)%44%,%#/--5.)#!4)/.

33%2)%3
3ERIES%DITORS 0ROFESSOR#*(UGHES
0ROFESSOR**/l2EILLY
0ROFESSOR'7HITE

)NTELLIGENT
.ETWORKS
0RINCIPLESAND!PPLICATIONS
/THERVOLUMESINTHISSERIES
6OLUME 0HASENOISESIGNALSOURCES702OBINS
6OLUME 3PREADSPECTRUMINCOMMUNICATIONS23KAUGAND*&(JELMSTAD
6OLUME !DVANCEDSIGNALPROCESSING$*#REASEY%DITOR
6OLUME 4ELECOMMUNICATIONSTRAFÙCTARIFFSANDCOSTS2%&ARR
6OLUME !NINTRODUCTIONTOSATELLITECOMMUNICATIONS$)$ALGLEISH
6OLUME 0ERSONALANDMOBILERADIOSYSTEMS2#6-ACARIO%DITOR
6OLUME #OMMON CHANNELSIGNALLING2*-ANTERÙELD
6OLUME 6ERYSMALLAPERTURETERMINALS63!4S *,%VERETT%DITOR
6OLUME !4-THEBROADBANDTELECOMMUNICATIONSSOLUTION,'#UTHBERTAND
*#3APANEL
6OLUME $ATACOMMUNICATIONSANDNETWORKSRDEDITION2,"REWSTER%DITOR
6OLUME !NALOGUEOPTICALÙBRECOMMUNICATIONS"7ILSON:'HASSEMLOOYAND
):$ARWAZEH%DITORS
6OLUME -ODERNPERSONALRADIOSYSTEMS2#6-ACARIO%DITOR
6OLUME $IGITALBROADCASTING0$AMBACHER
6OLUME 0RINCIPLESOFPERFORMANCEENGINEERINGFORTELECOMMUNICATIONAND
INFORMATIONSYSTEMS-'HANBARI#*(UGHES-#3INCLAIRAND*0%ADE
6OLUME 4ELECOMMUNICATIONNETWORKSNDEDITION*%&LOOD%DITOR
6OLUME /PTICALCOMMUNICATIONRECEIVERDESIGN3"!LEXANDER
6OLUME 3ATELLITECOMMUNICATIONSYSTEMSRDEDITION"'%VANS%DITOR
6OLUME 3PREADSPECTRUMINMOBILECOMMUNICATION/"ERG4"ERG*&(JELMSTAD
3(AAVIKAND23KAUG
6OLUME 7ORLDTELECOMMUNICATIONSECONOMICS**7HEATLEY
6OLUME 4ELECOMMUNICATIONSSIGNALLING2*-ANTERÙELD
6OLUME $IGITALSIGNALÙLTERINGANALYSISANDRESTORATION**AN
6OLUME 2ADIOSPECTRUMMANAGEMENTNDEDITION$*7ITHERS
6OLUME )NTELLIGENTNETWORKSPRINCIPLESANDAPPLICATIONS*2!NDERSON
6OLUME ,OCALACCESSNETWORKTECHNOLOGIES0&RANCE
6OLUME 4ELECOMMUNICATIONSQUALITYOFSERVICEMANAGEMENT!0/ODAN%DITOR
6OLUME 3TANDARDCODECSIMAGECOMPRESSIONTOADVANCEDVIDEOCODING
-'HANBARI
6OLUME 4ELECOMMUNICATIONSREGULATION*"UCKLEY
6OLUME 3ECURITYFORMOBILITY#-ITCHELL%DITOR
6OLUME 5NDERSTANDINGTELECOMMUNICATIONSNETWORKS!6ALDAR
6OLUME /PTICALÙBRESENSINGANDSIGNALPROCESSING"#ULSHAW
6OLUME )3$.APPLICATIONSINEDUCATIONANDTRAINING2-ASONAND0$"ACSICH
 )NTELLIGENT
.ETWORKS
0RINCIPLESAND!PPLICATIONS

*OHN!NDERSON

4HE)NSTITUTIONOF%NGINEERINGAND4ECHNOLOGY
0UBLISHEDBY4HE)NSTITUTIONOF%NGINEERINGAND4ECHNOLOGY,ONDON5NITED+INGDOM
4HE)NSTITUTIONOF%LECTRICAL%NGINEERS
&IRSTPUBLISHED

4HISPUBLICATIONISCOPYRIGHTUNDERTHE"ERNE#ONVENTIONANDTHE5NIVERSAL#OPYRIGHT
#ONVENTION!LLRIGHTSRESERVED!PARTFROMANYFAIRDEALINGFORTHEPURPOSESOFRESEARCH
ORPRIVATESTUDYORCRITICISMORREVIEWASPERMITTEDUNDERTHE#OPYRIGHT$ESIGNSAND
0ATENTS!CTTHISPUBLICATIONMAYBEREPRODUCEDSTOREDORTRANSMITTEDINANY
FORMORBYANYMEANSONLYWITHTHEPRIORPERMISSIONINWRITINGOFTHEPUBLISHERSORIN
THECASEOFREPROGRAPHICREPRODUCTIONINACCORDANCEWITHTHETERMSOFLICENCESISSUED
BYTHE#OPYRIGHT,ICENSING!GENCY)NQUIRIESCONCERNINGREPRODUCTIONOUTSIDETHOSE
TERMSSHOULDBESENTTOTHEPUBLISHERSATTHEUNDERMENTIONEDADDRESS
4HE)NSTITUTIONOF%NGINEERINGAND4ECHNOLOGY
-ICHAEL&ARADAY(OUSE
3IX(ILLS7AY3TEVENAGE
(ERTS3'!95NITED+INGDOM
WWWTHEIETORG
7HILETHEAUTHORANDTHEPUBLISHERSBELIEVETHATTHEINFORMATIONANDGUIDANCEGIVEN
INTHISWORKARECORRECTALLPARTIESMUSTRELYUPONTHEIROWNSKILLANDJUDGEMENTWHEN
MAKINGUSEOFTHEM.EITHERTHEAUTHORNORTHEPUBLISHERSASSUMEANYLIABILITYTO
ANYONEFORANYLOSSORDAMAGECAUSEDBYANYERROROROMISSIONINTHEWORKWHETHER
SUCHERROROROMISSIONISTHERESULTOFNEGLIGENCEORANYOTHERCAUSE!NYANDALLSUCH
LIABILITYISDISCLAIMED
4HEMORALRIGHTSOFTHEAUTHORTOBEIDENTIÙEDASAUTHOROFTHISWORKHAVEBEEN
ASSERTEDBYHIMINACCORDANCEWITHTHE#OPYRIGHT$ESIGNSAND0ATENTS!CT

"RITISH,IBRARY#ATALOGUINGIN0UBLICATION$ATA
!NDERSON*2
)NTELLIGENTNETWORKSPRINCIPLESANDAPPLICATIONS
4ELECOMMUNICATIONSSERIESNO
!RTIÙCIALINTELLIGENCE#OMPUTERNETWORKS
)4ITLE)))NSTITUTIONOF%LECTRICAL%NGINEERS


)3".DIGIT 
)3".DIGIT     

4YPESETINTHE5+BY4RADESPOOLS,TD&ROME3OMERSET

&IRSTPRINTEDINTHE5+BY-0'"OOKS,TD"ODMIN#ORNWALL
 Contents

Preface ix

Acknowledgments xiii

1 Introduction to intelligent networks 1

1.1 The basics of intelligent networks 1
1.1.1 Origins 1
1.1.2 Extracting the service logic 4
1.1.3 Interactive voice functions 5
1.1.4 The ‘classical’ IN architecture and its elements 6
1.1.5 Locating the ‘intelligence’ 7
1.1.6 What is an ‘intelligent’ network? 9
1.1.7 The role of international standards 9
1.2 Why did we need IN? 11
1.2.1 The commercial drivers 11
1.2.2 IN as a universal integrator 13
2 The foundations of IN 15
2.1 The service switching function 15
2.2 Triggering to remote service logic 17
2.2.1 Introduction to triggering 17
2.2.2 Introduction to IN call models 21
2.2.3 Trigger detection points 22
2.2.4 Trigger types and trigger criteria 23
2.2.5 Trigger tables 24
2.2.6 Event detection points 27
2.2.7 The CS-1 basic call state model 28
2.2.8 Notification and response DP types 29
2.2.9 Triggering dynamics 29
2.2.10 The CS-1 feature interaction manager 31
2.3 Hosting and creating IN services 32
2.3.1 SCPs and SCFs 32
vi Contents

2.3.2 Service data considerations 32

2.3.3 Multi-service SCPs 33
2.3.4 Service logic execution environment (SLEE) 33
2.3.5 Service creation environment 34
2.3.6 Operational support systems for IN services 35
2.3.7 SCP ‘non-functional’ requirements 37
2.3.8 Future trends for SCPs 38
2.4 The intelligent peripheral 39
2.4.1 The Assist procedure for IP connections 41
2.5 INAP (the intelligent network application protocol) 43
2.5.1 ETSI Core INAP operations 43
2.6 IN CS-1 implementation issues 48
2.6.1 Core INAP as a compliance benchmark 49
2.6.2 INAP implementations and feature interactions 50
2.6.3 SSP data management 51
3 Signalling intelligence 53
3.1 Introduction 53
3.2 Common channel signalling – the beginning 53
3.3 The layered signalling model 54
3.4 The message transfer part 55
3.4.1 Signalling management intelligence 58
3.5 Telephony user part intelligence 59
3.5.1 Basic call set-up 59
3.5.2 The ‘intelligence’ aspects of basic calls 61
3.5.3 Further ‘switch-based’ intelligence 63
3.6 SS7 signalling for IN 63
3.6.1 Signalling connection control part (SCCP) 63
3.6.2 Global title addressing 64
3.6.3 Transaction capabilities application part (TCAP) 65
3.6.4 Signalling differences between PSTN and GSM networks 68
4 International standards for intelligent networking 71
4.1 Introduction 71
4.2 Early US standards for IN 73
4.2.1 INWATS, IN/1 and IN/2 73
4.2.2 AIN 75
4.3 ITU-T IN CS-1 76
4.3.1 IN CS-1 documents 77
4.3.2 IN ‘benchmark’ services 78
4.3.3 The IN conceptual model 80
4.4 ETSI Core INAP and CS-1R 85
4.5 IN and mobile systems 85
4.5.1 The role of IN with GSM 85
4.5.2 CAMEL 86
4.5.3 CAMEL Phase 1 87
4.5.4 CAMEL Phase 2 89
 Contents vii

4.5.5 Future CAMEL phases 89

4.5.6 Wireless intelligent networking (WIN) 89
4.6 ITU-T IN CS-2 90
4.6.1 Extra functions provided in CS-2 90
4.6.2 Service plane for IN CS-2 92
4.6.3 Global functional plane for IN CS-2 92
4.6.4 CS-2 distributed functional plane 93
4.6.5 IN CS-2 documents 95
4.6.6 Physical plane for IN CS-2 95
4.7 ITU-T IN CS-3 97
4.7.1 IN CS-3 documents 97
4.8 ITU-T IN CS-4 97
4.8.1 Alliance between IN and Internet telephony 98
4.8.2 Interworking between CS-4 and other networks 98
4.8.3 Service application gateway function 99
4.8.4 Service control gateway function 100
4.8.5 CS-4 service features 101
4.8.6 IN CS-4 documents 101
5 Call party handling 103
5.1 Introduction 103
5.1.1 Early ‘leg control’ network functions 105
5.2 The CS-2 call model 106
5.2.1 Legs and connection points 107
5.2.2 CS-2 core switch capabilities 108
5.2.3 The connection view model states 109
5.2.4 Information flows supporting CPH 113
5.3 Call party handling – A renewal of IN? 116
6 Distributed intelligence 119
6.1 Distribution of intelligence for network efficiency 120
6.2 Distribution of intelligence across open network interfaces 121
6.3 A model for distributed IN 122
6.3.1 Infrastructure APIs 126
6.4 Object request brokers for network intelligence 126
6.5 The Parlay API 127
6.5.1 Background 127
6.5.2 What is the Parlay API? 128
6.5.3 Parlay specifications 129
6.6 TINA 131
6.6.1 TINA roles and reference points 131
6.6.2 The TINA architecture 132
6.6.3 TINA specifications 136
6.6.4 Relationship between IN and TINA 137
7 Service examples 139
7.1 Simple number translations – Freephone service 139
7.1.1 Inter-network Freephone 141
viii Contents

7.1.2 More complex number translations – Geographical routing 143

7.2
Personal numbering 143
7.3
Incoming call screening 144
7.4
Least cost routing 146
7.5
A virtual private network (VPN) service example 148
7.6
A directory enquiry call completion service 151
7.6.1 IN distribution of operator calls 151
7.6.2 Automated response and call completion service 152
7.6.3 Call completion service summary 156
7.7 Call gapping 157
7.8 Activate service filtering 158
7.9 Simple CTI (computer–telephony integration) 158
7.10 CAMEL calls 160
7.10.1 An outgoing CAMEL call 160
7.10.2 An incoming CAMEL call 162
7.10.3 Pre-pay mobile service 163
7.11 IN control for Internet dial-up access 165
8 Concluding summary 167
8.1 IN today – How did we get here? 167
8.1.1 Financial justification 168
8.1.2 Vendor control 169
8.1.3 Legacy switch-based features 169
8.2 Tomorrow’s intelligent network 170
8.2.1 Future services 170
References 173
Appendix 1: List of ITU-T IN documents 177
Appendix 2: Additional CS-1 INAP operations 179
Appendix 3: Example SS7 message sequence coding 181
Glossary 205
 Preface

The purpose of this book is to give some practical insight into how public
telecommunications networks have advanced from the old days of offering just
simple telephony services to becoming today’s ‘intelligent’ networks. Telephone
companies now operate in an intensely competitive environment, and so they have
to constantly struggle to find more ways to encourage customers to make ever more
use of the installed network equipment. Network operators are therefore constantly
looking for new services that will make the telephone system more attractive, easier
to use and perhaps even more indispensable for its users. The overall goal is of
course to increase the revenue-producing ability of the existing phone systems. The
aim of the ‘intelligent network’ (IN) has always been to help the telephone
companies achieve just this, but in a rigorously standard manner, so that network
equipment from different suppliers can work together. IN seeks to provide the
internal network rearrangements that give the agility and flexibility needed to bring
almost any new services into life very quickly.
The intended audience for this book is really anyone who has an interest in this
subject without necessarily yet having any significant level of expertise in
intelligent networking. The background to IN is formal and complex, because its
fundamental driver is a set of agreed interfaces between different functions
provided in different vendors’ computing and telecommunications equipment. This
book therefore takes pains to give what is hoped to be an easily understandable
account of IN for readers who are unfamiliar with the background detail.
A typical target reader might therefore be a student or engineer with background
knowledge in computer hardware and software, but with little specialist experience
in IN. The book will have succeeded in its aim if it enables such a reader to
assimilate the essentials of the subject of IN with reasonable speed. Because of the
potential complexity, we have aimed to keep the text reasonably economical, and to
focus primarily on explanation and illustration of the principles rather than the
presentation of comprehensive coverage. For further information we have in
general referenced out to the plethora of background and specialist supporting
documents that are available from the international and regional standards
organisations and elsewhere.
x Preface

Over the years the term ‘intelligent network’ has been used in association with a
variety of meanings and nuances. This has inevitably brought into existence several
different notions of what an ‘intelligent’ network actually is, so the first task for a
book with IN in the title is to establish exactly what we are actually talking about.
What actually do we mean by an intelligent network?
We live in a world of fast-advancing technology, and we are adapting to the
various ‘convenience’ lifestyle aids that modern telecommunications equipment
gives us. We are quickly becoming used to the availability of home-shopping
through the Internet, automated voice-response bank clerks, the ability to be
constantly in contact wherever we are on the planet, and future prospects of virtual
reality holidays. It may actually be more pertinent to question what a ‘non-
intelligent’ network might be! A casual Internet search on the phrase ‘intelligent
networking’ yields information on an alarmingly wide array of widely disparate
topics. These include, for example, ‘intelligent’ house wiring and burglar systems,
greenhouse-sprinkling systems, neural road maps for chimpanzee brains and a host
of interesting academic papers on network design theory. We therefore need to tie
down our particular use of the phrase ‘intelligent network’ early in the book, and so
we will build further on the outline definition of the aims of IN that we gave in the
opening paragraph of this preface.
Many papers and standard recommendations have been written about the
progression of the IN vision. However most major network operators have
completed their IN implementations based on the standards laid down in the early
1990s. Even though the international standards have come a long way since the
early days network operators with IN systems have generally not upgraded in line
with the more advanced IN recommendations. New services are still often being
introduced using the early well-tried IN standards. This book therefore concentrates
more on the practical exploitation of the existing IN systems than on speculation of
how newer IN standards might be used. We focus on the use of the existing IN
systems for advanced services, illustrating the principles with practical examples.
Earlier chapters in this book concentrate on the internationally established
standards for intelligent networking in fixed networks. The ITU-T IN ‘capability
sets’1 establish the progression of development of the international standard for IN.
Within this contextual scope, regional standards bodies, such as ETSI in Europe
and ANSI in the USA, have been active in establishing the details of workable
frameworks for the design of IN implementations.
Later parts of the book are about newer standards and developments because,
whilst the introduction of new IN features using the initial systems in fixed
networks has been reasonably steady, the use of IN in other areas, particularly
mobile networks, is demonstrating fast take-up and rapid growth. Also, the IN
designed 10 years ago was optimised for voice telephony, and big questions now

1
The ITU-T (International Telecommunications Union — Telecommunications division), which was
called CCITT (International Consultative Committee on Telephony and Telegraphy) until 1994, is the
over-riding international standards body for telecommunications. The ITU-T has defined a series of IN
functionality releases known as ‘capability sets’.
 Preface xi

are to do with how these principles carry forward into the new generation of
multimedia and ‘Internet’ services.
One of the earlier pioneering texts on IN [1] stated that by the year 2000 the term
‘intelligent network’ would be history. Whilst the major standardisation work on IN
was indeed completed in the last century, we have to remember that the standards
formulation is just the beginning for a successful product! IN now survives because
it has entered the fundamental realms of ‘business-as-usual’, and we are pleased to
be able to report that IN is still very much alive. However, IN features do now tend
to be taken for granted as part of the standard set of today’s network switching
functions, and so the use of the term IN as a visionary concept is probably now
historical.
 Acknowledgments

This book has drawn heavily on practical experiences gained whilst launching IN
services into the British Telecom UK domestic network. This was a large-scale
team effort and much of the material in the book relates to work done in
conjunction with many colleagues. In assembling this work I am particularly
indebted to those co-travellers who provided generous assistance. The following
colleagues kindly provided me with extra source material: David Gardner (BT),
Richard Hennessy (Conformance Standards Ltd), John Shepherd (BT) and Richard
Swale (BT). I am also indebted to Nigel Dadge (BT), Andy Fisk (BT), Bob
Hinchliffe (BT), David Smith (mmO2), Graham Fitt (Conformance Standards Ltd)
and Dick Cullingford (Conformance Standards Ltd) for their invaluable assistance
at the technical review stages.
I would like to thank Charles Hughes, Gerard White and Roland Harwood at the
IEE for driving the project through from beginning to end with patience and
continuous encouragement, and Richard Swale again for organising the final
editorial process that brought the book into existence. However, top marks for
unlimited support and long-suffering patience throughout the project go to my
family, particularly to my wife Elizabeth – who gave up many hours to help me
with the proof-reading and organisation of the drafts.
 Chapter 1
Introduction to intelligent networks

The past is a foreign country: they do things differently there.

Hartley, L.P.: ‘The Go-Between’ (Penguin Classics, 1953)

1.1 The basics of intelligent networks

The world of telecommunications has changed dramatically since the ‘intelligent

network’ was defined, around 20 years ago. However, the early principles, which
were established at that time, have now become embodied in most public telephone
networks and are the driving force behind many advanced telecommunications
services.

1.1.1 Origins

Like many things, intelligence in telephone networks is not new. In many ways,
public networks were at their most intelligent perhaps 70 years ago, when every
subscriber had a local telephone operator, perhaps at the local shop, who knew all
her customers by name and probably in person as well. ‘Calling by name’ was a
natural service – you would ask to be connected to a particular person and the
operator would endeavour to track that person down and connect you to them. If
you asked her for the florist shop or the taxi service, she probably wouldn’t have
had to trouble you for the number. A service such as ‘calling name delivery’ was
naturally included for local calls – the operator would announce who the caller was
when she put the call through to you.
Call diversion (e.g. to the doctor currently on duty) was easy – the operator
would have the doctor’s current whereabouts on a piece of paper in front of her, if
she didn’t know it already, and she would put you through to wherever he was
currently located. Billing was done by writing out a ticket for the call, so ‘alternate
billing’ or ‘reverse-charge calls’ were very straightforward services – she would
2 Intelligent networks – principles and applications

just write down on the ticket the number, or the name of the person to be charged.
A ‘conference call’ service was possible. Early switchboard equipment could cross-
connect several simultaneous calls, and the operator could offer her services in
supervising the conference. Most of the services we now call ‘advanced’ were
feasible from the manual boards of 70 years ago. Tele-shopping was fairly
common, because the telephone operator was usually also the local post-mistress
and grocer. I remember our own post-mistress in Oxfordshire still provided a tele-
shopping service in the 1950s – she used to deliver groceries in her 1932 Austin 7
Ruby a few hours after receiving the telephoned order.
However, the days of ubiquitous operator control were soon to be numbered
when some Kansas City undertakers tried to cheat a rival undertaker by installing
an efficient incoming call screening service on his exchange lines. They bribed the
local switchboard operator to divert his incoming business calls to their own
establishments rather than his. However, they hadn’t reckoned on the inventiveness
of their victim, whose name was Almon B. Strowger. When he discovered what
was going on, Strowger set out to redress the balance. He invented an automatic
call-switching device, which was destined to revolutionise the telecommunications
industry. This device, the construction of which was based on his top hat, was
called a 2-motion selector. It enabled one electrical inlet to be connected to any one
of 100 outlets on receipt of a pair of pulsed digits. When two sets of digit pulses
from a mechanical dial arrived at his new switch, the first train of pulses would step
the vertical motor of his selector up to the required level and the second train of
pulses would step the rotary selector around to the required horizontal outlet. The
operator’s manual switching function could thereby be dispensed with and calls
routed automatically to their proper destination without unlawful diversion along
the way.
The invention of this ingenious device in 1888 heralded the advent of automatic
telephony, and the first public automatic exchange was opened in 1892 at La Porte,
Indiana, USA [2]. Large armies of switchboard operators were no longer needed
and network traffic growth rocketed. However, whilst automation increased,
‘intelligence’ in the network, in terms of the variety and range of personal services
available to customers, gradually diminished (Figure 1.1) with the demise of the
hitherto ubiquitous switchboard operator.
The 1920s, 1930s and 1940s were periods of intense inventive activity, brought
on partly by military necessity. Many ingenious devices were invented to enhance
the automation of telephony – notably the electromechanical ‘register translator’
which could translate dialled digits into routing digits. This greatly enhanced the
embedded ‘intelligence’ of the network, leading to the automation of trunk calls in
the early 1950s.
After this followed the concept of centralised control of telephone exchanges,
and ‘stored program control’ (where pre-programmed network routings were held)
introduced more captive intelligence. These innovations have developed into
today’s digital exchanges, which use fast common channel inter-exchange
signalling to provide complex network services.
 Introduction to intelligent networks 3

Figure 1.1 The rise and fall of intelligence in networks

For a long time, therefore, the central theme of ‘intelligence’ in networks has
been to do with translating a dialled digit string into a different string of digits,
forming an address routing. The obvious illustration is the Freephone service,
where a caller dials an advertised number which is then translated by the network
into a routing number, and the particular translated number can vary depending on
factors such as time of day, location of origin of the call, etc. Similar principles
apply to ‘virtual private network’ calls, where the translated digits are also
dependent on the calling number. In countries such as the USA, where it is still
commonplace for telephones to have letters as well as numbers on the dials and
key-pads, there is far greater opportunity for advertisers to catch the public
imagination with a string of letters instead of numbers. An address such as 1800
CARPET tends to be rather more memorable, and thereby revenue producing, than
the equivalent digit string – 1800 116628. The potential high revenue generation of
these sorts of calls is of course one of the main reasons for the high interest in
intelligent network services.
The term ‘intelligent networks’ (IN) was introduced by Bellcore in the USA in
the mid 1980s as a consequence of the upheaval following the regulator’s break-up
of AT&T’s monopoly hold over the US telecommunications industry. Bellcore,
which has since been renamed as ‘Telcordia Technologies’, was the United States
telecommunications research body that was jointly funded by the seven regional
Bell operating companies (RBOCs). The RBOCs at the time were also keen to
provide the lucrative new additional services such as Freephone and Alternate
Billing, which had proved so successful for AT&T. The RBOCs therefore
commissioned Bellcore to develop the standard, which Bellcore called ‘intelligent
networks’. Consequently Bellcore led the way with IN standards for the latter part
4 Intelligent networks – principles and applications

of the 1980s. We discuss this early development of the IN standards in the US at

greater length later in the book.
Whilst many definitions now abound for the catch-phrase ‘intelligent network’,
technically the original underlying principle was that of separation of the logic
controlling services from the ‘basic’ call control function in existing telephone
exchanges. Telephone exchanges have of course always been ‘intelligent’ to some
degree, even if this just refers to an ability to translate a dialled digit string to an
outgoing routing number, or port number associated with a customer’s line. As well
as such translation facilities the current generation of digital exchanges has
standard facilities for services such as abbreviated dialling and call waiting, using
‘embedded’ service logic (Figure 1.2). However, the original essence of the
‘intelligent network’ standards was to separate the software for new services from
the basic call control software and to use a standardised interface between the two.
These pre-existing embedded services allow operators to achieve some degree of
competitive differentiation, but for new switch-based services there is a heavy
dependency on the switch manufacturer’s ability to provide these at a cost and
time-scale which is suited to the operating company’s commercial limitations and
opportunities. An appreciation of this commercial dilemma gives an insight into the
operating companies’ major interest in intelligent networks. The attraction is the
separation of the service logic into a separately controllable entity, where services
can be provided, modified and controlled by the operator rather than the switch
manufacturer (Figure 1.3).

1.1.2 Extracting the service logic

The switches in Figure 1.3, which have been upgraded to incorporate the IN
‘trigger’ functions, are renamed as intelligent network ‘service switching points’
(SSPs). An SSP ‘triggers’ when it detects a pre-set event during a call, such as a
particular string of digits dialled, or a particular line going off-hook, and it sends a
request message to a Service Control Point (SCP). This extra capability, which
transforms a telephone exchange into an IN SSP, acts as a sort of ‘go-between’, i.e.

Figure 1.2 ‘Embedded’ service logic

 Introduction to intelligent networks 5

Figure 1.3 ‘Separated’ service logic

a junction between the world of the transport network’s bearer circuit control and
the service logic running in the SCP.
The SCP’s function is to execute the service logic needed to implement the
required service. The SCP is shown, for simplicity, as a single node. It almost
certainly will have access to an internal or external database. In practice the SCP
may be a composite node, made up of several constituents, and the data, or service
logic, or both, can exist in a geographically distributed architecture.

