards in Next Generation Networks’, Ibid.,
pp. 16–41.
GOLENIEWSKI, L.: ‘Telecommunications Essentials’, Addison-Wesley,
Boston, MA, 2003, Chapter 14
Chapter 10
Numbering and addressing
Any form of communication between members of a population requires that each
person is identified by a name or a number, and an address signifying their location.
In the case of telecommunications, the set up of telephone calls or the delivery of data
packets is determined by the examination of the destination address and the application
of the routeing rules for that network. Fortunately, there is universal adherence to the
standards setting out the basic structures of the numbers and names for telephony
and data/Internet services, respectively. Thus, every telephone line is identified by
a unique number, potentially allowing any telephone to make unambiguous contact
with any other telephone in the World; a truly remarkable achievement.
Interestingly, subscribers often become quite fond of their telephone number and
there can be a pride in having particular initial digits, where these indicate a desirable
area in town, perhaps. Similarly, e-mail and web page names are often considered
prestigious. This means that there is value attached to certain numbers, particularly
those easily remembered – giving rise to the concept of ‘golden numbers’. Also,
subscribers do not like having to change their telephone numbers at the whim of
the national administrators of the numbering scheme. Indeed, for business customers
changing their telephone numbers can be an expensive undertaking involving new
signage on vehicles, reprinting advertising brochures, and so on. Consequently, the
seemingly dry subject of numbering and addressing does occasionally become quite
However, this chapter confines itself to considering how numbering and addressing is defined and applied to the various types of networks (both fixed and mobile).
After considering the two different approaches made in telephone and data networks,
particularly IP networks, we examine the newly introduced integrated system that
allows full interworking between telephone numbers and Internet addresses. It should
be appreciated that this is a surprisingly fast moving area of national regulation and
there may be changes made to the established principles of numbering and addressing,
250 Understanding telecommunications networks
particularly as new implementation of voice services, such as VOIP (see Chapters 8
and 11) are implemented.
Numbering and addressing in telephone networks
It is important to note that, traditionally, telephone numbers are used by the exchangecontrol system to not only route the call through the network, but also to determine
the charge category for the call. Subscribers are able to identify from the initial digits
in the number those calls which will incur long-distance, international or special tariff
rates, as described later.
Actually, in fixed telephone networks the number of each subscriber line also acts
as the address of that line within the network. That is not the case, of course, for
mobile networks since the mobile terminal identified by a number is not associated
with any line or even a permanent location. Thus, in a mobile network there needs
to be a translation for every call between the number and the current address of the
handset, as described in Chapter 9.
The internationally agreed basic format for all telephone numbers is described
in recommendation E164, issued by the ITU, (see also Appendix 1) [1,2]. Each
subscriber in the World has a unique international number with a maximum length
of 15 digits, made up of a country code of 1, 2 or 3 digits to identify the country,
followed by the national number. The length of the national number, and hence the
number of subscribers that can be identified, is limited by the length of the country
code – thus, larger countries are allocated shorter international codes. For example,
United States, Canada and the Caribbean Islands combined have country code 1 (also
known as the North American Numbering Plan, NANP, area); the UK has country
code 44; while the Irish Republic has country code 353.
Fig. 10.1 shows the allocation of the initial digits of all the country codes to the
various regional zones of the World. It is interesting to note that Europe, with its high
density of medium size countries, had the initial digits of 3 or 4 allocated to its country
codes. However, the allocation of country codes on the basis of regional zones has now
ceased. This is because country codes can now be assigned for global services, e.g.
+800 for international freephone service, +979 for international premium rate services and +808 for international shared cost services. It has also been recognised that
allocating the country codes strictly on a regional zone basis resulted in unbalanced
utilisation. The full set of assigned country codes is listed in Appendix 2 [3].
Within a country, the national significant number can either be organised on the
basis of geographic areas, e.g. for subscribers in a PSTN, or on a non-geographic
basis, e.g. for mobile subscribers. Fig. 10.2 shows how the geographical national
number for a PSTN subscriber’s line is made up of two parts: the area code, which
identifies the geographic region of the country, and the local number. The latter is
made up of the exchange code and the subscriber number. Examples of two UK local
numbers are shown: London and Ipswich. The area code for London (excluding the
initial 0) is 20, which is followed by a 4-digit exchange code beginning with either
8, 7 or 3 and a 4-digit subscriber number on that exchange. In the case of Ipswich,
Numbering and addressing 251
Figure 10.1
World Numbering Zones
International number
National significant number
Local number
Subscriber number
Examples of geographical numbers
Figure 10.2
Standard Format for Geographical Numbers (E.164)
the area code (excluding the initial 0) is 1473, i.e. 4 digits, followed by a 2-digit
exchange code (capable of identifying up to 100 local exchanges units in that area)
and a 4-digit subscriber number. Thus, in the United Kingdom up to 10,000 lines can
be identified on each exchange unit.
However, there are many variations on the way that national numbers are formatted in each country. In the United Kingdom there is a clear separation between
numbers allocated to the PSTN, which have geographical significance, and mobile
networks and special tariff numbers, which have no geographical significance. The
first digit of the national number (excluding the leading 0), known as the ‘Service’ or
252 Understanding telecommunications networks
Box 10.1
The UK Numbering Scheme [4]
(0) 1
(0) 2
(0) 3
(0) 4
(0) 5
Geographic numbers
Geographic numbers
Geographic numbers
Reserved for future services
Corporate numbers
Location-independent electronic communications
services (e.g. voice over broadband)
(0) 6
Reserved for future services
(0) 7
‘Find me anywhere’ services
70 Personal numbering service
76 Radiopaging service
77 Mobile service
78 Mobile service
79 Mobile service
(0) 8
Special Tariff services
80 (No charge to the caller) ‘Freephone’
844 Rate set by terminating party (up to 5 ppm∗ )
845 Charge at operator’s Local rate
870 Charge at operator’s National rate
871 Rate set by terminating party (up to 10 ppm)
Premium rate services
90–91 Special services at premium rate (e.g. up to 150 ppm)
908 Sexual entertainment service at premium rate
909 Sexual entertainment service at premium rate
∗ ppm = pence per minute
‘S’ digit, denotes the type of service involved. This gives users an indication of the
category of the call, whether it is to another fixed line number, a mobile number or
a special tariff number. So, all PSTN lines (i.e. geographically based) have numbers
with an S-digit of 1 or 2. (From the example of Fig. 10.2, London numbers have an area
code of 020, with an S-digit of 2; Ipswich numbers have an area code of 01473, with
an S-digit of 1.) Whereas, all mobile numbers in the UK are easily distinguished by
their S-digit of 7, i.e. they begin with 07. Similarly, the specially tariffed services are
in the 08-range, and the premium rate services are clearly distinguished by beginning
with 09. Box 10.1 shows the full allocation of the S-digits in the UK national number
format. Many European countries are progressively introducing a similar numbering
However, in the United States and Canada there is no use of a distinguishing
S-digit, so subscribers are unable to identify whether a number is for a mobile or
Numbering and addressing 253
(a) International call
International number
e.g. From UK to France:
From France to UK:
National number
01 45 32
(b) National call
In the UK:
(c) Local call
e.g. In London
Figure 10.3
Dialling Procedures for Geographic Numbers
fixed-line phone. This mixing of geographical and non-geographical numbers in one
range does, however, make maximum use of the capacity offered by the number
length – unlike the UK system where the S-digit format tends to block parts of the
number range from full exploitation. Nonetheless, the North American numbering
scheme does contain the 800 range for special tariff (‘freefone’) services.
An important point to note is that the 0 usually written in front of UK national
numbers is actually only a prefix used in national-dialling procedures, and is not part
of the number. Thus, the 0 is added by a subscriber when dialling a national call,
which is defined as a call between subscribers having different area codes. (In the
United States and Canada a 1 rather than 0 is used as the prefix for calls between
different area codes.) The dialling procedure for a local call in the United Kingdom,
i.e. calls between numbers having the same area code, is to dial just the local number.
International calls are made by prefixing the full international number of the foreign
subscriber by the international prefix as used by the originating network – the ITU
recommendation for this, which is now adopted by many countries, including the
United Kingdom, being 00. Fig. 10.3 illustrates the various dialling procedures for
(a) international, (b) national and (c) local calls. Correct application of the dialling
procedures relies on the correct presentation of telephone numbers in correspondence,
advertising, directories, etc. For example, calls between two exchanges in the London
(0) 20 area should use the local dialling procedure, i.e. using just the local number
(7869 XXXX). The London number should therefore be presented as 020 7679 XXXX
(and not the erroneous 0207 679 XXXX), so that the area code is clearly identified.
In the case of non-geographic numbers the procedure is always for the whole
number to be dialled, as illustrated in Fig. 10.4 for a call to a mobile number.
254 Understanding telecommunications networks
e.g. Mobile call in UK:
Figure 10.4
Box 10.2
National number
Dialling Procedures for Non-Geographic Numbers
UK Numbering and Dialling Formats
(01AB) CDE XXXX: 01 range with 7-digit local dialling
(01ABC) DE XXXX: 01 range with 6-digit local dialling
(01ABC) DE XXX: 01 range with 5-digit local dialling
(02A) BCDE XXXX: 02 range with 8-digit local dialling
07XXX-XXXXXX: 07 range with no local dialling
08XX-XXXXX/X: 08 range with no local dialling
09XX-XXXXX/X: 09 range with no local dialling
Box 10.2 summarises the range of telephone number formats in the UK numbering
scheme and the corresponding dialling procedure. The area codes, which can range
from 2 to 4 digits, plus the trunk prefix 0, are shown in brackets for the various
lengths of geographic numbers. This mix of 6-, 7- and 8-digit local numbers is a
characteristic of the UK numbering scheme, giving as it does the ability to cope
efficiently with differing densities of telephone lines within the area codes. Other
countries, for example the USA, have a constant length for all numbers across the
country. The three sets of non-geographic number formats are also shown; the 08 and
09 range having the option of 8 or 9 digits (not counting the trunk prefix) to cope
with differing capacities.
We can now consider how the local numbering range can be used. As discussed
earlier and shown in Fig. 10.2, the local number is made up of the exchange code and
subscriber number. Calls from other numbers in the area code are made by dialling
only this local number. However, not all of the local numbering range can be used
for local numbers because certain lead digits are needed for other activity. First, of
course, digit 0 cannot be used as the first digit of an exchange code because this could
not be distinguished from the 0 or 00 used by subscribers when beginning a national
or international call, respectively. Also in the United Kingom, codes in the range 1XX
or 1XXX are used to gain access to special features, services and indirect routeings
to other operators. Thus, exchange codes beginning with 1 cannot be used in the local
numbering scheme. Three types of 1XX(X) codes are used, as follows [4].
Numbering and addressing 255
Type A: These codes are widely recognised by subscribers and used by all operators
for equivalent services, e.g. 100 for access to operator assistance, 123 for the speaking
clock, 112 for the emergency services (the European standard code used in addition
to 999).
Type B: A set of codes used to enable customers to choose a service from another
network operator (e.g. 161 for indirect access from BT’s network to Energis). Each
operator can allocate the codes appropriately.
Type C: This set of codes is available for independent use by network operators
for their subscribers or employees (e.g. 151 for fault reporting on BT consumer lines,
174 for BT engineers’ test access).
In addition, the 118XXX range is used in the United Kingdom by the companies
offering directory enquiry services. This was introduced as a result of an initiative
to standardise the 118 range for access to directory services across all European
countries. (Similarly, it is likely that 116 will be introduced as a European standard
for social services, such as medical and safety services.) Also, in the UK code 999
has for many years been the standard code for access to the emergency services. This
means that network operators tend not to allocate initial digits 99 to their numbers,
so as to avoid inadvertent mimicking the emergency number.
Thus, only exchange codes beginning with digits 2–8 (and 90–98) are available
to a network operator – thus reducing the potential capacity to 790,000 6-digit local
numbers, 7.9M 7-digit numbers, or 79M 8-digit numbers [5]. However, in practice
the capacity of the numbering ranges is far less than these theoretical limits due to
the format of the local numbers, with the initial two, three or four digits being used
to identify the relatively small number of local exchanges within the national code
area. Thus, typically it is only the last 4 or 5 digits which can be fully used (i.e. giving
a maximum capacity of 10,000 lines or 100,000 lines, respectively). The net result
is that numbering ranges can in practice not support the large amount of lines that
the overall number length would suggest. (As we shall see in Section 10.7, a similar
situation has occurred with Internet or IP numbering.) Actually, occupancy of some
20–30 per cent of the telephone numbering ranges is generally considered good!
10.3 Administration of the telephone numbering range
Although the overall length of the significant international number and the country
code is set by the ITU-T Recommendation E.164, as described earlier, the allocation
of the capacity within the national number, viewed as a national resource, is governed
by a national administration within each country. In the United Kingdom, it is Ofcom,
the regulator of telecommunications and broadcasting, that takes this role. Similarly,
the national numbering scheme for the USA is administered by the Federal Commission for Communications (FCC). The role of the administrator is to ensure that the
national numbering scheme provides a framework for numbers to be allocated to all
the network operators in the country so that they can provide telephone services to
their subscribers. This allocation must be in accordance with the government policy,
256 Understanding telecommunications networks
particularly in the way that competing network operators are treated [6]. The numbering administration at Ofcom in the United Kingom has traditionally allocated the
available capacity to network operators in blocks of 10,000 numbers, i.e. groups of
the final 4 digits in the number. This relatively large minimum size means, of course,
that operators requiring, say, 500–1,000 numbers would utilise only a proportion of
the allotted capacity – another reason for the utilisation factors of national numbering
schemes being generally low. However, the greater decoding capabilities of modern
switching systems and the increasing demands on the numbering scheme are likely
to cause this procedure to change, and in the United Kingdom smaller allocations of
numbers are already being assigned in some ranges. Thus, the national numbering
range is progressively apportioned by the administration to all the network operators
and service providers in the country; they all own their piece of the numbering range.
The national numbering scheme also needs to have spare capacity to allow for the
requirements of new operators in the foreseeable future, as well as providing room
for expansion into new services beyond basic telephony. Ideally, to minimise the
disruption to subscribers incurred through enforced telephone number changes, the
national numbering schemes are designed to last for some 25 years.
However, in the United Kingdom since the 1960s there have been several changes
to the numbering scheme, each requiring some level of number changes for subscribers [7]. The first, and probably the most unpopular, change occurred in 1966
with the introduction of all-figure numbers to replace the use of letters to represent
the exchange names. Prior to then the national number comprised two letters and a
number to represent the trunk exchange (e.g. LE1 for Leicester) followed by a numerical local number. This use of letters was introduced in the late 1950s and early 1960s
to enable subscribers to dial their own trunk calls (i.e. ‘subscriber trunk dialling’ –
STD) using easily recognisable codes. Furthermore, in the six major metropolitan
areas, i.e. London, Birmingham, Manchester, Liverpool, Edinburgh and Glasgow, it
was the local exchanges that were represented by the first three letters of their name.
For example, numbers within the Hampstead local exchange area were presented
as HAM followed by the 4-digit subscriber number; Swiss Cottage numbers began
with SWI, and so on. (Similarly, to ascertain the time subscribers accessed the speaking clock by dialling TIM!) However, by the late 1960s this extremely convenient
naming/numbering arrangement was becoming unsustainable:
• In London the limit had been reached of the number of three letter codes that could
be used for exchange names, since groups of three or four letters correspond to
the same number on the dial, and only certain combinations of letters make sense.
• With the increase in customer-dialled international calls the inconsistent allocation
of letters to numbers on the dials in various countries was causing a growing
number of dialling errors on incoming calls to the United Kindom.
• The London Sector plan, described in Section 10.4, required changes to some
local exchange codes which would have created unacceptable exchange letter
It is interesting to note that the consequence of directly changing the telephone
numbers from letter–number combinations to figures-only gave rise to the current
Numbering and addressing 257
allocation of trunk area codes as an apparently random spread across the country.
This can be understood by considering the following examples, recognising that the
leading 1 was added only recently: area code 1473 relates to Ipswich (in the East of
England), previously coded as IP3; while 1472 is allocated to Grimsby (in the North
of England), previously coded as GR2 and 1474 is allocated to Gravesend (in the
South East of England), previously coded GR4 – whereas, area codes 1470 and 1471
cover the Isle of Skye in Scotland! In contrast, some countries, e.g. Germany, have
allocated the area codes on a systematic geographical regional basis.
In 1989, the trunk area code for all of London was changed from 01 to 071
for central London and 081 for the outer areas [8]. Then, beginning in 1995, there
was a major restructuring of the national numbering scheme to the current structure
described earlier and shown in Box 10.1, giving rise to the use of 0171 and 0181 for
inner and outer London, respectively. This was shortly followed by the re-introduction
of a single code, this time 020, for all of London [4]. These changes have been
necessary to cope with the rapid increase in the number of operators (mobile and
fixed), each requiring their own segregated numbering range. In addition, many new
services have been introduced, most of which require non-geographic numbers and
need to be differentiated from the standard geographical numbering ranges.
Routeing and charging of telephone calls
As discussed in the beginning of this chapter, numbering (and naming) of subscribers
is used by the network to determine how a call should be routed and what the charge
should be.
10.4.1 Numbering and telephone call routeing
In Chapter 1 we considered a simple call through the network in which the dialled
number was examined by the exchange-control system at each exchange in the routeing. The structure of the geographical telephone number gives the identity of the
called party in descending granularity reading from left to right, i.e. country code,
followed by the area code (which identifies the trunk exchange) followed by the local
exchange identity, and finally the subscriber’s line number on that exchange. Thus, an
exchange-control system needs to only examine the significant dialled digits (reading
from the left) depending on whether it is at the early or later stages of the call routeing, e.g. only examining the area code of the dialled digits at the originating trunk
exchange. Each exchange then uses the appropriate routeing table to determine how
the call should be progressed.
Box 10.3 gives a brief case history of the London Sector plan to illustrate how the
examination of the significant part of geographical numbers was used in the routeing
of calls from the rest of the United Kingdom into London [9]. The scheme, introduced
during the mid 1970s, relied on the grouping of the 350 or so local exchange units
in London into seven sectors, each served by a trunk unit and a central zone with
several trunk units. This arrangement enabled distant trunk exchanges to determine,
258 Understanding telecommunications networks
Box 10.3 Case Study of Traffic Routeing and Numbering: The London
Sectorisation Plan
In 1965 it was decided that a new structure was required for the handling of
trunk and international calls into and out of London [9]. At that time all such
calls for the capital’s 350 local exchanges were switched at trunk exchanges
located in the centre. The plan provided for a new set of trunk sector switching
centres (SSC) to be established in seven sectors radiating from the 6 km radius
central area. The sectors extend to the edge of the area code for London, some
20 km from the centre of London. At the start of the implementation of the
project in 1970 there were about 2.2M telephone lines within the London area.
Advantage had been taken of the move to all-figure numbers (see Section 10.3)
to ensure that the first two digits of the 3-digit exchange codes for London
formed into seven sectors and one central group, as shown in Fig. 10.5. As
mentioned earlier, this did require the introduction of some new exchange
codes where simply swapping the corresponding figures for the letters was not
At that time, the national number for London subscribers was of the format:
01-abc XXXX
where 01 was the area code for London and abc was the 3-digit local exchange
code. At the distant trunk exchange, examination of just the first 3 digits – 1ab
– enabled the traffic to be directly routed to the serving SSC or a central area
trunk unit. Fig. 10.6 shows an example of how two subsequent calls from a
distant exchange can be sent on the appropriate route to the North West Sector
by examining 186 from dialled number 01-864 XXXX, and to the West Sector by
examining 189 from dialled number 01-894 XXXX. This plan for London then
conformed to the standard arrangement for determining the required routeing
at a trunk exchange by the examination of the 3 digits following the 0, since at
that time all trunk codes were three digits in length.