1.1.3 Interactive voice functions

Many intelligent network services need the assistance of peripheral devices, such as
automatic voice announcements tailored for specific service situations, under the
control of the separated service logic (Figure 1.4). These devices are known as
‘intelligent peripherals’ (IPs).
An IN IP is a platform that supplements the transport layer’s abilities to conduct
dialogues with end-users. It acts as an interpreter between the SCP and the user,
enabling the service logic to communicate with the service user in order to obtain
more information than is passed across in the initial request (a set of dialled digits)
from the caller. Procedures have been standardised which enable an SCP to instruct
an IP to conduct a dialogue with a user and return the required information to the
SCP. This dialogue uses automatic voice guidance, requesting a response from the
caller in the form of a vocal reply, which can be recognised automatically, or in the
form of DTMF (dual tone multi-frequency) digits, keyed on the key-pad of the
telephone.
6 Intelligent networks – principles and applications

Figure 1.4 Automatic voice announcements included

1.1.4 The ‘classical’ IN architecture and its elements

Figure 1.5 shows a block diagram of the main network elements of the traditional
intelligent network architecture. It is often called ‘classical’ because its overall
structure has changed little over the past 15 to 20 years, and it is still relevant as a
guiding reference to the framework of the IN vision, which has endured throughout.

Figure 1.5 A ‘classical’ IN architecture

 Introduction to intelligent networks 7

Central to the architecture is the SS7 common channel signalling system [3].
SS7, sometimes known by the former title ‘C7’ (CCITT No. 7) was designed for the
transport of telephony call set-up messages, but it has been modified, by the
addition of non-circuit-related signalling capabilities, to carry IN messages between
the IN elements shown in Figure 1.5. With these modifications the SS7 signalling
system effectively provides the ‘glue’ that holds the IN elements together, and in
many respects is the most important intelligence component of all. The
enhancements needed to the SS7 specification to achieve this are the addition of
the signalling connection control part (SCCP) and transaction capabilities
application part (TCAP). These functions sit alongside the telephony user parts,
using the same underlying signalling data link capability. TCAP running over
SCCP supports a computer–computer communication framework and an IN
messaging protocol, known as intelligent network application protocol (INAP).
The SS7 network contains much distributed ‘intelligence’, both in terms of
information about the availability of the rest of the nodes in the network and in
terms of flexible ‘global title’ routing tables positioned at intermediate locations in
the network. The SS7 network enables the SSPs and IPs to communicate with the
SCPs reliably and in real-time. These aspects of signalling intelligence are
discussed further in Chapter 3.
Figure 1.5 also shows the IN service management system (SMS) and service
creation environment (SCE). A new service is developed, assembled and tested
using an IN SCE before being down-loaded, via the SMS to the various SCPs.
Typically SCEs use software components known as service independent building
blocks (SIBs), with graphical tools for linking and activating them. Service and
user data are stored in the service data point (SDP), and there are standards that
allow this network element to be remote from the SCP, although for convenience
the data are often held in the SCP itself. However, a separate SDP allows data to be
centralised and accessed by several SCPs. It also allows the SCP to access data in a
different network, and this could be in a service provider’s, or customer’s, system.
As well as organising the distribution and download of the new services, the
SMS also controls the network data population issues for new and existing services.
Service data is needed in the SSPs as well as the SCPs, although the former is likely
to use existing OSS (operational support systems) mechanisms for managing switch
data. Figure 1.5 only really shows the new elements for an IN-enabled network.
Most of the essential day-to-day background support functions, such as customer
provisioning, billing, performance monitoring and network management, all of
which need to be upgraded for new IN features, are not shown in Figure 1.5.

1.1.5 Locating the ‘intelligence’

A recurring point of interest with the choice of architectural arrangements is where

to put the intelligence. Before IN, network ‘intelligence’ was dispersed around the
network, and the call control was distributed along the speech path, and did not
divert to seek out specialised points of intelligence. For instance, number
8 Intelligent networks – principles and applications

translation facilities were usually provided at appropriate points along the ‘natural’
call chain, and the format of the network numbering scheme used to dictate the
physical path taken by the call.
However, with greater centralisation, the ‘intelligence’ tended to become more
concentrated and sophisticated. There was a consequential drive to provide the
intelligence in central locations, making for easier management than dispersed
intelligence schemes.
With centralised intelligence the service logic and data are usually physically
separate from the preferred speech path of the calls and we need to bring the two
into communication. As we have described, the standard IN method is to use
remote signalling to communicate between the switch and the central intelligence.
However, there are some circumstances where it is more practical to actually route
the call itself to a point of intelligence and then onward route to the required
destination. This is known as the ‘service node’ (SN) concept, where the transport
switching control and the intelligence centre are combined in a single specialist
node. SNs are sometimes referred to as ‘point’ solutions, or even ‘IN-in-a-box’
solutions, and are sometimes used to enable IN services to be provided quickly and
cheaply. The disadvantages are firstly that such solutions are not good for
scalability and secondly that they are provided on the call speech path. This means
that the speech circuits will almost certainly have to traverse greater distances, as
the calls now have to ‘trombone’ into and out of the SN before reaching the
ultimate call destination. With a long distance call this can lead to the need for
complicated network designs involving echo-cancellation devices as the transmis-
sion propagation delay becomes noticeable. The SN concept becomes attractive in
situations where IN triggering functions do not yet exist. In this case an option for
fast introduction of a Freephone service (for instance) is to introduce a single
specialist node, or sub-network of nodes, which either include service control
functions, or are capable of triggering to an external SCP. As we have said, this
leads to speech-path inefficiencies, but this solution may be feasible for an initial
low level of demand for a start-up service. This is sometimes known as ‘overlaying’
IN.
As suggested earlier, another commonly deployed ‘interim’ solution is the use of
a network node known as a ‘service switching control point’ (SSCP). In this node
an upgraded digital exchange is provided with an ability to break out from normal
call processing to an internal service logic module which behaves in a similar way
as an external SCP. There is an obvious performance advantage, because fast inter-
process messaging methods can be used instead of an external messaging system
such as SS7. A disadvantage is the increased data distribution task – data changed
centrally need to be relayed to the internal database on the SSCP. If the required
service logic is not locally resident then the node contains normal SSP functions
and so can trigger out to an external SCP for assistance in special cases.
The SSCP approach is understandably favoured by established equipment
vendors, because SSCPs offer the benefits of IN solutions to operators whilst still
broadening the market product range for the vendor. SSCPs in fact represent one of
 Introduction to intelligent networks 9

the earliest ways of implementing IN. They were actually used on occasion as
dedicated service nodes in central points in smaller networks.

1.1.6 What is an ‘intelligent’ network?

This chapter reviews some of the historical origins and concepts behind the
Intelligent Network ‘vision’ whilst continuing to introduce some of the network
elements such as service control points and service switching points, which are
needed in the embodiment of this vision. We also discuss some aspects of
introducing IN principles to existing digital switching networks, but first we must
define what we really mean by IN.
In the quest for the basic principles of IN we constantly return to this question of
what IN actually is. Whilst the principles have held, the intelligent network concept
has appeared in a variety of guises over the past 15 to 20 years. Also, because the
term was held in high regard in business cases, IN has sometimes been over-used as
an attention seeker. In fact the ‘de facto’ definition of the vision has broadened so
that the range of meanings now attributed to the term is so wide that the definition
of IN is a somewhat daunting task! However, in the main we restrict ourselves to
the original meaning of IN, which is the separation of service logic from bearer
control in telephone exchanges.
From a service viewpoint, the aim should be to organise services around the
users, rather than the other way round. A perceptive early articulation of the core
concept of the IN was ‘the creation of highly functional and highly customer-
specific services through software manipulation in a multi-vendor, multi-network,
environment [4].
With the recent meteoric rise in popularity of mobile telephony, another
viewpoint insists that, with a modern popular desire for people to be always in
contact, regardless of where they might be physically located, the most notable
capability of IN is that of universal mobility. The notion of IN has naturally
expanded to embrace the idea of intelligence ‘everywhere’, which is a theme that
we return to in later chapters.

1.1.7 The role of international standards

It is sometimes said that the advantage of standards is that there are so many of
them to choose from! Obviously this alludes to the problem of competing standards
and the dilemma that equipment manufacturers find themselves in when trying to
design their product range in advance of their competitors. How does one know to
which emerging standard to attach one’s allegiance? This is obviously a big
problem, and one that can have major financial consequences either way.
The international standards organisation surrounding the telecommunications
industry is vast – there is a large hierarchy of study groups, working parties and
committees. The input and output reports of these groups are formidable in their
10 Intelligent networks – principles and applications

volume, depth and detail. Decisions made at standards meetings can substantially
affect the major investment programs of the industry’s equipment manufacturers,
network suppliers and the operating companies, all of whom therefore take part in
the international standards debates with the intention of influencing the outcomes
towards their preferred directions. Consequently, much of the expertise in the fine
details of a complex topic such as intelligent networks resides within these
international standards working groups and sub-committees. The big challenge for
the industries supporting the standard-making process is to balance their investment
in this natural drive towards greater standards complexity against the practical
usefulness of the outcome of the process in their own organisations. A common
practical issue is the fact that the people who attend the standards working groups
are not always the same people who are tasked with solution implementation. Both
functions are usually full-time occupations, and the two disciplines need to find
time to communicate, which is not always as easy as it might seem.
We discuss the various standards bodies in Chapter 4, but in general, with our
detailed examples of working IN solutions, we adhere to the ETSI Core INAP
variant of the ITU-T Capability Set 1. This is because of the popularity of this
variant, which is now used widely in Europe and elsewhere. It is also the case that,
because of its relationship with the wider ITU-T standard [6], it shares common
roots with the US AIN standard and Asian standards. However, the aim of this book
is really to introduce the concept of IN, illustrating basic principles without
extensive comprehensive coverage of the detail. We therefore explain the
terminology we need to achieve this, but will not attempt to explain the larger
part of the detail which is to be found in the standard documents themselves.
The reason for the heavy emphasis on international standards for IN is that
digital switch technology has been developed over the past 30 years largely on a
proprietary and competitive basis. Different vendors and operators have worked to
provide the features they individually regard as important in their digital switches
and their controlling software suites. Therefore there has been little commonality of
internal structure, messaging, control methods, storage and retrieval systems and
suchlike, so there has been little scope for standardisation on the internal
architectural arrangements inside the switches.
Digital telephone exchange systems are still amongst the largest and most
complex real-time software systems in the world. To compound the complexity,
these designs have been (and still are) subject to regular modifications and upgrades
in response to intense commercial and regulatory pressures to provide new service
features. Consequently there are in existence many huge and monolithic telephony
switching systems, perhaps in the later years of their life cycles, with programs
which have been implemented in out-of-date programming languages and which
have been substantially modified, or patched, along the way. With many
modifications being made to these large systems, which were not originally
designed for the scale of upgrade to which they have been subjected, great care is
needed to ensure that maintainability is not compromised [5].
For this reason a large-scale overhaul and upheaval of the fundamental call
processing software in order to introduce the complexity needed for the IN
 Introduction to intelligent networks 11

breakout points was necessary. The network operators viewed this with trepidation.
In practice, the upgrades have generally been introduced gradually for this reason.
This goes some of the way to explain the slowness that has characterised the
introduction of IN features in the networks over the past 10 years. New builds on
switch software might be loaded every 6 or 12 months, and a limited amount of IN
triggering might be introduced each time. The amount has often been limited by
financial factors – usually business case justification is needed, in terms of new
revenue-producing service predictions, for each step along the way. The amount is
also limited by practicalities on how much change a new software build can
contain. IN features will be included alongside many other (non-IN) changes, bug
fixes, etc., and there has to be a limit on how much change an operational switch
can absorb without putting normal operation at risk. The period between builds is
usually unhurried for operational reasons – often a new build needs several months
to ‘settle down’. Inevitably there will be a certain amount of unexpected behaviour
in a new build, and time is needed for maintenance routines to adjust to this.

1.2 Why did we need IN?

We don’t like their sound, and guitar music is on the way out.
Decca Recording Co. rejecting the Beatles, 1962

1.2.1 The commercial drivers

From the network operator’s viewpoint, the sort of advantages promised by IN

which caught the imagination were:
. vendor independence
. fast service deployment
. customised services
. customer control of service data.

1.2.1.1 Vendor independence

After the major network operators had invested heavily in different varieties of new
digital networks of SPC exchanges in the early 1980s they began to discover that
introducing new service features was an expensive business. The industry entered a
phase where network upgrades were slow and expensive, and the operators had to
gear up for tortuous contract negotiations with their vendors, requiring the
production of water-tight specifications and rigourous test schedules before a new
service could be introduced for public service. Equipment vendors were generally
in a dominant position because the operator had no choice but to obtain expensive
proprietary network solutions from the suppliers of their existing network
equipment. Also, large amounts of patience were often needed. It sometimes
took several years to negotiate and specify a new service, capture the detailed
12 Intelligent networks – principles and applications

requirements, generate the design details and complete the testing. Then the new
service had to compete with other upgrades and features to be included in the
planned build release cycle.
Against this background, the prospect for the network operators of a new way of
introducing new services independently of the established process was very
attractive.

1.2.1.2 Fast and flexible service deployment

IN held the promise of a network operator inventing its own services, developing
testing and integrating them, and quickly rolling them out into the network, perhaps
just a few months after their conception. Without IN systems this could take several
years, by which time the original driver for the service might have disappeared,
resulting in the loss of a marketing opportunity.
Hence a generic network enhancement, such as IN, which might help end the
regime of service-specific network enhancements, was viewed with great interest
from operators’ business and marketing viewpoints. The procurement of facilities
for ‘in-house’ development of services using ‘service creation’ functions
(introduced earlier in this chapter in discussion of the SCE) made good business
sense.
Similarly, the opportunity for faster and cheaper methods of service introduction
meant that there was more chance of catching fleeting ‘windows of opportunity’ for
specialist services. For instance a voting service could be developed and installed
ready to run alongside a popular television programme.
As the process of service introduction became cheaper using IN features, it
became possible to be more experimental and inventive with new service ideas. If a
particular service IN does not generate sufficient revenue it can be disabled and
removed without heavy financial losses. In any case, services can be launched as
trials using IN, so that their potential demand can be assessed, before full-scale
introduction.
A basic characteristic of IN is the re-use of generic software functions for
different service scenarios, yielding obvious economies of scale. For example, a
‘number translation’ generic function will be re-used in different guises again and
again in a multi-service IN.

1.2.1.3 Customised services

The notion of a service creation environment (SCE) under the control of the
network or service provider, rather than the network supplier, became a
fundamental raft of the IN concept. A versatile SCE would enable the service
administrator to tailor services for particular customers. For instance, a user might
like a particular variant of a VPN (virtual private network) service whereby his
home phone is part of a VPN group during daytime business hours, but then would
be reverted to a normal private line in the evenings and at weekends. In principle
this sort of customisation is straightforward when the IN holds a service profile
database for the user. The SCP database would be checked for every out-going call
 Introduction to intelligent networks 13

anyway, so the extra feature of a ‘time-of-day’ routing plan is a relatively small

incremental feature.
Ultimately one can envisage that a customer’s set of services could be provided
by logic and data that are tailored to the customer’s own set of requirements. This
implies a movement towards a ‘customer-centric’ rather than ‘service-centric’
approach where the network operator, or the service provider, sells sets of service
features tailored to particular users. In other words, service logic and data provision
revolve around the customer’s own service profile rather than a set of stand-alone
service logic programs (SLPs).

1.2.1.4 Customer control of service data

A natural extension of customised services is to allow customers to have direct
access to a sub-set of their service data. This is commonly provided by allowing
users to have Internet access to a ‘diary’ profile in order to update their current
locations. Alternatively, the data could reside on the customer’s, or a third party’s,
computer and access provided from the network operator’s SCP.

1.2.2 IN as a universal integrator

A recent perspective is that IN might be the panacea that integrates all networks,
and incorporates all surrounding and ‘off-network’ intelligence. This is noticeable
where broadband architectures are concerned, where network intelligence has a role
to play with the background management tasks, such as linking session set-ups with
available bandwidth.
It is important to understand that IN features are added as enhancements to
networks that are not entirely stupid in the first place. Traditional telecommunica-
tions networks already have a certain amount of automatic ‘intelligence’ by virtue
of their signalling systems, by virtue of their switch-based services provided by the
switching equipment supplier and by virtue of ISDN features. It is useful therefore
to assess the position before the IN enhancements, to understand how existing
‘intelligent’ features work before adding in the IN features. This is particularly
important because there is a set of potentially complex problems, known as ‘feature
interactions’, which are to do with IN and non-IN features working together and
maybe against each other. To analyse and deal with these interactions it is
necessary to carefully study the origins and mechanics of both. For this reason we
spend some time (in Chapter 3) looking at the ‘intelligence’ in telecommunications
that is not actually provided by IN. In fact IN upgrades could not operate without
this ‘prior’ intelligence in public switched telephone networks (PSTNs).

1.2.2.1 Some other technical characteristics

We previously listed and described some of the commercial drivers for the IN
vision; IN has been driven as much by pragmatic network possibilities as by
commercial requirements. We have described the essential features of the IN
14 Intelligent networks – principles and applications

architecture; we now summarise some of the other technical features that have
characterised IN development:
. The construction of IN services has been based on the concept of generic
software modules, which we have already introduced as ‘service independent
building blocks’ (SIBs). Whilst there is valuable guidance on the use of SIBs in
the ITU-T standards, this is really an ‘off-shoot’ from the main purpose of the
standards deliberations, which was the definition of a standard real-time interface
between the service switching software in the exchanges and the separated call
control functions. Hence, whilst there are many ‘service creation’ systems, there
are no over-riding standard recommendations for the implementation details or
the interface with other IN elements. SCEs in use today are therefore
fundamentally proprietary to the different manufacturers.
. GSM (global system for mobile communications) radio networks use what is
essentially an application-specific IN for mobility management and for customer-
controlled call-management functions. This is discussed, as well as the specific
IN enhancements for the CAMEL (customised application of mobile networks
enhanced logic) system in later chapters.
. Service logic independence from underlying network technologies is provided by
‘middleware’ application programming interfaces. The IN aims to provide a
service-oriented network view, with the intention of concealing bearer-level
practicalities from the service level as far as possible.
. IN is a ‘coming together’ of the computing world and the traditional
telecommunications worlds. It makes use of ‘open’ computing interfaces, using
results from computer technology. An example of this is the ‘telecommunica-
tions information network architecture’ (TINA), which is an object-oriented
architecture, based on DPE (distributed processing environment) technology. In
fact IN can be viewed as presenting a generic API (application programming
interface) for network-transparent service provision, turning the network into a
‘programmable entity’ which integrates different connection network technol-
ogies. This is discussed in Chapter 6.
. As discussed earlier, much of the ‘intelligence’ of an IN is in fact provided by the
SS7 signalling system. For instance there are translation tables with the
signalling system (global title translations) which direct messages to appropriate
IN SCPs. Some Freephone implementations use the dialled number itself as a
global title and the translation tables can be updated in real-time to ensure that IN
queries always go to the appropriate SCP node. This is discussed further in
Chapter 3.
 Chapter 2
The foundations of IN

Drill for oil? You mean drill into the ground to try and find oil? You’re crazy.
Drillers who Edwin L. Drake tried to enlist to his
project to drill for oil in 1859.

2.1 The service switching function

As Chapter 1 intimated, the term ‘triggering’ in intelligent networks refers to the

process of breaking out from normal call-control procedures in digital telephone
exchanges to remote service logic. Triggering is fundamental because the original
purpose of the IN was to enable call control software to recognise pre-assigned
events and conditions and suspend call processing whilst it obtained further call
processing instructions from an external service control point.
However, when IN was introduced, the software which handles this ‘breaking
out’ process could not simply be added alongside existing call processing functions.
It had to be built in as an integral part of the basic switch call control. Because
different telephone exchange manufacturers used different design philosophies for
their call control, achieving a standard for the ‘break-out’ procedures was difficult.
In order to achieve progress in building the standard recommendations for
enhancing telephone networks with IN features, it was necessary to introduce a
hypothetical modelling device known as the ‘service switching function’ (SSF) to
the telephone exchanges.
The illusion was created of an apparent ‘separation’ between the new IN access
function (the SSF) and the existing call control function (CCF). In fact, because the
introduction of IN triggers could only be achieved by fundamental modifications to
existing call processing software, including the ‘embedding’ of the ‘break-out’
points in the CCF, there could not be a real functional split to correspond to this
modelling construct. For this reason, the specification of a standard interface
between the CCF and the SSF was not possible.
16 Intelligent networks – principles and applications

Although this issue caused some confusion early on (particularly when regulators
made attempts to enforce an open CCF – SSF standard), it did enable the modelling
process to progress to the completion of the development of a set of workable IN
standards.
This theoretical split between CCF and SSF is illustrated in Figure 2.1, which
shows a separation between the bearer control and the basic call process. The
bearer control is to do with the handling of the circuit-switched transport
connection, including the seizure and release of circuits. The basic call process is
concerned with functions such as the handling of address digits, call routing, and
any extra facilities provided, such as call forwarding or the presentation of CLI
(calling line identification) information.
Traditionally a single circuit-related network signalling messaging protocol has
provided both bearer control and call control signalling. This signalling protocol is
commonly ISUP (ISDN user part of SS7) [7]. However, bearer independent call
control (BICC) standards are becoming available that will enable the functions to
be handled separately [8].
The ‘customer data’ function in Figure 2.1 refers to switch-based information
held against particular lines. This could be data to do with the states of switch-
based services (such as call barring and call waiting) but it can also contain
information about addresses to be used during call set-ups. For instance, a
‘presentation’ number may be held against a line if a customer wants something
other than the network number of the calling line to be displayed at the called end’s
terminal for the CLI service. For incoming calls, a line may be on diversion to
another number, in which case the diverted number would be stored in the customer
data. This is essentially ‘non-IN’ data; it is data used for services provided by the
switches themselves without any IN assistance.

Figure 2.1 The CCF/SSF ‘split’ in the SSP

 The foundations of IN 17

Therefore, services such as ‘call forwarding’ can either be ‘embedded’ in the

switch call-control software or provided using IN features. However, the IN
features have to be aware of any related switch-based services and take account of
them. This is a facet of the ‘feature interaction’ issue, which was introduced briefly
in Chapter 1 and is discussed further in this chapter as well as at later stages in the
book.
Once triggering has occurred, the ‘IN call process’ function above the dotted line
in Figure 2.1 supports any dialogues required with external service logic. The
‘trigger table’ is accessed from the CCF to determine the conditions in force for a
particular trigger to occur, and it is also accessed from the IN call process to
determine which switch-resident data elements should be included in messages sent
to the remote service logic.

2.2 Triggering to remote service logic

The trigger (or ‘break-out’) points are implanted at strategic points in the basic call
control process. This then provides the opportunity for dialogue to take place
between the exchange and the control point that hosts the service logic.

2.2.1 Introduction to triggering

Once a trigger has occurred, there is a repertoire of 26 different INAP operations in

ETSI Core INAP2 that can be used for the communications between an SSF and an
SCF. The term ‘operation’ is used to mean either a request made by one end for
action to be taken at the other end, or to provide a response to a previous request.
There are five INAP operations available for communications between an SRF and
an SCF. The latter are used to obtain or provide further call-related information,
usually from the caller. This can be achieved, for example, by playing automatic
announcements and menu-driven invitations for the caller to respond verbally, or
by pressing handset keys to generate DTMF signals. The INAP operations are listed
later in Table 2.3.
Most CS-1 IN services are invoked by a ‘trigger’ event in the SSF, and an
InitialDetectionPoint (I_DP) operation is used to convey information about the
trigger event to the SCF. This information is carried in parameters, and the I_DP
operation has parameters such as service key (a numerical reference to the set of
trigger criteria that has been met by this particular trigger), address digits and a
trigger detection point number. The ETSI Core INAP protocol specifies the service
key as mandatory, and 16 other parameters as optional. The parameters are listed

2
Note: The relationship between the various standards is explored in Chapter 4, but for practical
purposes of explanation of IN principles we use the ETSI core INAP specification [10]. As we discussed
in Chapter 1, ETSI core INAP is representative, closely related to ITU-T CS-1, and has proved to be a
widely popular basis for IN implementations to date.
18 Intelligent networks – principles and applications

later in Table 2.2. Some of these parameters will nearly always be sent in any I_DP
(such as the called number – and probably the CLI) but others (such as the ISDN
bearer capability) will only be needed for more specialist services. It is the service
key that indicates to the SCP which particular service logic program is required to
run this service.
A simple example of an IN trigger is illustrated in Figure 2.2. This is a ‘dialled
digit’ trigger (formally known in the CS-1 standard [9] as the ‘analysedInforma-
tion’3 [DP 3] trigger detection point) and is in practice the most commonly used IN
trigger. There are many variations possible. It could be set to be an identifiable
dialling prefix, such as 080x (indicating a Freephone service in the UK) as shown
here. It could equally have been set for ‘digits ‘080815’ or just ‘08’. The request
message (the I_DP message) to the SCF could be sent immediately these digits
have been received, or it could be delayed until the remaining digits have been
received from the calling user.
It is often the case that the complete number length is unknown by the triggering
SSP. This is because many networks do not have a standard number of digits in
their customers’ telephone addresses and it is often only the destination parent local
exchange software that knows exactly how many digits are needed to uniquely
identify the called user or, in this case, the called service.
In the example of Figure 2.2, a minimum number length, say eight digits, can be
set in the data so that the I_DP could be sent to the SCF when the first eight digits
(‘08081570’) have been received by the call processing software. In this case the
SCP service logic may have information on the numbering scheme digit length,
and, after analysing the first eight digits it may know that just two more digits are
needed. It is able therefore to reply to the SSP with a CollectInformation message
(described in Section 2.5.1) requesting the two remaining digits and the IN request
can be processed when the SCP has the complete dialled address string.
However, if the digit length is unknown either in the SSP or SCP, then the
network must normally wait for a period of time (known as the ‘inter-digit time-
out’, which is typically 4 to 6 seconds) in case another digit should arrive. Another

Figure 2.2 A dialled-digit trigger

3
This is the ASN/1 (abstract syntax notation number 1) name for the analysed information trigger. More
information on ASN/1 is given in Section 2.2.5.3.
 The foundations of IN 19

way of avoiding this delay, which is used in some network implementations, is to

allow the user to key a delimiter digit (e.g. a ‘#’) to inform the network that the
dialling phase is complete.
Having received all the digits, the SCF will typically return an instruction (the
‘Connect’ message, which is described in Section 2.5.1) to connect the call onwards
to a translated number supplied by the SCF. This translated number, which is a
network routing address, might for instance depend on the time of day, the caller’s
line identity (CLI) or a host of other factors. It could be that part of the service
algorithm is to prompt the caller to enter an authorisation code using the DTMF
key-pad on the phone. The routing destination might then depend on the code
entered at this point.
Another common IN trigger is illustrated in Figure 2.3. This is the ‘off-hook’
trigger (formally known in the ITU-T CS-1 standard [9] as the ‘origAttemp-
tAuthorized’ [DP 1] trigger detection point). This trigger causes the IN request
(I_DP) message to be sent to the SCF when a user picks up the phone handset. This
trigger could be set, for example, for a phone where the account has been
discontinued because the previous tenant or owner has vacated the premises. The
phone company might want to keep the line connected, hoping to attract the custom
of the next occupant of the premises. An ‘off-hook’ trigger could therefore be set so
that when the new incumbent picks up the phone handset they are connected to an
announcement machine providing a welcome message.
In this case the SCP would reply to the I_DP with a message containing a ‘Play
Announcement’ operation (Section 2.5.1), which instructs the SSP to connect the
caller to an internal announcement machine, providing a reference number for the
particular announcement to be played.
In a refinement of this service the SCF might use a ‘Prompt and Collect User
Information’ operation (Section 2.5.1) to invite the caller to respond to an
automated dialogue, perhaps by entering a digit on the key-pad, or by responding
verbally. This mechanism is illustrated later in this chapter. The IN functional
entity that is used to play the announcements and conduct user-dialogues is an IN

Figure 2.3 An ‘off-hook’ trigger

20 Intelligent networks – principles and applications

SRF (specialised resource function). Depending on the capabilities of the particular

implementation this could be internal to the SSP,4 as discussed here, or externally
located in an IN IP (intelligent peripheral) node. Procedures for connecting and
communicating with IP nodes are described later in this chapter.
This same trigger point might be used to construct a ‘hot line’ service, whereby a
dedicated line in a public location (such as an airport terminal or railway station) is
connected directly to a pre-chosen destination – maybe a taxi company. In this case
the SCP would return a ‘Connect’ message containing a destination number address
for routing, instead of a ‘Play Announcement’ message, as in the previous example.
Without IN, ‘hot line’ services are available, but usually at the expense of renting a
private circuit from the network operator.
A further variation is the delayed off-hook trigger, where a call is automatically
routed to a pre-arranged number if no digits are dialled within, say, 5 seconds of a
phone going off-hook. This might typically be useful for an elderly person living
alone, where the pre-arranged number could be that of a community help-line.
This variation of service type that can be achieved with the same trigger point
illustrates the service independence that the IN infrastructure strives to achieve.
The same basic function of an off-hook trigger has been used to construct several
quite different services.
A third important trigger type, shown in Figure 2.4, is the terminating trigger,
which applies to incoming calls to a particular subscriber. This could be used for a
call forwarding service, as shown in Figure 2.4, where an incoming call is diverted
unconditionally to another line. The trigger could equally be set to activate on
‘busy’ or ‘no reply’, so providing a conditional call forwarding service.

Figure 2.4 A terminating trigger

4
Note: In an overview discussion such as this, the terms SSP and SSF are often interchangeable. Strictly,
the SSF is the IN access function, whereas the SSP is the point at which that function (and possibly
others) is located. Later on we are more careful when we need to distinguish between the physical SSP
network element and the associated SSF software functional entity. The same goes for SCP and SCF.
 The foundations of IN 21

2.2.2 Introduction to IN call models

State transition diagrams are normally used to chart the progress of a call through
its various stages in call processing software, and a representative simplification of
such an algorithm is shown in Figure 2.5.
Figure 2.5 shows an event sequence for a simple call, illustrating the call model
concept and the consequent ISUP messaging. A caller lifts the handset, causing the
state machine in the SSP to invoke an instance of the call model, starting it in the
‘off-hook’ state. When digits are dialled the call model goes to the ‘receive digits’
phase, then onto the ‘analyse digits’ state. When the call control software has
decided where to route the call, the call model goes to the ‘route call’ state, and,
assuming the call is destined for a user at a different exchange, an ISUP IAM
(initial address message) is launched on an appropriate telephony route. When the
far-end exchange has checked the address digits and has started to ring the
destination user’s phone, an ACM (address complete) message is returned,
followed by ANS (answer message) if the called user answers. The call is now in
the speech phase, and it is released when the users replace their handsets. This
instance of the call model has now completed its job and can be terminated after
any background housekeeping tasks, such as updating statistics and charging
records, have been completed.

Figure 2.5 A simple state transition diagram for a call set-up

22 Intelligent networks – principles and applications

It must be emphasised that Figure 2.5 is a simplification, intended merely to

illustrate the principles. It does not show all the states, nor does it use their correct
names. The formal state model for ETSI Core INAP [10] is shown in Figure 2.8,
where it will be seen that there are in fact simultaneous call model instances for the
originating and terminating halves of the exchange call processing. In Figure 2.5
we only show the state diagram for the originating side; in other words we see the
call from the caller’s, not the called party’s, viewpoint.

2.2.3 Trigger detection points

Triggering can occur at several places during the progress of a call, and Figure 2.6
shows four trigger detection points added to the simple call sequence shown in
Figure 2.5. The larger rectangular boxes in the diagram represent stages of relative
stability in the real-time call processing. They are, for this reason, termed ‘points in
call’ (PICs) and they have been renamed here with the formal CS-1 modelling
titles. The small boxes, the ‘detection points’ (DPs), are the points which generally
indicate a transition between states, and are where the IN ‘triggers’ can be
implanted. The full list of the ETSI Core INAP DPs is shown in Table 2.1.
IN triggers can be placed at several different points in the progress of a call set-
up, but triggering need not automatically occur on every call that encounters these
points. We can set various conditions that must be met before the processing is
suspended and a request message launched to the remote service logic. If the
conditions are not met then we can specify that normal call processing can
continue, or be terminated, depending on the circumstances. Similarly, if the SCF is

Figure 2.6 Trigger detection points

 The foundations of IN 23

Table 2.1 ETSI core INAP detection points

DP number DP name

1 OriginatingAttemptAuthorised
2 CollectedInfo
3 AnalysedInformation
4 RouteSelectFailure
5 OCalledPartyBusy
6 ONoAnswer
7 OAnswer
8 OMidCall
9 ODisconnect
10 OAbandon
11 Reserved (not yet used)
12 TerminatingAttemptAuthorised
13 TCalledPartyBusy
14 TNoAnswer
15 TAnswer
16 TMidCall
17 TDisconnect
18 TAbandon

currently unreachable, we can fall back to normal processing, or apply a default

routing, or fail the call after playing a suitable announcement to the caller.

2.2.4 Trigger types and trigger criteria

Most of the detection points listed above can be armed as Event Detection Points
(E_DPs) as well as Trigger Detection Points (T_DPs – commonly known as
‘triggers’). We discuss E_DPs later, and continue to discuss T_DPs here.
There are four categories of T_DP. These are:

& line-based (applicable to particular customers’ lines)

& office-based (applicable across the entire exchange)
& private-facility-based (e.g. only applicable within a particular Centrex group)
& trunk-group-based (i.e. applicable to a single telephony route).

All of these, except office-based T_DPs, can be unconditional, meaning that they
can be placed at many (but not all) DPs without any criteria being specified for
them to trigger. Obviously this could be a dangerous privilege to give to office-
based triggers – if every call passing through the exchange were to trigger, the
recipient SCFs would be at risk of immediate overload, as would the exchange’s
call processing software itself.
Factors used as DP criteria include dialled digit strings, nature of address, CLI,
class of service and call failure reasons. The set of relevant criteria obviously
24 Intelligent networks – principles and applications

depends on the DP. For instance, a ‘call failure reason’ can only be expected at a
DP which deals with call failures. The use of these as criteria is defined in (9), but
in practice much depends on what is feasible in different implementations. There is
scope for manufacturers to differentiate their products on different criteria selection
availability, but much depends on performance and memory overheads imposed by
an extensive range of criteria, as well as the complexity of software changes
needed. It also of course depends on what the network operator can afford to
specify, and this is likely to be based on service business case justifications for
particular features.

2.2.5 Trigger tables

Trigger data is stored in SSPs in data tables. As well as the DP criteria data just
discussed the SSP needs to hold information about what data to export when the
SSF generates IN requests to an SCF.

2.2.5.1 Trigger criteria data

Whilst the SSP does not contain any of the service logic for an IN service, it does
need to be populated with data to enable the CCF to recognise at every DP whether
the conditions and criteria exist that are necessary for a trigger to fire. This pre-set
triggering data will have been set up as part of the IN service provisioning process,
using the normal on-line data update mechanisms for exchange data.

2.2.5.2 Trigger ‘export’ data

The SSP needs to send some basic information in the request message to the SCP
for the service logic to perform correctly. The most important pieces of information
are usually the addresses of the called and calling parties, although other
information will be needed for different service circumstances. For instance the
service logic might need to know the category of the line, in terms of whether it is a
public phone box, or a business or residential line. This information may well have
a bearing on the logic to be carried out for this particular service instance. For
example, it may be inappropriate to offer a menu option of connection to a high-
cost premium rate service, chargeable to the calling line, for a call attempt from a
coin-box phone. If the trigger does occur in this case, the controlling service logic
may be required to flag an error report to a log file or a console, and to instruct the
exchange to connect dial tone and proceed as normal.
Similarly, the service logic may not be aware of the range of announcement
capabilities in the requesting exchange and may need this information to be
indicated in the request message. If the required announcement cannot be provided
by the originating SSP, then the service logic may need to instruct the exchange to
route the call to a separate network node – an IP, or maybe another SSP, which can
deliver the correct announcement.
The set of parameters to be forwarded from a particular trigger situation will be
defined in a trigger export-table in the SSP. To illustrate the data items that can be
 The foundations of IN 25

exported by an SSP in an IN message, Table 2.2 lists the 17 parameters specified in

the ETSI Core INAP for the I_DP. As indicated in the table, only the first of these
parameters is obligatory. Other parameters will be selected for inclusion depending
on what information is needed by the SCF service logic to handle particular trigger
requests.
An important feature of ETSI Core INAP is that an ‘extension’ mechanism is
provided to enable extra fields to be included in any message. These are to cater for
network-specific additions or for minor specification updates. To simplify the
illustration the I_DP extension field is not included in Table 2.2.

2.2.5.3 I_DP parameters5

Just six of these 17 I_DP parameters were designed specifically for INAP. The
remaining parameters have equivalents in the ISUP [11] or DSS16 [12] protocols,
so if the underlying network signalling is ISUP then the parameters are simply
copied across to the INAP message without change. We now describe these
parameters individually.
Service Key. This parameter links the trigger event in the CCF to the relevant
service logic program in the SCF. It enables the SCF ‘front end’ to route the request
to the appropriate application SLP (service logic program) within the SCP. This

Table 2.2 I_DP parameters

Parameter maps onto Optional (O) or manda-
I_DP parameter ISUP or DSS1 equivalent? tory (M)

serviceKey: No M
calledPartyNumber: Yes O
callingPartyNumber: Yes O
callingPartysCategory: Yes O
originalCalledPartyID: Yes O
locationNumber: Yes O
forwardCallIndicators: Yes O
bearerCapability: Yes O
eventTypeBCSM: No O
redirectingPartyID: Yes O
redirectionInformation: Yes O
iPAvailable: No O
iPSSPCapabilities: No O
cGEncountered: No O
additionalCallingPartyNumber: Yes O
serviceInteractionIndicators: Yes O
highLayerCompatibility Yes O

5
The parameter titles in Table 2.2 are presented using their ASN/1. (Abstract Syntax Notation No. 1)
names. ASN/1 is the formal language used by the ITU-T and ETSI for defining application-level IN
messaging.
6
Digital Subscriber Signalling System Number 1.
26 Intelligent networks – principles and applications

linkage of a set of trigger data to a particular service key represents the extent of the
CCF’s ‘knowledge’ about the service details that are held in the SCF.
Called Party Number. For an originating local exchange trigger, this is
normally the set of address digits dialled by the caller. In the case of a trunk
exchange trigger it could be the contents of the called party number parameter
received in an ISUP IAM on an incoming route. The purpose of the service logic
could be to translate these digits into another routing address to be returned to the
CCF in a consequent INAP message.
This parameter will obviously not be present in the case of an ‘off-hook’ trigger,
because no dialled digits would be available at this stage of the call.
Calling Party Number. This identifies the call’s origin, in terms of the
originating user’s network identity. In the case of a trunk trigger it would have been
received in the signalling set-up message (e.g. the ISUP IAM as before).
Calling Party’s Category. This indicates the type of the calling termination.
Possible options include:

& operator (speaking a specified language)

& ordinary subscriber
& priority subscriber
& payphone.

Original Called Party Identity. This contains the digits the caller originally
dialled. These may have been changed earlier in the call if a ‘call forwarding’
service had been encountered.
Location Number. This applies to calls from mobile subscribers. Because the
Calling Party Number for a call from a mobile user gives no information about the
caller’s geographic whereabouts, the Location Number parameter is provided to
carry data that can represent the geographical location of the call’s originator.
Forward Call Indicators. This is a two-octet set of indicators containing
information about how the call should be routed at intermediate exchanges. The
indicators are as follows:

National/international call indicator. This tells the destination whether a call

should be treated as an international or national call.
End-to-end method indicator. This specifies which, if any, method is to be used
in the network for conveying end-to-end information. If this information is to be
transferred, there are two options – the ‘pass-along’ method (using a dedicated
ISUP message) or a connectionless network signalling procedure.
Interworking indicator. This says whether or not SS7 signalling has been used
throughout the entire network connection for this call.
End-to-end information indicator. This indicates whether or not there is further
end-to-end information available at the sending end. If there is, the receiving
exchange will request it before alerting the called party.
 The foundations of IN 27

ISDN user part indicator. This tells the recipient exchange that ISUP has been
used all the way so far in the call connection set-up.
ISUP preference indicator. This says whether or not ISUP is preferred for the
complete connection.
ISDN access indicator. This says whether the originating terminal is ISDN or not.
SCCP method indicator. This is for end-to-end transmission of information via
SCCP, and says whether connectionless or connection-oriented (or both) methods
are available.

Bearer Capability. This is a Q.931 information element that describes the

connection type and the network signalling capability. It is carried in ISUP in a
parameter called ‘User Service Information’ and provides values for items such as
coding standard, information transfer rate and Layer 1 protocol type.
Event Type BCSM. This parameter is the detection point (DP) number for the
trigger occurrence. Figure 2.8 shows the DP number on the BCSM.
Redirection Party Identity. This gives the network address from which the call
was last redirected (diverted).
Redirection Information. This carries information such as redirecting reason,
redirecting counter, etc.
‘IP Available’ Indicator. This says whether or not an IP is attached and
available at the triggering exchange.
IP SSP Capabilities. This gives information on whether the triggering SSP has
an internal SRF that could assist with the triggered service.
Call Gapping Encountered. This indicates if the triggered call has encountered
(and survived) call gapping.
Additional Calling Party Number. This parameter is for the CLI provided by
the calling user’s access signalling system (rather than that provided by the serving
local exchange).
Service Interaction Indicators. This parameter is intended for passing
indicators about relevant switch-based services to the SCP. As the contents are
not standardised they can vary greatly between implementations.
High Layer Compatibility. This gives information about the technical
characteristics of the ISDN-Teleservice. Examples of teleservices follow (ITU-T
recommendations I.240 and I.241) are:

& Group 4 fax

& Teletex (for exchanging text documents)
& Videotex
& 7kHz telephony (for high quality voice services).

2.2.6 Event detection points

So far we have been concerned with T_DPs, which are pre-armed trigger detection
points, waiting to be ‘fired’ by call control software during normal call processing.
28 Intelligent networks – principles and applications

IN also has ‘event detection points’ (E_DPs), which are dynamically armed during
a call set-up by instructions from SCFs. For example, the SLP may need to be
advised if a call attempt meets a called party busy signal.
The INAP operation used by the SCF to arm an E_DP at an SSP is the
RequestReportBCSMEvent. If the CCF subsequently detects this E_DP event, the
SSF will respond to the SCF with an EventReportBCSM operation.

2.2.7 The CS-1 basic call state model

The purpose of the basic call state model (BCSM – defined in Q.1214 [9]) is to
provide the SCF with a representative view of the call processing in the SSF. The
SCF obviously has no need for the details of the call processing implementation. It
just needs to know enough to enable it to process the service logic correctly.
The CS-1 BCSM postulates that call-control software in an exchange can be
modelled as if a call has two halves – an originating half and terminating half (Figure
2.7). Two separate but interconnected call models, namely the O_BCSM and the
T_BCSM, represent the call. The full Core INAP BCSM is reproduced in Figure 2.8.
Principally, what Figure 2.8 shows is that the originating (left-hand) side
(O_BCSM) monitors the call at the incoming trunk or line, and the terminating
(right-hand) side (the T_BCSM) monitors the outgoing side. The T_BCSM has
fewer states than the O_BCSM. This is because the T_BCSM is not invoked until
the O-BCSM is some way into the processing of a call, and it so exists for a shorter
period of time.
In fact the majority of the common IN services can be constructed using only
originating-side triggers. Hence, for an originating call from a local exchange, the
O_BCSM handles the bulk of the set-up stages. It covers the initial authorisation
(PIC 1), where the software checks that the caller who has just lifted his handset is
in fact authorised to make calls and can indeed be supplied with dial tone. It then
collects (PIC 2) and analyses (PIC 3) the digits the caller has dialled. It then
launches the ISUP IAM (or equivalent) with the intention of alerting the called
party (PIC 4).
PIC 4 is the stage at which the terminating half of the call model is invoked. This
is indicated by the (dotted) communication line (1) in Figure 2.8. At the terminating
side, the call enters the formidably named ‘T_Null & Authorise Origination_

Figure 2.7 ‘. . .. a call has 2 halves . . .’

 The foundations of IN 29

Figure 2.8 The combined BCSM for CS-1

Reproduced, with permission, from Figure 4–9/Q.1214 ITU-T publication ‘Distributed Functional Plane
for Intelligent Network CS-1’ [10/95].

Attempt’, moving through the subsequent states as the call progresses into the
speech phase, where both sides of the call model achieve the stable ‘active’ state.

2.2.8 Notification and response DP types

It is possible to set a CS-1 detection point merely to inform the service logic that an
event has occurred. This type of DP is called a T_DP-N (trigger detection point –
notify), distinguished from a T_DP-R (trigger detection point – response), which is
the one we have generally discussed so far. Call processing is suspended at a T_DP-
R, awaiting a response, whereas CCF processing continues with a T_DP-N, and no
reply is expected.

2.2.9 Triggering dynamics

The various triggering functions we have been describing are illustrated in Figure
2.9.
This diagram shows the interaction of the three main IN call processing
functional entities, the CCF, SSF and SCF, for different T_DP types. It shows the
basic call manager (BCM) functions in this role. The basic call manager
30 Intelligent networks – principles and applications

Figure 2.9 An illustration of the dynamics of triggering

Reproduced, with permission, from Figure 4–10/Q.1214 ITU-T publication ‘Distributed Functional
Plane for Intelligent Network CS-1’ [10/95].

encompasses the BCSM and the associated DP processing functions. Also

introduced here is the CS-1 feature interaction manager (FIM).
Figure 2.9 shows five CCF trigger events occurring sequentially. The first is an
unarmed DP (really a non-event), where a message notionally goes from the CCF to
the SSF, where it is checked for arming criteria and found to be not armed. The
next illustration is a T_DP-N trigger, where the call processing resumes
immediately and the SCF is informed when the event occurs. This might, for
instance, be used for a service that needs to log the number of calls sent to a
particular destination address. Next is a T_DP-R, where call processing is
suspended until a reply returns from the SCF. In this case the reply initiates a
 The foundations of IN 31

BCSM instance in order to arm the required E_DP-R (event detection point –
response), by sending an appropriate EventReportBCSM operation. When this DP
is encountered, an E_DP-R response is sent to the SCF. Lastly, an E_DP-N (event
detection point – notify) is returned to the SSF and when this event is encountered,
the event is reported via the SSF, but call processing continues.

2.2.10 The CS-1 feature interaction manager

The FIM has the following main functions:

& Traffic management. This function checks to see whether any controls such as
call gapping or service filtering are active before passing received messages to
the SCF. Call gapping and service filtering are IN mechanisms (discussed later
in this chapter and illustrated by examples in Chapter 7) that aim to reduce
congestion problems at SCPs by reducing the quantity of messages sent from
SSPs.
& Feature interaction control. Rules for controlling interacting features are
applied here. The aim is to manage the current service request in the context of
other outstanding service requests, or active service features, particularly
switch-based service features that may have been invoked independently. For
this reason, close co-operation with CCF feature management is needed here.
Also there are DP processing rules to be applied. For instance, a detection
point could conceivably be armed both as a T_DP and an E_DP with the same
trigger criteria, so consistency rules are needed to ensure that there is always a
single point of control from the SSF. Obviously the ‘executive’ control point in
the SSF should only communicate with a single SCF service-logic program-
instance at a time in this event. Generally, Notification events and triggers take
higher priority than Request events and triggers, and an E_DP takes higher
priority than a T_DP if both are armed at the same DP. In this case, the T_DP is
processed after the E_DP if (and when) the latter results in termination of the
existing SCF control relationship.
A control relationship persists as long as there is at least one E_DP-R armed
for at least one CS (call segment) of a CSA (call segment association). If only
E_DP-Ns remain, the active control relationship changes to a monitor
relationship, and it can coexist with a separate control relationship on the
same call.
& Message formulation. This is the identification of the correct destination
service control function and the formulation of the request message (e.g. an
I_DP operation), with the required export data, to the appropriate signalling
handling function.
32 Intelligent networks – principles and applications

2.3 Hosting and creating IN services

Whilst a major part of the cost of an IN implementation lies in the switch re-
arrangements we have been discussing up till now, we now need to look at the
functions outside the SSP that actually direct the service itself.

2.3.1 SCPs and SCFs

For early IN applications, in the late 1980s, the prime requirement was for an
efficient external database for number translations. So the first SCPs tended to be
service-specific, specifically designed for translating numbers such as those
prefixed with digits such as 1800 (in the USA) for the universally popular
Freephone services. In these implementations the transaction response from the
SCP was a reply message containing data which was to be translated into a specific
action by the SSP, which had knowledge of the specific context of the invoked
service.
The term ‘service control point’ actually refers to a physical network node
containing the hardware on which the control functions run, but it is often used to
loosely represent the external service logic, and supporting structure, with which the
SSP communicates when it activates a trigger request. The IN function providing
the remote service logic is more properly termed the SCF (Service Control
Function), and it is commonly located in a seperate SCP in traditional IN solutions.
However it can be located in other places, including the SSP node itself. In this
case, where the service logic is co-located with the call control function, the SCF is
still ‘remote’ because it is functionally external to the normal call processing
software.
The prime purpose of the SCF is to provide the software environment that allows
the execution of the service logic programs whilst providing the necessary support
functions such as signalling access and transaction control, logic program selection,
provisioning and management. The SCF must be able to control interactions
between, and simultaneous invocations of, multiple SLPs and SLP instances.

2.3.2 Service data considerations

Early IN services were provided on service control points in blocks of logic and
associated data, which would include service housekeeping data and network
routing data along with customer-specific data such as account details, service
preferences and redirection numbers. A danger here is that when customers
subscribe to several services their identity data (CLI, VPN groups, etc.) might be
provided several times over. This can come about if service ‘products’ are marketed
and implemented separately within a sales department of an operator’s
administration.
 The foundations of IN 33

A more flexible approach can be to store ‘profiles’ of customer data, so that when
a customer is involved in any service invocation, that customer’s particular profile
is accessed, rather than a block of service-specific data. Using this scheme, data are
organised into customer-specific, rather than service-specific, data. These ‘user
profiles’ are then available for all IN service invocations. The focus switches from
the service to the customer, who benefits from a unified set of services. This is
easier to use than a set of discrete service packages, with perhaps different
authentication passwords for different service applications.
ITU-T defines the service data function (SDF) for the management of IN data, and
identifies an SCF–SDF interface for accessing the data. The ETSI Core INAP
standard SCF–SDF interface is defined in [13]. The SDF can be located in an SCP (or
other) node, with the associated SCF service logic, or in a dedicated service data point
(SDP), where the data is to be accessed from several different SCPs in the network.
The SCF–SDF interface is sometimes used to access switch-based data from an
SCP. An example would be where an SCF service needs to make a decision based
on the state of a switch-based call diversion service. The service logic may require
the SCF to retrieve the address digits of the customer’s diversion number. In this
configuration the SSP can be viewed as including an SDF, and the SCF–SDF INAP
interface is in parallel with any existing SSF–SCF interface.

2.3.3 Multi-service SCPs

The growing interest in virtual private networks led to the use of specific instructions
from the SCP to the SSP. This started the evolution of the SCP from a relatively
simple database facility into a wider network control point, which housed service
logic as well as the database. This trend continued towards what was later realised to
be unrealistic limits with the Bellcore proposals for IN/2, which are further
discussed in Chapter 4. IN/2 attempted to move all service logic to the SCP, ignoring
existing switch implementations that also contain supplementary service logic.
A typical architectural structure for a multi-service SCP is shown in Figure 2.10.
This shows the SLEE, which acts as the ‘front-end’ of the SCP, passing requests and
responses onto the correct (‘back-end’) service logic program (SLP). The SLP is
referenced by the ‘service key’ (SK), which is inserted into I_DP messages by the
SSF (see Section 2.2.5). The SLEE is discussed in more detail below.

2.3.4 Service logic execution environment (SLEE)

Externally, the SLEE appears as the ‘front-end’ for the SCP, because, as already
discussed, it arranges for incoming requests to be handled by the appropriate
service logic program. It actually does far more than this, as its prime role is to host
the SLPs. The SLEE provides the operating system for the execution of the SLPs,
and presents everything that is needed in a platform for service execution to its user
applications.
34 Intelligent networks – principles and applications

Figure 2.10 A multi-service SCP

The SLEE provides management and housekeeping interfaces for SLPs, enabling
their associated objects to be instantiated, activated, deactivated and withdrawn. It
provides and manages local resources needed by the SLPs, such as transient call
data, threads, clocks and synchronisation utilities. The SLEE will co-ordinate
service execution, providing scheduling and prioritising services; it will monitor for
errors and issue fault reports where needed.
The SLEE also organises the database services needed by the SLPs for their
service data, which is provided by an SDF. The SDF can be co-located with the
SCF at an SCP, as indicated in Figure 2.10, and data can also be requested from an
external SDF, hosted in a dedicated SDP. This would be the case where, for
instance, customer profile data have shared access from SCFs at several SCPs.

2.3.5 Service creation environment

Chapter 1 discussed the original ‘jewel in the crown’ for the IN vision, which was
the prospect for network operators to be able to control the creation and
modification of services, and to be able to introduce them to their real-time
networks quickly. This would do away with the traditional long and expensive
process of ordering and negotiating new services from equipment suppliers and
enable operators to corner niche market opportunities before competing operators.
The term ‘service creation environment’ (SCE) refers to any platform designed to
meet this need for crafting new services that can be hosted on an IN SCF.
The standard IN architectures (such as Figure 1.5) show the SCE function (the
SCEF) with an interface to a service management function (SMF), which, in turn,
interfaces to the SCF. The SMF has a secure interface to the SCF, and so SLPs that
have been created and tested at the SCEF are downloaded to the SCF under
 The foundations of IN 35

carefully controlled conditions by the SMF after the SMF is satisfied that the SLP is
network ready.
Traditional telephony services undergo rigourous testing before being brought
into service. In the SCE approach to service development therefore it is necessary
to provide tools for off-line checking and simulation before the service is allowed
to operate in the live network. As well as SSP and SCP simulators, network
operators will typically have captive installations for pre-service testing, and
operational SCFs are likely to have facilities for enabling the service to be trialled
in a controlled environment before full launch.
From a user’s viewpoint, the SCE function usually takes the form of a graphical
interface on a computer screen. This enables the designer to build up a
representation of the service logic pictorially, using a set of defined building
blocks and procedures. The service building blocks are called SIBs and they are
discussed further in Chapter 4.
A common technique in service creation systems is the use of a ‘tree’ structure
for expressing the flow of the service logic. An example of a segment of a simple
incoming call management service, where calls to a customer’s number are
diverted to a voice mail system at certain times, is shown in Figure 2.11.