In addition to limiting the growth of capacity on traffic routes into the centre
of London, the sector plan enabled other improvements to the structure of the
network, and the routeing of traffic within London. The result was an improvement in transmission performance, and a build-up of the cable infrastructure
in sector areas of London. The SSCs also formed the foundation for the later
digitalisation of London’s exchanges and transmission routes.
from the first 2 digits of the 3-digit London local exchange code as dialled, to which
of the sector or central incoming trunk units in London the call should be directly
routed. A new simplified structure for the London PSTN based on digital exchanges
and digital optical fibre transmission has since been introduced which reduced the
number of exchanges and routes, making planning and operations easier and cheaper
[10]. However, the grouping of the local exchange codes in London still conforms to
the original sector plan.
Numbering and addressing 259
20, 42, 45,
86, 90, 95, 96
56, 57, 74
75, 84, 89,
34, 36, 44, 80, 88
47, 50, 51, 52,
53, 55, 59, 98
Ea st SSC
South West
33, 39, 54
78, 87, 94
Central area
21, 22, 23, 24, 25, 26,
27, 28, 32, 35, 37, 38,
40, 43, 48, 49, 58, 60,
62, 63, 70, 72, 73, 79,
82, 83, 92, 93
South East SSC
29, 30, 31,
46, 69, 85
64, 65, 66, 67,
68, 76, 77
SSC =Sector switching centre,
with I/C & O/G trunk units
Central trunk units:
either I/C or O/G
Figure 10.5
London Sectorisation Code Allocation (’ab’ Digits of Local Exchange
North West
01-864 XXXX
01-894 XXXX
West Sector
Figure 10.6
London Sector Routeing
SSC = Sector switching centre
LE = Local exchange
260 Understanding telecommunications networks
In the case of calls to non-geographical numbers, e.g. a mobile subscriber or a
‘freefone’ call, the dialled number must first be translated into a destination network
address or number, as described in Chapters 9 and 7, respectively. For the ‘freefone’
call the translation of the dialled number is made by the IN, the control is then handed
back to the trunk exchange so that the call can be routed to the new destination number
through the PSTN in the normal way.
10.4.2 Number portability
The rise of the level of competition in the telephone call market has created the
pressure for customers to be able to keep their original telephone number when they
change their subscription to another network operator – a facility known as ‘number
portability’. This facility is being increasingly introduced into the fixed networks for
both geographical and non-geographical numbering ranges, notably in the United
Kingdom and the United States and is now a requirement within the European Union.
In the case of geographical numbers, the extent of the portability is for subscribers to retain their number given by operator A when changing to operator B,
whilst remaining at the same location. Thus, a new access line needs to be installed at
the premises by the new operator B. (Operator A may or may not physically remove
their defunct access line.) Since the subscriber’s number is still part of the numbering range for the local exchange of operator A, all incoming calls, wherever they
originate, will be routed to that local exchange. Therefore, the number portability
facility involves the detection by the so-called donating exchange (on operator A’s
network) that an incoming call is for a number that has been ported to operator B’s
‘recipient’ exchange. This information is held in the data store of the control system
of the donating exchange, whose role is to redirect all incoming calls for the ported
number to the recipient exchange. There are two main methods for routeing ported
calls, as shown below [11,12]:
• The incoming call is routed to the donating exchange in the normal way through
operator A’s network. On detecting that the number is ported, the donating
exchange sets up a ported-redirection call back to its parent trunk exchange using
a prefix (5XXXXX in the United Kindom) to identify the donating network, followed by the ported national number. The trunk exchange then routes the call
to the point of interconnection with operator B’s network. Since the call goes
into and out of the donating local exchange, this method is known as ‘trombone
• A more elegant and cost-effective solution now widely employed is for the call to
progress only as far as the donating trunk exchange. The donating local exchange
then returns a ‘call drop-back’ signalling message (SS7) in response to the normal
initial-address message from the trunk exchange (see Chapter 7). The call dropback message contains the porting prefix (e.g. 5XXXXX) plus donating number,
as described above, so that the call can be sent directly from the donating trunk
exchange to the point of interconnect with the recipient network. The latter then
routes the call to the ported subscriber.
Numbering and addressing 261
An alternative method for handling ported calls is for reference to be made to a
computer data base containing the details of ported numbers, i.e. the use of an IN system (see Chapter 7). This can be organised on the basis of a national data base owned
and operated by an independent agency for the benefit of all network operators in the
country. Individual network operators may then establish their own data bases within
their own IN, accessing the national data base periodically for updates of new and
changed portings. However, in order to reduce the need for all calls to invoke a (relatively expensive) data-base enquiry, this IN method is usually limited to supporting
portability of non-geographic numbers – especially those in the 08XXXXXX freefone,
etc., range (800 in the United States) – since such calls are easily identified by the
first digit (minus the leading 0). As the proportion of geographical numbers that are
ported increases, there is an economic break-even point where the cost of having to
access the data base for all geographic calls becomes equivalent to that of tromboning
or call drop-back for just ported calls. At that stage the IN approach is economical
for geographical as well as non-geographical fixed network number portability.
There is also a potential need to introduce number portability for mobile networks.
The call set-up procedure for mobile networks using the HLR and VLR systems, as
described in Chapter 9, lends itself to the use of an IN approach for handling number
porting. As telecommunications providers look to introduce NGNs using IP-based
technology, as described in Chapter 11, it is already envisaged that other ways of
providing number portability, such as ENUM (described later in this chapter) come
to the fore.
10.4.3 Numbering and telephone call charging
Generally, charging for calls within the PSTN is based on distance between the calling
and called subscriber and the duration of the call. There may also be different tariff
rates depending on the time of day. Rather than determining the distance-based charge
rate for every combination of exchanges in the country – some 42M for the 6,500
exchanges in the United Kindom – the exchanges are instead collected into about 680
charge groups (CGs) for this purpose. By definition, calls between subscribers within
a CG are at the local rate. The tariff for calls between different CGs is determined
by the distance between the defined charge points in each CG. Each CG is identified
by one or sometimes two area codes, there being some 840 area codes in the United
Kingdom. Thus, by examining the area code in the dialled digits the originating local
or trunk exchange can determine the appropriate charge rate for the call.
These CGs vary in size, but their boundaries have been drawn with the intention
of recognising communities of interest and containing roughly equal numbers of subscribers within a region. Of course, due to the distribution of the population and the
geography of the United Kingdom the characteristics of the charge groups are varied: the largest CG being the London (0)20 area with some 5M subscribers, and the
smallest being based on individual Scottish islands with a few hundred subscribers.
However, the key requirement with telephone call charging is that it appears reasonable to customers, thus calls between subscribers in close proximity are deemed to
be local and those further apart are charged at various trunk rates.
262 Understanding telecommunications networks
CG = Charge group
Figure 10.7
Charging in the PSTN
One of the major anomalies that can occur with telephone distance-charging is
that calls from one end to another within a charge group might be deemed local,
while calls between subscribers a shorter distance apart, but on either side of the CG
boundary, would be charged at the trunk rate. In the United Kingdom this anomaly
is avoided by defining the local-call fee area as being not only within a CG, but also
encompassing all adjacent CGs. Only calls between non-adjacent CGs are deemed to
be trunk rate. This is illustrated in Fig. 10.7, where the local-fee area for a subscriber
in charge group A comprises CG-A, CG-B, CG-C and CG-D; calls to subscribers in
CG-E are at the trunk or national rate.
Charging for international calls is determined at the originating international gateway exchange by examining the country code at the beginning of the dialled digits
(after the international prefix, usually 00). Since, as described at the beginning of
this chapter, the country codes were initially grouped on a geographical-region basis,
the charge rates for international calls from the country can easily be grouped – thus,
simplifying the billing software for the network operator and making the rates easier
for customers to understand. More information on telephone call charging principles
is available in Reference 11.
It is important to note that the increase in competition for telephone calls and
the progressive reduction in prices means that many operators now depart from the
principles of charging described earlier through the use of bundled tariffs and flat-rate
charging, often without discriminating on distance or duration. Also, the introduction
of voice-over-IP over broadband, described in Chapter 8, introduces the notion of
free calls, even between countries, for subscribers on the same system. This gives
rise to the need for operators and service providers to develop new charging models –
often based solely on subscription – in order to recover the network costs.
Numbering and addressing 263
Data network
identification code (DNIC)
Network terminal
number (NTN)
10 Digits
Network digit
Data country
code (DCC)
1 Digit
3 Digits
Figure 10.8
Routeing code
Local number
BT Packet-switched service
International Numbering Plan for Public Data Networks ITU-T Recommendation X.121
Data numbering and addressing
Data networks (see Chapter 8) also require recognised schemes for numbering/naming
and addressing. As with telephony described earlier, the numbering and addressing is
used to determine the network routeing, but it is not generally used in the assessment
of the charging rate, since normally no usage or distance fees are incurred – instead,
the charging is subscription based.
The first number scheme standardised for public data networks is defined in
the ITU-T Recommendation X.121 [13] and was introduced in the 1970s. Fig. 10.8
shows the X.121 number structure. This is based on a data network identification
code (DNIC) of four digits followed by a 10-digit network terminal number (NTN).
The DNIC comprises a 3-digit country code (similar in concept but different to those
used for telephone numbers) and a single digit to identify up to nine networks within
the country. As an example of the application of this standard, the number format
of BT’s packet switching service, based on the ITU-T X.25 protocol and known as
packet switch stream (PSS), is also shown in Fig. 10.8. For this service the NTN is
split into a 3-digit routeing code (RC), which identifies the parent packet switch, and
a 5-digit local number to identify the data line (i.e. terminating point), leaving the
final 2 digits to act as a sub-address for use within the customer’s premises [1,10].
As with telephony numbers, the country codes for X.121 addresses are defined by
the ITU-T and the network identity codes are allocated by Ofcom to operators in the
United Kingdom, who then allocate the network terminating numbers.
10.6 ATM addressing
With the introduction of ATM, frame relay and other connection-orientated data network services (see Chapter 8), new addressing schemes were specified. Public and
264 Understanding telecommunications networks
and format
Domain specific part
Indicates domain
Indicates type
of address
Figure 10.9
ATM Addressing: Basic Format of ISO NSAP
Encapsulated E.164 Format
AFI = 45
E.164 number
Domain specific part
Encapsulated X.121 Format
AFI = 37
Figure 10.10
X.121 number
Domain specific part
ISO NSAP: Types of ATM Addressing Using Encapsulation
private (i.e. corporate) ATM networks can use either the E.164 telephony numbering
scheme, as described earlier, or one of the three versions of an ISO format address.
The so-called ATM end-system addresses (AESA) are based on the ISO NSAP (network service access point) format, shown in Fig. 10.9. The address comprises two
main parts: the initial domain part (IDP) and the domain specific part (DSP). The IDP
contains a 2-digit authority and format indicator (AFI), which is used to specify which
type of addressing is being used, and an initial domain indicator (IDI), which gives
the most significant part of the address. The remainder of the structure is known as the
DSP, which is used by the network to determine the routeing to the data terminal [14].
The various formats of the ISO NSAP ATM schemes are summarised in Box 10.4,
Fig. 10.10 and Fig. 10.11.
In the case where the ATM network is carrying IP packets, a system is used to map
the IP addresses to the ATM addresses; this is known as the ATM address resolution
protocol (ARP) [14].
Numbering and addressing 265
Box 10.4
The ISO NSAP ATM Numbering Scheme
(i) Encapsulation of E.164 numbers: With an AFI set to 45, the IDI is formed
from an E.164 telephony-type number (although separate from the allocation
of numbers for the PSTN), as shown in Fig. 10.10.
(ii) Encapsulation of X.121 numbers: Alternatively, an X.121 number can be
used as the IDI; this is indicated by setting the AFI to 37, as shown in Fig. 10.10.
(iii) Data country code ATM format: With this scheme the significant part
of the address is given by a data country code (DCC), allocated by ISO, which
gives geographical identification of a data network – the DCC for the UK is
826. The domain specific part is then made up of an organisation code, allocated
in the United Kingdom by the Federation of Electrical Industries (FEI), which
identifies the ownership of the private data network. The final part of the DSP
is allocated by the organisation. Fig. 10.11 shows the format of this scheme,
which is indicated by an AFI of 39.
(iv) International code designator (ICD) ATM format: The generally recommended scheme for private ATM networks is the use of the ICD format, with
AFI set to 47, in which the IDI comprises an organisation code identifying
the ownership of the private network. The British Standards Institute (BSI)
administers the allocation of codes to organisations in the United Kingdom on
behalf of ISO. The domain specific part is then allocated by the private network
organisation, as shown in Fig. 10.11. The alternative use of the ICD format,
which is recommended for public ATM services, uses an international network
designator (IND) specified by the ITU-T in place of the organisational code in
the IDI.
IP numbering/naming and addressing
As with telephone numbering and addressing on the PSTN, every device (e.g. routers
and data terminals or computers/hosts) on the Internet needs a unique address. Since
the conventions were set by the IETF in 1984, 32-bit addresses adhering to the IP
version 4 (IPv4) standard have been used throughout the World. A new enhanced
version – IPv6 – is gradually being introduced, although mass migration to this new
protocol has still to occur. For convenience, we will not consider this new version
until the end of this chapter. These IP 32-bit source and destination addresses specify
the entry and exit points, respectively, on the Internet and are inserted in the IP packets, so that they can be routed appropriately, as described in Chapter 8. We will come
back to discussing the IP addresses later in this section, after we have considered the
more-familiar and user-friendly Internet names that are used in e-mails and web site
identification. Clearly, these names need to be translated to Internet (or intranet)
addresses before the IP packets can be dispatched, which we will also consider
266 Understanding telecommunications networks
Data county code (DCC) format
Domain specific part
Initial domain part
AFI = 39
IDI = Country code
Allocated by
(UK = 826)
Organisation code
Allocated by FEI
End system ID
Allocated by
International code designator (ICD) format
AFI = 47
IDI = Oganisation code
Higher order DSP
by BSI on behalf
of ISO
Figure 10.11
End system ID
Allocated by
ISO NSAP ATM Addressing Formats
10.7.1 Internet names
IP names and web site addresses are based on the use of mnemonics that are enduring,
easily identifiable and memorable. The structure of the Internet name is hierarchical,
but with the granularity in the reverse order to the international telephone number,
i.e. moving from specific to general reading left to right. The levels of the hierarchy
are in descending order as follows:
• Top level domain (TLD);
• Domain;
• Sub-domains;
• Host name. (The term ‘host’ means the device or computer attached to the Internet
or private intranet, i.e. the item being addressed.)
There are two categories of top level domains, namely:
(i) Country code TLD (ccTLD), of which there are some 450;
• Examples: .ca = Canada
.es = Spain
.fr = France
.uk = United Kingdom
.us = USA
Numbering and addressing 267
(ii) Generic TLD (gTLD)
• Examples: .com
Fig. 10.12 shows the Internet domain-name hierarchy, with an example of the
analysis of the host address ee.ucl.ac.uk (with TLD of uk, domain of ac and subdomain of ucl and host machine ee). The base of this tree-like structure is known
as the root. E-mail names are created by adding the individual name in front of
this Internet name, e.g. person’s name@ee.ucl.ac.uk. The host in this case could be a
server on which web pages are stored, accessed by an Internet name based on the URL
(uniform resource locator) format: e.g. www.ee.ucl.ac.uk/page number. This format
can be expanded to name individual sections (i.e. ‘pages’) out of a vast collection
stored on the server/host.
With the rise in popularity of the Internet and web-based services, such as on-line
purchasing of everything from airline tickets to books and even cars, the various
top-level domain names have become well recognised. There is much value in terms
of prestige for a company in having a web site name with a well-known gTLD such
as .com. In general, short names such as company.com have greater marketing value
than longer names such as company.co.uk. In fact, the new phenomenon of ‘cyber
squatting’ has arisen through people registering Internet names of well-known companies in the expectation that those companies would some time later be prepared
to buy the name from them at an inflated price. Finally, of course, the use of names
Figure 10.12
Internet Domain-Name System Hierarchy e.g. ee.ucl.ac.uk
268 Understanding telecommunications networks
rather than numbers for the Internet gives the opportunity for people to impersonate well-known companies by registering the same name with a different TLD or
mischievously similar names, creating the phenomenon of ‘cyber fraud’.
Top-level domain names are administered by the Internet Corporation for
Assigned Names and Numbers (ICANN), which is an internationally organised, nonprofit corporation that also has responsibility for IP address space allocation, and
root server system management functions. As a private–public partnership, ICANN
is dedicated to preserving the operational stability of the Internet and promoting competition. In order to address some of the concerns over cyber squatting and fraudulent
use of domain names it has introduced a uniform domain-name dispute resolution
policy (UDRP), which is applicable across all gTLDs and now forms the basis for
resolving such issues.
10.7.2 Internet addresses
An example of an IPv4 address is shown in Fig. 10.13. By convention, the 32-bit
address is more conveniently shown in documentation as a set of the four decimal
numbers representing each byte (8 bits). In theory, a 32-bit binary number gives:
2∧ 32 = 4, 294, 967, 296 potential addresses.
However, the original allocation of this address space in the mid 1980s has severely
reduced the actual capacity available in practice. This allocation was associated with
partitioning of the address space. To understand the need for the partitioning we should
appreciate that the (binary) IPv4 address in the destination field of an IP packet (see
Fig. 8.9 of Chapter 8) is used by all the intervening routers in the Internet or intranet
to determine how to onward route that packet. This is achieved by using the first
part of the address to define the identity of the destination IP network – which is
examined by the routers within the Internet – leaving the remainder of the address
to define individual hosts on that network. The latter part is not examined until the
packet reaches the ingress gateway router of the destination network, and it is then
examined by any subsequent routers on that destination network.
Figure 10.13
IPv4 Addresses
Numbering and addressing 269
bit #
Class A
0 1
Host number
126 networks, 16,777,214 hosts addressed using 50% of IPv4 space
bit #
Class B
0 1 2
15 16
Network number
Host number
16,384 networks, 65,534 hosts addressed using 25% of IPv4 space
bit #
Class C
2 3
23 24
Network number
Host number
2,097,154 networks, 256 hosts addressed using 12.5% of IPv4 space
Figure 10.14
IPv4 ’Classful’ Addressing
The partitioning scheme is based on three sets of classes, each designed to accommodate different sizes of IP networks. Fig. 10.14 shows the so-called Classful address
format [15]. Class A is designed for the very largest IP networks, having address space
for some 16,777,214 hosts; Class B is designed for medium sized organisations, having address space for some 65,534 hosts; finally Class C, with its capacity of 256
hosts, is designed for the smaller organisations. As Fig. 10.14 shows, the type of class
being used is indicated by the initial binary digits of the address – 0, 10, 110, for
Classes A, B and C, respectively. For completeness, we should mention that there are
also two further special classes (omitted in Fig. 10.14 for clarity), namely: Class D,
which is used for multi-cast addresses enabling broadcasting of an IP packet to many
hosts, indicated by initial digits 1110; and Class E, indicated by initial digits 1111,
which was reserved for future use.