2.3.6 Operational support systems (OSSs) for IN services

Intelligent networking is really about the fast introduction and efficient delivery and
management of a multitude of services, and this depends as much on the
operational ‘surround’ as on the complexities of the service logic in the SCP and
the associated real-time IN network components.
There are two broad categorisations for OSS, and these are service management
and network management – although the two are of course strongly inter-related.

Figure 2.11 A service creation tree

36 Intelligent networks – principles and applications

We return to service and network management several times during the course of
this book, but for the time being some of the main functions provided by the two
disciplines are illustrated in Figure 2.12.
Service management has overall control of the initial loading of a new service
and its data, rollout of new services, taking orders for the services from potential
customers, organisation of billing and revenue-apportionment schemes and,
generally, all customer and business-facing issues.
Network management is concerned with the smooth operation of the network
itself and it has responsibilities associated with network planning, configuration and
upgrade (to provide the infrastructure required for the IN services). It also has
responsibility for network data updates – e.g. for SCP number translations, SSP
trigger tables7 and network routing data – and for in-service support issues such as
maintenance, software and hardware routining, fault diagnosis and fault clearance.
Network management is also responsible for the performance aspects of network
control, such as statistics production and traffic congestion control. Last, and
certainly not least, the successful transfer of call charging records from their points
of origination to the appropriate billing centres is also a network management
responsibility.
As we imply in Figure 2.12, the service management function pervades most IN
elements, although, as we indicate, it is likely to have lesser involvement with the
IN IP because this does not generally hold volatile service-related data. However,
an efficient service management data update route is needed for service data held at

Figure 2.12 Management environment for IN network elements

7
Section 2.2.5 describes the SSP IN trigger data and Section 2.6.3 further discusses some particular
issues relating to SSP data management complexities.
 The foundations of IN 37

the SCP and SDP nodes, and, for changes to service triggering parameters, the
service management needs to have access to trigger table data held on the SSP.
Network management functions are mainly targeted on the high performance
real-time elements. These are primarily the SSP, IP, SCP and, to a certain extent,
the SDP (for the case where a service’s data is stored separately from the service
logic in the SCP and needs to be accessed in real-time). However, the SCE is likely
to have less involvement with the network management systems, so a reduced
linkage is shown between SCE and the network management in Figure 2.12.
Generally speaking, standards are available for management interfaces in the
lower part of the diagram (for the real-time elements) [49–51]. However, the
applicability of standards decreases as we ascend through Figure 2.12. This is
because, as we move upwards into the SCE area, the variety of the proprietary
interfaces in existence tends to increase because we are approaching the domain of
business-specific and necessarily non-standardised methods and policies for service
strategy, revenue generation and customer relationships.
In most cases the design of the service logic and the real-time IN architecture is
actually a smaller task that that of organising the support ‘surround’ needed for the
new service to work in the operational environment suggested by Figure 2.12. OSS
covers an extremely wide range of network support activities, as suggested by the
example lists of functions in the diagram. The details of operational support
systems (OSSs) are outside the scope of this book, and the reader is referred to [48,
49] for coverage of service and network management systems and their standards.
Chapter 4 below does include some further discussion on the service management
recommendations that were published for the ITU-T IN CS-2.

2.3.7 SCP ‘non-functional’ requirements

We have so far described the ‘functional’ requirement of SCPs, i.e. the things they
are expected to do to perform their main task of responding to requests for
assistance from the underlying switched network. However, there are several ‘non-
functional’ expectations for SCPs; these are necessary supporting attributes, or
behaviour characteristics.

2.3.7.1 Continuous availability

Telecommunication systems are traditionally renowned for their rigourously high
standards of availability – a factor that has always fundamentally influenced the
design of central processor units in digital exchanges. Transaction processors used
by banking systems, for instance, must have high standards of security in terms of
data integrity, and must have comprehensive methods of roll-back and roll-forward
to ensure that data are preserved to the last detail under failure conditions.
However, even these systems are allowed ‘down-time’ at quiet periods, for
upgrades and maintenance activities. In the case of telecommunications equipment,
even a 5 minute planned outage in the middle of the quietest night is often regarded
as intolerable. A distinctive feature therefore of the design of telecoms control
38 Intelligent networks – principles and applications

equipment is the provision of mechanisms which allow real-time on-line data

updates. These mechanisms allow data updates to be made whilst the system is
operating, usually at normal load. They also allow software enhancements to be
carried out while the system is operating. In practice, a system ‘restart’ period of
perhaps 30 seconds (usually at night) is often tolerated, to allow the changeover to
the new software, but this is usually conditional on there being no calls in progress
lost.
Unsurprisingly, therefore, performance criteria for SCP nodes are usually equally
rigid. However, for SCPs, this can be achieved by redundancy, using alternative
network routing arrangements, multiple sites, diverse routes, redundant commu-
nication links, or a workable combination of all of these. In the early years of
intelligent networking, SCPs were based on digital exchange designs, because of
their proven availability standards. Following this, there was a move towards
banking-community transaction processing machines. However, because of their
extremely high-standard security requirements, transaction processors are expen-
sive. It was gradually realised that data in telecommunications does not perhaps have
quite such high security needs, and that although the ‘fault-tolerant’ transaction
processors are often advertised as ‘non-stop’, they were not continuously available –
requiring down-time for system maintenance and upgrades. Nowadays the trend is
towards achieving full availability through over-provisioning of cheaper, less fault-
tolerant but high performance computing units.
However, the critical issue for continuous availability is software reliability,
particularly in complex software systems. The current trend is towards object-
oriented design, because experience with systems which have been developed using
these techniques has generally demonstrated higher productivity and lower fault
rates.

2.3.7.2 High performance

The aim is to optimise the system’s performance, particularly at the high end of its
design load. Overload protection mechanisms should not be allowed to compromise
system performance. On the other hand, an application’s performance achievement
should not suffer because a neighbouring service application receives a large
amount of transaction requests. Flow control mechanisms should be sufficient to
protect the non-offending application process. There is often a need for resource
reservation schemes to preserve processing capacity for high revenue-earning
services, or for particularly important customers. The provision of a single resource
pool for the whole node is usually more efficient than providing local ‘pockets’ of
such functionality.

2.3.8 Future trends for SCPs

So far we have mainly described IN architectures with centralised service logic,

implicitly hosted on stand-alone, monolithic, SCP processors. The configurations
discussed have, for simplicity, shown SCPs connected to many SSPs via SS7
 The foundations of IN 39

networks using relay points and signal transfer points. Data and programs would be
loaded via dedicated service management systems, directly connected to the SCPs,
which are often configured in pairs for security. Data would be mastered on a
separate database, or one of the SCP’s internal databases would act as ‘master’ and
the other SCP’s database would follow in step with data changes.
However, SCP evolution trends have generally tended towards more flexible
computing structures with arrays of co-operating servers hosted on a bus and rings.
This introduces the concept of multi-processing to SCPs, providing easy scalability,
enabling the addition of new functionality servers when required. The design trend
has continued towards the prospect of fully distributed computing systems, when
the ring structures have expanded and effectively become wide-area networks
connecting geographically dispersed servers and databases. Distributed service
logic is discussed further in Chapter 6.

2.4 The intelligent peripheral

Much of the discussion so far in this chapter has centred on the relationship
between the SSF and SCF and how we can reach out from normal call processing to
supplementary service logic which can change the course of the call, depending on
data held there. We saw in Chapter 1 that many IN voice services also need some
interaction with the caller, either to play an announcement or to collect some
further information to supplement the dialled digits to allow the service logic to
proceed to completion. For example, a caller may be offered a choice between
long-distance carriers, so a menu of options would be presented verbally and callers
invited to indicate choices verbally, or by pressing key-pad digits.
To provide this facility the IN IP was introduced in order to temporarily
terminate a speech channel on behalf of the IN service logic and to provide a means
of collecting further instructions from the user about the subsequent call handling.
The SSP will already have a repertoire of announcements, such as those that can be
played following receipt of call rejection messages in normal call handling. It may
be that the IN service logic needs to instruct the SSP to play one of these standard
messages, or it may be that a specialist message is required. In the latter case the
call could be routed to an IN IP.
POTS users have a limited signalling repertoire available. Traditional dials used
make-and-break pulses, which can only communicate with the network using the
digits 0 to 9. The more common DTMF key-pads are slightly better off, having *
and # keys so that codes for activating switch-based services such as call diversion
can be used. In the early days it was simply not envisaged that users would want to
use their signalling interfaces to do anything more that dial destination telephone
numbers comprising digits between 0 and 9.
Standards are not specific about whether the IP should be seen as a network node
or terminal equipment, and both options are included in the ITU recommendations.
If it is to be a network node, the natural interface to the SSPs is ISUP, which is a
network signalling system; if the IP is to be terminal equipment, a customer
40 Intelligent networks – principles and applications

signalling system (such as DSS1) is appropriate, and is more convenient for

terminal manufacturers. The question hinges on whether the IP is seen as a public
or private resource, and there are identified needs for both.
Whilst we have introduced the IP as a new network node for these user-
interaction features, the functionality needed (termed the ‘specialised resource
function’ – SRF) could physically be located in an SSP. In fact SSPs almost always
have a degree of SRF built into them, e.g. for playing general-purpose
announcements. However, for more specialist dialogues the SRF can be located
in a separate IP node, as illustrated in Figure 2.13.
The SRF acts on behalf of the SCF. It is the SCF that arranges for an SRF to be
brought into a call and the SRF in turn reports its results back to the controlling
SCF. When the SCF service logic decides that it needs further information from the
caller in order to proceed, it instructs the SSF to arrange for the caller to be
connected to an appropriate SRF, which in Figure 2.13 is located in an external IP.
There are several variations on the mechanisms used to bringing an IP into play
for a call. One of the commonly used sequences is known as the ‘Assist’ procedure,
which is described in the next section. The Core INAP operations available for the
SCF to instruct the SSF to make a connection to an SRF are the ‘establish
temporary connection’ (ETC) and the ‘connect to resource’ (CTR) operations. The
CTR is typically used for connections to SSP internal functions, and the ETC
(illustrated in Section 2.4.1) is generally used for connections to external IPs.

Figure 2.13 The IP’s role

 The foundations of IN 41

2.4.1 The Assist procedure for IP connections

Figure 2.14 shows the network messaging relationships8 and Figure 2.15 shows the
sequence of messages between the elements. We assume that the IN service logic is
invoked by an I_DP message from an SSP (message 0 in the diagrams).
After the appropriate SCP has been invoked by message 0, the SCP instructs the
SSP to establish a temporary connection with a nominated IP node (message 1), and
the SSP sets up a bearer connection (message 2). The ETC message contains two
essential co-ordinating parameters that must consequently be passed forward to the
SRF in the IP node in the bearer signalling. These two parameters are the ‘SCF
identity’, which enables the IP to identify the controlling SCP and the ‘correlation
ID’, which is a reference number sent directly from the IP to the SCP in the ARI
message (number 3). This enables the SCF to associate this new transaction opened
with the IP with the original transaction with the SSP.
Having made this association, the SCF sends a ‘prompt and collect’ message
(message 4) to the SRF instructing the speech resource equipment to present
audible choices to the caller. It also instructs the SRF to prepare to receive a
response from the caller over the transmission path and to return this to the SCF.
Because there will be a delay whilst the user–IP interaction takes place, the SCP
instructs the SSP (message 5) to reset its covering timer for a response to the
original I_DP.
After information has been obtained from the caller by the SRF, it returns
(message 6) a ‘return results’ message in reply to the ‘prompt and collect’ of
message number 4. The SCP hopefully now has enough information to instruct the
SSP on what to do next, such as route the call to a particular destination address.
The SCP therefore ends the dialogue with the IP (message 7) and instructs

Figure 2.14 Assist procedure messaging relationships

8
Acknowledgment is made here to the ITU-T recommendation for the Assist procedure description in
Q.1219 Appendix 4.2.
42 Intelligent networks – principles and applications

Figure 2.15 The Assist procedure – message flow

(message 8) the SSP to disconnect the ‘call’ to the IP and to set up a call attempt to
the required destination (message 9).
The bearer signalling transport protocol between SSP and IP depends on
individual network arrangements. Commonly used transport protocols are ISUP
(used in our example), TUP derivatives and Q.931. ISUP has the co-ordinating
parameters specified and they can be carried in the initial address message (IAM)
but this is not the case with TUP derivatives, such as IUP. In this case the parameter
contents might be included as a prefix to the address digits in the IAM (or IFAM).9
It will then of course be necessary to work around any local limitations caused by
number length restrictions.
In summary, the numbered stages in Figure 2.14 correspond to those in the
message flow of Figure 2.15, representing the following actions:

0. I_DP from SSP to SCP.

1. ETC from SCP to SSP.
2. The SSP routes the call to the IP.
3. IP sends ARI (Assist request instructions) to open a transaction with the SCP.
4. SCP sends P&C (prompt and collect) back to the IP.

9
This may also be the case with some implementations of ISUP. Although the field specification for SCF
Id. and Correlation Id. are included in Q.763 [7] corresponding procedures were not specified in earlier
standards.
 The foundations of IN 43

5. The SCP instructs the SSP to reset its application timer.

6. The IP returns the collected information to the SCP.
7. The SCP terminates the transaction with the IP.
8. The SCP sends DFC (disconnect forward connection) to the SSP to disconnect
the temporary connection to the IP, and Connect is also sent to instruct the SSP
to route the call to the required network destination.
9. The call is onward routed.

The use of the ‘Assist’ procedure described here is further illustrated in Chapter 7,
where we show how it can be used several times in the same call to offer a ‘call
completion’ facility for a directory enquiry service.

2.5 INAP (the intelligent network application protocol)

The real-time IN architecture elements (SSP, IP and SCP) and the associated
functions (SSF, SRF and SCF) have now been introduced and so we next return to
the INAP messaging system that links these IN components together.

2.5.1 ETSI Core INAP operations

We provide here brief accounts of the 29 CS-1 INAP operations that were chosen
for inclusion in the ETSI Core INAP specification. Complete descriptions of the
procedures associated with these operations can be found in Section 3 of [10].
There are 26 ITU-T CS-1 operations that were not included in ETSI Core INAP.
For background information these are listed separately in Appendix 2. The ETSI
Core INAP operations are listed in Table 2.3 and then explained in the subsequent
text.
ActivateServiceFiltering (SCF ? SSF)
This operation was designed particularly for mass-calling applications, such as
telephone voting (‘televote’), where there are likely to be many similar requests to
the SCF in a short space of time, all of which need to be counted. For more efficient
traffic handling it is possible for the SCF to prescribe the standard action to be taken
within the SSP, rather than trigger to the SCF for each call that meets the triggering
criteria for this service. The procedure effectively devolves much of the repetitive
part of the service handling to the SSF by allowing the SCF to instruct the SSF to
only trigger for this service on a proportion of calls. The triggering might occur
perhaps on every 20th or every 1000th call, depending on the level of demand
expected for the particular voting scenario.
The calls to be filtered can be identified by using the triggering service key, or a
combination of the triggering service key and the dialled digits. The criteria can be
further narrowed to a particular geographic area by specifying some of the calling
address digits. The action prescribed by the SCF could be typically to increment a
counter and play a specified announcement to the caller. The values of the counters
44 Intelligent networks – principles and applications

Table 2.3 ETSI Core INAP operations

29 ETSI Core IN operations SSF — SCF direction SRF — SCF direction

ActivateServiceFiltering /
ActivityTest /
ApplyCharging /
ApplyChargingReport ?
AssistRequestInstructions ? ?
CallGap /
CallInformationReport ?
CallInformationRequest /
Cancel / /
CollectInformation /
Connect /
ConnectToResource /
Continue /
DisconnectForwardConnection /
EstablishTemporaryConnection /
EventNotificationCharging ?
EventReportBCSM ?
FurnishChargingInformation /
InitialDP ?
InitiateCallAttempt /
PlayAnnouncement /
PromptAndCollectUserInformation /
ReleaseCall /
RequestNotificationChargingEvent /
RequestReportBCSMEvent /
ResetTimer /
SendChargingInformation /
ServiceFilteringResponse ?
SpecialisedResourceReport ?

are then periodically returned to the SCP using the ServiceFilteringResponse INAP
operation (explained below). Further illustration of the use of the ASF
(ActivateServiceFiltering) feature is given in a televoting service example that is
included in Chapter 7.
ActivityTest (SCF ? SSF)
The ActivityTest operation is used to check that a previously established
relationship between the SCF and SSF is still in existence. If it is, then the SSF will
respond. Otherwise the SCF deduces that the relationship has failed in some way
and it will close the outstanding transaction. This procedure might be used, for
example, where an SCF had armed an E_DP for ‘no reply’ (at detection points 6 or
14 on the BCSM, Figure 2.8) and the call is a long duration one. As a safeguard the
service logic may check that an EventReportBCSM response has not been lost by
sending an ActivityTest operation to the SSF.
 The foundations of IN 45

ApplyCharging (SCF ? SSF)

The ApplyCharging operation provides a way for the SCF to interact with
switch-based charging mechanisms.
ApplyChargingReport (SSF ? SCF)
This operation is for the SSF to report back to the SCF in response to a
previously received ApplyCharging operation.
AssistRequestInstructions (SSF ? SCF, or SRF ? SCF)
The ARI operation is used after a connection has been set up to an IP (or another
SSP) as a result of a ConnectToResource or an EstablishTemporaryConnection
operation being sent from the SCF to the original SCP. The assisting SRF, or SSF,
will contact the SCF with the ARI operation in order to obtain information on how
to proceed with the call.
CallGap (SCF ? SSF)
The SCF uses this operation to tell the SSF to reduce the rate of I_DPs for a
specified service key or calling/called address (or both).
CallInformationReport (SSF ? SCF)
The CallInformationReport message is used as a reply to a previous
CallInformationRequest operation.
CallInformationRequest (SCF ? SSF)
This operation is used by the SCF to instruct the SSF to keep a record of
specified information (about a single call) and respond, with the results, to the SCF
at the end of the call using a CallInformationReportoperation. The requested
information could be any of the following:

& call set-up elapsed time

& call stop time
& call connection elapsed time
& called address (before translation)
& call release reason.

Cancel (SCF ? SSF or SCF ? SRF))

This operation can be used to either cancel all outstanding requests or to
particularly cancel a previous PlayAnnouncement or a PromptAndCollectUser-
Information operation.
CollectInformation (SCF ? SSF)
The CollectInformation operation is intended for use in conjunction with a
RequestReportBCSMEvent operation in the arming of DP2. It specifies how many
digits should be collected.
Connect (SCF ? SSF)
The SCF sends a Connect operation to instruct the SSP to route the call to a
specified destination. It is one of the most commonly used INAP operations and it
enables the SCF to exert considerable influence over the CCF call processing. It can
have the following parameters:
46 Intelligent networks – principles and applications

& destination routing address

& correlation identity. This is used for the ‘hand-off’ or ‘Assist’ procedures, and
the intention is that the SSP should pass it forward (with the SCF identity
indicator) in the network signalling system to the next SSP or IP node, which
should in turn pass it up to the SCF. This will enable the SCF to correlate the
new request with the original transaction with the first SSP.
& SCF identity (ditto).
& cut and paste. This allows the SSP to replace some of the leading digits of the
address dialled by the caller with translated digits returned from the SCF.
& route list. This allows the SCF to over-ride the CCF’s normal routing function
by specifying a list of the telephony routes to be attempted to reach the call’s
destination.
& calling party’s category. This allows the SCF to change the value (operator,
payphone or ordinary customer) of this parameter from what it might have been
previously set to, either in the exchange data for the customer, or as received in
an incoming network signalling message.
& original called party identity. This is for the original number dialled by the
caller.
& redirecting party identity. This is the address from which the call was last
redirected en route to the SSP.
& redirection information. This contains information on the call forwarding
history of the call, such as redirecting counters.
& alerting pattern. This is applicable for terminating local exchanges only, and
allows different ringing patterns to be specified – assuming the exchange
supports them.
& service interaction indicators. This is intended to provide a means of
controlling interactions between switch-based network services and services
provided at the SCF. The content, however, is largely unspecified, and so is
highly implementation-dependent.

ConnectToResource (SCF ? SSF)

The ConnectToResource (CTR) operation is used to instruct the SSP to connect a
call to the physical entity (e.g. an external intelligent peripheral) containing an
SRF. If this physical entity is internal to the SSP, the SCF should provide the call
leg identifier for the leg to be connected to the resource (the A leg is the likely
default) as a parameter. If the SRF is in an external IP, then an IP routing address
will be needed.
Continue (SCF ? SSF)
‘Continue’ is used by the SCF to request the SSF to proceed with call processing
from the point at which it stopped in order to send the request to the SCF.
DisconnectForwardConnection (SCF ? SSF)
This operation is used to disconnect a temporary connection, which had been set
up (using EstablishTemporaryConnection or ConnectToResource operations), to a
separate network resource such as an IP.
 The foundations of IN 47

EstablishTemporaryConnection (SCF ? SSF)

This operation, ETC, establishes a connection to a separate network resource,
such as an IP, in order to play an announcement or to collect user information.
EventNotificationCharging (SSF ? SCF)
This operation is used to notify the SCF of a charging-related event previously
requested by the SCF in a RequestNotificationChargingEvent operation.
EventReportBCSM (SSF ? SCF)
This is used to notify the SCF of the occurrence of a call-related event (such as
‘busy’ or ‘no reply’) that had been armed using a RequestReportBCSMEvent
operation.
FurnishChargingInformation (SCF ? SSF)
This operation can be used to send charging information to the SSF for inclusion
in the switch-based call record to be sent for off-line processing at the end of the
call.
InitialDP (SSF ? SCF)
This is the first operation in an INAP interchange instigated by the SSF. It is sent
when an active trigger has been encountered at a T_DP when all triggering
conditions have been met and the call is not gapped or service-filtered. The ‘service
key’ parameter is mandatory and there are 16 optional parameters (e.g. calling and
called numbers); the available parameters are further described in Section 2.2.5.3.
InitiateCallAttempt (SCF ? SSF)
This operation is used to request the SSF to create a new call leg, using address
information provided by the SCF. It could be used to connect a destination number
to an announcement device for an alarm call service.
PlayAnnouncement (SCF ? SRF)
The PlayAnnouncement operation instructs the SRF to play a specified
announcement, or tone, over the speech path. The speech path is set up using
ETC or CTR operations. For standard tones and announcements the SRF can be
collocated with the SSF in the network switch.
PromptAndCollectUserInformation (SCF ? SRF)
This operation, P&C, is used to enable the SCF to interact with the user by
instructing an SRF to prompt for and collect information such as in-band dialled
digits, or voice responses, over the speech path.
ReleaseCall (SCF ? SSF)
This operation requests the SSP to release an existing call, which can be at any
phase of the set-up. A ‘cause’ parameter is provided.
RequestNotificationChargingEvent (SCF ? SSF)
This operation is used to request the SSP to monitor for a charging-related event,
then to respond to the SCF when the event occurs.
RequestReportBCSMEvent (SCF ? SSF)
This operation is used to arm E_DPs, thereby instructing the SSF to arrange for
the CCF to monitor for the occurrence of call-related events (such as destination
busy or no reply). Several E_DPs can be armed simultaneously, but individual
EventReportBCSM operations must be returned when events are detected.
48 Intelligent networks – principles and applications

ResetTimer (SCF ? SSF)

This operation is used when SCF processing is taking longer than expected, and
the SCF sends ResetTimer in order to instruct the SSF to refresh its application
timer to give the SCF sufficient time to finish the processing for this request.
SendChargingInformation (SCF ? SSF)
This operation requests the SSP to modify the real-time charging with
information passed from the SCF. Because charging mechanisms vary significantly
between vendors, the parameter content detail is operator-specific.
ServiceFilteringResponse (SSF ? SCF)
This operation is used to reply to the SCF with the values of counters specified in
a previous ActivateServiceFiltering operation. An example of the use of SFR
(ServiceFilteringResponse) is given in Chapter 7.
SpecialisedResourceReport (SRF ? SCF)
This operation is the reply to the PlayAnnouncement operation when the
‘announcement completed report’ indicator is set.

2.6 IN CS-1 implementation issues

As we have discussed, every exchange design is different, and so the ease of

upgrading to IN varies between exchange types, and the consequent effect on
overall performance overhead is important. The addition of triggers and the trigger
criteria is a call processing upgrade in the switch in just the same way as any other
call processing upgrade. A new switch build containing IN features is therefore
subject to normal upgrade procedures, and IN features are likely to be competing
against business-as-usual items like bug fixes, maintenance improvements and new
(non-IN) service features. There are usually too many candidate functions for a
single new build to handle and so the choice of what features go in a new build is
inevitably a question of business priorities.
Because of these issues to do with upgrading existing switches, IN was typically
introduced in a piecemeal fashion. The international standards allowed operators
and equipment vendors to negotiate controlled sub-sets of the ETSI CS-1 functions,
and the contents of these IN ‘phases’ depended on short-term requirements and
budget allowances.
Because many useful IN services could be provided with just a limited set of SSP
functionality, an early delivery of IN functions might have included, say, 5 of the
‘originating’ side trigger points in the BCSM and perhaps 12 or so of the most
useful INAP operations. Obviously service ‘walkthroughs’ would be needed to
make sure that the planned services would work with the chosen subset. Such a
subset would have been designed to support the features needed for the more
popular IN services involving number translation variants. More detailed
terminating functions needed for, say, call queuing and call distribution controls
could then be added in future builds in the light of ongoing service and business
demand and budget availability.
 The foundations of IN 49

2.6.1 Core INAP as a compliance benchmark

Discussions in this chapter are based on the CS-1 ETSI Core INAP, because, within
Europe, this still remains the most widely used basis for IN implementations. The
ETSI standard was itself derived from the ITU-T CS-1 standard, which in turn
provides a wider framework, enabling linkage with world-wide trends in IN, in
particular with US AIN standards.
The CS-1 BCSM was chosen by the ITU-T IN standards community as
representing a generally acceptable common starting point for IN design. However,
in practice, software designers of the industry’s different digital exchange variants
had previously developed their own pragmatic design models that would meet their
particular digital exchange requirements.
The BCSM is a theoretical model, devised independently and after the designs of
the digital exchanges themselves. Several modelling philosophies were possible,
and so the choice became a theoretical compromise, but the BCSM was finally
agreed as a suitable starting point for the standards production. The main purpose of
the standards process was to equip the telecommunications industry with a usable
basis for IN designs with as much international commonality as possible. The main
goal was to develop an IN signalling protocol that handled the separation of the
service logic from the legacy call control software in the exchanges. SS7 INAP
filled this role, and the BCSM provided the necessary level of agreement across the
international standards community.
It must be recognised that, from the SSP viewpoint, the IN standard in many
ways still remains at the level of a useful theoretical construct, and cannot really be
taken as a blueprint for the way things actually work in detail. Different switch
manufacturers adopted the standard and amalgamate it with their particular call
processing designs in their own ways. This then presents difficulties for operators
when trying to assess which is the best product option for their network upgrade to
IN.
At the SCP these concerns do not of course apply. Generally the BCSM is
sufficiently detailed for the new SCF designs. The BCSM provides a ‘birds eye’
view of the essential switch functions, with the purpose of enabling the remote
service logic in the SCF to be aware of the context of IN request transactions. The
SCF obviously has no need for the details of the call processing implementation; it
just needs to know enough to process the service logic correctly.
ETSI Core INAP [10] provided network operators with a practical text against
which to check compliance of the various vendors’ CS-1 IN implementations. For
instance, ETSI CS-1 specifies 17 DPs (most of which can be T_DPs or E_DPs) and
29 INAP operations. Whilst it might possibly be feasible for an individual
implementation to justify including all the 29 INAP operations, the justification is
unlikely to extend to a full set of all the possible T_DPs and E_DPs. Therefore the
standard has an invaluable subsidiary role as a compliance benchmark against
which to check different manufacturers’ available product sets.
50 Intelligent networks – principles and applications

2.6.2 INAP implementations and feature interactions

Within a chosen set of INAP operations there are still decisions to be made about
which parameters are needed in the controlled sub-set. For example, the ETSI
INAP Connect operation has 12 possible parameters, and these are detailed in
Section 2.5.1 above, but some of these may not be relevant to some switch designs
or may be expensive to implement. For instance, the parameter ‘alerting pattern’
specifies different ringing patterns for different calls, but how far this can be used of
course depends entirely on the underlying switch implementation. Whilst the
inclusion of a certain parameter in a message might be straightforward, providing
the underlying supporting functions to make proper use of this parameter may be
less straightforward.
The workings of many parameters are left for the network operator to specify.
For instance, ‘Service Interaction Indicators’ in the Connect operation are for the
SCF to convey instructions about feature interaction control between IN services
and the network-based services. However, this is heavily dependent on switch
design and on how the operator wants the services to interact (or not). The CS-1
feature interaction manager (FIM – Section 2.2.10) addresses the interaction of IN
service features, but the bigger issues are to do with the interaction of IN-provided
services and switch-based ‘embedded’ network services. Extensive predictive
testing of possible interaction scenarios is needed, and if necessary services should
be broken down into their component parts and tested individually with network
services. As discussed later in Section 4.6.2, part of the CS-2 specification (Q.1224)
lays out the principles for IN vs. switch based interactions using the UFM formal
methods, but switch-based services are generally so proprietary that in most cases
there is still no real alternative to comprehensive testing of individual scenarios.
To allow for different legacy implementations and different operators’ particular
service requirements, most of the IN operations specified include an ‘extension’
field. These extension fields provide carriers in the message for items that are
operator-specific. For example, the extension fields in the PlayAnnouncement and
P&C operations could be used to invoke custom dialogues at an SRF, enabling
advanced speech recognition devices and software to be used.
As another example of the use of extension fields, a particular IN implementation
may specify that the EventReportBCSM operation should carry a time-stamp for
the event in question. This could be achieved using an extension field parameter.
Extension fields have proved invaluable in supplementing standard features with
operators’ specific needs, but they need to be looked at carefully when ensuring
interworking compatibility between implementations.
Lower level signalling interface agreements must also be negotiated in detail,
because some of the standard is not mandatory. Therefore different vendors may
choose subtly different formats, yet still remain within the confines of the standard
recommendation. An example of this is the TCAP ‘contents length’ field (defined in
[14]), which specifies three possible forms for this field – ‘short’, ‘long’ and
‘indeterminate’. The choice of such options obviously needs to be agreed for
successful interworking between two IN network elements.
 The foundations of IN 51

2.6.3 SSP data management

The sets of T_DPs, E_DPs and trigger criteria specified in the standard are likely to
translate into a non-trivial data provisioning mechanism for the switch, particularly
if the new features become successful and many IN services, needing differing
triggering conditions, co-exist. Some of the data may be relatively transitory, with
some services provided for short trial periods only, but other data will be more
durable.
There is potential for different services to interact adversely with each other if
the SSP data are incorrectly provisioned. Access safeguards are necessary; it is
possible to achieve catastrophic results by misplacing IN trigger data in the SSP.
For instance, if someone (inadvertently) inserted an office-based T_DP 1 trigger at
a busy time of day, the spectacular result would be that every call on the exchange
would trigger before the dial tone. This would result in near-immediate focused
overload at the SCP and the exchange itself.
Tight control of the SSP triggering data-builds is therefore necessary, and the
importance of easily comprehensible and up-to-date user-guide documents cannot
be under-estimated!
 Chapter 3
Signalling intelligence

The wireless music box has no imaginable commercial value. Who would pay for a
message sent to nobody in particular?
David Sarnoff’s associates, in response to his urgings for
investment in the radio in the 1920s.

3.1 Introduction

Whilst this book is primarily about IN we have to remember that IN was introduced
into the context of existing voice networks, which already exhibit a fair degree of
‘intelligence’, both in terms of ‘embedded’ switch-based services and the common-
channel signalling system and the associated call control procedures. Furthermore,
without some of these intelligence features already installed in voice networks, the
IN enhancements would actually not have been possible.
In order to underpin the discussion on the incremental intelligence features
provided by IN we now review the common-channel signalling system and the
traditional call processing procedures.

3.2 Common channel signalling – the beginning

The introduction of common channel signalling, i.e. the CCITT Signalling System
No.7 (SS7), was revolutionary. Until then signalling systems were mainly in-band
and line-based, using ‘make-and-breaks’ on the transmission path, or various
combinations of tones on the speech path [15].
Implementations of SS7 started to appear in the early 1980s, along with SPC
(stored program control) digital telephone exchanges [16, 17]. The new common
channel signalling enabled the switch processors to communicate directly, sending
call set-up information between each other using labelled messages. The advent of
SS7 allowed the signalling to be removed completely from the speech path, to the
54 Intelligent networks – principles and applications

enormous benefit of both. Speech circuits could be provided with complete freedom
from danger of interference from signalling noise, and SS7 combined the signalling
for many speech circuits into a dedicated 64 kbit/s-channel [18].

3.3 The layered signalling model

Figure 3.1 gives an overview of the main constituents of the SS7 family of
protocols. The basis is a layered model, which has a loose association with the OSI
model of the early 1980s.
The diagram shows the relationship between the SS7 model and the 7-layer OSI
model (which actually post-dates the SS7 model definition). It can be seen that the
match between the SS7 levels and the OSI layers is not exact.
The OSI model was defined by the International Organisation for Standardisation
(ISO) to promote open networking environments with the aim of allowing
equipment from multiple computer equipment vendors to interwork in a standard
manner [19]. Each of the OSI layers is designed to communicate independently
with its peer at the distant computer. The lowest layer provides the actual physical
bit-stream communications between the two computers, and the next layer up, the

Figure 3.1 The SS7 layered signalling model

 Signalling intelligence 55

‘data link layer’, provides the link control functions such as error control and flow
control. Addressing and routing of messages is carried out at the ‘network’ layer,
presenting the ‘transport’ layer (Layer 4) with an end-to-end network service. Layer
5 is responsible for organising and managing ‘sessions’ between communicating
application processes. In some cases a transaction relationship is needed; in others
an extended dialogue is more appropriate – this is the concern of Layer 5. The
‘presentation’ layer ensures that the syntax of the data exchanged is understandable
to each end and the ‘application’ layer represents the end user in specifying the
nature of the required communications. When data is transmitted, each stage, from
Layer 7 down to Layer 2, encapsulates the data received from the layer above along
with its own header, and passes it to the layer below. Layer 1 takes the data from
Layer 2 and sends it bit by bit to the transmission line.
As Figure 3.1 suggests, SS7 signalling evolved from an original 4-layer model,
where the underlying message transfer part served just the ISUP, TUP (telephony
user part) and TUP-derivatives [21]. We return to these call control protocols in
Section 3.5.1.
The signalling model was extended for the protocol requirements of IN and MAP
(mobile application part), which both needed reliable and real-time computer–
computer transaction-handling communications independent of telephony circuits.
The additions to the SS7 model that allow this are SCCP and TCAP, which are
discussed later in this chapter.
The foundation component of the SS7 hierarchy is the message transfer part
(MTP), which provides the actual messaging service, and is usually regarded as the
cornerstone of the SS7 specification set. As indicated in the diagram, MTP has
Level 1, 2 and 3 functions (explained in the next section), and can have several
users. Some users, such as TUP, employ MTP to transfer ‘circuit-related’
signalling, where the messages are closely tied to a particular speech (or other
bearer transport) circuit. Other users, such as IN, need the SS7 network to conduct
computing transactions that are independent of any particular bearer channel. In
this latter case, the SCCP is used as an intermediary. We shall return to these higher
level components later.

3.4 The message transfer part

The purpose of the MTP is to send messages, with guaranteed accuracy and
reliability, between user applications across the signalling network. These messages
can be delivered directly between two signalling points (SPs), or indirectly by
making use of intermediate signalling nodes. The intermediate signalling points are
generally known as signalling transfer points (STPs). The SS7 network is often
represented as a packet-switched signalling network of SPs and STPs overlaid on
the transport network, which physically interconnects the exchanges. This is
illustrated in Figure 3.2.
MTP is a reliable packet switching system, where the packets are despatched
over synchronous signalling links. In the UK these signalling links are usually
56 Intelligent networks – principles and applications

Figure 3.2 The separated signalling network

multiplexed into a 32-channel 2 Mbits/s PCM transmission system, typically

occupying time-slot 16. For short distances, these 2 Mbits/s systems can be carried
over coax cables pairs (one cable each for the ‘go’ and ‘return’ directions), but they
are multiplexed into higher order systems for transmission over longer distances.
Where a signalling point in the network acts as an STP, its prime concern is to
transfer packets from incoming links to the correct outgoing links, in the same way
that an IP router switches packets from one route to another. However, if the
signalling point also has user applications, such as the call processing software in a
telephone exchange, the MTP’s main purpose is to provide the signalling for the
user processes. A network node can of course perform both these functions, and
Figure 3.3 shows the various functions within MTP which make this possible.
The overall purpose of MTP is to transfer messages between users (top left of
Figure 3.3) in a controlled manner using the interconnecting signalling links
(bottom right). The message routing function ensures that messages from users
reach the correct outgoing signalling link. When an MTP user application requires a
message to be delivered to one of its peers at a distant node, it will send the
information, fronted by a routing label containing originating and destination point
codes (OPC and DPC), to the message routing function for packaging and onward
transmission [20].
On the basis of the DPC, the message routing function will direct the outgoing
message towards the appropriate signalling route and thence to a particular link in
that route. At this point the MTP Level 2 procedures will add the forward and
backward sequence numbers and checksum information, which is needed to ensure
 Signalling intelligence 57

Figure 3.3 MTP functions

that the messages are securely delivered in the correct order without loss or
duplication.
At MTP Level 1 the signalling links operate synchronously, at 64 kbits/s,10 and
when there is no user traffic to transfer, FISU (fill-in signalling units) are sent.
Against this synchronous background, user-signalling traffic arrives spontaneously,
and is included in the form of MSUs (message signalling units) in the signalling
stream. Link status signalling units (LSSUs) are used to exchange status
information between the ends of the link.
For incoming messages, the Level 2 MTP functions will check the sequence
numbering and the checksum and will pass the content of the message up to the
message discrimination function (shown in Figure 3.3). The message discrimina-
tion function checks the DPC of the incoming message. If the DPC corresponds to
this node the message is passed up to the message distribution function, which
directs it to the appropriate MTP user. The MTP user is determined from
examination of a service information octet (SIO). If the DPC does not correspond to
this node, and the message is a valid one, then this signalling point is acting as an
STP and the message discrimination function sends the message to the message

10
. . . . . . or 56 kbits/s in USA standards.
58 Intelligent networks – principles and applications

routing function. The message is then routed toward its required destination using
the outgoing routing tables, and is treated in the same way as a normal outgoing
message from this SP.
Figure 3.3 also shows the MTP management functions, which, as well as
providing the background functions essential for the correct working of the
signalling messaging handling, also contribute the major part of the ingenuity of
MTP. The signalling management functions ensure that MTP provides a reliable
service to its users, even in the face of network failures and disruptions to signalling
routes and links. Every signalling point therefore needs to have current knowledge
about the reachability of local and distant parts of the signalling network. To
provide for this, the signalling management software processes at different
signalling nodes send each other information messages to allow their routing data
tables to be kept up to date as changes occur.
A signalling point will be aware through the normal Level 2 procedures (see
below) when a directly connected signalling point becomes unavailable. If,
however, it has a signalling relationship with a more remote node that becomes
unreachable, it will receive information about the far end’s unavailability through
MTP ‘transfer prohibited’ messages. The node can then take appropriate action,
such as re-routing the signalling messages, or refusing to accept more traffic until it
receives a ‘transfer allowed’ message telling it that the problem has been cleared.
The routing data tables provide information on the availability of signalling
routes and signalling links. Signalling links are provided in ‘link-sets’, usually
comprising two (for security) or more links, depending on signalling traffic volume
requirements. Link-sets are provided co-operatively between signalling points, and
on each link they continually exchange signal units, sending FISUs in the absence
of MSU payload traffic. The signalling points at each end of the link-set are
therefore always aware of the availability of the links. If a link goes out of service
there are link management procedures for redirecting traffic onto a good link, and
messages which had just been sent on the failed link can be recovered from a buffer
and retransmitted on the good link.
Signalling routes are assigned to link-sets, and so if a link-set becomes
unavailable then any signalling route assigned to that link-set also becomes
unavailable. To provide signalling security, it is common practice to allocate
alternative signalling routes for a user traffic route, using intermediate signal
transfer points. Signalling routes with a common destination grouped together in
this manner comprise a ‘route-set’, and it is the route-set as a whole that is made
available to provide a secure signalling facility to the Level 4 user.

3.4.1 Signalling management intelligence

The ability of the signalling network to automatically ‘re-route’ to a second or third

signalling route (a collection of signalling links) at times of congestion or
subsequent node unavailability gives some insight into the built-in distributed
intelligence possessed by the SS7 network. Another, and perhaps more pointed,
 Signalling intelligence 59

indication of this built-in intelligence is the ability of the MTP network to distribute
information about node unavailability. Each SS7 signalling terminal at an exchange
runs a ‘watch-dog timer’, and if it detects a loss of communication with the
exchange central processor, it will autonomously signal ‘processor outage’ to the
node at the far end of that link. The far end will, in turn, inform any surrounding
nodes, which might use that node as a signal transfer point, that it cannot transfer
messages to the unavailable node.
The existence of this distributed intelligence, automatic signalling re-routing
coupled with the call processing software’s (normal) ability to automatically re-
route calls (independently of any signalling re-routing) sometimes is a source of
unease. Implementations need to be sufficiently robust to control the generation of
large quantities of automatic management messages that might otherwise swamp
the network at difficult times.

3.5 Telephony user part intelligence

We now move up from the MTP layers in Figure 3.1. This section illustrates the
ISUP and TUP telephony handling functions that use the services of the lower MTP
layers.

3.5.1 Basic call set-up

The original purpose of SS7 signalling was to enable telephone calls to be set up
and cleared down across digital telephone networks. The first major recommenda-
tions for SS7 were published by the ITU-T in 1980. These recommendations
included the telephone user part (TUP) [21], which provided the framework for the
messaging needed to set up, connect and clear down calls.
Whilst TUP defined the common standard, in practice individual network
operators defined additions to the standard to suit their own particular network
flavours. This was because different administrations had inevitably developed their
own way of handling the more complex call types, such as operator calls.
These local variations were called NUPs (National User Parts). However, with
the introduction of a multiplicity of network operators in the 1990s, it became
important to re-establish the standard for interconnecting public networks. In the
UK, an initial pubic telephony protocol for telephony interconnections, known as
IUP (Interconnect User Part), was derived from BT’s NUP. The latest standard for
interconnecting networks in the UK is the UK variant [35] of the ISUP protocol.
As an illustration of how the telephony protocols work, Figure 3.4 shows the SS7
level 4 (call control) messages that are typically exchanged between two local
exchange calls using the ISUP protocol in a simple telephone.
IAM (initial address message). This is the first (and the most informative)
message of a call set-up. It carries several parameter fields that describe the calling
terminal and the call handling required. In particular it carries the dialled address
60 Intelligent networks – principles and applications

digits, and an implicit request to the next exchange to establish a bearer connection
to the required destination. If all the address digits are available at the start of the
call set-up (for instance if the complete address has been received in an ISDN set-
up request message) then the originating exchange can forward the complete digit
string in the IAM. This is known as ‘en bloc’ signalling, and this is the scheme that
is illustrated in Figure 3.4. After the destination exchange is satisfied that it has
collected the correct number of digits needed to establish a bearer connection it
returns an address complete message (ACM). This indicates to the originating
exchange that the destination phone is ringing, and that the switch path can be
switched through in the reverse direction to enable the caller to hear the ringing
tone over from the far end.
If there is a time lag in sending the digits then an IAM containing just the initial
digits can be sent, and subsequent digits are sent in subsequent address messages
(SAMs). This is known as ‘overlap’ signalling and can also be used when the
originating node does not know the number length of the complete address digit
train – the far end will have this information and can therefore send the address
complete message when sufficient digits have been received.
When the called party picks up the phone, an ‘answer’ message (ANS) is
returned and the originating DLE can switch through the speech path in both
directions and start the call charging mechanism. Finally, at the end of the call, one
end sends a ‘release’ message, call charging is stopped and the other end returns a
‘release complete message’.
The sending of address digit information constitutes the bearer establishment
control phase, which is followed by the call establishment phase when the

Figure 3.4 ISUP message exchange for a simple call

 Signalling intelligence 61

destination needs to determine whether it can support the call protocol requested in
the IAM. The messages for a particular call set-up are correlated by a ‘circuit
identity code’ (CIC), which appears in every message for a call and relates to the
time-slot chosen in the 30 channel bearer. The IAM includes calling line identity
(CLI), allowing the destination DLE to provide services such as ‘calling line
display’ and ‘selective call barring’.

3.5.2 The ‘intelligence’ aspects of basic calls

Several ‘intelligent’ services can be provided in conjunction with a basic call set-up
using SS7 signalling without adding any new features. A particular example of this
is the set of CLI-based services (see Figure 3.5).
A common feature of SSP telephony signalling systems (ISUP and TUP) is to
convey CLI information from the originating exchange to the destination exchange.
This is commonly done for administrative reasons, such as accounting, detection of
malicious call perpetrators, presentation to a help-desk, etc., but also several
supplementary services build on this ability of SS7 to pass CLI information across
the network.

3.5.2.1 ISDN CLI services

The supplementary ISDN CLI-based services can apply to ‘connected’ (i.e.
terminating) lines as well as calling lines. A call originator may wish to display the
number of the line to which they are actually connected, as this may be different
from the number dialled.
Common CLI-based ISDN supplementary services are:

. CLIP (calling line identity presentation). This service enables customers to

display callers’ line identities on incoming calls.
. CLIR (calling line identity restriction). This allows the originating end to mask
the display of calling line identity at the terminating end, and so it over-rides a
terminating user’s CLIP service.

Figure 3.5 CLI-based services

62 Intelligent networks – principles and applications

. COLP (connected line presentation). For outgoing calls this enables users to see
a display of the terminating line identity when a call is connected.
. COLR (connected line restriction) allows the terminating end to prohibit the
display of his line identity at the originating user’s display, so over-riding the
COLP service if the caller has it enabled.

In Europe the CLI-based services are subject to ETSI standards and guidelines.
Interactions between these and other ISDN supplementary services have recently
been standardised in Reference 22.
In the case of ISDN, the CLI can be provided either by the network, using data
held in the customer data tables at the originating local exchange, or provided by
the user in the DSS1 signalling message. The interface from a standard POTS
terminal only conveys dialled address digits, so a caller cannot send a ‘user-
provided’ number in the way that an ISDN user can.

3.5.2.2 POTS CLI services

For outgoing POTS calls it is common to be able to prevent CLI from being sent
forward by dialling a prefix code such as ‘141’. This is equivalent to the ISDN
CLIR service. At the terminating end, it is common for customers to be able to
subscribe to a service that enables the CLI to be logged at a display device before it
is answered. This is equivalent to the ISDN CLIP service. If the originator has
withheld the CLI, ‘number withheld’ will usually be shown on the display at the
terminating end.
However, there are cases where, for technical or regulatory reasons, the CLI is
unavailable for forwarding to the destination terminal. The signalling procedures
distinguish between this and the case where the caller actually opts to withhold the
CLI. Further, a ‘presentation’ number is sometimes preferred to the ‘network’
number, which is the public number assigned to a termination. A presentation
number might for instance be a Freephone number, which might be a more useful
number for customers to call back on than the network number of the originating
terminal.
There are other variants on the CLI services. A common feature offered is for the
incoming exchange to automatically reject calls where the CLI has been withheld.
Call recipients can also interrogate the exchange, by dialling a code such as ‘1471’,
about the origin and time of missed call attempts, and then automatically return the
call by dialling another code. For calls to busy numbers, a ‘ring-back’ feature is
usually available, where the network can be instructed to automatically connect the
call when the destination number becomes available. A European standard version
of this service is known as CCBS (call completion to a busy subscriber).
Whilst these features illustrate the intelligence that resides in the SS7 signalling
network, they are all in fact made possible by the network signalling procedures,
and do not, in their basic forms, necessarily need any IN involvement, although the
IN could further embellish the services.
 Signalling intelligence 63

3.5.3 Further ‘switch-based’ intelligence

This transfer of CLI information was one of the main service features of early SS7
TUP implementations. However, because early digital exchanges usually had a set
of extra services, and there were slight messaging differences for operator calls,
there were several additions and variations to the simple call sequence of Figure
3.4. More recently, services such as ‘ring-back when free’ were introduced and
further signalling modifications were made. Because detailed international
standards were not available for these extra services, and because there were
different implementations in different administrations, these signalling modifica-
tions were often implemented as upgrades to national variants of TUP. Typically,
non-standardised values of the service handling protocol (SHP) parameter have
been used to instigate more complex messaging dialogues for such services. The
parallel development of ISDN and non-ISDN signalling specifications led to
differences in terminology and procedures for the supplementary services.

3.6 SS7 signalling for IN

So far this chapter has only looked at ‘circuit-related’ SS7 signalling, where the
identity number of the speech circuit is used as the reference key for the signalling
interchange. Right from the early days of SS7 design there were plans to also use
SS7 for computer-to-computer transactions (with no associated speech circuit) [3],
and so provision was made for non-circuit-related signalling for applications such
as INAP and MAP.

3.6.1 Signalling connection control part (SCCP)

The first CCITT SCCP recommendations were published in 1984 in the so-called
‘Red Book’. These were updated four years later, along with significant TUP
changes [21], when the C7 ‘Blue Book’ was published. The ‘Blue Book’ series of
signalling standards also included details of the network management aspects of
SCCP. The SCCP was designed primarily to support transaction-based messaging
using connectionless signalling on the established common-control signalling
network, for which it is now used extensively, with many implementations still
based on the Blue Book recommendations. However, SCCP standards were
completed well in advance of widespread adoption of connectionless signalling and
so from an implementer’s view much of the detail was still quite theoretical.
Anticipated applications for SCCP at the time included [23]:

& interrogation of centralised databases by telephone exchange processors

& updating of location registers for mobile communications
& remote activation of supplementary services
& transfer of data between network management nodes.
64 Intelligent networks – principles and applications

Of these applications, SCCP is now in extensive use for the first two, and the third
has not been standardised and is in limited use in proprietary systems. The fourth
was dropped because of the risk of swamping the real-time signalling channels with
large quantities of background data, adversely affecting the performance of the
network.
Current ITU-T SS7 SCCP recommendations are given in Reference 24.

3.6.2 Global title addressing

At the lower level, MTP uses point codes for message addressing, but a major
advantage of SCCP is that it can use global title (GT) addressing, which allows
greater flexibility. A GT represents a destination user or software process in terms
of an agreed format, which could, for example, be E.164 address digits. Before the
message reaches its destination it will encounter a GTT (GT translation) function,
which will map the GT address to a routable address, which could be an SS7 point
code and an SCCP subsystem number (SSN). An SSN is a sub-address representing
the SCCP user-function.
Figure 3.6 shows a network scenario using global title translation. A Freephone
call triggers from an SSP and there is a choice of SCPs that can handle the request.
Because there is an intervening SCCP translator node that can direct the request to

Figure 3.6 SCCP routing – from SSP to SCP

 Signalling intelligence 65

whichever SCP is currently the first choice, the SSP does not have to be aware of
the network identity of the target SCP. The SSF in the SSP addresses the request to
a GT that identifies the service. In this case we choose to identify the service by the
most significant eight digits that have been dialled by the caller (historically it is
common in many implementations for the dialled digits to be used as the global
title identity in this manner). At the MTP level the SSP does not need to have a pre-
established signalling relationship with the destination SCP – it just needs to have a
signalling route to the translator node, so it routes the message to the translator
node’s point code.
On receipt of the message, the translator node looks up the called party global
title address in its translation tables. In this case the resulting translation is the point
code for the first choice SCP and the sub-system number representing IN
applications at the SCP. The sub-system number (SSN) represents the user of
SCCP. SCCP management messages will be directed to and from a separate SSN,
and there are different numbers for other sets of applications, such as MAP
services, which may be hosted on the same control point.
When the request message arrives at the SCP the GT enables the request to be
routed to the appropriate service logic and a Freephone translation will then be
returned to the sending SSP. Figure 3.6 shows the SCPs sharing an external SDP,
but this may not always be the optimum architectural arrangement. It will depend
on many factors to do with a particular service implementation, such as reliability,
speed of response, database size, messaging loading and the capabilities of the
database management system.
Figure 3.7 shows the return message from the SCP to the SSP. Because the SCP is
now aware of the point code of the originating SSP, it can use this in the SCCP
address for the return message. If there is a signalling route configured between the
SCP and SSP the SCP can direct the message to the SSP itself (almost certainly via an
STP, which may be the translator node itself). However, Figure 3.7 assumes there is
no pre-set signalling relationship with the SSP, so at the MTP level the SCP must
route the message towards the translator node’s point code even though it puts the
target SSP’s point code in the SCCP called party address field (with the SSN). This
time the translator node discovers the SSP point code in the translation look-up, and
so it is able to instruct the outgoing MTP to route the message to the SSP’s point code.
This example explores one particular routing scenario, in order to explain the
global title translation function, but many variations are possible, depending on
individual network needs.