One way of accommodating large organisations with several individual departmental LANs supporting many hosts, e.g. in the case of a university campus, within
the Classful structure is the use of sub-netting. This technique relies on a single network address, which is used by all routers in the Internet to reach the ingress router
for the organisation, and the host address space is split into a sub-net address and its
host addresses. An example of an organisation with three sub-nets associated with
one network address of 156.5 is shown in Fig. 10.15. Let us consider the routeing
of an IP packet sent over the Internet to a host machine on sub-net A, with the destination address of Now, in general the control software of routers has
look-up routeing tables bifurcated into addresses of all the IP networks on the Internet and the addresses of their own dependent hosts or sub-networks. For each entry
there are two fields: the IP network address and the appropriate outgoing route, as
described in Chapter 8. In our example, the network gateway router has an IP address
of – since it is only the first two bytes that are used by the Internet routers to
270 Understanding telecommunications networks
Extended network prefix
Network address
Host address
Sub-net A
Sub-net B
Sub-net C
Figure 10.15
reach that network. So, when the packet is received from the Internet with a destination address of the network router software recognising 156.5 as its own
address then examines the next bytes to determine the required sub-net, i.e. 32. The
subnet router(s) then route to the destination host (.4). In general, routers within a
sub-net environment use the network address together with the sub-net address when
exchanging packets between hosts on the various sub-networks – this is known as
‘Extended Network Prefix’ working, as shown in Fig. 10.15.
Unfortunately, the Classful partitioning scheme does not make good use of the 32bit address space, since it accommodates only 126 Class A networks (which occupies
50 per cent of the total address space), 16,384 Class B networks (25 per cent of
total address space) and 2,097,154 Class C networks (12.5 per cent of total address
space). This inherent inefficiency has been exacerbated by the allocation of addresses
in the early days of the Internet without regard for the conservation of the addressing
resource. Of course, the enormous rise of the popularity of the Internet, far beyond
the original concept of a specialist network linking academic institutions around the
World, has created unforeseeable demands for Internet addresses. It was soon realised
that the Classful scheme, particularly with the Class B offering too much capacity
and Class C too little capacity for hosts for most medium-sized organisations, was
unnecessarily restrictive. Even as early as the beginning of the 1990s exhaustion
of the IPv4 addressing space was predicted as rapidly approaching. Thus, in 1993
Numbering and addressing 271
Table 10.1
Examples of CIDR Internet Addressing
Number of
Number of
11111111 11111000 00000000 00000000
11111111 11111100 00000000 00000000
11111111 11111110 00000000 00000000
11111111 11111111 00000000 00000000
11111111 11111111 11110000 00000000
11111111 11111111 11111000 00000000
16 k
32 k
64 k
128 k
256 k
512 k
256 k
128 k
64 k
allocation of new Classful addresses was ceased and the alternative classless interdomain routeing (CIDR) scheme was introduced to maximise the occupancy of the
remaining available address space.
With CIDR, there is a close match between the capacity of the address space
allocated and the required numbers of hosts. This is achieved through defining the
position of the moveable boundary between network and host addresses. A CIDR
address therefore needs two components: the full 32-bit IPv4 address, followed by
an indication of the length of the network address portion. Both these pieces of
information need to be held in the routeing tables of all the Internet routers. (Thus,
the routeing tables need three fields of information per entry, i.e. IP network address;
CIDR address length and outgoing route.) A CIDR address is written in the format: /20, indicating that the network address occupies the first 20 bits of the
full 32-bit Internet address (leaving 12 bits to identify the hosts). In practice, the
routers employ a binary mask which is ‘ANDed’, using Boolean algebra, with the
full 32-bit address in order to extract the portion representing the network address.
11000010 00100010 00001110 00000000 (binary of
ANDed with:
11111111 11111111 11110000 00000000 (20 binary 1’s)
Network address: 11000010 00100010 0000
Table 10.1 gives some examples of CIDR addresses and for each length of mask the
resulting number of network and hosts addresses that can be accommodated. (It should be
noted that the numbers of addresses shown are derived from the number of binary bits in
the address, and are therefore based on powers of two. Thus, 10 bits = 1,028 (i.e. 2∧ 10),
written as 1 k; 11 bits = 2,048 (i.e. 2∧ 11), written as 2 k, 12 bits = 4,096 (i.e. 2∧ 12),
written as 4 k, etc.)
10.7.3 Translating Internet names to addresses
Now we need to think about how the Internet name (e-mail or web page) is associated
with an appropriate Internet address. As stated earlier, the Internet addresses and names
272 Understanding telecommunications networks
Further DNS
[12. 57.6.1]
www.microsoft.com ?
www.microsoft.com ?
Figure 10.16
How IP Terminals Interrogate DNS
are administered worldwide by ICANN, who in turn delegate parts of the address and
name space to various regional and national bodies for allocation to ISPs, companies and
In the case of e-mails, each subscriber’s name, in the form of name@ISP.com needs to
be allocated an Internet address by the ISP from their set of allocated addresses. In the case
of dial-up subscribers, the ISPs usually minimise the overall need for the valuable Internet
addresses by using a computer server to manage a pool of addresses, allocating them on
a temporary basis to subscribers for the duration of their Internet session. However, the
always-on nature of broadband subscribers means that they require a permanent allocation
of Internet addresses from the ISP. Allocation of addresses to users of corporate networks
(e.g. LANs) is undertaken by the company’s network administrator, again using a server
to allocate on a real-time temporary or permanent basis, as appropriate. The ISPs or
Corporations obtain their allocation of Internet addresses from larger ISPs, who in turn
obtain their allocation from one of five regional Internet registries covering Europe, North
America, Asia, Latin America and Africa, respectively.
Now to consider how the addresses for web sites are determined. As mentioned above,
web sites are identified by the user-friendly URL mnemonic-based names. The translation
between a URL and the Internet address of that site is made by a directory service known as
the domain name system (DNS). This service may be provided by the subscriber’s ISP or by
independent DNS operators. Fig. 10.16 illustrates a simple example of a user attempting
to access www.microsoft.com. Assuming that an IP session already exists between the
terminal and the ISP, a request for the address of Microsoft’s web site is sent in an
IP packet with source address (the terminal) and destination address of
the ISP ( If the DNS of the ISP does not have the required information it
generates a request of the next DNS in the hierarchy (address, as shown. In
the unlikely event of this DNS not having the address, a request for a further DNS up
the domain hierarchy is generated. This process continues until the definitive source of
addresses, the root DNS (see also Fig. 10.12), is reached. However, assuming the remote
DNS can provide a translation, it sends the required address for Microsoft’s web site – – in an IP packet to the ISP’s Local DNS, where it is temporarily cached
in case there are further requests for this address, before advising the terminal. A session
Numbering and addressing 273
can now be established with Microsoft’s web site from the terminal by the insertion of the
Internet address in the destination field of the outgoing IP packets.
10.7.4 IPv6
Work has been undertaken by the IETF since the early 1990s to define a new version
of IP, IPv6, which is designed among other things to increase vastly the capacity of
Internet addresses, and improve the efficiency of network routers by better structuring
of the network addresses (see also Chapter 8). A need for a significant increase in the
number of Internet addresses had for some time been predicted, not only because of the
current shortage in the supply of IPv4 addresses, but also more importantly because of
an expected mushrooming in demand in the future. This is driven by the introduction of
IP-based services, e.g. third-generation mobile networks, and communications between
intelligent devices using Bluetooth and other radio technologies (see Chapter 9).
The first thing to note about IPv6 is that it is based on 128-bit addresses, organised
as eight 16-bit bytes – giving a potential capacity of 64 billion times the capacity of the
IPv4 address space! Needless to say, the proposed partitioning of this 128-bit address
into a rigid format significantly reduces the actual capacity available for use, but it is
still enormous and it is thought unlikely to exhaust this century. The notation for IPv6
addresses is based on the representation of each 16-bit byte using hexadecimal digits (see
Box 10.5). For example:
For convenience the leading zeros in each 16-bit byte or block can be suppressed. Also,
a double colon can be used to represent consecutive zeros. So, the above example can be
simplified to:
The general format of the IPv6 address for a single interface (or host) – known as a ‘global
unicast address’ – falls into three main parts, as shown below.
(64−m) bits
Global routeing prefix
m bits
subnet ID
64 bits
interface ID
The value of m, defining the size of the subnet identification, is commonly set to 16 bits.
Thus, the global routeing prefix, which defines the network address, is commonly 48 bits.
The specification for IPv6 is still relatively immature and the formats of the address
as well as some of the other features are likely to be refined as experience is gained. For
this reason, and given the huge established base of IPv4 equipment across the World, the
move towards IPv6 has been relatively slow so far. Also, it can be expected that the rate
of adoption of IPv6 will be different within the public Internet, the corporate intranets,
enterprise networks and public WiFi, and within the fixed and mobile networks, depending
on the commercial and service issues pertaining [16].
274 Understanding telecommunications networks
Box 10.5
Hexadecimal Numbers
This table summarises the three numbering or counting systems used in telecommunications and data networking for values of 0–15. As is commonly known, the
counting scheme of decimal is structured around a base of 10, i.e. the use of the
10 digits 0 to 9, with each position to the left in the number having a weighting of
increasing powers of 10. Similarly, the counting scheme of hexadecimal is structured around a base of 16, i.e. the use of 16 digits 0 to F, with each position to
the left in the number having a weight of increasing powers of 16. For binary the
base is 2, using digits 1 and 0, with each position to the left in the number having
a weight of increasing powers of 2.
(base 10)
(base 16)
(base 2)
Inter-working of Internet and telephone numbering and
So far in this chapter we have considered E164 numbering and addressing in the context
of telephone and other services supported by the PSTN (fixed) and mobile networks,
and Internet naming and addressing in the context of IP data services. However, the two
Worlds are no longer separate. We introduce the concept of providing telephony calls
over an IP data network in Chapter 8, identify the move to an all-IP interface for later
versions of 3G mobile networks in Chapter 9, and we investigate the eventual large-scale
replacement of the PSTN by (VOIP) in the so-called NGN in Chapter 11. These are
all examples of the much-heralded convergence of voice and data services. Clearly, an
important requirement of this convergence is the smooth translation between telephone
numbers and Internet addresses.
Numbering and addressing 275
Query: ?
4.4 (UK)
IP Network
1. (NANP)
3.3 (France)
020 7679 8135
IP Network
Figure 10.17
Use of the ENUM System
In Chapter 8 the sequence of a voice call between two telephones on different PSTNs
carried over an intermediate IP network (the Internet or an intranet) was described
(Fig. 8.17 refers). The inter-working of the PSTN and IP network require the originating
PIG (PSTN-to-IP Gateway) to seek the IP destination address of the distant PIG so that
the voice IP packets can flow between gateways to create a call connection. This involves
the originating PIG obtaining, from a DNS, the IP address of the PIG serving that numbering range of the PSTN. Given that the DNS also holds name-to-address information
for the Internet, as described earlier, the addition of translation details for PSTN numbers
creates a significant extra storage load. This need for extra capacity not only increases the
cost of the DNS but it also affects the performance greater interrogation times involved.
Consequently, as the quantity of VOIP calls increases, a more elegant solution for the
translations of PSTN numbers to Internet addresses is needed. Such a solution is offered
by the ENUM system currently being specified by the IETF in collaboration with the
ITU-T [17].
ENUM is a DNS-based system that converts a E.164 telephone number into an Internet
name or address for routeing through an IP network. The design of the system reflects the
fact that the use of telephone numbers is so universally accepted that even when a high
proportion of phones are IP-based, and circuit-switched PSTNs are replaced by VOIP
equivalents, people will be still be using such numbers.
Fig. 10.17 shows an example of the use of ENUM, in which PSTN subscriber A wishes
to call a LAN-based subscriber B who has an IP phone. However, subscriber B is still
recognised as a telephony subscriber with an E.164 number 020 7679 8135. On dialling
this number the call from subscriber A is routed through the PSTN to the appropriate
IP gateway (i.e. PIG). An enquiry for an IP address for this number is then sent from
276 Understanding telecommunications networks
the PIG to its DNS. ENUM has a dedicated top-level domain – expected to be .arpa –
encompassing all the translations for the World’s E.164 numbers. The enquiry message
to the ENUM system is created as follows:
Dialled number: 020 7679 8135
Convert to standard international number: +44 20 7679 8135 (i.e. a UK number)
Eliminate spaces and non-digits: 442076798135
Reverse and insert dots:
Add TLD and domain name:
This standard DNS enquiry format is then passed to the ENUM directory for translation
into an Internet address using a decoding tree, as shown in Fig. 10.17. The call is then
routed through the IP network to the destination subscriber phone B.
A variant on this example is for the case of a SIP-controlled call (see Chapter 7),
where the ENUM directory translates the enquiry into an IP name rather than an address,
e.g. name@ISP.com. This is then subsequently converted at a DNS to an Internet address
for routeing through the IP network in the normal way. Actually, the ENUM system is
more complex than described earlier because it also copes with the many forms of service
that use E.164 numbers other than telephony, e.g. fax and e-mail. Thus, a single E.164
number can be translated within the ENUM system into several outputs: Internet numbers
or names, depending on how the node in the Internet dealing with that service is to be
In this chapter we have examined the structure of the internationally recognised telephone
number scheme, and noted how the dialled number is used by the control systems in
the exchanges to determine the appropriate call charges and how it should be routed
through the network. The application of similar numbering schemes was then applied to
connection-orientated data services such as X25, frame relay and ATM. The chapter then
looked at the Internet, noting that the names are used in place of numbers and that there
is a clear distinction between numbers/names and addresses – unlike the arrangement for
the PSTN, where the subscribers’ numbers are also used as the network addresses. Finally,
we looked at how ENUM helps manage the inter-working between networks using either
form of addressing, thus allowing increasing convergence of services in the future.
McLEOD, N. A. C.: ‘Numbering in Telecommunications’, British Telecommunications Engineering, Vol. 8, Part 4, January 1990, pp. 225–231.
2 ITU-T Recommendation E.164: ‘Numbering for the ISDN Era’ (ITU, Geneva).
3 Complement to ITU-T Recommendation E.164: ‘List of ITU-T Recommendation E.164 Assigned Country Codes’, (Annex to ITU Operational Bulletin No.
805, Geneva)
4 OFCOM: ‘The National Telephone Numbering Plan’, 2004, (www.ofcom.gov.uk).
Numbering and addressing 277
OFCOM: ‘A User’s guide to Telephone Numbering’, originally published by
Oftel and subsequently adopted by Ofcom, (www.ofcom.gov.uk).
BUCKLEY, J.: ‘Telecommunications Regulation’, IET Telecommunications
Series No. 50, The Institution of Electrical Engineers, Stevenage, 2003,
Chapter 6.
UNDERWOOD, H.: ‘National Code and Number Change – Technical Solutions for BT’s Network’, The Journal of the Communications Network, Vol. 1,
Part 1, April–June 2002, pp. 107–113.
Change’, British Telecommunications Engineering, Vol. 8, Part 3, October 1989,
pp. 134–143.
WHERRY, A. B., & BIRT, J. F.: ‘The London Sector Plan: Background and General Principles’, The Post Office Electrical Engineers Journal, Vol. 67, Part 1,
April 1974, pp. 3–9.
WOOD, S. & WINTERTON, R.: ‘ANew Structure for London’s Public Switched
Telephone Network’, British Telecommunications Engineering, Vol. 13, Part 3,
October 1994, pp. 192–200.
FLOOD, J. E.: ‘Numbering, Routing and Charging’, Chapter 9 of ‘Telecommunications Networks’, Second edition, edited by FLOOD, J. E., IET
Telecommunications Series No. 36, Stevenage, 1997.
BUCKLEY, J.: ‘Telecommunications Regulation’, IET Telecommunications
Series 50, The Institution of Electrical Engineers, Stevenage, 2003, Chapter 7.
ITU-T Recommendation X. 121: ‘International Numbering Plan for Public Data
Networks’, ITU Geneva.
ROSS, J.: ‘Telecommunication Technologies’, Prompt Publications, Indianapolis, IN, 2001, Chapter 8.
TANENBAUM A. S.: ‘Computer Networks’ Fourth edition, Prentice Hall PTR,
Amsterdam, the Netherlands, 2003, Chapter 5.
DAVIDSON, R. ‘IPv6: An Opinion on Status, Impact and Strategic Response’,
The Journal of the Communications Network, Vol. 1, Part 1, April–June 2002,
pp. 64–68.
NEUSTAR: ‘ENUM: Driving Convergence in the InternetAge’, www.enum.org.
Chapter 11
Putting it all together
In the first ten chapters of this book we considered various aspects of telecommunications networks: their components, how they are constructed, and the way that voice
and data is carried. Now, in this final chapter we take a holistic view of how all the
pieces are put together, how the end-to-end service provided to the users is managed,
and how the networks are progressively enhanced and developed to take advantage
of new technology. Specifically, the so-called NGN concept is examined. It is hoped
that by the end of this chapter readers will have gained an insight into the intriguing
question raised in the Foreword to the book, i.e. whether it is right to assume that all
communications will eventually move onto the Internet.
11.2 Architecture
As we have seen in the earlier chapters there are many components used in telecommunications networks, which require appropriate linking in order to provide services for
customers. This situation is shown conceptually in Fig. 11.1, with the network components of access, circuit-switching, core transmission, intelligence and signalling
shown as pieces in a jigsaw puzzle, residing within the domain of the applicable commercial and regulatory environment, which together provides services. The classical
method of understanding in a systematic way how all the pieces fit together is through
the use of architectures. In every day life we understand how the use of architecture
helps to describe the structure of a house – giving the relationships between the walls,
windows, floors, roof, etc. It also shows the deployment of the various services,
such as water, drainage, heating, air-conditioning, electricity and gas. Architecture is
applied in much the same way in telecommunications networks.
However, architecture in telecommunications networks is more than just a diagram showing how all the ‘boxes’ are joined together, important though this is! An
280 Understanding telecommunications networks
Commercial environment
Figure 11.1
Putting it all Together
Commercial model
Logical view
Techno / regulatory view
User (person
or machine)
Physical view
Usernetwork interface
Figure 11.2
The Many Views of Architecture Views
architecture also needs to encompass the commercial and regulatory conditions applying to the network, and the services it receives and supplies to other networks/service
suppliers, as well as the services provided to its users (‘customers’). The architecture of an operator’s network, for example, must comply with the strategic policy of
the company in terms of the business objectives and services offered [1,2]. Therefore, typically four architectural views are required to adequately define a network
operating in today’s environment, as shown in Fig. 11.2 and described below.
Putting it all together 281
11.2.1 Commercial or service model view
This view shows the various roles involved in getting a service to a customer, and
the players that undertake those roles. In addition, most importantly this view needs
to identify the flow of money between the roles. One might characterise this view as
showing who does what to whom and who is paying for it! A simple example is that
of a subscriber to BT’s network accessing the Internet via dial-up using a ‘freefone’
(0800) number to an ISP (Chapter 2). The subscriber pays a line rental to BT for the
local line, which is also used for normal telephony calls (with appropriate call charges
to BT). The subscriber also pays a fixed (non-usage based) monthly subscription to
the ISP for access to the Internet. Since the ISP has provided the subscriber with free
call access for the Internet by virtue of the freefone number, BT does not receive any
money from the subscriber for such calls; instead, the ISP pays BT a rental for the
0800 freefone number. There is thus a set of commercial relationships between the
subscriber, the ISP and BT covering services and payment.