3.6.3 Transaction capabilities application part (TCAP)

SCCP provides the flexible data transfer mechanism needed for IN, and a powerful
routing and addressing capability. To enable transaction-based protocols users such
as INAP and MAP to use the services of SCCP, SS7 provides transaction
capabilities [26] that interface directly with SCCP. These capabilities are presented
to the SSP as the TCAP application. As shown earlier in Figure 3.1, TCAP and its
66 Intelligent networks – principles and applications

Figure 3.7 SCCP routing – from SCP to SSP

users (such as INAP and MAP) reside at the top level of the SS7 stack [25]. This is
equivalent to Layer 7 (the application layer) on the OSI stack working directly to
Layer 4 (the network layer) without requiring the services of the intermediate
presentation, session and transport layers. As Figure 3.1 shows, ISUP and TUP
derived protocols implicitly contain these intervening layers.
TCAP’s role in IN is effectively to organise and manage the transactions used by
the INAP operations and parameters that are exchanged between SSFs, SCFs and
SRFs. TCAP performs a similar function for MAP, which enables a mobile
switching centre (MSC)’s VLR (visitor location register) to request service profile
details from the mobile customer’s HLR (home location register).
TCAP messages are carried in the SCCP part of an MSU (an MTP message
signalling unit) over an SS7 link. A TCAP message is made up of a transaction
portion and a component portion [26]. Information elements in TCAP use ‘name,
length, value’ encoding, which gives TCAP its flexible applicability, although the
messaging overhead can be appreciable.
The following sub-sections briefly describe the SS7 TCAP transaction and
component portions. However, the signalling mechanism is best described by an
example message sequence, and so we illustrate the action of the TCAP elements in
 Signalling intelligence 67

Appendix 3, where we show the sequencing of the messages for a CAMEL mobile
network call. CAMEL (customised application of mobile networks enhanced logic)
uses the SS7 CAMEL Application Protocol (CAP), which in turn is a user of
TCAP, SCCP and MTP signalling in just the same way as INAP. Therefore the
behaviour of SCCP and TCAP in transporting the CAP messaging is identical to
that for the transport of INAP messages. In fact, as we shall show in later chapters,
there is close equivalence between CAP and INAP, because – as we shall see – the
former was derived from the latter.

3.6.3.1 Transaction portion

The transaction (or dialogue) portion of TCAP co-ordinates and correlates
messages using transaction identities. This allows the existence of multiple
transactions between end-points. A message can have an originating transaction
identity (OTID) and a destination terminating identity (DTID), but not necessarily
both, and the messages are grouped into one of four types:

^ Begin This initiates a transaction with a remote node.

^ Continue This continues an existing transaction.
^ End This is a normal termination of a transaction.
^ Abort This is an abnormal termination for a transaction.

The transaction Id. is locally assigned by the sending node. The ‘continue’ message
type is the only one that has both OTID and DTID. The ‘begin’ will only have
OTID and the ‘end’ and ‘abort’ message types only have the DTID.

3.6.3.2 Component portion

The component portion of TCAP contains one or more components and is
responsible for correlating responses to requests and handling error detection. The
components are based on the application protocol data unit (APDU) concept of the
ITU-T ROSE (remote operations service element) defined in ITU-T X.229.
Components are grouped into one of five types:

Invoke This requests the invocation of an action at the remote end. Invoke
components have an invoke identity and contain an operational code
and (optionally) related parameters, which for ITU-T IN are the INAP
operations and parameters described in Chapter 2.
Return result – This contains a response to an earlier Invoke.
‘not last’
Return result – This contains a final response to an earlier Invoke.
‘last’
Return error This indicates that the associated invocation was not successful, and
an error code and further error information is provided.
Reject TCAP, or the TC user application, sends this when previously sent
information is not understood.
68 Intelligent networks – principles and applications

3.6.4 Signalling differences between PSTN and GSM networks

Historically the technical differences between fixed and mobile networks have been
big enough to justify the evolution of separate network structures. As a result,
separate switching network methodologies have established themselves in the fixed
network and mobile worlds.
However, in the core of both networks there is commonality in the use of ISUP
(or TUP-based) SS7 for interchange of call set-up information between switching
centres. INAP is also found in both network types, but in the case of mobile,
equipment vendors have generally supplied the network operators with tailored
proprietary variants to cater for the networks’ particular needs. However, where
CAMEL applications are used, the standardised CAP interface is also used, and this
has introduced opportunities for multi-vendor interworking of equipment.
Mobile networks generally make richer use of the various SS7 protocols. As well
as using ISUP for call set-up signalling, CAP (and proprietary INAP versions), for
IN access, SS7 has several other uses in mobile networks.
SS7 MAP is used extensively for the mobility functions between MSCs, HLRs
and VLRs. In the ‘A interface’ between the switching centres and the base station
controllers, the BSSMAP manages resources and connections, MAP relays
supplementary service information and DTAP carries the radio interface messages
(RIL 3). Figure 3.8 compares the variety of SS7 protocols used in mobile calls
compared with fixed line calls.
Apart from the obvious differences in access technologies, mobile call
processing is more complex because authentication and mobility management

Figure 3.8 Comparison of PSTN and mobile network signalling architectures

 Signalling intelligence 69

procedures are needed in addition to the call control function. There are therefore
more signalling messages needed for a mobile call than for a fixed network call.
This is the reason why mobile phone users have become used to waiting several
seconds between dialling a number and hearing a call progress tone (such as ring
tone). For a fixed network phone this waiting time is of course typically well under
a second.
Of course, technical differences are not the only reasons for the differences
between fixed and mobile networks. Business drivers for mobile phones have
generally been more urgent in the race to compete with fixed network operators and
so less time is available for in-depth analysis of technical options, which used to be
normally expected with the incumbent fixed-network operators. In more recent
years, speedy evolution has been essential in the race to win customers and so there
has been little incentive to seek a common evolution path with fixed networks.
As for intelligent network developments, Figure 3.8 shows two separate IN
SCPs. This is because, during the formative years of mobile networks, there was
little effort made to converge the mobile IN service control functions with those in
neighbouring fixed networks. Mobile network evolution paths were driven
primarily by the procurement of vendor implementations that could deliver the
required service features. Consistency across fixed-network and mobile technol-
ogies was rarely a significant factor in procurement decisions on either side of the
fence, because operators tended to specialise in either fixed or mobile networks,
and so were not overly concerned with the other side.
Different switch designs have necessitated different triggering arrangements, and
the use of different IN standard variants has ensured that services are generally
implemented differently in the SCPs.
 Chapter 4
International standards for
intelligent networking

While considering which of two things you should teach your child first, another
child has learned them both.
Dr Samuel Johnson (1709–1789)

4.1 Introduction

IN was once heralded as a network re-arrangement that would provide any

telecommunications service need that any user could conceivably want. In practice,
over the years, commercial priorities, financial constraints and inter-working
limitations have inevitably limited the vision’s potential. Whilst in the early days
the IN concept had high aspirations, reality had to intervene sooner or later to
remind us what might be feasible and what might not be. For example, call routing
decisions based on the time of day are perfectly feasible, whereas a routing decision
based on the colour of a user’s shirt may be less feasible!
As already discussed, AT&T and then Bellcore, in the USA, made the running on
early exploratory work on IN specifications and the European operators watched
the results carefully. In the early 1990s there was a lot of activity in the USA on IN
specification production, although these standards being formulated by Bellcore
were obviously designed to meet US needs and US switching system requirements,
some of which were less relevant outside the US.
In the early 1990s, the CCITT (now the ITU-T) set up a study group hierarchy
for intelligent network standards, and the first result was the CCITT IN CS-1
(Capability Set 1) in 1993. CS-1 was planned to be the first of several steps towards
the introduction of intelligent network concepts into telephony networks. It was
seen as a relatively modest set of early IN capabilities for voice calls. It deals with
single-ended, single point of control, IN service features.
‘Single-ended, single point of control’ means that there is just one controlling
SCP acting on a call for any one service invocation. What it excludes is the
72 Intelligent networks – principles and applications

possibility of having co-operative service invocations, controlled by different SCPs

and acting on different SSPs, affecting the same call. The important point is that all
the invocations must be independent – there is no dialogue between separate SCPs
about the same service invocation. This means that the SSF is not able to interact
with more than one SCF at any given time in order to complete a sequence of
queries and responses (forming an IN transaction) on behalf of a calling or called
party. As we cannot have more than one SSF–SCF relationship taking part in an IN
transaction an SCP is unable to orchestrate a call across a network by simultaneous
SCF–SSF transactions. Figure 4.1 shows two separate SCP invocations on one call,
which is allowable. These could, for instance, be a user authorisation service, where
a check on a caller’s identity is made with the first SCP followed by a request to a
second SCP later in the call for a ‘least cost routing’ service. The two SCPs are
working in isolation and so the ‘single-ended, single point of control’ rule is not
broken.
The use of the term ‘SCP’ is representational here. Strictly the ‘control point’ is a
single computer housing the requisite service logic. In fact, as we shall see later, the
service control functions may be distributed geographically. Also, in the stricter
analysis to follow, the term ‘service control function’ (SCF) is used, indicating that
the control functionality can be hosted at a variety of locations – not necessarily in
SCPs.
Another important limitation for CS-1 is that it is really for two-party calls only,
and there are no IN functions provided that will allow manipulation of conference
calls involving three or more users.
In the early ITU-T standards deliberations there was a lot of discussion about the
form of the IN call model that should be included in CS-1. In particular, some
delegations did not want CS-1 to include some of the more advanced ideas of ‘leg-
control’ (which had already been put forward in the Bellcore IN/2 proposal in the
US but had not been received favourably by switch vendors). Leg control enables
IN service logic to manipulate the different parties involved in a call, enabling
provision for IN controlled conference calls. Whilst the ITU-T CS-1 contents were

Figure 4.1 Separate service invocations on the same call

 International standards for intelligent networking 73

being decided, Bellcore was standardising the AIN set of recommendations in the
US in parallel with the ITU activities, although there was co-operation between the
two activities.

4.2 Early US standards for IN

In the US, the catalyst for early rapid IN development was the ‘1800’ Freephone
service. IN really originated in the middle of the 1980s, when the seven US regional
Bell operating companies (RBOCs) expressed interest in centralising the service
control for services such as Freephone and Calling Card. Because the US had
retained letters on telephone dials and key-pads, entrepreneurial ingenuity was
ensuring the success of the 1800 service. Advertisements such as ‘dial 1800
FLOWERS’ (for flower deliveries) were becoming common. The calling card
verification service, also known as ‘alternate billing’, enabled users to make calls
from public payphones and have the call charges accrue to an account independent
of the normal exchange charging mechanisms.
More recently there have been a series of obligations imposed by the US Federal
Communications Commission (FCC) concerning Local Number Portability (LNP).
The 1996 Telecommunications Act demanded that the US local exchange carriers
(LECs) must allow customers to take their phone numbers with them when they
moved from one carrier to another whilst maintaining existing standards of service
quality and reliability. This mandate was subsequently extended to include wireless
carriers. Much of the recent investment in AIN implementations has therefore been
to meet these requirements.

4.2.1 INWATS, IN/1 and IN/2

The early AT&T toll carrier exchanges in the 1970s were designed to offer an
important feature known as Inward Wide Area Telecommunications Service
(INWATS). This was the original Freephone service. The switches were designed
to route the ‘1800’ service request messages, using the new common channel
signalling system, to a central database, which would translate the digits dialled by
the caller into a standard NANP (North American numbering plan) routing address
and return it to the controlling switch. This was followed in the early 1980s by the
addition of the Calling Card verification service using similar network principles.
Evolution of these principles at Bell Labs led to the DSDC (direct services dialing
capabilities) architecture. The DSDC access switches were known as ACPs (action
control points) and the central database was known as the NCP (network control
point). Bell Labs also added the NSCX (network services complex) for DTMF and
voice interaction with the caller, and this became the forerunner of the IN
Intelligent Peripheral.
A system based on the DSDC, from AT&T Network Systems, was installed by
BT in the UK in 1988, where it became known as the derived services network
74 Intelligent networks – principles and applications

(DSN). The DSN provided the basis for BT’s Freephone service for several years.
Similar systems were installed elsewhere in Europe in the early 1990s.
In 1984 the giant Bell company was broken up, and AT&T, the toll carrier
company, became separated from the Regional Bell operating companies (RBOCs)
who owned the local switching networks and customer access switches. As a result
of this, the DSDC technology could not be applied to local switches, and the
application of DSDC-type functions became an important sphere of interest of
Bellcore (Bell communications research – since renamed Telcordia Technologies).
Bellcore began to use the term ‘intelligent network’ for providing DSDC-like
services in the local carrier context.
That period in history was a time of fast growth in the uptake of these additional
services, and there was a lot of interest in the market. As well as the RBOCs’
interest in networks that could efficiently handle them, there was pressure, through
the regulators, from external service providers who wanted to be able to offer their
own variants of these enhanced services over the network operators’ established
telephone networks to their customers.
In response, Bellcore developed the IN/1 and IN/2 architectures. IN/2 was well
ahead of its time, and it was not well received by the equipment vendors because it
was too wide-ranging. It included advanced leg-control features such as ‘split’ and
‘join’, which appeared several years later in ITU-T IN CS-2 specifications, but even
today these remain largely unimplemented.
On the other hand, IN/1 (Figure 4.2) was relatively limited, but more realistic in
that it could be implemented in practical time-scales to meet the market demand. It
was designed for a small set of specific services, whereas IN/2 was generic. IN/1
introduced software upgrades to the network switches to allow them to use the
existing SS7 STP network to send a request to an SCP when they detected a user-
request for one of the ‘special’ calls such as Freephone or Calling Card.
As Figure 4.2 illustrates, the characteristic of IN/1 was that it was service
specific. In other words it was built around particular services. Dedicated 1800 and
alternate billing triggers were implanted in the switches, a separate SCP was needed

Figure 4.2 Bellcore IN/1

 International standards for intelligent networking 75

for each service, and separate service creation, management and provisioning
hierarchies were needed for each.

4.2.2 AIN

After the IN/1 and IN/2 initiatives the next set of IN standards from Bellcore was
termed ‘advanced intelligent network’ (AIN). The AIN specification aimed to move
away from IN/1’s reliance on service-specific components that could not be re-used
for other services. The AIN proposal, illustrated in Figure 4.3, was comprehensive
and wide-ranging, calling for generic management systems, control points and
switch software for triggering. AIN included an IN IP for voice announcements and
voice/DTMF recognition resources. The IP could also enter into dialogue with the
end-user, on behalf of the service logic at the SCP, in order to collect more
information from the user. The IP connected with the SSP using ISDN primary or
basic rate interfaces. Also included was an ‘adjunct’, which provided SCP-like
functions but was directly connected to the SSP over a high-speed interface rather
than SS7.
However, like the earlier IN/2 proposal, the AIN Release 1 specification was too
ambitious for vendors to use to meet the prevailing short-term demand for IN
services. Compared with the ITU-T specifications, it was a superset of CS-1,
including mid-call trigger detection points and multi-party handling. For this reason
phased releases of AIN were established. Bellcore defined AIN 0.1 (broadly
equivalent to ITU-T CS-1) as an initial scaled-down set of functions in order to
provide an implementation target. In practice the RBOCs set about to define sub-
sets to achieve their individual needs, and this resulted in AIN Release 0. AIN
Release 0 implementations consisted of proprietary specifications that differed
between the RBOCs.

Figure 4.3 Bellcore AIN

76 Intelligent networks – principles and applications

4.3 ITU-T IN CS-1

Whilst the pioneering work was in progress in the US, and regional standards were
being formed, the ITU-T began to take an interest in IN. General ideas about the
sort of hooks and the sort of services that could be provided were already
formulated and in 1989 the ITU-T started to develop the IN vision, which was so
successful in the US, into a potential international standard. The concept of
providing ‘hooks’ in traditional telephone exchange call processing software in
order to invoke the real-time assistance of external service logic was progressed
and developed into the bottom-up, top-down mechanism illustrated in Figure 4.4.
Within the scope of what was initially likely to be possible in terms of network
upgrades, a set of benchmark services was drafted. A framework process was then
assembled in which a protocol could be produced for the first tranche of the ITU-T
IN standard.
This 4-layer framework, on the right-hand side of Figure 4.4, became known as
the intelligent network conceptual model (INCM), and it provided a backdrop
against which to co-ordinate the different IN working groups in the ITU-T whilst
components of the overall picture were being assembled. The goal of the entire
exercise was to achieve an interworking standard against which network operators
could procure IN components and upgrade their networks. The first release of this
standard was called IN Capability Set Number 1 (IN CS-1)[27].
The term ‘capability set’ is actually used as a generic term throughout ITU-T
standards, although in this book we use the term ‘CS-1’ to refer to the first ITU-T
IN standard. There is generally not a connection between the capability set versions
across the different standards disciplines. For instance, at around the same time as
the IN CS-1 recommendation was being assembled a parallel ITU-T study group
was forming the initial release of the broadband ISDN (B-ISDN) recommendation.
This was also called CS-1. The B-ISDN activity was concerned with converting the

Figure 4.4 The IN standards-generation process

 International standards for intelligent networking 77

existing ISDN point-to-point services to run over ATM, and it was generally based
on the familiar narrowband ISDN (N-ISDN) Q.931 standard.

4.3.1 IN CS-1 documents

CS-1 was intended as the first of a series of capability sets, and so, in order to
indicate an evolution path, it was set against a background of generic
recommendations. The ITU-T documentation authorities set aside the Q.12xy
block of numbers for IN, where the ‘x’ digit was allocated to the capability set
number, with the value 0 for the generic range of IN documents. The INCM was
fundamental to the documentation structure, and Table 4.1 below summarises the
documents that were allocated for the generic range.
The CS-1 specific recommendations were numbered as shown in Table 4.2.
We now briefly summarise the contents of these CS-1 documents.
Q.1210 shows the organisation structure of the CS-1 document series, providing
a summary overview for the individual recommendations.
Q.1211 describes the scope of the CS-1 standard, outlining the network functions
and their relationships. It also describes a set of ‘benchmark’ services that were
used for the IN CS-1 recommendation formulation process. These benchmark
services are individually described in Section 4.3.2.
Q.1213 gives details of the ‘global functional plane’ (GFP), which translates
service requirements into the top-level theoretical structure for IN. Part of this
process is the definitions of a set of service independent ‘building blocks’.
Q.1214 is a substantial document that shows how the high-level functions of the
GFP are distributed into the various IN functional entities (FEs). It describes the
operation of the IN CS-1 state model and the resulting information flows.
Q.1215 shows how the IN FEs can be populated into different physical entities.
Q.1218 describes the INAP protocol details.
Q.1219 is a detailed guide to the understanding and the implementation of IN
CS-1. It includes some example IN service scenarios.
The set of CS-1 documents was approved for publication in March 1993.
Appendix 1 provides a list of all the ITU-T IN documents, with individual
publication dates, which were in force at the time of publication of this book.

Table 4.1 ITU-T IN generic documents

Q.1200 General Series Intelligent Network Recommendation Structure
Q.1201 Principles of Intelligent Network Architecture
Q.1202 Service Plane Architecture
Q.1203 Global Functional Plane
Q.1204 Distributed Functional Plane
Q.1205 Physical Plane
Q.1208 General Aspects of the Intelligent Network Application Protocol
78 Intelligent networks – principles and applications

Table 4.2 The ITU-T IN CS-1 documents

Q.1210 Q Series Recommendations Structure for IN CS-1
Q.1211 Introduction to IN CS-1
Q.1213 Global Functional Plane for CS-1
Q.1214 Distributed Functional Plane for CS-1
Q.1215 Physical Plane for CS-1
Q.1218 Intelligent Network Application Protocol for CS-1
Q.1219 User Guide for CS-1

In October 1995, an update to CS-1 was issued. This was called CS-1R (for
‘refined’). It incorporated solutions to issues raised by industry during the process of
defining the ETSI Core INAP standard. CS-1R is discussed further in Section 4.4.

4.3.2 IN ‘benchmark’ services

To establish the formal starting point, a set of ‘benchmark’ services was agreed and
documented in Q.1211. These services were assembled as a list of examples against
which to test the IN constructs as they were developed. It must be stressed that
these are examples only – IN is not service-specific and it provides the
infrastructure for many other services and service variations. Many of the services
on the list can (and often are) provided using CPE-based functions, or switch-based
software on the PSTN, and they do not necessarily need the extra features of IN.
However, as explained earlier, IN provides the opportunity for making significant
efficiency improvements by centralising the control, data and management for these
services.
Q.1211 contains the precise definitions for the benchmark services. The
following comments summarise the essence of the listed services. Services with
close similarities have been grouped together in these descriptions.
Abbreviated dialling. This service allows commonly used address digit strings
to be represented by one or two digits, which can be dialled instead of the complete
digit string to reach the required destination address.
Account card calling and automatic alternative billing. These are variations
on the same service, enabling callers to bill call usage to their own accounts or their
employers’ corporate accounts. The service variants allow the user to either dial the
card number and PIN as prefix digits before the required address digits, or to
‘swipe’ the card through a reading device attached to the terminal equipment. The
cost of the call is thereby debited to the user’s account rather than to that of the
renter of the telephone line. This is particularly useful for making calls from pay
phones without having to use coins. The credit card calling service extends this
capability to include direct billing to a bank credit card.
Call distribution. Incoming calls to a customer can be distributed to various
answering points depending on the subscriber’s own pre-set rules. The destination
 International standards for intelligent networking 79

call routing service allows the calls to be directed to different destinations

depending on conditions such as time-of-day or call-origination point. The call re-
routing distribution service additionally allows the call to divert to a different
destination if it encounters a problem such as congestion in trying to reach the
customer.
Call forwarding. Incoming calls can be forwarded to a chosen alternative
destination. If the service is provided using IN, a mechanism could be provided for
the customer to set up the call forwarding parameters automatically under IN
control. Selective call forwarding (on busy or no reply) is a refinement where calls
are forwarded to an alternative destination if the customer’s line is engaged, or
there is no reply within a preset time-out period.
Completion of calls to busy subscribers (CCBS). The network automatically
arranges for a call that does not succeed (because the called party is engaged) to be
retried when the line becomes free.11
Conference calling. Several users can take part in the same conversation.12
‘Follow-me’ diversion. Customers should be able to invoke call diversion from
wherever they are so that they can receive their calls on any convenient phone.
Freephone service. This is the most well known of all IN services and is subject
to much discussion in other parts of the book. It effectively provides a reverse-
charge call for anyone who knows the Freephone customer’s number.
Malicious call identification (MCI). This is a service that is offered to users
who complain of being subjected to abusive incoming calls. If the user invokes the
service (by pressing the recall button or a key on the key-pad) then the network
operator is alerted. The calling line identity is then recorded and the line held for
tracing if necessary. This is another service which, although it was included on the
potential list of IN service examples, was available long before IN and so is
traditionally provided using switch-based and network signalling functions.
Mass calling service. This typically applies to large corporate Freephone
customers where the target destination numbers may need many thousands of ports
to handle high volumes of incoming calls. An essential feature is the availability of
statistics to enable the customer to manage and plan the call reception points.
Originating call screening. Before an outgoing call is completed a profile is
checked to ensure that the renter of the phone allows the type of call which has
been requested. If the call is disallowed, maybe because of its cost, a rejection
announcement or tone is played. Conversely, terminating call screening can be
applied to incoming calls, where calls can be rejected if specified criteria are not
met. For instance, calls from a particular CLI could be rejected.

11
Note: whilst CCBS was incorporated in the list of benchmark services as a challenge to the CS-1
standards-makers, it actually needs some IN functions which are more advanced than CS-1. In the
meantime the service is therefore commonly implemented using non-IN (switch-based) functions, with
proprietary network signalling messaging sometimes used to inform the caller when the line becomes
free.
12
Note: as with CCBS, it is not possible to provide IN controlled conference calling without more
advanced, post-CS-1, IN features. Again, conference calling is a common PSTN service, generally
provided with non-IN capabilities until CS-2 features become available.
80 Intelligent networks – principles and applications

Premium rate service. The premium rate service provides access to sought-after
information for which a high charge can be levied. The charge is recovered from
the user’s phone bill and the revenue is apportioned between the content provider,
the service provider and the network operator as appropriate. Distinctive number
range codes should be reserved for premium rate services so that a user is aware of
the potential high cost before making a premium rate call. The revenue
apportionment of premium rate services is an example of the ‘split charging’
feature, where IN charging functions can be used to allocate the cost between
calling and called parties (as well as to third parties, where appropriate).
Security screening. Security screening can be applied at an SCF to check that
requests made to a service provider’s application have originated from authorised
sources.
Televoting. IN functions can be used to count the number of calls to particular
network addresses that have been previously advertised as options for voting.
Callers dial the numbers advertised for their particular voting preferences and calls
are registered in the counting service logic. A unit fee is typically raised and the
calls are terminated on announcements that assure callers that their votes have been
registered. The specific IN network ‘activate service filtering’ feature, devised for
this service application, has been introduced in Section 2.5.1 and is illustrated in a
service example later, in Chapter 7.
Universal access number. This is sometimes known as a ‘personal number’
service, where a customer can be allocated a ‘personal’ number that is separate
from the network termination port address (telephone number) of a particular
telephone line. This feature allows the universal personal telecommunications
service to move between different fixed (and mobile) lines and still be reached. The
customer can control how incoming calls are handled so that calls can be directed
to different network addresses at different times of the day. Calls can automatically
be sent to answering machines at certain times, data calls can be sent to data
terminals and fax calls to fax terminals. Other criteria, such as the CLI of the caller,
can be used to determine where to terminate the call.
User-defined routing. A customer is able to specify the network routing to be
used for outgoing calls. This is commonly used to express preference in the
selection of a particular long-distance carrier network for a call. If the first carrier is
not available then the network should attempt the call via a customer-selected
second choice.
Virtual private network (VPN). This allows a corporate customer to set up
what appears to be a private network using public network resources. Customer
benefits include a private numbering plan, using short numbers, and reduced call
charges to other ‘on-net’ destinations.

4.3.3 The IN conceptual model

The INCM, which was shown in the overview in Figure 4.4, is shown in greater
detail in Figure 4.5. Essentially, the analysis starts with the service examples going
 International standards for intelligent networking 81

into the top of the process with the completed protocol specifications, meeting all of
the network needs of the example services, coming out of the bottom.

4.3.3.1 The services plane

At the top level is the services plane, where services are described in terms of sets
of service features. Working down from the 25 ‘benchmark’ services described in
Section 4.3.2 above, CS-1 defined [27] a set of 38 separate service features that
could support these example services. It must be remembered of course that the
purpose of this set of service features was not just to enable the 25 benchmark
services to work, but to provide the ingredients for a far wider range of IN services.
The actual list of service features defined for the ITU-T CS-1 is not reproduced
here, but as an example, an ‘alternate billing’ service implementation might make
use of (at least) the following three service features:
Originating user prompter – this feature is to connect an announcement to the
calling line in order to invite the caller to enter extra information, such as a PIN
code, using DTMF digits or speech.
Authorisation code – this is the ability of the network to collect and process the
PIN code when the caller provides it.
Call logging – this prepares a call record to be prepared for each call. The record
can then be relayed to a billing centre for processing.