We can elaborate the simple example of the dial-up call to the ISP by considering what happens if the subscriber makes a purchase over the Internet using a web
page of a trader. There is then a further commercial relationship between the trader
and the subscriber. However, the situation could become even more complex if a
search engine or web portal is used by the subscriber when seeking the trader on
the Internet, because the trader will then need to pay the portal service provider for
the customer referral. Such use of the Internet (World Wide Web) for purchasing,
often called ‘e-purchase’, is part of a general phenomenon known as ‘e-commerce’
or ‘e-business’. This itself is part of the bigger so-called information industry, which
embraces the provision of information (e.g. share prices and local weather forecasts)
over the Internet or corporate networks, broadcast television via satellites, etc., as
well as e-business applications. During the mid-1990s the terms ‘Global Information
Highway’ and ‘Information Superhighway’ were coined to describe the route to this
World of electronically accessible information.
In 1995, the European Telecommunications Standards Institute (ETSI) undertook
a strategic review of the implications of the emerging infrastructure for the information industry, and any consequent need for new standards. They produced a generic
reference diagram for the roles in the information industry, separated into ‘structural
roles’ providing the information routeing, supported by ‘Infrastructural roles’ providing the supporting infrastructure of terminal and network equipment. Network
operators provide the infrastructural role of ‘Telecommunication Service provision’.
This ETSI Strategic Review reference diagram is shown in Fig. 11.3; it represents a
good example of a commercial or service architectural view.
11.2.2 Techno-regulatory view
The second architectural perspective shown in Fig. 11.2 is the techno-regulatory view.
This indicates the boundary conditions of the network: the user-network interfaces
on one side and the network–network interfaces, where interconnections to others
networks are made, on the other side. Generally, these two sets of interfaces are
defined under the control of the national or state regulator to ensure competitive
282 Understanding telecommunications networks
and networking
of information
of information
processing and storage
service provision
Equipment supply
Figure 11.3
ETSI SRC 6 Roles in Information Industry
openness. Thus, users can employ any make of terminal devices provided they comply
with the users–network interface specification. Similarly, other network operators and
independent services providers will be assured of technical inter-operability when
connecting to the network.
In addition to the physical and electrical interface characteristics, there may be
application – or software – levels of interfacing that have to be specified. Examples of
these for network–network interfaces are the message sets for signalling between the
networks, e.g. SS7 or SIP and API (application programming interface), as described
in Chapter 7. These APIs may also apply as software/interfaces at the user–network
11.2.3 Functional or logical view
The functional or logical view (Fig. 11.2) describes contributions made by the various elements of a network. In this perspective the elements are deemed to provide a
‘service’ to other elements of the network. Generally, each element is supported by
the service of a lower-level element and, in turn, it provides a service to a higher-level
element. Therefore, these elements tend to be considered as residing in layers. (Engineers love to analyse complex situations into multi-layered functional architectural
views!) Examples of this form of architectural view that are described in this book
are the OSI 7-layer model (Chapter 8 and Fig. 8.4) and the 4-layer model of SS7
(Chapter 7 and Fig. 7.8).
Putting it all together 283
11.2.4 Physical view
The fourth view is that of the physical perspective, as shown in Fig. 11.2, which is
probably the most commonly understood manifestation of architecture. This view
shows the physical joining of the various boxes of equipment. Most of the blockschematic diagrams used throughout this book are presenting physical views.
These views are really just different ways of looking at the same thing. They may
be likened to looking at a mountain from four directions – say, North, East, South
and West – each view being a different way of seeing the same mountain. We return
to the use of architectural views several times in this chapter.
11.3 A holistic view of a telecommunications network
We will now use two architectural perspectives – logical and physical – to understand
how all the component networks considered in this book fit together to form a PSTN.
Such holistic pictures are important for the understanding of how the overall network
performs, how it provides services, and how it can be developed in the future.
11.3.1 Logical multi-layered network views of a PSTN
Fig. 11.4 presents a multi-layered logical architectural view of the component networks of a typical PSTN [3]. The lowest layer, the Transmission-Bearer network,
provides the transmission capability used by all higher layers – thus acting as a transmission utility. This layer contains the copper and optical fibre cables, as well as
Control and monitor
SS7 NCR links
SS7 links (CR)
Traffic routes
Circuit switching,
ATM switching,
IP routeing networks
Figure 11.4
The Multi-Layer View of a PSTN
284 Understanding telecommunications networks
any radio links forming the access network. It also contains the junction and trunk
circuits between the network nodes of the core transmission network. As described
in Chapter 5, the nodes of the Core Transmission Network (i.e. the CTSs) provide
points for interconnection of circuits or groups of multiplexed circuits (this is known
as ‘transmission flexibility’). About half of the CTSs are co-sited with the switching
equipment, i.e. both are in the exchange building; the remainder are transmissiononly nodes, housed in transmission equipment buildings. This is reflected in Fig. 11.4;
when tracking the nodes in the bottom two layers it can be seen that there are more
nodes in the transmission bearer network than in the switching layer above.
The switching layer contains all the switching nodes: remote subscriber concentrators, local exchanges with co-sited subscriber concentrators, junction tandem
exchanges, trunk exchanges and international exchanges. Between the nodes, shown
as dotted lines in Fig. 11.4, are the traffic routes. Of course, the transmission paths
for these traffic routes are physically carried in the bottom layer; although logically
the traffic routes are in the switching layer. This separation of physical and logical
functionality is an important architectural concept.
The next layer up is that of the signalling network. Between the subscribers and
the serving concentrator (whether co-sited or remote) are the various subscribersignalling systems (Chapter 7). Between the exchange nodes are SS7 links. Again,
logically the SS7 links are in this third layer, even though physically their transmission
is within the time slot 16s of 2 Mbit/s digital systems carried in the TransmissionBearer Network Layer.
At the fourth layer of Fig. 11.4, we have the synchronisation network, comprising
synchronisation nodes located at all the digital exchange and data nodes, linked by a
set of synchronisation links. The role of this network layer is the control of the speed
of the digital clocks at each of the nodes. As described in Chapter 4, digital TDM
networks require all participating switching and multiplexing nodes (voice or data)
to be synchronised. The actual arrangements for synchronising may differ between
national networks, ranging from disseminating the timing from a central atomic clock
through a mesh of synchronisation links and control nodes [4,5] to the reception at
each network node of the global positioning satellite (GPS) system timing pulses.
Again, the transmission of the synchronisation links or satellite links is physically
carried in the transmission utility of the bottom layer.
Intelligence for advanced call control, in the form of IN SCPs (intelligent-network
service-control points), and the control links to the switching nodes is provided by
the fifth layer of Fig. 11.4, (Chapter 7). The control links are provided by SS7 noncircuit related signalling, carried physically in time slot 16s of 2 Mbit/s systems in
the transmission bearer network of layer one.
Finally, overseeing all of these component networks to the PSTN shown in
Fig. 11.4 is the administration network at layer six. This network comprises the set of
network-management centres and their communication and control links, which perform the function of managing the operation of all the lower network layers. We will
consider the network management function in a little more detail later in this chapter.
The importance of this logical model of a typical PSTN is that it clearly demonstrates the role of the component networks and their functional inter-relationships.
Putting it all together 285
In practice, PSTN operators tend to manage the component networks separately – with
different planning and operating organisations and cultures dealing with each level.
However, of course, it is the combined operation of all the component networks – i.e.
across all six layers of the architectural model – that produces the services provided
to the PSTN operator’s customers.
11.3.2 Physical view of the set of a Telco’s networks
Fig. 11.5 presents a physical view of all the networks associated with a typical telecommunication network operator, i.e. a so-called Telco. The view shows on the left-hand
side the various forms of subscribers or customers: mobile, residential and small business (fixed), and medium and large business fixed customers – with the first vertical
line indicating the user-network boundary for each of the categories. The second and
third vertical lines indicate the boundary between the access network and the core
networks, i.e. the location of the aggregator nodes. In the case of the PSTN, the subscriber’s concentrator unit takes this role, being either remotely located or co-sited
with the parent local exchange (LE). Within the Core Network there are two sets of
nodes located at the third and fourth vertical line. For the PSTN, the first group are
the junction tandems (JT) or trunk exchanges (TE); in the second group are just TEs.
Finally, the TEs are connected to international exchanges. At the extreme right-hand
side is the final vertical line indicating the important network–network boundary,
the interface between the Telco’s network and the other PNO and service providers
S *
Other operators
ATM data
and large
Res. and
Digital private circuits
and JT
Other operators
IN and
Junction SDH
Junction PDH
Access SDH
Access PDH
Intelligent network
* Note: DSLAMs may also be located at Remote Concentrators (omitted for clarity)
Figure 11.5
Typical Set of Telco Networks
286 Understanding telecommunications networks
(e.g. ISPs). The PSTN comprises the network terminating point (at the user-network
boundary), the Access Network, LE, JT, trunk, and international exchanges.
As Fig. 11.5 shows, the parent trunk exchange acts as a gateway to many other
networks and non-PSTN exchanges. These include other operators (PNO), the mobile
switching centres (MSC), auto-manual centres (which provide operator-assistance,
including emergency [999 or 112] and directory enquiry services), as well as the
Intelligent Network (IN) service control points (SCP). Within the Core Network all
links between nodes are carried over digital SDH or PDH transmission on optical fibre
cables or microwave point-to-point radio. These forms of transmission are also used
where high capacity links are required in the Access Network, e.g. for high-speed
private circuits, ISDN links to digital PBX (private branch exchanges) or high-speed
data services – all provided to business customers’ offices.
Also shown are the ATM switches and IP routers providing the core data network
capacity, used by a variety of access methods: ‘broadband’ over ADSL (with DSLAM
at the aggregator node), optical fibre and copper cables. High-speed digital private
circuits are directly carried over PDH or SDH transmission links on the access and core
networks with manual patching or automatic cross-connecting at the CTSs. Lowerspeed digital private circuits are automatically patched via digital cross-connection
units (DXC) located in exchanges , shown as ‘PC’ in Fig. 11.5.
For clarity, some of the other nodes that might be in a Telco’s existing set of networks have been omitted from Fig. 11.5, e.g. special business networks for Centrex
and VPN services (see Chapter 2). However, despite the inevitable simplifications,
this physical view does convey the complexity and variety of the set of Telco’s networks serving a range of customer types, with a wide portfolio of services. We will
consider this complexity and variety when we look at the issues around enhancing
and modernising the networks, later in this chapter.
Quality of service and network performance
A primary objective for a network operator or service provider is to provide its customers or users with the required quality of service (QOS). Of course, users want as
high a QOS as possible, but the operators need to balance the manpower and equipment costs of meeting certain standards against the price that customers are prepared
to pay. Usually, there is a clear statement by the operators in the contracts for service
describing the quality that customers can expect – there may even be penalty clauses
or promises of compensation when these levels are not met. Broadly, QOS covers
both network-related as well as non-network related aspects. Examples of the latter
include: the time between receiving an order and the provision of a service to a new
customer, time to repair faults, and accuracy and convenience of billing. Networkrelated aspects (e.g. congestion, clarity of a call connection) are dependent on the way
that the network performs. As indicated earlier in this chapter, the performance of
the network perceived by the user results from the contribution of all the component
networks involved in providing the service end-to-end.
Putting it all together 287
Quality of service
Support systems
and processes
Transmission performance
Figure 11.6
Grade of service
Factors Affecting QOS
Fig. 11.6 illustrates the inter-relationships of the factors affecting QOS, showing the non-network and network-related factors [6]. The network performance is
dependent first on the extent that the network is available for service, known as
the ‘availability’, measured as a proportion of a year (e.g. 99.99 per cent). This is
influenced by the inherent reliability of the equipment used in the network and the
extent of redundancy in the network structure and the use of protection switching (see
Chapter 5). The availability is also dependent on the speed with which service can be
restored following equipment failure, i.e. the average time to repair faults or provide
an alternative way of carrying the traffic.
The other major factor influencing network performance is the so-called trafficability, which is a measure of the ease with which telephone calls, data packets or even
permanent transmission paths for private circuits, can be routed through the network.
The level of trafficability for telephone service is set by measures such as GOS, giving the probability of calls failing because of network congestion (see Chapter 6). In
addition, the extent of interconnectivity of the network, i.e. the degree of meshing
between network nodes and the opportunities for patching or switching in alternative
paths for private circuits or transmission routes between network nodes, has a major
impact on trafficability.
We now briefly consider the main parameters used in measuring network
performance [7].
11.4.1 Transmission loss and loudness in the PSTN
Probably the most obvious performance parameter for a PSTN is the loudness of
the telephone call connection, as perceived by the users. The loudness of a telephone
call is set by the characteristics of the telephone instruments at either end and the
288 Understanding telecommunications networks
Network loss = x dB
Network loss = x dB
Figure 11.7
Transmission Loss in the PSTN
transmission loss between them through the network connection. Fig. 11.7 shows
the arrangement, with the two directions of transmission separated. Each type of telephone terminal has a characteristic inherent loudness as generated by its microphone –
the sending level – and by its earpiece – the receiving level. Since loudness is a subjective measure for humans, panels of observers are used in a testing environment
to evaluate the sending and receiving loudness ratings for each type of telephone
terminal – known as ‘SLR’ and ‘RLR’, respectively.
Referring to Fig. 11.7, it can be seen that the overall loudness rating as heard by
subscriber B is made up of the sending rating of telephone A, SLR(A), the network
loss across the network, x dB (see Box 4.1 in Chapter 4 for an explanation of dB
units), and the receiving rating of telephone B, RLR(B). In the reverse direction, the
loudness rating as heard by subscriber A is made up of: SLR(B) + x + RLR(A) dB.
Operators design their networks (whether fixed or mobile) to work within an
acceptable range of overall loudness, recognising the range of loud and soft speakers
in the populations of users, as well as the range of telephone terminals available.
Again, based on the views of panels of speakers and listeners the acceptable range of
overall loudness is determined to be between the upper levels, where 50 per cent of
customers have difficulty because of it being too loud, and the lowest level, where
50 per cent of customers would have difficulty because of it being too quiet.
11.4.2 Transmission stability
Fig. 11.8a shows simplified view of a PSTN, with 2-wire copper local loops at each
end of a 4-wire (i.e. separate Go and Return paths) switch and transmission core
Putting it all together 289
6 dB loss
12 dB
local loop
local loop
6 dB loss
Figure 11.8
network. At the point of the two-to-four-wire conversion (provided by the ‘hybrid
transformers’, as described in Chapter 1) there is, in practice, due to imperfect setup the possibility of some of the send power breaking into the receive direction,
and vice versa. If the 4-wire network is operated at too low a loss overall, the breakthrough electrical signals would circulate in the loop created between the two 2/4-wire
conversion points. Subscribers would hear an annoying screech – confusingly known
technically in electrical engineering parlance as ‘ringing’ – and the circuit would be
deemed unstable.
Stability in a PSTN – or any network using a mix of four and two wire circuits –
is ensured by operating the 4-wire portion of the network at a 6 dB loss (i.e. 6 dB).
This ensures that any circulating break-through signals are reduced, (‘attenuated’) by
12 dB per revolution, rapidly eliminating any ringing. This 6 dB loss is introduced at
the subscriber line card in the serving digital local exchange (see Chapter 6). Similarly,
if two-wire circuits are used with VOIP equipment 6 dB of loss needs to be inserted
by the codec in the media gateway device (see Chapter 8).
11.4.3 Echo and delay
Apart from the screeching effect (i.e. ringing) caused by unattenuated circulating signals, there is the problem of echo paths also being set up. Echo becomes annoying
for the telephone users when the propagation delay between the directly transmitted and break-through circulating voice signals becomes appreciable. Two sets of
paths are established: the talker echo-path and the listener echo-path, as shown in
Fig. 11.9. Normally, the loss inserted at the subscriber line card for stability purposes,
described above, also reduces the level of both types of echo signals sufficiently to
give satisfactory quality for telephony service.
All transmission of an electrical current incurs a propagation delay, the extent
depending on the type of medium and the distance involved. Box 11.1 shows details
of typical transmission system propagation delays. In addition, nodal equipment such
as multiplexors, digital circuit switches, packet routers, etc., introduce their own
delays. However, if the level of propagation delay through the network is above that
of a standard digital PSTN, say due to the use of ATM packet switching for voice
290 Understanding telecommunications networks
Listener echo path
Talker echo path
local loop
local loop
Figure 11.9
Box 11.1
Echo Paths
Contributions To Network Delay [8]
2 Mbit/s pair cable
140 Mbit/s optical fibre
140 Mbit/s microwave radio link
Digital multiplexor–demultiplexor
Digital trunk exchange
Digital local exchange
Geostationary satellite
4.3 μs/km typical
4.9 μs/km typical
3.3 μs/km typical
1 to 2 μs typical
450 μs (International recommendation)
1.5 ms (International recommendation)
260 ms
(see Chapter 8) or satellite transmission, then special echo-control equipment is
normally required. This is usually located with the offending systems in the network.
It should be noted that codecs used in mobile handsets and many of the voice over
IP (VOIP) terminals use special computer processing systems to reduce the required
bit rates below the normal standard of 64 kbit/s. Such processing introduces extra
delay which is perceptible to telephony users. With certain combinations of mobile
handsets and VOIP networks, the resulting total delay can cause annoyance to users –
although, it will not introduce echo if no 2-wire copper is used in the connection since
2-4 wire conversion is avoided.
11.4.4 Digital errors
An important performance characteristic of any digital system used in telecommunication networks, whether it deals with data or encoded voice, is that of digital errors.
These can result from a variety of causes, including minor timing and sampling
glitches, impulsive noise causing distortion to the signal and equipment malfunction.
Digital errors may occur in just one bit, i.e. a binary 0 being corrupted to a binary 1, or
vice versa, or a cluster of bits being in error. The latter is often referred to as an ‘error
burst’. There are techniques using extra information inserted at the sending end to the
Putting it all together 291
Box 11.2
Digital Network Error Standards
G821 Recommendation
Errored seconds
A time period of 1 s with 1
or more errors
<1.2% of 1 s intervals to
have any errors
Severely errored
A time period of 1 s with
65 or more errors
99.935% of 1 s periods with
error ratio better than 10∧−3
Degraded minutes
A time period of 1 minute
with 5 or more errors
98.5% of 1 minute periods
with error ratio better
than 10∧−6
data packets or transmission stream, which enable errors caused by the transmission
medium to be detected, and even corrected, at the receiving end, as described in
Chapters 4, 7 and 8. However, these error-detecting and correcting systems can cope
with only a limited range of errors. When the error rate exceeds these limits, whole
packets need to be resent in the case of packet switched networks, or the whole system
has to be shut down in the case of TDM digital switching and transmission systems.
For example, the latter would occur when the error rate on a digital transmission
system over an optical fibre degraded to one error in every 100,000 bits (i.e. 1 in
10∧ 5) or worse.
Box 11.2 presents the ITU international recommendation (or standard) G821 on
the values of various digital error performance parameters at which the networks
should operate. Specific causes of error resulting from timing problems between the
sender and receivers on a digital transmission system include:
Jitter. The very-short term variation in timing (equivalent to ‘flutter’ in an audio
HiFi system).
Wander. The medium-term variation in timing (equivalent to ‘wow’ in an audio
HiFi system).