4.3.3.2 The global functional plane

The plane below the top-level services plane in Figure 4.4 is the global functional
plane (GFP). Here, the logic of the service features is expressed in terms of the pre-

Figure 4.5 The IN conceptual model

82 Intelligent networks – principles and applications

defined service independent building blocks (SIBs). Just as the GFP is an

abstraction of the complete IN-structured network as a single entity, the SIBs are
abstract representations of the network’s capabilities.
CS-1 defined 13 SIBs [28] for the previously defined service feature set. The
SIBs are listed in Table 4.3. CS-1 also defines a ‘special’ SIB called the basic call
process (BCP), whose purpose is to provide an underlying representation of the call
control ‘thread’. The BCP co-ordinates the invocations of the other SIBs.
Figure 4.5 illustrates how SIBs can be chained in order to emulate service
features. SIBs are independent of both the services they are creating and the
underlying implementation technology.
Whilst the GFP was part of the overall standards-making model, the anticipation
was that the arrangement of standard SIBs in the GFP could be re-used as a
standard tool for service designers to create service features from standard network
components using a standard programming environment. This has been partially
successful, although SIB definitions are high level and there are differences
between different implementations of service creation systems. For this reason
multi-vendor working involving SCEFs and other IN FEs has not generally been
achieved.
We do not pursue the theory of SIB chaining and the GFP any further. Service
creation was discussed in some detail in Chapter 2 but for more information the
reader is referred to Reference [29].

4.3.3.3 The distributed functional plane

The distributed functional plane (DFP), which is the next plane down in the INCM
(Figure 4.5), breaks the SIBs down into functional entities (FEs). Six FEs are
modelled in the CS-1 DFP; they were all introduced in Chapter 2 and are listed
below:

^ Call control function (CCF)

^ Service switching function (SSF)
^ Specialised resource function (SRF)
^ Service control function (SCF)
^ Service data function (SDF)
^ Call control agent function (CCAF).

Table 4.3 IN CS-1 SIBs

& Algorithm & Screen
& Charge & Service data management
& Compare & Status notification
& Distribution & Translate
& Limit & User interaction
& Log call information & Verify
& Queue
 International standards for intelligent networking 83

CS-1 models these six FEs [30], but additionally it describes another three FEs, in
preparation for more detailed treatment in later capability set releases.
The CCAF, which represents a network user device that is more sophisticated
than an analogue POTs phone, does not trigger directly to an SCF, but conducts all
communication through the CCF.
The additional FEs, described by but not modelled by CS-1 are:

^ Service management function (SMF)

^ Service management access function (SMAF)
^ Service creation environment function (SCEF).

These extra functions have already been introduced in Chapter 2. The complete set
of the functional entities prescribed by the ITU-T for the generic DFP are described
in Reference 40.
Figure 4.6 reproduces the Q.1214 CS-1 DFP, showing the FE relationships
applicable for IN CS-1. The two sets of SSF/CCF combinations have identical
functionality but are shown because it is possible for one SSF to act as an ‘assisting
SSF’ – where the ‘Assist’ procedure is used in the same way as it is with an
assisting SRF. There are therefore two possible SSF – SCF INAP relationship
types.
The FEs (the ‘network ‘objects’) communicate with each other using IN
‘information flows’ (IFs). An IF is a message that passes between two FEs. Most
IFs in the DFP map directly onto operations in the physical plane, and have similar
names in both planes.

4.3.3.4 The physical plane

The physical plane contains the mapping between the FEs of the previous section –
the DFP and physical entities. The mapping is illustrated in Figure 4.7. Seven of the

Figure 4.6 IN distributed functional plane model for CS-1

Reproduced, with permission, from Figure 3-1/Q.1214 ITU-T publication, ‘Distributed Functional Plane
for Intelligent Network CS-1’ [10/95].
84 Intelligent networks – principles and applications

Figure 4.7 IN physical elements for CS-1

Reproduced, with permission, from Figure 1/Q.1215 ITU-T publication ‘Physical Plane for Intelligent
Network CS-1’ [10/95].

possible physical entities (PEs) are shown, and some of the FE mappings are
optional. It is possible (as indicated in the diagram) to include SDF and SCFs in an
SSP, in which case it is normally called an SSCP (service switching control point),
which was introduced in Chapter 1. These mappings are representational and other
PE–FE mappings are not excluded. It is also possible for FEs to be located in PEs
other than those illustrated here [30].
 International standards for intelligent networking 85

Except for the NAP and AD, the physical entities shown in Figure 4.7 were
introduced in Chapter 2. The NAP contains only CCAF and CCF functions and has
no means of communicating with an SCP. The AD (adjunct – which was introduced
earlier in the chapter in connection with the Bellcore AIN standard) is equivalent to
an SCP but is more closely coupled to the SSP, using a higher speed connection
than SS7.

4.4 ETSI core INAP and CS-1R

After the ITU-T defined CS-1 in 1992 it was found to be too wide-ranging and
incomplete in detail to use for detailed implementation. ETSI therefore set about
extracting and defining a pragmatic sub-set that could be used as a basis for IN
installations in Europe. This work was completed in 1994 and was called the ‘core
INAP’ standard. It was then adopted by the ITU-T and released as an international
standard in 1995 under the title CS-1R.
As an indication of the difference between the original ITU-T CS-1 and CS-1R,
CS-1 initially defined 55 INAP operations, but only 29 survived in ETSI core
INAP. These core INAP operations were described in Chapter 2, and the remaining
26 omitted from core INAP are listed and described in Appendix 2.

4.5 IN and mobile systems

An area where IN is having a significant impact today is in mobile telephony.

Several standard supplementary services are available in GSM, including features
such as conditional call forwarding and call barring, but until the past few years
there has been little standards support for some of the more advanced voice
services, such as VPN. This has resulted in the development of a variety of non-
standard solutions tailored for particular services. Whilst these sorts of solutions are
optimised for particular sets of requirements, evolution without agreed standards
can obviously be problematic and expensive. Another difficulty which has been
emerging with operator-specific solutions for advanced services for mobile users,
and one which is becoming increasingly significant, is that familiar services do not
always work correctly for international travellers who are temporarily served by a
foreign mobile network.

4.5.1 The role of IN with GSM

To help with these problems, GSM Phases 2 and 2þ included standard

specifications for sets of supplementary services. However, tight standardisation
presents operators with a converse dilemma in that if all operators can offer the
same services then there is limited scope for competitive differentiation.
86 Intelligent networks – principles and applications

For this reason manufacturers started to offer new mobile switches and upgrades
that are also able to function as intelligent network SSPs, based on the CS-1
functions. The incentive for introducing IN functions is to give mobile operators the
potential for implementing centralised advanced services at reduced costs, and with
the flexibility to customise standard services with extra features to distinguish their
particular product set.
An option for extending the set of available services, beyond that provided by
switch-based signalling procedures, is to provide access from the MSCs to an IN
service control point. Hence, some advanced services have been provided in mobile
networks by using INAP. However, because the ITU-T IN standards were designed
for fixed networks they do not transfer directly to mobile networks. For this reason,
proprietary extensions to the INAP protocol, and individual adaptations and
modifications of the IN standard network element concepts, were needed, and so
inevitably a variety of ‘proprietary’ INAPs were introduced.

4.5.2 CAMEL

In order to make IN solutions available in the mobile world in a more standardised

fashion, the GSM CAMEL specification was published. The prime driver for
CAMEL was to give mobile network operators the ability to offer their own sets of
advanced services to their customers even when roaming into other operators’
networks.
CAMEL originated in ETSI and represents the first co-ordinated approach of the
IN standards into the world of mobile networks.
CAMEL is the leading IN standard for GSM. Because it originated from ETSI it
is derived from the ITU-T IN CS-1 specifications. There is therefore a degree of
compatibility between ITU-T fixed network IN systems and GSM CAMEL
systems, although they have developed separately to a large extent, so inter-
working between the two is not entirely straightforward. Phase 1 of CAMEL
(published in 1997) provided a basic set of IN capabilities. The aim was for the
interface between the GSM SSP (termed a ‘gsmSSP’), which is an upgraded MSC,
to follow the ITU-T CS-1 INAP specification as closely as possible, even though
the MSC procedures relating to the IN messages are mobile-specific. Figure 4.8
shows the GSM network architecture enhanced with the IN elements.
CAMEL Phase 2 (1998) added the network features needed to support
interaction with SRFs (specialised resource functions). These are provided in the
IN intelligent peripheral (IP) node, illustrated in Figure 4.8. This allows
announcements to be provided, for example, during Pre-pay services, so rather
than cutting off a call with no warning when the caller’s credit has expired, a
suitable announcement can be played to the caller first. Examples of the Pre-pay
mobile service are given in Chapter 7.
The IP may be located either in the home network or in a foreign network. For
international travellers served by a foreign network, access to an IP located in the
 International standards for intelligent networking 87

same foreign country can be considerably more efficient since no international

transmission circuits are used, resulting in lower costs.
CAMEL support is provisioned in the HLR on a per-subscriber basis. When a
customer is marked for CAMEL service in the HLR, the switch-based SSP
functions and SCP interactions can be invoked for all circuit-switched calls except
emergency calls. When the SSP functions are invoked, events, such as call set-up,
answer and disconnect can be reported to the SCP. The SCP can modify call
handling at call set-up time, and can instruct the SSP to release a call at any stage.
Forwarded calls are treated as mobile-originated calls.
The sets of detection points in the CAMEL Phase 1 and 2 standards are listed in
Table 4.4.

4.5.3 CAMEL Phase 1

The CAMEL Phase 1 standard was finalised by ETSI in 1997 and it was part of the
GSM Release ’96 package, with conforming equipment becoming available around
three years later.
The CAMEL Phase 1 call model has three trigger detection points (T_DPs) in the
originating half of the basic call state machine (O_BCSM) and three in the
terminating half of the basic call state machine (T_BCSM).
The gsmSSF in the MSC (or G-MSC) uses these trigger points to send requests
for further instructions on call handling to the gsmSCF in an IN SCP. The T_DPs
are statically armed, as with fixed-network IN procedures, but, unlike fixed-
network IN, they are not pre-set in the SSP data – they are armed by virtue of the
fact that they are provisioned in the HLR.

Figure 4.8 GSM enhanced with CAMEL IN features

88 Intelligent networks – principles and applications

Table 4.4 CAMEL IN detection points

CAMEL CAMEL
CS-1 (Q.1214) detection points (DPs) Phase 1 DPs Phase 2 DPs

In the originating BCSM:

1 – Orig_Attempt_Authorised
2 – Collected Info \ \
3 – Analysed Info
4 – Route_Select_Failure \
5 – O_Called_Party_Busy \
6 – O_No_Answer \
7 – O_Answer \ \
8 – O_Mid_Call \
9 – O_Disconnect \ \
10 – O_Abandon \
In the terminating BCSM:
12 – Term_Attempt_Authorised \ \
13 – T_Called_Party_Busy \
14 – T_No_Answer \
15 – T_Answer \ \
16 – T_Mid_Call \
17 – T_Disconnect \ \
18 – T_Abandon \

When the gsmSCF receives a request from a gsmSSF, it is can respond in order
to instruct the gsmSSF to do one of three things:

. connect the call to a specified destination

. continue with normal processing
. release the call.

The gsmSCF can also set Event Detection trigger Points (E_DPs) for ‘answer’ and
‘disconnect’. E_DPs are dynamically armed by Request Report BCSM
(RR_BCSM) instructions sent from the gsmSCF to the gsmSSF.
The SS7 IN signalling protocol for CAMEL is called the CAMEL Application
Part (CAP), and is based on ETSI core INAP, but with a reduced set of operations.
Whilst CAMEL Phase 1 has fewer INAP operations and triggers than CS-1, it is
still a powerful capability. The key events in a call (call set-up, answer and
disconnect) are included.
Finally, a new interface to the HLR, using additional and modified SS7 MAP
operations, was provided in CAMEL Phase 1. This enabled the SCP to find out
about mobile status, such as busy or idle, and mobile location information, such as
VLR number or geographic co-ordinates of the serving cell base site.
The IN feature set of CAMEL Phase 1 is a modest sub-set of IN CS-1, and
because many operators had already installed proprietary versions of INAP, take-up
of CAMEL Phase 1 has been steady rather than swift. However, Phase 2 provides a
big increase in the number of detection points supported.
 International standards for intelligent networking 89

4.5.4 CAMEL Phase 2

CAMEL Phase 2 [31] is backward compatible with Phase 1 and uses a more
comprehensive BCSM, with a greater range of options for charging of Pre-pay
calls. The INAP Apply Charging and Apply Charging Report operations were
included in the enhanced CAP for Phase 2.
However, the main addition for Phase 2 is a set of procedures for including IN
Intelligent Peripheral nodes (containing gsmSRF) in mobile calls. This enables the
gsmSCP to control transactions in which users can buy more credit or make balance
enquiries.
Also, Phase 2 extended the list of E_DPs. For instance it included the
O_No_Answer E_DP, which allowed the CAMEL service environment (CSE) to
provide a diversion on no reply service.
In summary, CAMEL Phase 2 provided the following extras on top of Phase 1:

. improved set of IN detection points and enhanced CAP signalling

. conditional E_DPs and additional E_DPs such as abandon, busy and no answer
. announcements, tones, voice prompting and information collection using an IN
IP
. controls for interactions between network and SCP-based services
. new capabilities for tariff control (for Pre-pay services) and charging information
control
. alerting pattern control.

4.5.5 Future CAMEL phases

Phases 3 and 4 of CAMEL address GPRS and SMS interworking with IN functions.
Some IN CS-2 functions are also included to enable leg manipulation, e.g. for
playing announcements during a call [47].

4.5.6 Wireless intelligent networking (WIN)

CAMEL, developed by ETSI as an extension to GSM, is the European regional

standard for applying IN to mobile systems. The equivalent regional standard in the
USA is WIN (wireless intelligent network). WIN was developed by the TR45.2
work group of the US standards authority, TIA (Telecommunications Industry
Association), using ITU-T IN principles as a starting point. This was in
acknowledgment of the widespread international success of the ETSI GSM
standard and a desire to ease the future interworking between the two.
Whilst there are underlying differences in air interfaces and network signalling
protocols, the WIN and CAMEL standards groups aimed to harmonise as closely as
possible in order to ease international standardisation of the IN interface across
GSM and non-GSM networks (e.g. ANSI-41 networks in the US).
90 Intelligent networks – principles and applications

CAMEL is based on the ITU-T Q.1214 CS-1 BCSM and WIN is based on the
ITU-T Q.1204 [40] generic BCSM. Although the latter contains more PICs and
DPs than the former, this common standards ancestry ensures a shared pedigree for
the CAMEL and WIN standards whilst additionally providing WIN with linkage
with the US AIN standard, which uses a similar call model to that in Q.1204.

4.6 ITU-T IN CS-2

The ITU-T significantly widened the scope of the IN standard for CS-2 in several
directions, but the prime feature, and the one which received the most attention in
terms of detailed definition, was ‘call party handling’ (CPH – otherwise known as
‘leg control’). This is an enhancement to the call modelling principles of CS-1, and
it enables us to use the IN to assist with the provision of multi-party calls, such as
conference calls, call waiting and call transfer services. The CS-1 BCSM was
principally concerned with influencing the sequential flow of a call as it progressed
through the different set-up phases.
With different emphasis, CPH seeks to allow the remote service logic to
manipulate the individual legs of a call. This is made possible because the
originating and terminating call segments of CS-1 are now supplemented with a
third ‘associated’ call segment, which makes it possible to model call party
handling.

4.6.1 Extra functions provided in CS-2

The CPH mechanisms are discussed in further detail in Chapter 5, and the other
extended functions specified for CS-2 [32] are summarised here.
End-user access to call and service processing. Greater attention is given in CS-
2 to providing for IN service access for ISDN (primary and basic rate users) than
was the case with CS-1.
Service invocation and control. As with IN CS-1, CS-2 limited its scope to
single-ended, single-point-of-control services (called ‘type A’ services). Otherwise,
IN-CS2 aimed to provide a ‘super-set’ of the IN-CS-1 capabilities, extending the
range of service possibilities, but still predominantly attending to voice networks.
Internetworking. As illustrated later in the DFP diagram for CS-2 (Figure 4.10),
support for data communication between networks is included, using SMF–SMF,
SDF–SDF and SCF–SCF interfaces. The SCF–SCF interface recommendation is
designed for the exchange of data or distribution of service logic and not for service
scenarios involving distributed service control. The new mechanism will allow an
SCF to query another SCF, or to hand over the call to another SCF.
Out-channel call-related user interaction (OCCRUI). For communication
between the user and the service control function in CS-1, an IP (intelligent
peripheral) interaction was necessary to collect DTMF digits, or spoken responses,
in answer to an audible menu choice. CS-2 defines a mechanism for exchanging
 International standards for intelligent networking 91

service-independent information elements directly between the user’s equipment

and the SSF for onward transfer to the SCF. This feature makes it possible for a
terminal to communicate directly with an SCP, for example to provide further
address information, or respond to requests for choices of call processing options.
Out-channel call-unrelated user interaction (OCUUI). This feature allows IN to
assist in cases where non-call-associated features need to be invoked independently
of call contexts. Examples are terminal registration, location updating and message
waiting indicator transfer. OCUUI can therefore assist with convergence between
fixed and mobile networks. It can enable SSPs to provide access to home location
registers (HLRs) in the same way that GSM dedicated switches do. To model the
required state changes for OCUUI, CS-2 introduces a new state transition model
called the Basic Call Unrelated State Model (BCUSM).
The new functions introduced for OCUUI are the CUSF (call unrelated service
function) and the SCUAF (service control user agent function). The BCUSM (in
the CUSF) exists in parallel with the BCSM (in the SSF), located typically in a CS-
2 SSP. Typically the SCUAF would be expected to reside in an ISDN CPE device,
using D-channel signalling for network access. The CUSF contains the functions
that allow the SCF and SCUAF to interact, detecting call-unrelated triggers and
supporting the ensuing SCF transactions.
Wireless access. IN CS-2 also provides other enhancements for the support of
wireless and mobility systems. New functions for radio systems include:

^ CCAFþ (call control agent function plus)

^ CRACF (call-related radio access control function)
^ CURACF (call-unrelated radio access control function)
^ RCF (radio control function).

Feature interactions. As the complexity of IN systems expands, the scope for

adverse interactions amongst IN and non-IN features expands with compound
interest. Q.1224 and Q.1222 lay out principles for the avoidance of interference
between features and services, but the application of these principles is very much
implementation-dependent and will vary across different network operators’ IN
environments. The handling of feature interactions, which is fundamental to
successful IN implementations, has been discussed earlier (Chapter 2). The CS-2
proposed formal method for feature interactions is described below in the context
of the IN CS-2 service plane.
Extra trigger point. To supplement the CPH functions an additional BCSM PIC
(O-suspended) and DP gave improvements in ‘follow-on’ call handling. This allows
three-party calls (e.g. for call waiting) to be handled more satisfactorily than in
CS-1.
Service creation and service management. In addition to an extra six SIBs for
new CS-2 functions, CS-2 also provided some added coverage for service creation
and service management procedures, as well as for inter-SMF (service management
function) communications. However, by comparison with the detailed text on CPH,
these are high-level reviews. Because of the practical predominance of proprietary
92 Intelligent networks – principles and applications

solutions in these areas, they have turned out to generate less attention in multi-
vendor scenarios than the CPH network recommendations provided by CS-2.
The SMF is responsible for the control of the real-time IN entities, such as SSF,
SCF and SRF, including service provisioning, data updates (to service data as well
as customer-specific data), etc. The CS-2 SMF is based on the TMN functional
architecture [50, 51]. Appendix B3 of Q.1224 explains the mapping between the
service creation and service management functions for CS-2 and function blocks in
the TMN functional architecture.
CS-2 also defines a service management access function (SMAF) for interfaces
between SMF and the end-user and between SMF and the network or service
operator.
The service creation environment function (SCEF) provides the toolkit for the
preparation for installation of new IN services and for modifying existing ones.
After the SCEF has prepared the new service (or altered an existing one), it
dispatches it to the SMF (Figure 4.10), which then has responsibility for deploying
the service. This responsibility includes version control, compatibility tracking,
maintenance and fault handling aspects. The SMF also looks after real-time
triggering issues, such as service key allocation, number plan management and SLP
identification in the SCF.

4.6.2 Service plane for IN CS-2

Unlike CS-1, CS-2 aimed for a disciplined ‘top-down’ architecture approach, which
would characterise and define the required services first and then work downwards
through the INCM planes towards suitable implementation recommendations. This
did not happen so emphatically with CS-1 because at that time the drive was to
evolve existing network structures towards the concepts of IN. What could be done
was highly dependent on what was possible in existing networks, so the approach
was essentially ‘bottom-up’.
As well as specifying this top-down process, the CS-2 service plane document
[33] provided some initial steps towards a rigourous methodology for dealing with
feature interactions using UFM (unified functional methodology). With UFM, all
services (and eventually network architectures) are described using recommended
functional descriptions and information flows. Basically, by carefully following the
CS-2 ‘top-down’ ideal, all services descriptions would use the same recognised
methods, and so the identification of potential interactions between the features
would become automatic.

4.6.3 Global functional plane for IN CS-2

New SIBs were included in CS-2, including some for CPH leg control operations as
well as SIBS for starting, ending and controlling parallel SIB chain execution (the
last four on the list in Table 4.5). The idea of parallel processing was new in CS-2,
 International standards for intelligent networking 93

Table 4.5 The extra SIBs in IN CS-2

& Join
& Split
& Create service process
& Send
& Wait
& End

allowing more than one SIB chain to be executed at once. To keep the execution of
SIB chains in step with each other, points of synchronisation were also introduced.
Two new SIB groupings were introduced, namely ‘high level SIBs’, which are
groups of lower level SIBs, and ‘service processes’, which are groups of high level
SIBs and lower level SIBs. These new groupings are used to simplify the logic
descriptions when parallelism is used [34].
The global functional plane contains the ‘global’ (i.e. abstract) service logic
(GSL), which defines the operation of the SIBs, and their interactions with the basic
call process model to provide particular services. The GSL defines the point of
initiation in the BCP where the logic of the specified SIBs is executed. The GSL
also defines the point of return to the BCP.

4.6.4 CS-2 distributed functional plane

As with CS-1, the CS-2 DFP describes the CS-2 network functions independently
of their physical locations. CS-2 represented a major expansion to the coverage of
the ITU-T IN recommendation, and there were many candidate network features
competing for inclusion. However, the functions that could be included in the CS-2
release had to achieve a balance between the CS-2 service expectations and the
limitations imposed by the capability of the installed base of network technology to
evolve with sufficient flexibility to make the upgrade feasible (Figure 4.9).
The actual DFP model for IN CS-2, showing the CS-2 functional entities and
their relationships is reproduced in Figure 4.10. The five new FEs modelled in

Figure 4.9 The DFP scope constraints

94 Intelligent networks – principles and applications

Figure 4.10 The IN CS-2 DFP architecture

Reproduced, with permission, from Figure 3-1/Q.1224 Fascicle 1/3 ITU-T publication, ‘Distributed
Functional Plane for Intelligent Network CS-2’ [09/97].
 International standards for intelligent networking 95

CS-2, additional to those in the CS-1 DFP, are explained in Section 4.6.1 above.
These are the SCUAF, CUSF, SMF, SMAF and the SCEF.
The CS-2 distributed functional plane recommendation [32] is a large document
and mainly comprises detailed exposition of the functions and relationships
represented in Figure 4.10. It explains the information flows, in terms of operations
and parameters, and includes separate annexes for the coverage of mobility, TMN,
network and fault management, and testing.
In addition to the new FEs introduced, Figure 4.10 differs from Figure 4.6 (the
equivalent CS-1 diagram) in three other general respects:

1. Peer relationships now exist for SCFs, so that service logic can be distributed
onto multiple SCFs using the new SCF-SCF interface. Similar peer relationships
are provided for SDF and SMF functional entities.
2. Internetworking. As an extension to these new peer relationships, connection to
other networks is now recognised as being standardisable. An ‘intelligent access
function’ (IAF) is introduced for connections between IN and non-IN networks.
3. User to network signalling (known as ‘out channel’ signalling) is included. This
uses the new SCUAF and CUSF functional entities.

4.6.5 IN CS-2 documents

The CS-2 recommendations were finalised in 1997. The 4-plane INCM, which
provided the framework for the IN CS-1 document structure (Section 4.3.3 above),
applies equally for the CS-2 recommendations. Therefore the numbering for the
recommendation document series follows the same pattern as the IN CS-1 series,
except that the third digit is a ‘2’, for CS-2. Table 4.6 lists the CS-2 document set.
The covering document is Q.1220, which explains the IN CS structures and includes
the index for each Q122x document. Q.1290 contains terms and definitions.

4.6.6 Physical plane for IN CS-2

Q.1225 defines the CS-2 physical plane, which is reproduced in Figure 4.11. This is
similar to Figure 4.7 (the CS-1 physical plane) except that it shows the extra

Table 4.6 The ITU-T IN CS-2 documents

Q.1220 Q series recommendations structure for IN CS-2
Q.1221 Introduction to IN CS-2
Q.1222 Service plane for IN CS-2
Q.1223 Global functional plane for CS-2
Q.1224 Distributed functional plane for CS-2
Q.1225 Physical plane for CS-2
Q.1228 Interface recommendations for CS-2
Q.1229 User guide for CS-2
96 Intelligent networks – principles and applications

Figure 4.11 Physical plane architecture for IN CS-2

Reproduced, with permission, from Figure 1/Q.1225 ITU-T publication, ‘Physical Plane for Intelligent
Network CS-2’ [09/97].
 International standards for intelligent networking 97

physical entities associated with the new IAF, TMN and call-unrelated functions
included in IN CS-2.

4.7 ITU-T IN CS-3

In December 1999 a scoping document (Q.1231 – Introduction to IN CS-3) for IN

CS-3 was produced, as well as a description of the IN CS-3 management
information model. Six months later, the series of Q.1238 interface recommenda-
tions was released.
In general terms CS-3 consolidates many of the principles of CS-2, with some
modest expansion in certain areas. For instance, there is some support, albeit
limited, for broadband IN using the CS-2 call model. An IN-controlled number
portability service is addressed, there are enhancements to the CS-2 CUSF (call-
unrelated service function) and IN-ISDN interworking is improved. Multiple points
of control are allowed, in the sense that CS-3 permits multiple triggering from the
same SSF to different services, which can then run concurrently (this was
permissible in CS-1 and CS-2 if the triggers occurred in separate SSFs). To allow
for multiple points of control the FIM was enhanced to deal with ensuing
interactions between the SSFs involved in the invocations.

4.7.1 IN CS-3 documents

The new documents for CS-3 are listed in Table 4.7. Where no new document
appears, the CS-2 recommendations are carried forward for CS-3.