Slip. Where complete TDM frames (Chapter 4 refers) are dumped or repeated
(i.e. ‘slipped’) in order to compensate for the accumulated long term differences
in timing between the node and the rest of the network [8].
There are many other parameters of network performance used by operators to
help manage the network at an acceptable quality level. For example, the performance
of call processing by an exchange control system is measured by the delay to dial tone
and connection-establishment delay. The performance measurement of ATM packet
switched service uses several parameters, e.g. cell transfer delay (CTD) and cell-error
ratio (CER).
292 Understanding telecommunications networks
Telecommunications network
E Local
Local E
Elements contributing to network performance
Overall network performance
10 dB
6 dB
10 dB
8 ms
7 ms
8 ms
0.3 ms
Figure 11.10
0.1 ms
0.9 ms
4.4 ms
0.9 ms
0.1 ms
0.3 ms
Apportionment of Performance Parameters
11.4.5 Apportionment of performance impairments
Finally, we come to the, perhaps strange, concept of apportioning during the design
of a network the performance impairments that each of the various network elements
may incur. The network is then operated and maintained at this performance level.
This apportionment must be based on the eternal balance between acceptable cost
and required quality. The overall network performance offered by a network operator
(which may or may not match customer’s expectations!) is usually based on international standards and historical levels of achievements. Each overall performance
parameter is then allocated to the contributing network elements in a way that minimises the total capital cost of equipment and the operational cost of the technician
manpower to maintain the network to that level of performance. This usually means
that the Core Network, with its higher-loaded circuits and more centralised operations, is required to perform to higher levels than the disparate more voluminous
Access Network (e.g. 20 to 30 millions lines).
Fig.11.10 illustrates the apportionment of two important parameters, loss and
delay, to the PSTN, as discussed in the following section.
The overall reference equivalent (ORE) for acceptable telephone calls is around 37 dB,
of which the two telephone terminals contribute a total of 11 dB. As explained earlier,
the overall loss of a 4-wire digital PSTN is set at 6 dB to ensure stability and to
minimise echo problems. This leaves some 20 dB maximum loss for the sum of the
copper local loops at each end. The local loop contains a wide range of cable lengths
and gauges, with subscribers at differing distances from the exchange (or remote
concentrator). Typically, the maximum loss of a local loop is set at 10 dB – although
Putting it all together 293
many subscribers’ lines incur lower losses. With telephone terminals typically having
a sending reference equivalent of 10 dB SRE and a receiving reference equivalent of
1 dB RRE , this causes the worst case loss end-to-end to be 20 + 6 + 1 + 10 = 37 dB
RE. However, the majority of telephone calls would incur some 10 dB less overall
loss than this worst case [8].
The overall limit of acceptable end-to-end delay for a telephone call through a PSTN
is 23 ms (i.e. 23 thousandths of a second). The allowed contribution of all the network
elements for the most complex call (i.e. highest number of components incurred in
the call) is shown in Fig. 11.10 [8].
It must be emphasised that the situation shown in Fig. 11.10 assumes just one
network involved in the call connection. However, many calls are carried over several
different networks, e.g. from a mobile (‘cellphone’) via an alternative network carrier
and terminated on the incumbent operator’s PSTN and local loop. The apportionment
of performance to the various participating networks, of course, should total to an
overall level which is acceptable to users. In practice, given the wide variety of
mobile and fixed telephone terminals and network operators, there may occasionally
be combinations that cause difficulty for certain users (e.g. too much delay or not
loud enough for listening clarity).
Undoubtedly, the main attribute that differentiates one network operator from another
is the quality of their management of the operations of the networks and support systems, and the day-to-day activity of dealing with customer-service interactions. This
is because all network operators have the ability to use the same range of technology in their equipment, so it is the way that the networks and services are operated
that makes the difference. Curiously, this seemingly obvious factor can be missed
by successive generations of managers working for network operators who tend to
concentrate on the choice of network technology and neglect the consideration of the
less glamorous aspects of how a new service or network is to be managed. This has
resulted many times in cases where new services or network enhancements have been
launched with much publicity, only to be temporarily withdrawn shortly afterwards
for some months while the operational aspects are sorted out. By the same token, the
lead time (see Chapter 6) to introduce a new service, network or new technology in
an existing network, can be dominated by the time taken to develop the necessary
new operations-support systems.
Clearly, we need to include a view on operations in this chapter’s consideration of
putting all the pieces together. Not surprisingly, there are many books and technical
journals dedicated to this vast subject (e.g. References 8–12) and we can provide only
a brief overview here.
Operations covers all the day-to-day activities involved in running a public or private telecommunications network. This includes the planning, design and installation
of equipment into the network and its upgrades and expansion – generally referred
to as ‘design and build’ functions. It also includes the providing of service to new
294 Understanding telecommunications networks
customers (known as ‘provisioning’) and the management of the performance of the
network – known as ‘provide and maintain’ functions. These two sets of functions,
which are focused on the network, are normally considered to provide the so-called
role of ‘network management’.
For any national operator, network management is a vast undertaking. It involves
the control of many thousands or millions of lines, multiplexors, subscriber concentrator switches, digital transmission line systems, local exchanges, junction tandems,
trunk exchanges, etc. Each piece of equipment has to be configured, assigned to a
customer or for common use in the network, its performance monitored, faults have
to be repaired and the equipment brought back into service. Just keeping an up-to-date
inventory of all the equipment identities, their location, their status, etc., is a huge task
in itself. Although historically the records used were paper-based with manual tracking, increasingly now operators use computer-based network management systems.
Thus, the network management function is delivered through a range of large computer systems running programs that enable technicians at several dispersed centres
within the country to control remotely whole regions of the network. Such systems
require large data bases to hold all the inventory and status information of the equipment in the catchment area. Real-time monitoring and remote control is provided by
the extensive deployment of control links from the various network elements (i.e.
equipment) to network-management centres. Typically, the latter are referred to as
‘operations and maintenance centres’ (OMCs). The control-links from the equipment
and the OMCs are deemed to reside in the top layer (Administrative Layer) of the
multi-layered model of Fig. 11.4, described earlier in this chapter.
OMCs provide a range of functions, typically including:
(i) Remote monitoring of alarms from exchanges and Core transmission systems.
(ii) Remote access to the exchange-control systems to:
• change the status or features of a subscriber’s line;
• initiate a new subscriber’s line;
• monitor a subscriber’s line;
• change the telephone number of a subscriber’s line;
• change the contents of the exchange routeing codes and tables;
• set software changes and upgrades;
• manage software restoration actions;
• install software builds.
(iii) Monitoring of unmanned exchange and Core-transmission buildings for
intruder and fire alarms, etc.
(iv) Remote collection of traffic usage information (to be used for dimensioning
and forecasting of growth in demand, traffic dispersion, etc.).
(v) Remote collection of call-record data from the exchanges for forwarding to
separate billing centres.
In addition, there are separate network management centres which monitor the whole
national network, the links to other operators and the international links to other
countries. These national or regional control centres have the responsibility of overall
control of the network performance. Importantly, it is the technicians at these centres
Putting it all together 295
that have the ability to initiate remedial action across the range of transmission systems
and exchanges to cope with major breakdowns or traffic overloads in the network.
Where there is advanced warning of likely telephone traffic surges – e.g. as a result
of televised telephone voting, where a massive number of calls can be expected to a
single number during a short time of day – technicians at the network control centres
can initiate re-routeing of calls and other measures, such as call gapping, to limit the
extent of overload and ameliorate the effects on the QOS for the customers. In the
case of call gapping, the control systems of the offending local exchanges are set to
switch only a limited proportion of calls to the overloaded destination, e.g. one call
every 5 seconds [13].
In practice, the management of the network is undertaken at two levels. The
first level is that of the management of individual pieces of network equipment or
‘elements’, e.g. cables, multiplexors, line systems, cross-connects, exchanges, signalling systems and intelligent network data bases. So-called element managers are
control systems, usually computer-based, that are specific to the particular elements’
technology. For example, an SDH add-drop multiplexor (ADM) controller is used
to configure the ports on all the ADMs supplied by a particular manufacturer. Normally, element controllers are able to extend remote control to all the many elements
within an area – typically a region within a country, as set by practical constraints
or organisational boundaries of the network operator. In addition to managing the
configuration of the equipment, element managers usually also monitor one or more
performance parameters (e.g. digital error rate) and any fault alarms or system error
messages. Element managers are usually located in operational buildings, such as
exchanges or Core Transmission Stations.
The second level of network management is at the overall network level, having
end-to-end control for that particular network. Examples include: the full network
view of private circuits (or leased lines), telephone calls and ATM cell routeing.
Generally, network managers, which are also computer-based systems, coordinate
the outputs from all the element mangers involved in the network so that a total
overview is obtained. It is these network management systems that are located in the
network management centres described above.
So far we have considered only the management of the network itself, but there is a
further range of operations associated with managing the interactions with customers –
usually referred to as ‘customer service’or ‘service management’. There are four main
areas of service management, namely: order taking, fault management, provisioning
and billing. The key aspect of service management is that it involves providing an
interaction with customers. This is provided by service centres which are contacted
by customers through telephone calls, e-mails, fax, web sites or even in person. The
support systems for service management employ large-scale computing with massive
data bases.
Fig. 11.11 presents a summary top-level view of the widely accepted logical
architecture for operations management, which is structured as a five-layer hierarchy. At each layer a distinct set of operational activities are undertaken by groups of
people, using dedicated computer support systems, associated with the relevant data
bases, following prescribed processes and providing outputs for different recipients.
296 Understanding telecommunications networks
• Budget
• Cost tracking
• Business planning
• P and L management
• Project management
• Order handling
• Fault report management
• Billing
• Account management
• Customer records
• Provisioning
• Network Operator’s
• Customers and users
Support Processes
Network Control Layer
• Performance management
• Routeing table build
• Remedial action
• Restoration and fault mgt. Support
• Resource management
• Software builds
Element Control Layer
• Alarm management
• Configuration management
• Assignment management
• Network-element controllers
• Other network operator’s
• Network elements
Support Processes
Network Elements
Figure 11.11
Operations Management Hierarchy
At the base of the hierarchy are all the network elements, i.e. the network itself. As
described above, these are managed by technicians at network buildings using element
controllers, shown as the second layer in the architecture. The element controllers
are, in turn, managed on an end-to-end basis by the network management centres,
which are deemed to sit in the third or network-control layer. Above this is the servicemanagement layer providing the interface to the customers of the network service.
Finally, there is a top layer which comprises all the activities associated with managing the operator’s business. This includes functions such as budget build, financial
tracking of expenditures within the organisation – particularly expenditure on network equipment! – human-resource management, payment of salaries, invoicing and
treasury functions, etc. Generally, the element-control layer and the network-control
layer are assumed to act as a combined network-management function.
Whilst the five-layer architectural view of Fig. 11.11 helps define the various
categories of activities involved in managing a telecommunications-network-operator
business, it does not provide a structure for the design of the vast range of support
systems and their data bases and the process associated with execution. However,
the Telecommunications Management Forum (TMF), which includes representation
from network operators and equipment manufacturers worldwide, has addressed this
problem. The TMF have developed the so-called FAB model to help the industry
agree on how the set of activities or processes involved in providing network services
should be structured. Fig. 11.12 presents the model, which identifies three sets of
processes: those associated with customer care, service development and operations
and network and systems management. Its name is derived from the three fundamental
Putting it all together 297
Customer QoS
Invoicing /
Customer Care Processes
and planning
Rating and
Service development and operations processes
and planning
Network and systems management processes
Figure 11.12
TeleManagment Forums’s FAB Model
categories of action that the processes support:
• Fulfilment, covering all that is involved in provisioning service to a customer.
• Assurance, covering the management of the QOS received.
• Billing, including all that has to be done in the network and the support systems.
The TMF have incorporated the FAB (Fulfilment, Assurance, Billing) model into
a comprehensive functional view covering all the processes of a network operator,
known as the ‘Extended Telecommunications Operations Map’ – or ‘eTOM’. This
gives an all-embracing view of the processes associated with element management,
network management, service management and business or enterprise management –
covering not only the operational (day to day) aspects, but also the longer-term strategy, planning and development aspects. Fig. 11.13 presents the generic picture of
eTOM, with three distinct domains identified: the first is Strategy, infrastructure and
product, the second is Operations and the third is Enterprise (or business) management. For the first two domains, the processes are grouped into those that specify
what is required in interfacing to customers (shown as the top pair of horizontals);
those processes associated with service management, i.e. day-to-day interactions with
customers (shown as the second pair of horizontals); those processes associated with
managing the computer support equipment and the network equipment (the third
pair of horizontals); and finally the bottom pair of horizontals contain the processes
associated with managing the relationship with equipment suppliers and operational
298 Understanding telecommunications networks
Strategy, infrastructure and product
Strategy and
support and
Marketing and offer management
Customer relationship management
Service development and management
Service management and operations
Resource development and management
(application, computing and network)
Resource management and operations
(application, computing and network)
Supply chain development
and management
Supplier/partner relationship management
Figure 11.13
Strategic and
Brand management,
market research and
Financial and asset
Human resource
Stakeholder and external
relations management
Research and
Disaster recovery,
security and fraud
Enterprise quality
management, process and IT
planning and architectures
Extended Telecommunications Operations Map (eTOM)
All the relevant processes in the horizontals are fully described in the eTOM
(although for clarity not shown in Fig. 11.13). For example, service management and
operations processes for billing are located in the intersection of the second horizontal
and final column of eTOM. It is through commonly agreed and understood models like
eTOM that the network operators are able to buy from a number of suppliers standardised off-the-shelf software support systems for their business, service and network
management processes, rather than incur the cost and complexity of developing their
own bespoke systems. As a result, most network operators have undertaken the huge
task of rationalising their operations processes and support-system architectures so
as to conform to eTOM, and there is a general preference for the use of off-the-shelf
operational support systems.
Network evolution
In looking at how all the pieces fit together we need to consider the way that all the
technologies used for the network elements go through a life cycle from inception
and small scale deployment, through progressive growth to the mass deployment
of a mature technology, ending with obsolescence and withdrawal from the network. The shape of these life cycles is similar to that of the well-known products
life cycle, described in all text books on marketing. As with products, the useful
life of a technology as an element within the network is usually extended by progressive enhancements to keep it competitive compared to the newer technology
Putting it all together 299
network era
Analogue automatic
network era
Digital network
Figure 11.14
Technology Product Life Cycles
alternatives constantly becoming available to the industry. Eventually, of course, one
of the emerging new technologies becomes so attractive in terms of what it can do
(i.e. its ‘functionality’) or its costs, that it starts to be adopted by network operators in
preference to continuing to install established technology for growth in capacity. Once
the new technology becomes mature it is also used to replace the old (i.e. previous
technology) equipment in the network.
Fig. 11.14 shows a succession of life cycles for different generations of telephone
switching systems, showing how each is eventually superseded by a new technology,
and the overlap period when the established technology is replaced by the new technology. For illustration purposes the relevant years are shown in the figure, but these
are indicative only. Each of the technologies – manual switch-boards, two-motion
selector automatic analogue switching (using electro-mechanical technology), semielectronic analogue (using reed-relay switches with electronic control), and digital
SPC switching (using semi-conductor electronic switching and control) – has undergone significant enhancements to perpetuate their useful lives. A corresponding set of
technology life cycles exists for all the different network elements: Access Network
and Core Transmission Systems, signalling systems, intelligent network systems, etc.
The set of technology life cycles for all the network elements applying simultaneously
for a period leads to the concept of ‘network eras’. For example, during the periods of
analogue two-motion selector and semi-electronic switching the transmission systems
were predominantly also analogue (e.g. FDM) with multi-frequency tone signalling
between exchanges – hence this is referred to as the ‘Analogue Network Era’. The
range of technologies described in this book fall into either the established ‘Digital
Network Era’ or the rapidly emerging ‘Data Network Era’, as shown in Fig. 11.14.
Thus, the technology used in the World’s telecommunications networks is currently
being changed from that of one era to the next, as examined in the section below.
Telecommunication networks are constantly in a state of evolution, as new technology is progressively used to improve the cost, quality and functionality of the
300 Understanding telecommunications networks
services offered to the operator’s customers. Of course, wholesale changes of the network equipment are highly disruptive to the day-to-day network operations and the
conversion costs can outweigh the possible advantages in using the new technology.
Thus, there is normally only gradual introduction of new technology, creating a slowly
evolving network, which is interrupted periodically by major shifts in technology,
when there are sufficient reasons for a radical change as a new era is started [14].
Next generation network
The term commonly used to describe the technology of the new data network era is
NGN. Most network operators are in the process of changing or planning to change
their networks progressively towards an NGN. The generally-accepted reasons for
making this transition are as follows.
(i) The adoption of the Internet by people at work and at home has radically
changed the communications needs and expectations of the network operators’
customers. Most day-to-day business transactions are now conducted via email, and the World Wide Web, with the readily available information found
by search engines, provides research facilities, on-line purchasing and transactions, as well as sources of recreation. This combination of computing – or
information technology (IT) – and communications functions are often referred
to as ‘information and communications technology (ICT)’. Descriptions of the
way that the industry and business has changed as a result of ICT include such
expressions as ‘The Internet Economy’ and the ‘Digital Networked Economy’.
An NGN is expected to be better suited to supporting ICT services for the
predominantly data networking needs of customers in this new economy.
(ii) Credible IP-technology-based alternatives to a digital telephone switching systems, as used in the PSTN, have become available. (These are described in
Chapter 8.) An NGN based on this technology therefore offers the advantages of
conveying voice and data directly over a common platform, giving convergence
(i.e. ICT) as well as cost-saving opportunities for network operators.
(iii) The digital circuit-switched telephone exchanges of the PSTN, described in
Chapters 1 and 6, are reaching the end of their economic lives. Network operators therefore need to replace this obsolescent equipment. Thus, an NGN is
viewed also as a way of modernising the PSTN.
The general consensus within the industry on the necessary attributes of an NGN
may be summarised as follows:
(i) Voice and data traffic are supported on the same network;
(ii) Voice and data traffic are integrated at the user’s premises, ideally using multimedia terminals;
(iii) Broadband access;
(iv) Customer services are accessible everywhere – i.e. ‘Any time, any place’;
(v) Capable of supporting several service providers on a single NGN, in a multioperator environment;
Putting it all together 301
(vi) Overall capital and operational cost savings compared to existing generation
Although the details of the technical design of the various versions of NGN differ
between network operators and the equipment manufacturers, there is one overarching
technology that dominates – namely, the use of VOIP as the switching technology
for telephony [15] (Chapter 8 refers). This use of the major alternative to the existing
PSTN technology of circuit switching enables the NGN to be structured around a
network of IP routers which carry not only voice but also the wide range of data
services on the one ‘platform’. (The latter term is frequently used to describe a
deployment of equipment and its network management systems that act as a selfcontained and complete piece of network infrastructure.) Network operators therefore
have the opportunity to replace the PSTN exchanges as well as some of the specialised
data network nodes by deploying an NGN. The resulting reduction in the number
of network platforms, with the consequent savings in operating expenses, meets
necessary attribute (i) listed earlier.