4.8 ITU-T IN CS-4

The CS-4 ‘introduction’ [41] and ‘distributed functional plane’ [42] documents set
out the scope of the topics that are expected to be central to IN CS-4

Table 4.7 The ITU-T IN CS-3 documents

Q.1231 Introduction to Intelligent Network Capability Set 3
Q.1236 Intelligent Network Capability Set 3 – Management Information Model
Requirements and Methodology
Q.1238 Interface Recommendation for IN CS-3
Q.1238.1 Interface Recommendation for IN CS-3 – Common Aspects
Q.1238.2 Interface Recommendation for IN CS-3: SCF-SSF Interface
Q.1238.3 Interface Recommendation for IN CS-3: SCF-SRF Interface
Q.1238.4 Interface Recommendation for IN CS-3: SCF-SDF Interface
Q.1238.5 Interface Recommendation for IN CS-3: SDF-SDF Interface
Q.1238.6 Interface Recommendation for IN CS-3: SCF-SCF Interface
Q.1238.7 Interface Recommendation for IN CS-3: SCF-CUSF Interface
98 Intelligent networks – principles and applications

implementations. Many new service features and corresponding network

capabilities [46] are outlined; however, the prime focus of CS-4 is the
establishment of IN’s role against a background of converging IP (Internet
protocol) and PSTN networks. The scope of CS-4 includes support for IN access
from H.323 Gatekeepers and SIP Proxy Servers, although this support is limited to
the BCSM approach, and does not to date extend to the CS-2 call party handling
functions.

4.8.1 Alliance between IN and Internet telephony

The CS-4 documents available at the time of writing demonstrate an increasing

desire to align with the Internet standards coming from the IETF. In particular, CS-
4 acknowledges and builds on the Internet-driven initiatives such as PINT [43] and
SPIRITS [44].
PINT (‘PSTN/Internet interworking’) transactions are characterised by requests
that originate via an IP network and result in the automatic invocation of telephone
calls. An example could be an Internet user clicking on an icon on a travel agent’s
Web page with the intention of eliciting a return telephone call from the travel
agent giving more information.
The interface from an IP server to a PSTN could be through a PBX, an ISDN
device or a mobile phone network. However, the main IN interest is the connection
of an SCF (through the IAF) to a PINT gateway. This then allows calls to be set up
automatically across the PSTN, using IN network features, on receipt of PINT
requests.
SPIRITS (‘services in the PSTN/IN requesting Internet services’) is concerned
with requests in the other direction – from the IN into an IP network. The SPIRITS
IP client could be built into an SCF, located in an SCP or SN, for instance, and
originate requests to a remote IP server at significant stages during a PSTN call.
Internet call waiting (where a user whose only phone line is engaged on an Internet
connection is alerted to the arrival of an incoming call attempt to the phone) is a
typical SPIRITS application.
In another example of the use of SPIRITS technology, a corporate Freephone
customer might subscribe to a service that provides statistics at an IP server on
failed incoming calls. In this case the IP server could be alerted by a SPIRITS
transaction that is invoked whenever a ‘routeset fail’ E_DP (event detection point)
is triggered in an incoming call to that customer.

4.8.2 Interworking between CS-4 and other networks

The emphasis of CS-4 is on interworking with surrounding networks. CS-4 expands

on the use of the IAF (intelligent access function), which was introduced in CS-2,
in order to interwork with non-IN structured networks. As outlined above, it also
 International standards for intelligent networking 99

includes preparatory information for interworking with SIP, H.323, PINT and
SPIRITS systems.
Specifically CS-4 also defines gateway functions for service applications and
service control. These are:

^ SA-GF (service application gateway function)

^ SC-GF (service control gateway function).

4.8.3 Service application gateway function

The SA-GF provides access to distributed service logic via an open API, including
the necessary firewall and security mechanisms. Typical API technologies are
CORBA, JAVA and JAIN. Figure 4.12 shows SA-GF as embedded in an SCF,
which is then able to provide the open API with access to the PSTN via INAP
interfaces to SSFs and SRFs, etc. However, another CS-4 option is for a new
interface to be defined between separated SCFs and SA-GFs. ITU-T CS-4 is also
linked into the OSA API definition [37], which is introduced in Chapter 6 in
connection with the Parlay API evolution.

Figure 4.12 Service application gateway function

100 Intelligent networks – principles and applications

4.8.4 Service control gateway function

The SC-GF allows access to existing IN service logic from an IP13 network node. It
could be co-located in either the SCF or IP (Internet protocol) server, or separately
implemented, as shown in the interworking example of Figure 4.13. This diagram
illustrates an IP network call control executive on the right-hand side, accessing
traditional IN functions on the left-hand side. This call control executive in the IP
network is typically an SIP proxy, an H.323 gatekeeper or a media gateway
controller (MGC), so these are the logical VoIP components that would be
modelled by an IN SSF as shown.
The SC-GF provides the necessary lower-level mapping functions for INAP
messages between the two networks.
At the call and bearer control level, speech trunks terminate on a media gateway
and the ISUP signalling terminates on a signalling gateway, which then interacts
with the IP control element. Because the ISUP signalling is no longer associated

Figure 4.13 Service control gateway function – an example scenario

13
At this stage of the book there is scope for confusion with the abbreviation ‘IP’, which is used for both
the IN ‘intelligent peripheral’ and ‘Internet protocol’. For instance, in Figure 4.13 both terms appear. We
therefore take care to clarify usage each time, but where the term ‘IP network’ is used this invariably
refers to Internet protocol.
 International standards for intelligent networking 101

with the bearers that terminate on the media gateway, a channel control interface is
required between the IP call control entity and the media gateway. ITU-T H.248
(based on Q.931 and also known as MEGACO) is the international standard
interface for this. For further information on VoIP, including the MEGACO, H.323,
SIP and BICC aspects, the reader is referred to Reference 52.
The different IN functional entities, including the ‘new’ SSF in the IP network,
are managed by the IN SMF. This includes, for instance, the population and update
of locally held IN data such as the SSF trigger criteria.14

4.8.5 CS-4 service features

Q.1241 describes several new service features, mostly drawing on the recent
convergence work between IN and Internet protocol technology. Q.1244 [42]
provides descriptive examples of how such services can be provided using PINT,
SPIRITS and SIP in conjunction with traditional IN network entities. The new
service features include:

^ End-User Service Data Customisation (this is via an IP network, using PINT

access).
^ Request-to-Call-back (an SCP sets up a call, instigated by a user clicking on a
Web page icon).
^ Request-to-Call-back via IP (in this case the call proceeds as a VoIP call).
^ Internet Call Waiting.
^ Web-controlled PSTN/IP Conferencing.

4.8.6 IN CS-4 documents

The following recommendations (Table 4.8) apply specifically to CS-4. The service
plane, GFP and physical plane have not changed, so in these cases the previous CS-
2 or CS-3 documents still apply to CS-4.

Table 4.8 The ITU-T IN CS-4 documents

Q.1240* IN CS-4 Recommendation Structure
Q.1241* Introduction to IN CS-4
Q.1244* Distributed Functional Plane for CS-4
Q.1248* Interface Recommendation for IN CS-4 – Common Aspects
Q.1248.2* Interface Recommendation for IN CS-4: SCF-SSF Interface
Q.1248.3* Interface Recommendation for IN CS-4: SCF-SRF Interface
Q.1248.4* Interface Recommendation for IN CS-4: SCF-SDF Interface

* Approved pre-published versions of new recommendations, July 2001.

14
For diagrammatic clarity the speech circuits between SSP and IP are not shown in Figure 4.13. The
SSP–IP connection is shown as ISUP – it could equally well have been Q.931 (Section 2.4 refers).
 Chapter 5
Call party handling

Airplanes are interesting toys but of no military value.

Marechal Ferdinand Foch, Professor of Strategy, Ecole Superieure de Guerre.

5.1 Introduction

The ITU-T CS-1 IN standard, along with the subsequent ETSI core INAP
modifications, was a resounding success. It led to IN installations in most major
operators’ networks because the equipment vendors generally supported the
engineering content of the recommendations, or at least a significant sub-set of
them. However, in the period leading up to the original CS-1 publication there had
been lively debates about the basic network model that should be chosen for the
standard. The contenders were the sequential call logic approach, which later
became known as the ‘basic call state model’, and the connection view state (CVS)
model, based on ‘snapshot’ views of call-states including the connection
configurations. The CVS model included transition rules that prescribed the
possible connection states between the ‘legs’ that represented the end-points. In the
end the CS-1 standard makers settled on the basic call state model, which cycles
through the logical sequence of events on the originating and terminating side, and
this approach has generally been the subject of the earlier chapters of this book.
Essentially the debate was about how to represent the mechanics of the
underlying switch network to the IN service logic in the SCF. The SCF does not
need to be aware of too much of the detail of the telephone network; on the other
hand it needs a logically consistent view, comprehensible by the SSF as well as the
SCF, of the service logic’s context. This representation is known as the ‘call
model’, as we have discussed earlier.
The debates over the nature of the CS-1 call model became complex and tended
towards the philosophical, but the upshot was that the BCSM approach was chosen
for CS-1. The IN switching manager, using the CVS approach, was deep-seated in
the original ITU-IN framework document series. It was for instance described in
104 Intelligent networks – principles and applications

overview terms in Q.1204, which is the IN generic distributed functional plane

architecture and it was introduced into the CS-1 recommendations [9]. However, it
was not fully adopted as a standard recommendation until IN CS-2 in 1997.
Whilst CS-2 is a preferred basis for the development of IN services, in the mid-
1990s the major European network operators could not wait for the culmination of
the lengthy standards process that was in place for CS-2. They therefore seized on
the newly available and vendor-endorsed ETSI Core INAP, and invested in the
necessary switch upgrades and in service control equipment. In practice the
network upgrades needed to achieve an IN architecture usually had to be justified
step-by-step in discrete business cases, each being justified by revenue-projections
of the services that could be realised at each step. The introduction of CS-1 was
therefore a staged affair, often taking several years. Hence, when the CS-2 standard
was finally agreed, and the agreed CVS model ratified as an international standard,
it was too late. The established network operators had, for the most part, embarked
on expensive CS-1 based programmes and another fundamental change to the basic
call model so quickly after the first one was unrealistic and could not be justified.
Also, most of the business case arguments that would have applied to IN CS-2 had
in fact already been used earlier to justify the implementation of CS-1 upgrades to
the existing switches!
However, new operators, without a legacy of investment in CS-1, are well placed
to take advantage of the more powerful service control features that IN CS-2 offers.
Otherwise, for a network operator with an existing set of legacy IN CS-1 services, it
becomes necessary for the IN versions to co-exist when the need arises for an
upgrade to IN CS-2 leg control capabilities. Some or all of the network’s SSPs need
to be enhanced with the new call model, and CS-2 enabled SCPs must be
introduced. A typical scenario might be as shown in Figure 5.1, where an SCP that

Figure 5.1 Coexisting CS-1 and CS-2 call models

 Call party handling 105

can handle both interface types is able to view dependent SSPs either directly
through the CS-1 BCSM or via the CVS modelling structures. As we shall see
shortly, the CVS approach also includes the BCSM, but at a more detailed level. In
the interests of clarity, Figure 5.1 does not show the detail of BCSM on the CS-2
call model side of the SCF.

5.1.1 Early ‘leg control’ network functions

In order to more substantially introduce the concept of call ‘legs’ before we go on

to explain details of call party handling mechanics, it is probably helpful to review
some of the familiar CS-1 INAP operations, which we are familiar with from
Chapter 2, in terms of their ‘leg-control’ aspects. Several of these earlier operations
implicitly involve the notion of call legs, as we now demonstrate:

. The Connect operation requests the SSF to attempt to connect a waiting A leg
(representing a caller’s line or an incoming speech circuit in an inter-exchange
trunk group) to a new B leg. The B-leg’s local identity is obtained by the SSP by
analysis of the destination address that is provided as a parameter of the Connect
operation from the SCF.
. The ConnectToResource (CTR) operation connects a waiting A Leg to a B leg
that represents a channel in a trunk group leading toward a physical entity –
typically an IN intelligent peripheral containing an SRF module. CTR is also
commonly used for extending a call to an announcement resource on an internal
channel within the SSP. In this case the B leg represents the internal channel
connection to the built-in hardware.
. The DisconnectForwardConnection (DFC) operation is sent from the SCF to
request the SSF to disconnect a B leg that had previously been set up (e.g. in a
CTR or ETC operation) to the channel resource hosting an SRF module.
. The EstablishTemporaryConnection (ETC) operation is sent from the SCF to
request the SSF to set up a temporary connection between the calling party’s A
leg and a B leg representing a connection to another network node, such as an IP,
or another SSP, hosting the target SRF module.
. The InitiateCallAttempt (ICA) operation is sent from the SCF to request the
SSF to create a new call to a call party identified by address information provided
by the SCF in the associated DestinationRoutingAddress parameter. The new
call leg can then be connected to an exchange announcement. This would be
carried out for instance by using the CTR and PA (PlayAnnouncement)
operations. However, in CS-1 this operation cannot normally be used (as it can
be in CS-2) to connect together two network destinations.
. The ReleaseCallPartyConnection operation is sent from the SCF to request the
SSF to release a specified leg of a call.
. The RequestReportBCSM operation requests the SSF to monitor (and report
on) certain call-related events. It does this by arming event detection points. The
leg to be monitored is specified within the operation’s parameters.
106 Intelligent networks – principles and applications

5.2 The CS-2 call model

The main components of the CS-2 call modelling activities in the SSP are shown in
Figure 5.2.
Towards the bottom end, the basic call manager (BCM) acts on the bearer
control function and is an abstraction of the fundamental call control functions and
event detection mechanisms in the switch. The BCM contains the BCSM, including
the detection points and trigger criteria needed for IN invocations. In a similar
manner to CS-1 (described in Chapter 2), the BCSM identifies the sequential sets of
call processing events that have relevance for external service control. The CS-2
BCSM is an expansion of the CS-1 BCSM within the context of the overall ITU-T
BCSM of the ITU-T IN generic framework distributed functional plane of Q.1204.
The feature interaction manager (FIM) is shown above the BCM in Figure 5.2,
and this performs similar functions to those described for CS-1 earlier, in Chapter
2.
At the upper level of the SSP model in Figure 5.2 is the IN switching manager,
which is responsible for providing the ‘aperture’ through which the SCF views the
call control activities of the switch. The IN switching manager provides the

Figure 5.2 The CS-2 call modelling architecture

 Call party handling 107

controlled access for external service logic to act on the lower level switch
resources of the SSP. The operating service logic for the switching manager is
located mainly in the IN switching state model (SSM).
As we mentioned earlier, the overall shape of the SSP call model in Figure 5.2 is
similar to that originally laid out for CS-1. However, as we show in the following
paragraphs, CS-2 expands on the detailed contents of the IN SSM, which is
effectively the ‘engine room’ of the switching manager. The IN switching manager
had already been outlined in the international standards before the publication of
the CS-2 recommendations in the context of the scope of the overall ITU-T IN
framework in the CS-1 recommendation. With this background, CS-2 provides the
essential operational detail required in order to lead into practical implementations
of the CPH principles.

5.2.1 Legs and connection points

The term ‘call party handling’ (CPH) is fundamental to the CS-2 connection view
approach that we are describing here. CPH refers to a system of modelling
telephone calls where call ‘legs’ (representing bearer channels or end-points) can
be added to, joined into or deleted from a connection point. The advantage of this
approach is that the manipulations required for telephony services such as ‘call
transfer’, ‘call hold’ and ‘multi-party calling’ can be fairly easily represented to the
service logic. The sequential call flow model of CS-1 can, by comparison, be
cumbersome in its attempts to model these types of supplementary services.
For CS-1, the only view seen by the SCF was the progress of the call as it stepped
through the different ‘points in call’, and this could be seen at both originating and
terminating sides. This meant, however, that SCF manipulations were more or less
limited to calls involving just two parties, unless multi-leg involvement was
modelled through sequential actions. However, this is not perfect because of the
intervening time delays. A service example (a ‘call completion for directory
enquiries’ service feature) that illustrates the way in which CS-1 has to emulate
multi-party connections with sequential state-machine operations is discussed in
Chapter 7.
With the CPH approach, the service logic has an additional viewpoint, and it now
becomes aware of a call as a whole, rather than in halves. A consequence of this
approach is that the addition of more ‘legs’, or call parties, for multi-party calls
becomes easier to model. CPH allows legs representing bearer channels to be
added, deleted, joined and separated from the other legs of a call using the functions
introduced in CS-2.
In CPH, CS-2 uses object-oriented techniques to describe the IN SSM and the
underlying CVS concept. The CVS model gives an overview representation of the
state of the underlying calls and their connections in terms of connection view
‘objects’. The connection view objects used in the model are:

. call segments
108 Intelligent networks – principles and applications

. connection points
. call segment associations
. basic call state machine (BCSM) instances.

Whereas a CS-1 call generally comprises one originating and one terminating
BCSM instance, CPH in CS-2 can now include multiple (more than 2) BCSM
instances in a call. This is why the BCSM instance is included in the above list of
basic objects for call modelling scenarios.
Each BCSM instance used in particular modelling situations then has devolved
responsibility for the recognition of IN-significant events (and their notification to
the SCF) for a specified call segment. The notion of ‘call segments’ is explained
further on in this chapter.
With the enhanced CS-2 call model, the SCF service logic now needs to be
aware of the connection-state viewpoint as well as the (possibly several) BCSM
instances for the legs involved in a call using IN features. The SCF therefore has to
be aware of substantially more of the complexities of the underlying switch
network than was necessary with a CS-1 BCSM-only approach. An indication of
the consequent potential for increased implementation complexity is given by the
relative number count for ITU-T CS-2 INAP operations. CS-2 specifies 74 SSF-
SCF INAP operations, which compares with 26 for the CS-1R standard
recommendation!

5.2.2 CS-2 core switch capabilities

To provide practical guidance for upgrading existing switches or SSPs to host the
new call party handling functions, the ITU-T CS-2 recommendations [32] defined
four basic capabilities. These are briefly summarised as follows:

1. Either of the parties involved in the call should be able to summon the attention
of the service logic during the speech phase. This is known as a mid-call trigger
detection, which could be expedited via a switch-hook ‘flash’ or the dialling of a
nominated digit string. This might for example be used in order to arrange for a
third party to be brought into the call.
2. It should be possible for any party in a call to be temporarily disconnected from
an established connection and, temporarily or otherwise, be reconnected to a
speech path resource, such as an announcement machine, under the control of IN
service logic in an SCF. The switch should then be able to correlate the
consequent transport connections, so these can be individually manipulated as
required.
3. The IN switching manager can present the SCF with an abstract view of an
isolated portion of a call, and this portion is a call segment, or a ‘half-call’. The
term ‘call segment’ refers to the collection of legs and connection points that
represent the physical transmission and connection resources and the call
processes (the BCSMs) that constitute that isolated portion of the call. This
 Call party handling 109

concept of restricting the SCF’s view to an abstraction of an isolated portion of a

call is illustrated in Figure 5.3.15
4. Lastly, the switch should be able to combine, on instruction from the SCF,
selected transport paths that have been independently manipulated or
transferred, into a call connection. Figure 5.4 shows a case where an SCF is
able to act on an association of two call segments. In this illustration user A is
part of an existing call with user B and a new call attempt from user C is being
set up towards user A. The controlling SCF is now able to see the developing
association between a new call segment for the incoming third party call
arriving at A’s termination in conjunction with the established call segment
representing the original outbound call setup towards user B.

5.2.2.1 The hybrid approach to call party handling

In order to provide an interim means of introducing some CPH features without
heavy switch-upgrade overheads, a ‘hybrid approach’ using CVS principles was
included in the CS-2 distributed functional plane [32] description.
We do not enter into the details of the hybrid approach here, apart from to say
that it provides a means of allowing IN CS-2 services to be provided without a full
CPH upgrade to the switches. This is achieved by using a bridging function in a
separate and dedicated IP network node that contains the required CPH capabilities.
This enables an expedient solution to be introduced while a reduced set of
connection view states exists in the SSP. Details of the network options offered by
the hybrid CPH solution are provided in Reference 32.

5.2.3 The connection view model states

We now look at the various connection view states that are possible in IN CS-2
CPH.
The enhanced CPH view of the SCF, which must now be aware of the connection
state of the relevant call segment as well as keeping track of any BCSM models
with which it has a relationship, is shown in Figure 5.5. As we have already
mentioned, this is referred to as the ‘connection view’ model.
Connection view states (CVSs) describe the connection configurations for
particular connection views. In order to illustrate the principles, Figure 5.6 shows
the progression of a call segment from the ‘null’ state to the stable 2-party state for
a simple 2-party call. This progression is achieved either through an originating call
set-up or by the arrival of an incoming call attempt.

15
This emphasises the fundamental difference, which we introduced earlier, between the new CS-2 call
party handling and the CS-1 approach. With CS-1, the SCF was directly aware of the BCSM state of the
call through the exchange of the CS-1 INAP operations. In the case of the expanded set of CS-2 INAP
operations, the SCF is additionally able to see the state of the individual call segments in terms of their
connection states.
 110
Intelligent networks – principles and applications

Figure 5.3 Call segments in a two-party inter SSF/CCF call

Reproduced, with permission, from Figure 4–13/Q.1214 ITU-T publication, ‘Distributed Functional Plane for Intelligent Network CS-1’
[10/95].
Figure 5.4 Multi-party calls and associated call segments
Reproduced, with permission, from Figure 4–14/Q.1214 ITU-T publication, ‘Distributed Functional Plane for Intelligent Network CS-1’
Call party handling

[10/95].
111
112 Intelligent networks – principles and applications

Figure 5.5 The connection view model

Reproduced, with permission, from Figure 4–18/Q.1224 ITU-T publication, ‘Distributed Functional
Plane for Intelligent Network CS-1’ [09/97].

Figure 5.6 Simple connection view states

 Call party handling 113

Figure 5.6 shows four CVSs, each with a single call segment, identified by the
number in the central circle, which represents the connection point. Call segments
are conventionally drawn as boxes with rounded corners. Legs are depicted by solid
or broken lines, and there is always just one controlling leg (labelled ‘c’ in the CVS
and always on the left hand side of the connection point) and relates to the
originating or terminating access. On the right hand side of the connection point is
shown the passive leg (or legs in the more complex call segments shown later).
These are labelled ‘p1’, ‘p2’, etc.
A leg is shown as either solid or broken. We have also shown arrowheads on
some of the lines to indicate call set-up direction. A solid line indicates ‘joined’
status, which means that the user attached to that leg is in communication with
other users through the segment. A broken line indicates that the leg status is
‘pending’, as is the case with the ‘Orig. set-up’ and ‘Term set-up’ CVSs in Figure
5.6, or is ‘shared’ or ‘surrogate’. ‘Shared’ means that a call segment is part of a
CSA, and the user is associated with the controlling leg of the associated segment.
An example is in the ‘M-party set-up’ CVS in Figure 5.7, where the controlling leg
is shared between the two call segments. A controlling leg can be a ‘surrogate’
when it represents a user who is not actively participating in a call, as in the case of
the ‘transfer’ CVS of Figure 5.7, where the controlling user has actually left the call
to the two passive legs.
Having introduced the CVS concepts for the simple 2-party transitions of Figure
5.6, we now illustrate the power of the IN CS-2 CPH features by showing, in Figure
5.7, some of the more complex CVS configurations.
Figure 5.7 shows most of the CS-2 call configurations that are possible. They are
labelled CC0 to CC12 (with some gaps). The diagram also shows several possible
transitions, representing potential IN service features for CS-2. For instance, it
shows how the ‘stable 2-party’ CVS of Figure 5.6 can progress to a multi-party call
via the ‘M-party set-up’ and ‘call on hold’ CVSs. These last two CVS
configurations each consist of two call segments and they are therefore CSAs
(call segment associations).
There is obviously a finite number of ways in which call parties can be connected
to each other, from the viewpoint of the SCF. IN CS-2, Reference 32 provides a
catalogue of 14 example CVSs, and each one is unambiguously described in terms
of its entry and exit events, and its relationship to the BCSM. Rules are specified,
using object-oriented techniques, in order to describe the allowable transitions
between the different connection view states. The resulting set of rules comprises
the IN SSM – introduced earlier in this chapter (illustrated in Figure 5.2).
The documented [32] CVS examples are listed, with brief summary explana-
tions, in Table 5.1.

5.2.4 Information flows supporting CPH

We now list, and briefly describe, the protocol information flows that are defined
for the SSF–SCF interface to enable the call party handling functions, which we
114 Intelligent networks – principles and applications

More complex CVSs

Figure 5.7
 Call party handling 115

Table 5.1 The connection view states

CVS Explanation

Null Call processing is inactive and no legs are connected to the

connection point.
Originating set-up A two-party call is represented in the set-up phase.
Stable two-party A stable two-party call – can be originating or terminating
from the controlling user’s perspective.
Terminating set-up A terminating two-party call in the set-up phase.
M-party set-up The controlling party has put one (or more) party on hold
while originating a new call to another party. Two (or more)
associated call segments are involved.
Call on hold The call has two associated segments. The controlling user
has one party on hold whilst participating in a stable two-party
call with a third party.
Call waiting A stable two-party call is in progress and a new call attempt is
targeted at the controlling user.
Stable M-party A multi-party call in the active (speech) phase.
Transfer The controlling leg has now disconnected, and has become a
‘surrogate’, whilst a stable call exists between the two passive
legs.
Forward The call is forwarded by the controlling party, who becomes
the surrogate leg, and set-up is being established between the
passive legs.
Originating set-up M-party The controlling user is connected to an SRF, which has
instigated a new call, which is in the set-up phase.
Active M-party The call being set up in the ‘originating set-up M-party’ CVS
has now reached the stable state.
One-party set-up A single party call is originated by the network and is in the
set-up phase.
Stable one-party The ‘one-party set-up’ has now moved to the stable phase.

have been describing, to be brought into play. Some of these operations are
illustrated against the example CPH transitions that were shown in Figure 5.7
earlier.
The comprehensive list of the extra INAP operations for CS-2, and details on
data values, are in Reference 32. The operations here are ordered alphabetically:
CreateCallSegmentAssociation – creates a new CSA. The ‘Move Call
Segment’ operation (described below) can then be used to populate the new CSA.
DisconnectLeg – enables the SCF to release a specified leg from a call
connection, leaving the remaining legs, and any connection between them, intact.
HoldCallInNetwork – allows a call to be queued in the set-up phase (e.g. to
provide a ‘call-completion to busy’ service the call would be queued until the
destination became free). The SCF can, in an associated parameter, provide
information on what announcement or tone is to be played to the caller in the
meantime.
InitiateCallAttempt – was specified earlier in CS-1, and it was intended for
services such as network-originated alarm calls. At first sight its potential might
116 Intelligent networks – principles and applications

have been be overestimated. For instance in CSI it was not normally used to
originate a ‘call’ between two external parties. However, its flexibility improved
with CS-2, where it became capable of being used to connect two external
destinations together in a call.
MergeCallSegments – will merge two call segments which are ‘associated’ (i.e.
they have a single controlling leg) into a single call segment. It can also be used to
re-establish the bearer connection between two legs within one single 2-party call
segment. We can therefore initiate a call (ICA) to an operator, whilst a caller is
listening to a voice announcement, in preparation for through-connection. This
could also be used for a ‘pop-up’ operator facility.
MoveCallSegments – moves a call segment from the source CSA to a target
association. The controlling leg party must be the same, so it can have the effect of
merging two associated call segments into a single call segment. However, this