It should be emphasised that the IP platform used in an NGN to support voice
and data services must provide an adequate quality of conveyance for all the services
carried. This means that the basic IP network must have additional equipment that
controls the flow of packets so that those services that do not tolerate delay or loss
are given priority over those that do. In practice this flow and quality management is
provided by the use of virtual paths at layer 2 of the OSI model underpinning the IP
platform – examples being MPLS, ATM, and Ethernet (Chapter 8 refers). A further
important point is that the IP/ MPLS, ATM or Ethernet network is dedicated to the
operator’s NGN and does not form part of the uncontrolled set of IP routers on the
(public) Internet.
Fig. 11.15 shows a generic architectural view for the way that voice is carried
over an NGN. The top of the figure shows how the existing PSTN carries telephony
calls (Chapter 1). It also shows how ADSL provides a separate data path bypassing
the PSTN, which can be used for Internet access via an ISP or, perhaps more profoundly, it provides an alternative to the PSTN by allowing voice in the form of IP
packets to be carried over an ATM data network. However, the NGN itself comprises
the access media gateway (MG), typically located at the local exchange buildings
(physically next to the PSTN subscriber concentrator switch-blocks that they will
eventually replace), and the enhanced edge and core IP routers, as described earlier.
The subscribers are permanently connected to the MGs. In the case of subscribers
using existing conventional telephone terminals, the MG needs to provide the A/D
conversion (using a codec) as well as the packetisation of the voice. Of course, neither of these functions is required for subscriber lines from so-called ‘IP phones’, i.e.
terminals that include codec and packetisation [16]. But for the many years during
the transition to the new era, an NGN needs to cope with customers using existing
phones as well as the new IP phones.
Also shown in Fig. 11.15 are the media gateway controllers (MGC), which provide
call control on the NGN. Any calls to or from existing subscribers on the PSTN are
passed via a trunk media gateway which converts the media flow between IP and
302 Understanding telecommunications networks
Existing Network
IP phone
ATM network
IP core network
IP phone
Figure 11.15
Generic NGN Architecture for Voice
TDM. The Trunk SS7 signalling gateway provides the signalling interconnection
between the MGC of the NGN and the PSTN. (A fuller description of all the above
is provided in Chapter 8.)
Finally, the alternative data path for voice calls routed over the broadband (i.e.
ADSL) link of the subscriber on the existing network can extend through the ATM
network to the enhanced IP network of the NGN, thus providing a further way of
delivering IP voice packets from the existing network to the NGN. Care is required
by the network operator to ensure that such hybrid routeings of voice calls over the
new and old networks do not introduce QOS problems, in the form of delay and
distortion, for the users.
The required broadband access for an NGN’s customers [attribute (iii) above]
is provided by a variety of technologies, depending on the geographical terrain, the
local availability of installed optical fibre, and the customer base being served (Chapters 4 and 5). In the case of serving the residential and small business customers, the
use of high speed electronics over already-installed copper cable between exchange
and subscriber, i.e. ADSL, to provide moderately high speed broadband access is
the favoured technology for most incumbent network operators. Higher speeds are
provided by the use of hybrid optical fibre and copper systems, in which the optical
fibre is installed between the local exchange and street-located electronics in locked
cabinets and the very-high-speed DSL (VDSL) signal carried over the short length
of copper wire to the subscriber’s premises. Business customers will continue to be
connected to the network, whether NGN or existing, via optical fibre point-to-point
or passive optical networking (PON) fibre systems. In other cases, fixed microwave
radio access systems give an alterative means of broadband delivery to the NGN.
Putting it all together 303
It is expected that the ready availability of multi-media and ICT services at high
speeds over the NGN will encourage users to adopt ever-more ‘bandwidth-hungry’
communications services from the network operators. This means that the capacity
of transmission between network nodes will progressively need to increase, making
it worthwhile to deploy new very high capacity optical fibre systems, using DWDM,
i.e. carrying many different colours of light. Furthermore, the core transmission networks could be converted to optical-only cross connects and add-drop multiplexors,
totally eliminating the use of electronics. This will give rise to the establishment of
wavelength or λ (lambda)-routed core transmission networks providing the required
huge digital capacities between major towns for the NGN (Chapter 5).
Although not shown in the simplified Fig. 11.15, NGNs include call and service
control systems that supply the range of ICT services, including customer authentication. Typically, this also involves the use of APIs for third-party service providers
(see Chapter 7). There may also be control links between the NGN control systems and the HLRs of 2G, 2 12 G or 3G mobile networks (Chapter 9), enabling some
new location-based services. The latter are examples of converged services derived
from fixed and mobile networks working closely together, i.e. so-called fixed-mobile
We can now pull together the various technology options described so far for
an NGN and create a generalised logical architecture, as shown in Fig. 11.16. This
presents an architectural view of the relationships between the different functional
OSI Model
(Layers 4–7)
tolerant data
Voice codec
Layer 3
Layer 2
ATM, Ethernet
Layer 1
Layer 0
Cu, Optical fibre, Wireless
Optical fibre, Wireless
Figure 11.16
Generalised Logical Architecture for NGN
304 Understanding telecommunications networks
layers in an NGN and the possible choices of technology for the components. The
architecture shows the different choices for the access and core parts of the NGN, to
reflect the differences in traffic densities and network structure, and hence economics
in the two areas. As Fig. 11.16 shows, the platform layers correspond loosely to
the OSI 7-Layer model (see Chapter 8), with the cables, optical fibres and wireless
links forming the basic Level 0, on which all the communications are carried. At
the top of the architectural view the range of services are characterised as voice,
real-time data and private circuits – all of which are delay or latency intolerant, and
the latency-tolerant data services. All services are carried either over IP packets (at
OSI Layer 3) or over cells or frames at OSI Layer 2. In the case of private circuits
(also known as ‘leased lines’), which in a conventional network are provided over
dedicated transmission paths, a system is required which will manage the spasmodic
nature of the conveyance of the packet or cell layers of an NGN in such a way that the
customer receives a close approximation to the required continuous transmission – a
technique known as ‘circuit emulation’. The private circuits themselves will be used
by the customers to link the nodes of the corporate private networks, carrying voice
and data, as appropriate.
The relationships between the services and platforms or between one platform
and another are indicated by the arrowed lines in the architectural view. It should be
noted that, where appropriate, upper layers may be supported directly by layers two
or three levels below. The way that the contents of one platform are translated into a
lower platform is often referred to as ‘mapping’ or ‘adaptation’. Thus, for example,
bits forming a large IP packet are ‘adapted’ into the many smaller fixed length cells
of an ATM stream.
The actual physical architectures chosen by the various network operators for
their NGNs will differ in detail, with different degrees of integration of services onto
a single network and the extent of replacement of existing platforms. More importantly, the timing and rate of deployment of the NGN will depend on the particular
circumstances applying. Given the current rapid rate of technology developments,
leading-edge operators who deploy NGNs sooner will be committed to technologies
and designs which will differ from those of NGNs deployed later. An example of an
actual NGN physical architecture, namely BT’s 21st Century Network [17], is given
in Box 11.3.
This chapter attempted to piece together all the various elements of a telecommunications network. To this end, the concept of the four architectural views –
commercial model, techno-regulatory, functional (or logical) and physical – was
introduced. The first application of an architectural perspective, viewing the various
component networks of a PSTN as a set of layers, was then used to give an holistic
description. Customers’ perceived QOS and the measures of a network’s performance
were then introduced. These measures cover telephony services as well as digital
transmission and data packet services. This led into the concept of network operations,
Putting it all together 305
Box 11.3 An Example of an NGN Physical Architecture
Fig. 11.17 presents a physical architectural view of the 21st Century Network
(21st CN) design, the NGN being implemented by BT. (This view is derived
from publicly available initial design information and may differ in some details
from the final design.) The logical architecture for 21st CN corresponds to the
generalised view given in Fig. 11.16. The main component is the aggregator
node, located at some 5,500 local or concentrator exchange buildings, i.e. at the
focal points of the copper access network. The aggregation is provided by multiservice access nodes (MSAN), on which all subscriber lines are terminated. The
MSAN contains several types of line card:
(i) A combination POTS and DSL card for converting telephone voice on
copper lines to digital (PCM) samples which are converted to IP packets
by a MG and for terminating the broadband ADSL signals on the copper
loop serving residential and small business subscribers;
(ii) A reduced functionality line card for interfacing to ISDN lines (copper
or optical fibre based);
(iii) Ethernet cards for business subscribers that have their services terminated
on lines using the LAN protocol;
(iv) Circuit-emulation cards for medium speed private circuits. The higherspeed private circuits are directly carried over SDH.
An important factor in the MSAN design is its ability to terminate lines from
subscribers using conventional telephones as well as new data-based terminals
such as IP phones and LAN equipment.
The various services are carried over a set of Ethernet virtual LANS on SDH
or gigabit Ethernet optical fibre transmission systems from the MSANs to parent
metro nodes, of which there will be some 100 around the country. It is these
Metro Nodes that contain the IP Edge routers, which provide the gateway to
the MPLS-enhanced IP core network at the heart of 21st CN. The Metro Nodes
also contain the trunk media gateway and signalling gateway, as described in
Chapter 8, providing the interconnection back to the existing PSTN and other
operators’ (‘PNOs’) networks. Call control and service intelligence is provided
at one of several Data Centres by IN servers – service control points (SCPs) –
and media gateway controllers (MGCs), as described in Chapters 7 and 8,
respectively. Call sessions are set up across 21st CN between the appropriate
MSAN line cards through the use of service-initiation protocol (SIP) signalling
BT’s 21st CN is intended to replace several of the existing network platforms,
including: the PSTN, digital private circuit network, frame-relay network, X25
packet switched network, and the PDH transmission network.
306 Understanding telecommunications networks
Intelligence node
(‘Data centre’)
over SDH
on optical
and SME
IP phone
IP phone
Figure 11.17
SDH over
optical fibre
Edge node
(‘Metro Node’)
Example of an NGN Physical Architecture
and in particular how an operator needs to use support systems to manage the network
equipment on a day-to-day basis – generally encompassed by the term network management – and the management of customer interactions, e.g. order taking, billing,
provisioning and fault reporting, known as service management. The chapter finished
by presenting a view of the way that telecommunications networks evolve and, in
particular, the concept of an NGN.
The Foreword to this book raised the intriguing question of whether the existing
well-established networks providing switched telephone service are to be replaced
by the Internet. We have learnt how the existing network operates, using digital circuit
switched technology and digital transmission over copper, optical fibre or radio, all
under the control of exchange-based processors. However, we have also learnt how the
newer packet-based data platforms can also provide the telephone switched service.
One such platform is the public Internet, which is really a constellation of many
thousands of loosely-linked IP networks made available through a widely adopted
naming and addressing system. Whilst VOIP can be successfully carried over the
Internet, the quality is inevitably dependent on the instantaneous loading by other
users, and so large-scale reliable use of VOIP requires the provision of a dedicated
IP network. Such networks, with suitable quality enhancements, will undoubtedly
form the basis of the NGN. However, as this book has demonstrated, NGNs will still
use much of the existing infrastructure, such as the copper access network and Core
Putting it all together 307
transmission systems. Thus, whilst IP, the technology of the Internet, will form part
of the solution for voice calls in the future there is a continuing need for dedicated
managed network platforms to form the public telecommunications infrastructure.
Perhaps, a bigger question for the future, however, is how will mobile networks,
wireless WiFi and fixed networks inter-relate? Early work on 4G Mobile Technology suggests that they might all be replaced by ad hoc groupings of wireless-based
networks owned and operated by the customers themselves. So, perhaps the more
relevant question is: Wither the network operator?
STEWART, H.: ‘Building an Architectural Framework’, British Telecommunications Engineering, Vol. 13, October 1994, pp. 186–191.
WRIGHT, T.: ‘An Architectural Framework for Networks’, British Telecommunications Engineering, Vol. 14, July 1995, pp. 141–148.
REDMILL, F. J. and VALDAR, A. R.: ‘SPC Digital Telephone Exchanges’, IET
Telecommunications Series No. 21, Stevenage, 1994, Chapter 21.
Ibid., Chapter 10.
BREGNI, S.: ‘Synchronization of Digital Telecommunications Networks’,
John Wiley & Sons Ltd., Chichester, 2002, Chapter 4.
P.: ‘Telecommunications Quality of Service Management: From Legacy To
Emerging Services’, IET Telecommunications Series No. 48, Stevenage, 2003,
Chapter 2.
COOK, G. J.: ‘Network Performance’. Chapter 17 of ‘Telecommunications Networks’, Second edition, edited by FLOOD, J. E., IET Telecommunications Series
No. 36, Stevenage, 1997.
OODAN A. P., WARD K. E. and MULLEE, A. W.: ‘Quality of Service in
Telecommunications’, IET Telecommunications Series No. 39, Stevenage, 1997,
Chapter 10.
FURLEY, N.: ‘The BT Operational Support Systems Architecture Framework’,
British Telecommunications Engineering, Vol. 15, July 1996, pp. 114–121.
GARRISON, R., SPECTOR,A. and DE GROOT, P. C.: ‘The BT Network Traffic
Management System: A Window on the Network’, British Telecommunications
Engineering, Vol. 10, Part 3, Special Issue on Network Management Systems,
October 1991, pp. 222–229.
FAIRLEY, J. M.: ‘Network Management’. Chapter 18 of ‘Telecommunications
Networks’, Second edition, edited by FLOOD, J. E., IET Telecommunications
Series No. 36, Stevenage, 1997.
AIDAROUS, S. and PLEVYAK, T. (Eds): ‘Telecommunications Network
Management into the 21st Century’, IET Telecommunications Series No. 30,
Stevenage, 1996.
FLOOD, J. E.: ‘Numbering, Routing & Charging’, Ibid., Chapter 9.
308 Understanding telecommunications networks
14 VALDAR, A. R., NEWMAN, D., WOOD, R. and GREENOP, D.: ‘A Vision of
the Future Network’, British Telecommunication Engineering, Vol. 11, Part 3,
October 1992, pp. 142–152.
15 PHILPOTT, M.: ‘Migration Strategies to an All-IP Network’, The Journal of
the Telecommunications Network, Vol. 1, Part 3, October–December 2002,
pp. 47–51.
16 WARDLAW, M.: ‘Intelligence and Mobility for BT’s Next Generation Networks’, Ibid., pp. 28–47.
17 BEAL, M.: ‘Evolving Networks for the Future – Delivering BT’s 21st Century
Network’, The Journal of the Communications Network, Vol. 3, Part 4, October–
December 2004, pp. 4–10.
18 REEVE, M. H., BILTON, C., HOLMES, P. E. and BROSS, M.: ‘Networks and
Systems for BT in the 21st Century’. BT Technology Journal, Vol. 23, No.1,
January 2005, pp. 11–14.
19 CRANE, P.: ‘A New Service Infrastructure Architecture’, Ibid. pp. 15–27.
20 LEVY, B.: ‘The Common Capability Approach to New Service Development’,
Ibid., pp. 48–54.
21 STRANG, C. J.: ‘Next Generation Systems Architecture – The Matrix’, Ibid.,
pp. 55–68.
22 BRAZIER, R. W. and COOKSON, M. D.: ‘Intelligence Design Patterns’, Ibid.,
pp. 69–81.
23 CATCHPOLE, A. B.: ‘Corporate Multimedia Communications’, Ibid.,
pp. 98–107.
24 NUNN, A.: ‘Voice Evolution’, Ibid., pp. 120–133.
25 INGHAM, A., GRAHAM, J. and HENDY, P.: ‘Engineering Performance for the
Future of Intelligence’, Ibid., pp. 134–146.
26 STRETCH, R. M. and ADAMS, P. M.: ‘Standards for Intelligent Networks’,
Ibid., pp. 154–159.
Appendix 1
Standards organisations
By its very nature, telecommunications depends on successful interaction between
distant systems and for this to be successful some degree of commonality is required
in specification of the equipment interfaces and protocols used. There are several
organisations addressing the areas of standards for telecommunications networks.
These are organised on a global, regional or national basis. In addition, there are
a range of other organisations, often consortium or forums of interested parties in
the industry that set ad hoc standards, many of which become adopted later by the
more formal regional and global organisations. The key organisations referred to
throughout this book are summarised below.
Global organisations
A1.2.1 International Telecommunications Union (ITU)
This is a specialist agency of the United Nations, comprising some 200 member
countries, responsible for telecommunications standardisation. The organisation is
split into two main areas of interest:
(i) ITU-T covering telecommunications (ITU-T was previously known by the
title of CCITT – Comité Consultatif International de Télégraphique et Téléphonique);
(ii) ITU-R covering radio (ITU-R was previously known by the title of CCIR –
Comité Consultatif International des Radiocommunications).
310 Appendix 1
A1.2.2 International Standards Organisation (ISO)
ISO is a joint organisation with the International Electrical Commission (IEC) responsible for standardisation of information technology (IT). The ISO 7-layer model is
probably the most famous output from this organisation.
Regional organisations
A1.3.1 European Telecommunications Standards Institute (ETSI)
ETSI is the main standards organisation covering the European Union countries.
It follows on from the work of the predecessor organisation CEPT (Conférence
Européenne des Administrations des Postes et des Telecommunications). A notable
output of CEPT and ETSI is the GSM standard for second-generation mobile systems,
now adopted worldwide.
A1.3.2 American National Standards Institute (ANSI)
ANSI is the overarching standards organisation in the United States covering many
areas in addition to telecommunications.
A1.3.3 Institute of Electrical and Electronics Engineers (IEEE)
The IEEE is one of the largest professional societies in the World; among its many
activities is the standardisation for the electrical and electronic engineering field,
including telecommunications. One of the most notable standards is the IEEE 802.11
series covering LANs.
Other organisations
A1.4.1 Internet Engineering Task Force (IETF)
The IETF, which is composed of manufacturers, operators, service providers and
academics, is the main body setting the technical standards for the Internet. Its most
famous output is the TCP/IP protocol.
A1.4.2 Internet Corporation for Assigned Names and Numbers (ICANN)
The ICANN controls the assignment of top-level domain names and other parts of
the naming and numbering scheme for the Internet worldwide.
A1.4.3 Network Management Forum (NMF)
This industry forum sets the standards for service and network management of
telecommunications networks.
Appendix 2
List of ITU-T recommendation E.164 assigned
country codes
Country Code
Country or Geographical Area
USA, Canada and the Caribbean Islands
Cote d’Ivoire
Burkina Faso
Sierra Leone
Central African Republic
Cape Verde
Sao Tome and Principe
312 Appendix 2
Equatorial Guinea
Gabonese Republic
Democratic Republic of the Congo
Diego Garcia
Rwandese Republic
Somali Democratic Republic
South Africa
Faroe Islands
Appendix 2 313
San Marino
Vatican City State
Serbia and Montenegro
Bosnia and Herzegovina
Group of countries, share code
The Former Yugoslav Republic of Macedonia
Czech Republic
Slovak Republic
United Kingdom
Falkland Islands
El Salvador
Costa Rica
314 Appendix 2
Saint Pierre and Miquelon
Argentine Republic
French Guiana
Netherlands Antilles
New Zealand
Democratique Republic of Timor-Leste
Australian External Territories
Brunei Darussalam
Papua New Guinea
Solomon Islands
Wallis and Futuna
Cook Islands
American Samoa
Appendix 2 315
New Caledonia
French Polynesia
Marshal Islands
Russian Federation & Kazakhstan
International Freephone Service
International Shared Cost Service (ISCS)
Korea (Republic of)
Democratic People’s Republic of Korea
Hong Kong, China
Macao, China
Laos (People’s Democratic Republic)
Inmarsat SNAC
Universal Personal Telecommunication Service (UPT)
Reserved for national non-commercial purposes
International Mobile, shared code
International Networks, shared code
Sri Lanka
Syrian Arab Republic
Saudi Arabia
United Arab Emirates
316 Appendix 2
International Premium Rate Service (PRS)
Trial of a proposed new international telecommunication public
correspondence service, share code
Azerbaijani Republic
Kyrgyz Republic
Reserved for possible future use within the
Telecommunications for Disaster Relief (TDR) concept
Authentication, Authorisation and Accounting
ATM Adaptation Layers
Automatic Alternative Routeing
Alternating Current
Automatic Call Distribution
Analogue-to-Digital (conversion)
Add–Drop Multiplexor
Asymmetric Digital Subscriber Line
ATM End-System Addresses
Authority and Format Indicator
Amplitude Modulation
Advanced Mobile Phone System
Access Point
Application Programming Interface
Automatic Protection Systems
ATM over Passive Optical Network
Address Resolution Protocol
US defence department project: the forerunner of the Internet
Autonomous System
Asynchronous Transfer Mode
Administrative Unit
Authentication Centre
Busy Hour Call Attempts
Broadband ISDN User Part
Battery, Over-load protection, Ringing, Supervision, Codec,
Hybrid, Test
Broadband Passive Optical Network
Broadband Radio Access
318 Abbreviations
Base Station
Base Station Controller
British Standards Institute
Base Transmitter Controller
Base-Transceiver Station
see SS7
Channel-associated signalling
Community Antenna TeleVision
Citizen’s Band (radio)
Constant Bit Rate
Common Channel Signalling
Code Division Multiplex
Code Division Multiple Access
Cellular Digital Packet Data
Conférence Européenne des administrations des Postes et des
Cell-Error Ratio
Classless Inter-Domain Routeing
Charge Group
Caller Line Identification
Connection Memory
Communication Network
Class Of Service
Customers Premises Equipment
Central Processing Unit
Carrier-Sense Multiple-Access/Collision Avoidance
Carrier-Sense Multiple-Access/Collision Detection
Cell Transfer Delay
Core Transmission Station
Digital-Advanced Mobile Phone System
Data over Telephone
Direct Current
Data Country Code
Digital Distribution Frame
Direct Dialling In
Digital Enhanced Cordless Telecommunications
Dynamic Host Configuration Protocol
Digital Local Exchange
Digital Junction Switching Unit
Digital Main Switching Unit
Discrete Multi-tone
Data Network Identification Code
Domain Name System
Distribution Point
Abbreviations 319
Digital Private Network Signalling System
Dual Queue Dual Bus
Digital Subscriber Line
Digital Subscriber Line Access Multiplexor
Digital Signal Processor
Domain Specific Part
Data User Part
Dense Wave-division Multiplex
Digital Cross-Connection (unit)
2 Mbit/s PDH digital transmission block
8 Mbit/s PDH digital transmission block
34 Mbit/s PDH digital transmission block
140 Mbit/s PDH digital transmission block
Echo Canceller
Enhanced Data Rates for Global Evolution
Error-Free Second
Extra High frequency
Exterior Gateway Protocol
Equipment Identity Register
Ethernet Passive Optical Network
Extended Telecommunications Operations Map
European Telecommunications Standards Institute
Fulfilment, Assurance, Billing
Frame Alignment Signal
Federal Commission for Communications
Frequency Division Duplex
Frequency Division Multiplex
Frequency Division Multiplex Access
Federation of Electrical Industries
Far End Crosstalk
Frequency Modulation
Fixed–Mobile Convergence
Fixed–Mobile Integration
Full Service Access Network
Frequency-Shift Keying
File Transfer Protocol
Fibre To The Home
Fibre To The Kerb
Fibre To The Office
2 12 G
Second Generation (of mobile networks)
Step between second and third Generation (of mobile networks)
Third Generation (of mobile networks)
Fourth Generation (of mobile networks)
320 Abbreviations
Gateway GPRS Support Node
Gateway Mobile Switching Centre
Grade of Service
General Packet Radio System
Gigabit/s Passive Optical Network
Global Positioning Satellite
Global System for Mobile [formerly Groupe Spécial Mobile]
Hand-over Distribution Frame
High Bit Rate Digital Subscriber Line
High Frequency
Hybrid Fibre-Coax
Home Location Register
High-Speed Circuit-Switched Data System
Hyper-Text Transfer Protocol
Hertz (cycles per second)
Initial Address Message
Internet Corporation for Assigned Names and Numbers
International Code Designator
Information and Communications Technology
Initial Domain Indicator
Integrated Digital Network
Initial Domain Part
Institution of Electrical and Electronic Engineers
Institution of Engineering and Technology [formerly Institution
of Electrical Engineers (IEE)]
Internet Engineering Task Force
Interior Gateway Protocol
Intelligent Network
Intelligent Networks Application Part
International Network Designator
Intelligent Network Database
Internet Exchange
Internet Protocol
IP Virtual Private Network
International Switching Centre
Integrated Services Digital Network
International Standards Organisation
Internet Service Provider
Intermediate Service Part
ISDN User Part
Information Technology
International Telecommunications Union
Interworking Function
Junction Tandem
Abbreviations 321
Local Area Network
Loop Disconnect (signalling)
Local Exchange
Light Emitting Diode
Low Earth Orbit (satellite)
Label Edge Router
Low Frequency
Local Multipoint Distribution Service
Label Switch Router
Line Termination Equipment
Line Termination Unit
Multiple Access
Medium-Access Control
Metropolitan Area Network
Mobile Application Part
Main Distribution Frame
Mobile Exchange
Medium Earth Orbit (satellite)
Multi Frequency (signalling)
Medium Frequency (radio)
Media Gateway
Media Gateway Controller
Media Gateway Control Protocol
Man–Machine Interface
Mobile Network Operator
Moving Picture Experts Group
Multi-Protocol Label Switching
Mobile Station
Multi-Service Access Code
Multi-Service Access Node
Mobile Switching Centre
Multiplex Section – Dedicated Protection Ring
Multi-Service Platform
Mobile Service Provider
Multiplex Section – Shared Protection Ring
Mobile Terminal
Message Transfer Part
Mobile Virtual Network Operator
North American Numbering Plan
Near End Crosstalk
Network File System
322 Abbreviations
Next Generation Network
Next-Generation Switch
Non-Line Of Sight
Network Interface Card
Nordic Mobile Telephone
Network Service Access Point
Network Termination
Network Termination Equipment
Network Terminal Number
Network Termination Point
Network Termination and Test Point
Network Termination Unit
Operations, Administration and Maintenance
Optical Carrier
Orthogonal Frequency Division Multiplex
Optical Line Termination
Operations, Maintenance and Administration Application Part
Operations and Maintenance Centre
Open-Network Provision
Optical Network Termination
Optical Network Unit
Overall Reference Equivalent
Open Systems Interconnection
Open Shortest Path First
Optical Cross Connect
Private Automatic Branch Exchange
Packet Assembler/Disassembler
Pulse Amplitude Modulation
Private Branch Exchange
Private Circuit
Pulse-Code Modulation
Personal Communication Network
Primary Connection Point
Personal Communication Service
Personal Digital Assistant
Plesiochronous Digital Hierarchy
PSTN-to-Internet Gateway
Public Land Mobile Network
Phase Modulation
Public Network Operator
Point of Interconnection
Passive Optical Network
Point Of Presence
Post Office Protocol
Plain Old Telephone Service
Abbreviations 323
Pulse Position Modulation
Point-to-Point Protocol
Phase-Shift Keying
Packet Switch Stream
Public Switched Telephone Network
Permanent Virtual Circuit
Quadrature Amplitude Modulation
Quality Of Service
Reverse Address Resolution Protocol
Routeing Code
Remote Concentrator Unit
Ringing Equivalent Number
Received Loudness Rating
Radio Network Controller
ReSource Reservation Protocol
Real-time Protocol
Signalling Connection Control Part
Service Creation Environment
Secondary Connection Point
Service Control Point
Synchronous Digital Hierarchy
Symmetric Digital Subscriber Line
Signalling Gateway
Serving GPRS Support Node
Super High Frequency
Self-Healing Ring
Subscriber Information Module
Session Initiation Protocol
Switched Label Path
Sending Level Rating
Speech memory
Service-Management Centre
Switched Multi-Megabit Data Service
Small–Medium Enterprise
Short Message Service
Short Message Service Centre
Simple Mail Transfer Protocol
Simple Network Management Protocol
Synchronous Optical Network
Stored-Program Control
Service Protection Network
Sending Reference Equivalent
Signalling System No. 7
324 Abbreviations
Sector Switching Centre
Service Switching Point
Set-Top Box
Subscriber Trunk Dialling
Synchronous Transport Module
Signal Transfer Point
Signal Unit
Switched Virtual Circuit
1.5 Mbit/s PDH digital carrier (USA)
6 Mbit/s PDH digital carrier (USA)
45 Mbit/s PDH digital carrier (USA)
274 Mbit/s PDH digital carrier (USA)
Total Access Communications System
Trans-Atlantic Telecommunications (cable)
Transaction Capabilities Application Part
Transmission Control Protocol
Time Division Duplex
Time Division Multiplex
Time Division Multiple Access
Trunk Exchange
Terrestrial Trunk Radio
Top Level Domain
Telecommunications Management Forum
Telephony over Passive Optical Network
Transmission Repeater Stations
Time Slot
Time–Space–Time (switch block)
Transmission Unit
TU Group
Telephone User Part
User Equipment
User Datagram Protocol
Uniform Domain-Name Dispute Resolution Policy
Ultra High Frequency
Universal Mobile Telecommunications System
Uniform Resource Locator
UMTS Terrestrial Radio Access Network
Variable Bit Rate
Virtual Circuit
Virtual Container (SDH)
Very-High Bit Rate Digital Subscriber Line
Voice Frequency
Very High Frequency
Very Low Frequency
Abbreviations 325
Visitor Location Register
Very Large Scale Integration (integrated circuits)
Voice over ATM
Video On Demand
Voice over Frame Relay
Virtual Path
Virtual Private Network
Virtual Terminal
Wavelength Add–Drop Multiplexor
Wide Area Network
Wide Area Tandem
Wide-Band Code-Division Multiple Access
Wave-Division Multiplex
Wireless Local Area Network
Wireless Regulation Congress
World Wide Web
eXternal Data Representation
generic Digital Subscriber Line
Metric prefixes used in this book
Power of 10
Less than One:
0.001 (one thousandth)
0.000001 (one millionth)
0.000000001 (one billionth)
0.000000000001 (one trillionth)
(10−3 )
(10−6 )
(10−9 )
(10−12 )
Greater than One:
(103 )
(106 )
(109 )
(1012 )
Binary prefixes used in this book
Power of 2
(210 )
(220 )
(230 )
(240 )
The above shows the normally accepted multipliers when applied to computer storage
of data. For example, 2 kilobytes (kB) means 2,048 bytes rather than 2,000 bytes.
Page numbers in italics show pages with tables or figures separated from text.
21st Century Network (21CN) 304–5
access networks 22–3, 99–115, 286
access point (AP) 212
add-drop multiplexors (ADMs) 118–21,
124, 295
address/addressing: see numbering and
advanced mobile phone system (AMPS) 223
aerial cable 23, 69, 87, 126
aggregation points 120, 205
A-law 75
American National Standards Institute
(ANSI) 310
amplitude distortion 63
amplitude modulation (AM) 62
analogue signal 2–4, 17, 44, 55–7, 135, 181
analogue-to-digital conversion (A/D)
55–9, 73
analogue transmission 59, 64, 80, 94, 184
application programming interface (API)
177, 303
about architecture 279–80
architectural diagrams 196, 197
cellular networks 225–6
commercial/service view 281
functional or logical view 282
multi-layered view 283–5
physical view 283, 285–6
techno-regulatory view 281–2
area codes 250–1
assistance, operator 255, 286
asymmetric digital subscriber line (ADSL)
88–9, 113, 205
ADSL2 109, 110
ADSL2+ 109, 110
and local loop unbundling 107–9
noise and crosstalk 109
VOIP over broadband 202–4, 302
asynchronous transfer mode (ATM) 34–5,
188–92, 191, 199, 204–6
ATM adaptation layer (AAL) 189–91
ATM addressing 263–5
ATM cells 195–6
ATM end-system addresses (AESA) 264
ATM over PON (APON) 94
attenuation (power loss) 63
authentication, authorisation and accounting
(AAA) 204
authentication centre (AUC) 231
automatic alternate routeing (AAR) 152
automatic call distribution (ACD) 177
automatic exchanges 8–9
automatic protection systems (APS) 125
autonomous system (AS) 194, 196
bandwidth 15, 48
base station (BS) 22–3, 226–7, 236
base station controller (BSC) 22–3, 226–7,
230, 236–7, 239
base-transceiver station (BTS) 230, 236
basic rate ISDN 88, 144, 145
batteries 1, 6, 101
328 Index
Bell, Alexander Graham 1
bells/ringing current 3–4, 5, 8
billing 24, 113, 165, 232, 286, 298, 306
binary number 58
Bluetooth (IEEE 802.15) 214, 245–6
BORSCHT functions 136, 141, 144, 226–7
British Telecommunications plc (BT) 12–13
interconnection with PNOs 26, 30
broadband 32, 107–12, 189
see also asymmetric digital subscriber line
broadband PON (BPON) 94
broadband radio access (BRA) 111
broadcast TV service 23
business service network (BSN) 37
busy hour call attempts (BHCA) 173
cable (cable TV) networks (CATV) 23, 23–4,
31–2, 154
cable modem 31–2, 107
call connection procedures 14–15
call control: see exchange-control systems;
intelligent network (IN)
call distribution 146, 147
call record 165
capital cost 56
carrier-sense multiple-access/collision
avoidance (CSMA/CA) 212
carrier-sense multiple-access/collision
detection (CSMA/CD) 211
carrier wave 62, 69, 221, 222–3
carrier-wave modulation 62
catchment areas 11
CCITT: see International Telecommunications
Union (ITU)
CCITT No7 (CCSS7, SS7 or C7) 163–9, 172,
176, 178, 201–2, 231, 284
cell hand-over 235–7
cellular networks/systems 223–8
digital enhanced cordless
telecommunications (DECT)
241, 246
enhanced data rates for global evolution
(EDGE) 241
high-speed circuit-switched data (HSCSD)
system 241
terrestrial trunk radio (TETRA)
system 241
see also general packet radio service
(GPRS); global system, mobile
(GSM); mobile networks/systems
central office 6, 12
see also exchanges
Centrex service 13, 37, 39
channel associated signalling (CAS) 161
charge group (GC) 261–2
charge rate 15, 26, 261–2
Internet 27, 30, 31, 32, 241
in the PSTN 24, 261–3
circuit emulation 304–5
circuits, 2-wire/4-wire/conversion 2–5, 17,
18, 61–2, 136, 289–90
circuit switched connection 182, 235
circuit-switching/switching systems
about switching systems 129–30
at DXCs 123
digital exchange structures 141–4
digital telephone switching systems
ISDN exchanges 144–5
private branch exchanges (PBX) 141
subscriber switching (local) units 130–2
classful (IP) addressing 269
classless inter-domain routeing (CIDR) 271
class of service (COS) 208
coaxial cable 69–72
codec (digital coder/decoder) 58, 75, 227–9,
290, 301
code division multiple access (CDMA)
228, 242
code division multiplexing (CDM) 51–2
commercial model 218, 304
common-channel signalling (CCS) systems
76, 162–9
communication providers (CPs): see public
network operators (PNOs)
concentration ratio 49, 149, 205
concentrator switches/switching/switch blocks
42–3, 130–2, 135, 141–4, 150
congestion 145–8, 286–7
connectionless (packet) mode 185–6
connection memory (CN) 138–9
connection-orientated packet mode 183–5
consolidating function 53
constant bit rate (CBR) service 190
contention ratio 205
conversion 2-wire/4-wire 2–5, 17, 18, 61–2,
136, 289–90
conveyance charge 26
copper pairs: see local loop copper pairs
Index 329
core transmission networks (CTNs) 61,
116–26, 284, 285
core transmission station (CTS) 116–21, 135
corporate private network 199, 304
country codes 250–5, 266
crosstalk 88–9, 109
data country code (DCC) 265
datagram service 185
data network identification code (DNIC) 263
data numbering and addressing 263
data services networks 34–5
data switching: see packet switching and
data user part (DUP) 168
Datel (data over telephone) 44
decibel/dBs/dBm 68
delay of signals 63, 293
delay and echo 289–90
dense wave-division multiplex (DWDM) 91,
92, 94, 121, 303
deployment (of networks) 239, 241, 244,
design period 153
digital blocks 76–8
digital coder/decoder (codec) 58
digital cross-connection (DXC) 34, 118–20
circuit switching at 123
digital distribution frames (DDFs) 69, 103–4,
digital enhanced cordless telecommunications
(DECT) system 241, 246
digital errors 290–1
digital exchange structures 141–4
digital junction switching unit (DJSU) 13
digital local exchanges (DLEs) 57, 141
digital main switching unit (DMSU) 12–13
digital microwave point-to-point systems 105
digital networks, advantages 55–7
digital pair gain systems 86
digital signal 2, 44, 54, 57
digital signal processors (DSPs) 87
digital space switching (S) 139–40
digital subscriber line access multiplexor
(DSLAM) 31, 107–8, 205
digital subscriber line (DSL) transmission
systems 87–90
digital telephone switching systems 133–40
digital time switching (T) 137–9
digital-to-analogue conversion (D/A) 58–9
digital transmission 56–9, 64, 73–5, 80
direct current (DC) 3, 101
direct dialling in (DDI) 38
directory enquiry 33
direct-to-tandem ratio (d/t) 151
direct wave (free space wave) radio
transmission 69
discrete multi-tone (DMT)
modulation 88–9
distribution point (DP) 66, 100, 103
diverse routing 123–4
domain name server (DNS) 198
domain name system (DNS) 272
Doppler shift 222
DSI rate, multiplexed payloads 76
DSL: see xDSL
dynamic host configuration protocol
(DHCP) 194
E.164 telephone number scheme/format 250,
255, 264–5, 276, 311–16
earth satellite systems/communications 72,
earth stations (satellite) 72, 120
echo cancellers (EC) 231
echo cancelling equipment 231
echo and delay 289–90
electrical circuits 3
electromagnetic spectrum 17
electromagnetic waves/waveforms 16
element-control layer 296
e-mail 198
emergency calls 33
encapsulation of X.121 numbers
264, 265
enhanced data rates for global evolution
(EDGE) 241
ENUM system 273–4
e-purchase/e-commerce/e-business 281
equipment identity register (EIR)
Erlangs 147–8, 149
E-side cables 102
Ethernet LAN 196, 209–11
Ethernet PON (EPON) 94
European Telecommunications Standards
Institute (ETSI) 281, 310
European Union (EU) 260, 310
exchange capacity planning 152–8
design periods 153
exchange codes 14, 25–6, 250–8, 259
330 Index
exchange-control systems 171–4
exchange manhole 102
auto-manual 9, 33
automatic 8–9
international 13, 24
junction tandem 13, 44
local 11–15, 24–6, 31, 56–7, 86
manual 8–9
rural 12
trunk 12–15, 24–6, 150–2, 164–6, 175–7,
178, 260–1
trunk tandem 12
very small 12
extended network prefix working 270
extended telecommunications operations map
(eTOM) 297–8
exterior gateway protocol (IGP) 194
far-end handover 26
fax 35, 113, 276, 295
Federal Commission for Communications
(FCC) 255
fibre to the home (FTTH) 110
fibre to the kerb (FTTK) 110, 112
fibre to the office (FTTO) 110
fixed–mobile convergence (FMC) 246–7
fixed–mobile (network) integration
(FMI) 247
fixed networks 246–7
fourth generation (mobile) system (4G) 246
frame alignment 74–8, 134
frame alignment signal (FAS) 77
frame relay (FR) service 35, 185
freephone calls 175, 281
free space loss 220–2
free space wave (direct wave) radio
transmission 69
frequency 16, 17
frequency division duplex (FDD)
227–8, 237
frequency division multiple access (FDMA)
227–8, 237–8
frequency division multiplexing (FDM)
48, 90
frequency modulation (FM) 62
frequency-shift keying (FSK) 63, 223
fulfilment, assurance, billing (FAB) model
fully provided (dimensioning) 152
G821 ITU error standards 291
general packet radio service (GPRS)
238–41, 246
spare capacity allocation 240
system outline 240
geographical numbers (PSTNs) 251–3
geo-stationary satellite 95
gigabit/s PON (GPON) 94
global positioning satellite (GPS) 142, 284
global system, mobile (GSM) 228–38
with GPRS systems 238–41
see also cellular networks/systems
grade of service (GOS) 148, 149
grooming function 52–5
ground reflected radio waves 69–70
ground or surface radio waves 70
group switches 44
see also route switch/route switching
GSM frame structure 237–8
H323 (ITU-T) signalling system 169–71
half circuit 24
hand-over, mobile networks 23, 26, 225, 229,
hand-over distribution frame (HDF) 109
handsets: see mobile terminal (MT) (handset)
hexadecimal numbers 274
high-bit-rate digital subscriber line (HDSL)
systems 90
high-speed circuit-switched data (HSCSD)
system 241
high usage (dimensioning) 152
home location register (HLR) 231,
234–5, 261
hot spots, wireless LANS (WLANs) 246
hot standby 125
hybrid fibre coax (HFC) 110
hybrid fibre copper pair 110
hybrid transformers 4, 5, 17, 18
hyper-text transfer protocol (HTTP) 195, 197
inductance 3
information and communications technology
(ICT) 300, 303
information technology (IT) 300
initial address message (IAM) 164, 166,
Institute of Electrical and Electronic
Engineers (IEEE)
about the IEEE 310
Index 331
IEEE range of standards 210
IEEE 802.3 209
IEEE 802.11 96, 213
IEEE 802.15 (Bluetooth) 214, 245–6
IEEE 802.16 (WiMax) 96, 107, 111
integrated digital network (IDN) 56–7
integrated services digital network (ISDN)
29–30, 57, 104, 144–5, 189
intelligent-network application part (INAP)
168, 169
intelligent network (IN) 33, 46, 169,
174–7, 261
interconnection of networks 24–7
interior gateway protocol (IGP) 194
intermediate service part (ISP) 169
international call charging 262
international call routeing 24, 25
international calls 24–7
international code designator (ICD) 265
international network designator (IND) 265
International Standards Organisation
(ISO) 310
international switching centres
(ISCs)/international exchanges 13,
24, 25
International Telecommunications Union
(ITU) 73, 197
about the ITU 309
E.164 telephone number format 250, 255,
264, 311–16
G821 error standards 291
ITU Common Channel Signalling System
7 (CCSS7, SS7 or C7) 163–9, 172,
176, 178, 201–2, 231, 284
ITU H323 169–70
ITU-R allocated frequency bands 94–5
ITU-T Q931 170
X25 data service 184
X.25 (packet switch stream (PSS))
protocol 263
X.121 number scheme 263
Internet 27–9, 196–9
access 29–32
charging 31
Internet telephony 199
names and addresses 265–76
see also Internet protocol (IP);
voice-over-IP (VOIP)
Internet Corporation for Assigned Names and
Numbers (ICANN) 268, 310
Internet Engineering Task Force (IETF) 170,
197, 265, 310
Internet exchange (INX) 29
Internet protocol (IP) 28, 30, 35, 177, 186,
192–6, 196, 204–6
IP packets 207
IP routers 192–3, 207
IPv4 192, 193, 268–70, 273
IPv6 192, 273
numbering/naming and addressing 265–76
Internet service providers (ISPs) 28–9
charging 31, 281
virus attacks 198
interworking functions (IWF) 231
intranets 198–202
IP over ATM 204–6
IP virtual private network (IPVPN) 208–9
ISDN user part (ISUP) 168, 231
ISO network service access point (NSAP)
264–5, 266
jitter problems 291
junction routes 9, 13
junction tandems (JTs) 13, 285–6
see also trunk exchanges (TEs)
keying 63
frequency-shift keying (FSK) 63
phase-shift keying (PSK) 63
label switch routers (LSRs) 207
latency 181, 196
leased lines, Internet access 32
light, as an electromagnetic wave 16
light emitting diodes (LEDs) 67, 90
line cards 141
line codes/coding 65
line-of-sight (LOS) microwave system 94–5,
line signals/signalling 160
line termination equipment (LTE) 104
line termination unit (LTU) 141
link (transmission) functions 53–4
loading coils 10
local area networks (LANs) 35, 209–12
Ethernet LAN 209–11
wireless LANs (WLANs) 96, 106–7
local exchanges (LEs) 11, 11–15, 12, 24–6,
31, 56–7, 86, 285
see also subscriber switching (local) units
332 Index
local loop copper pairs 5–10, 36, 66–7
access features 100–3, 110, 112
network unbundling 107–9, 112
resistance of 10, 102
local multipoint distribution service
(LMDS) 95
local networks 12
local number 250–5
location management 232–4
London sector plan 257–60
long wave (LW) 17, 18
loop: see local loop
loop disconnect (LD) signalling 160
loudness/loudness ratings 287–8
low Earth orbit (LEO) system 95–6
main distribution frames (MDFs) 100–1, 102,
103, 134–5
main switching units: see trunk exchanges
man-machine interface (MMI) 173
mapping 154, 304
media access protocol (MAC) 211
media gateway controller (MGC) 201–2,
media gateway (MG) 201–2, 301
medium Earth orbit (MEO) satellite 95–6
medium wave (MW) 17, 18
Megaco (media gateway control) 202
message transfer part (MTP) 168, 169
metallic-line transmission systems 86–7
metropolitan area network (MAN)
LANs 211–12
wireless MANs 96
MG control protocol (MGCP) 202
microphones 2
microwave frequencies 69, 71
microwave radio: see broadband
microwaves 16
point-to-point relay route 71, 72
mobile application part (MAP) 168, 231
mobile exchanges (MEs) 165
mobile network operators (MNOs) 219
mobile networks/systems 22–3, 26–7, 217–47
see also cellular networks/systems; fixed
mobile convergence (FMC); general
packet radio service (GPRS); global
system, mobile (GSM); radio
waves/transmission; second
generation (mobile) system (2G);
third generation (mobile) system
(3G); wireless scene
mobile service providers (MSPs) 218
mobile switching centre (MSC) 23, 26–7,
226–7, 230–1, 232, 233–5, 239
mobile terminal (MT) (handset) 218, 230,
mobile to mobile via PSTN 26–7
mobile virtual network operators
(MVNOs) 219
mobility management 218
modem 29–32
modulation, about carrier-wave modulation
58–9, 62–3, 222–3
Mu (μ) law 75–6
multi-frequency (MF) signalling 160, 171
multiplexed payloads 72–85
multiplexors/multiplexing 46–52, 117–21
multiplex section-dedicated protection ring
multiplex section-shared protection ring
multi-protocol label switching (MPLS) 35,
multi-service platform (MSP) 206, 218–19
national number 250–60
national regulator 26, 30, 105, 115, 246
native ATM 191
near-end handover 26
network components 41–60
see also nodal concepts
network control layer 296
network dimensioning 145–53
networked centrex 13, 37, 39
network evolution 298–300
network intelligence centres 13–14
network interface card (NIC) 212–13
network management 294–6
network optimisation 9–10
network performance 286–93
network planning, 3G systems 244
network resilience 121–6
networks 21–38
see also access network; cable (cable TV)
networks (CATV); data services
networks; intelligent network (IN);
interconnection of networks; Internet;
local area networks (LANs); mobile
networks; operator-services network;
public switched telephone network
Index 333
(PSTN); Telco’s networks; telex
network; wide area network (WAN)
network structures 11–12
network terminal number (NTN) 263
network terminating unit (NTU) 144, 145
network termination and test point (NTTP or
NTP) 113, 114
next generation network (NGN) 109, 169–71,
physical architecture example 305, 306
nodal concepts 41–53
with analogue and digital transmissions 64
impulsive with ADSL 109
quantisation noise 58
with radio 222
as a signal impairment 64
signal to noise ratio 64
nomadic working 246
non-circuit related (NCR) signalling 175–6
non-geographical numbers for PSTNs 253–4
Nordic Mobile Telephone (NMT) 223
numbering and addressing 249–76
number portability 176, 260–1
Nyquist rate
and pulse amplitude modulation (PAM) 58
and sampling rate 49
Ofcom 255–6, 263
off-hook 8, 29–30, 130, 171
Ohm’s law 3
on-hook 30, 171
open network provision (ONP) 175, 177
open shortest path first (OSPF) 193
open systems interconnect (OSI) reference
model (7 layer) 187–8, 196, 197
Layer One (Physical) 187, 197
Layer Two (Data link) 187, 191, 197, 211,
239, 304
Layer Three (Network) 187, 192, 197,
239, 304
Layer Four (Transport) 187, 194, 197
Layer Five (Session) 188, 197
Layer Six (Presentation) 188, 197
Layer Seven (Application) 188, 197
operational cost or expenditure (opex) 112
operational support systems 298
operations 293–8
operations, administration and maintenance
(OA&M) 173
operations, maintenance and administration
part (OMAP) 168
operations and maintenance centres
(OMCs) 294
operator-services network (OSN) 33
optical carriers (OCs) 63
in SONET 82
optical cross connect (OXC) 121
optical fibre cable 67–9, 103–5
optical fibre systems/networks
dense wave-division multiplex (DWDM)
91, 92, 94, 121, 303
dense wave-division multiplex system 91
optical fibre access network 103–5
passive networks 91–4
point-to-point 90–1
optical line termination (OLT) 93
optical network termination (ONT) 93
optical network unit (ONU) 93
opto-electrical conversion 65
orthogonal frequency division multiplexing
(OFDM) 88–9
packet switching and routing 44–6, 181–214
packet switch stream (PSS) 263
pair gain systems 86
Parlay consortium 177
passive optical network (PON) 91–4
PCM multiplex
24-channel 75–6
30-channel 73–5
peer-to-peer (working) 204
period 15
permanent virtual circuit (PVC) working
personal communication services (PCS) 229
phase distortion 63
phase modulation (PM) 62
phase-shift keying (PSK) 63
plain old telephone service (POTS) 108
platform 300–1, 304–7
plesiochronous digital hierarchy (PDH)
systems/networks 75, 78–82,
117–18, 286
point of interconnection (POI) 13, 25–6, 174
portal 197–8, 281
power feed, local lines 130
power management, GSM systems 235–7
prefix, international 24, 253
prefix, trunk 254
334 Index
primary connection points (PCPs) 23–4,
102–4, 111–14
primary rate ISDN 144, 145
private automatic branch exchanges
(PABX) 141
private branch exchanges (PBX) 38,
141, 142
private circuit 34–5, 37–8, 286, 287
processor exchanges 143
processor node 12
product life styles 299
PSTN-to-Internet gateway (PIG) 200–2,
public network operators (PNOs) 25–6,
108–9, 285–6
public switched telephone network (PSTN)
networks associated 36–7
pulse amplitude modulation (PAM) 58, 63
pulse-code modulation (PCM) 59, 63, 73–5,
pulse position modulation (PPM) 63
Q931 (ITU-T) protocol 162, 170
quality of service (QOS) 286–93
and ATM 192
core networks 122
and MPLS 207–8
and network dimensioning 146
quantisation noise 58
radio access networks 105–7
radio frequency planning 105–6
radio network controller (RNC) 243
radio spectrum 71, 223–4
radio-telephone systems: see mobile
radio waves/transmission 16, 69–72, 220–3
real-time processing 173
real-time protocol (RTP) 202
receiver 1–5
remote concentrator units (RCUs) 12, 143–4
repeater antennas, microwave systems 71, 72
resilience: see network resilience
reverse address resolution protocol
(RARP) 194
ringing current/bells 3–4, 5, 8, 130
ringing equivalent number (REN) 101
ringing/screeching, from instability 289
roaming 26, 218, 219
roll out of networks: see deployment (of
routeing code (RC) 163
routers/routeing 44–6, 151–2, 185–6, 192–3,
207–8, 260–1
core routers 192
edge routers 192, 206–7
label edge routers (LER) 207
label switch routers (LSR) 207–8
route switch/route switching 43–4, 131
route switch-block 150
sampling, time division multiplexing 49–51
satellites, communication satellites 72
S-digit 252–3
secondary connection points (SCPs) 102
second generation equipment
ADSL2 109, 110
ADSL2+ 109, 110
second generation (mobile) system (2G) 224,
241–2, 244
2.5 generation system (2.5G) 239,
241–2, 244
sector switching sector (SSC) 258, 259
service control point (SCP) 113, 175, 176–7
service creation environment (SCE) 175
service management centre (SMC) 113
service management (SM) 295–8
service protection networks (SPNs) 125–6
service providers 120, 177–8, 204
service switching point (SSP) 175, 176
serving GPRS support node (SGSN) 239
session initiation protocol (SIP) 170–1, 202
set-top box (STB) 23
short message services centre (SMSC) 232
short message service (SMS) 229, 232
signalling 159–71
see also common-channel signalling (CCS)
signalling connection control path
(SCCP) 169
signalling gateway (GC) 201
signalling system No7 (SS7) 163–9, 172,
176, 178, 201–2, 231, 284
signal to noise ratio 64
signal transfer point (STP) 168
signal unit (SU) 165, 166, 167
sinusoidal waves/waveforms 16
skype 200, 203
skywaves 70
slip problems 291
Index 335
soft switches/MGCs 201–2
SONET 80–5, 124
sound waves 2, 16–18
spare plant 114–15, 124–6
specialist networks 32–5
PSTN bandwidth allocation 48
waveform analysis 47–8
speech memory (SM) 138
standards organisations 309–10
stored program control (SPC) 133, 134,
173, 299
street cabinets (PCPs) 23–4, 102–4, 111–14
stuffing, PDH systems 78
sub-loop unbundling 108
submarine transmission systems 87, 90–1
sub-netting/sub networks 212, 269–70
sub networks 212
subscriber concentrator switch:
see concentrator
switches/switching/switch blocks
subscriber information module (SIM) 231,
subscriber line card 136
subscriber lines 12
subscriber number 250–4
subscriber switching (local) units 130–2
generalised local exchange 131, 132
route switch 131
supervisory signalling 124, 135, 160, 229
surface or ground radio waves 69–70
concentrator 226–7, 301
route 132, 135–7, 140–54, 172
switched label paths (SLPs) 207–8
switched multi-megabit data service (SMDS)
35, 186
switched traffic concept 145–6
switched virtual circuit (SVC) 184
switching: see circuit-switching/switching
symmetrical digital subscriber line (SDSL)
systems 90
synchronisation 141
synchronisation channels, and multiplexed
payloads 73–5
synchronous digital hierarchy (SDH)
systems/networks 80–5, 118–21, 124,
195–6, 284–6
synchronous optical network: see SONET
synchronous transport module (STM) 82,
Telco’s networks 36–8, 285–6
Telecommunications Management Forum
(TMF) 296–7
telegraph, telegraphy 1
telemetry 44
telephone demand 146
telephone exchange: see exchanges
telephone networks 9–14
telephone numbering: see numbering
and addressing
telephone user part (TUP) 168, 169
telephony 1–18
telephony over PON (TPON) 93–4
telex network 35
terrestrial trunk radio (TETRA) system 241
third generation (mobile) system (3G) 224,
239, 241–4
universal mobile telecommunications
system (UMTS) 242, 246
time division duplex (TDD) 228
time division multiple access (TDMA)
49–51, 237–8
for cellular networks 228
time division multiplexing (TDM) 49–51, 228
and multiplexed payloads 76–8
quality of service (QOS) 291
time frame 49, 76–7
time slots (TS) 137, 139
with 24-channel PCM multiplex 75–6
with 30-channel PCM multiplex 73–4
time switching 137
toll switches: see trunk exchanges (TEs)
top-level domain (TLD) 266–8
total access telephone system (TACS) 223
traffic engineering, with MPLS 209
traffic flow concepts 147–9
traffic routeing hierarchy 12
traffic routeing/routes 12, 149–52
transaction capabilities application part
(TCAP) 169
trans-Atlantic telecommunications (TAT)
optical cable system 91
transceivers 26, 229, 230, 236
acoustic 4
optical 67, 90
transmission control protocol (TCP)
194–6, 197
336 Index
transmission loss/impairments 63–4, 287–8
transmission networks: see networks
transmission principles 61–6
line codes/coding 65
transmission repeater station (TRS): see core
transmission station (CTS)
transmission stability 288–9
transmission systems 61–97
transmission units (TUs)
with SONET 83
transmission unit groups (TUGs) 83
transverse screen cable 54, 67, 86, 99
tropospheric or troposcatter radio systems
trunk exchanges (TEs) 12, 12–15, 24–6, 130,
150–2, 164–6, 175–7, 178, 260–1,
two wire/four wire circuits/conversion 2, 4–5,
UK telephone numbering 250–7
UMTS terrestrial radio access network
(UTRAN) 242–4
uniform domain-name dispute resolution
policy (UDRP) 268
uniform resource locator (URL) 170, 267, 272
universal mobile telecommunications system
(UMTS) 171, 242–4, 246
USA, country number code 250, 255
user datagram protocol (UDP) 195–6, 202
UXD5 143
variable bit rate (VBR) service 191
very-high bit rate digital subscriber line
(VDSL) systems 89, 108, 302
video 66, 69, 88, 94, 105, 190–2
virtual circuit (VC) 183–4
virtual container (VC) 83, 85
virtual path (VP) 183, 186, 189, 191–2,
205, 207
virtual private network (VPN) 13, 37, 38,
175, 208
virus attacks 198
visitor location register (VLR) 231,
233–5, 261
voice over ATM (VOATM) 199
voice over frame relay (VOFR) 199
voice-over-IP (VOIP) 169–71, 195, 199–202,
209, 290, 301
VOIP over broadband 202–4
voltage 3
Vonage 203
wander problems 291
waveform analysis, speech 47–8
waveforms 15–18
wavelength 16
wavelength add-drop multiplexors
(WADMs) 121
wide area network (WAN) 35, 212
wide area tandem (WAT) 13
wideband CDMA (W-CDMA) 242
WiFi: see wireless LANs (WLANs)
WiMax system (IEEE 802.16) 96, 107, 214
wireless LANs (WLANs) (WiFi) 96, 106–7,
wireless MAN 96
Wireless Regulation Congress (WRC) 223
wireless scene 244–6
World Wide Web (WWW) 28, 197, 281
X25 (ITU) data service 184, 263
X.121 (ITU-T) number scheme 263
xDSL 87–8
X-Rays 16
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF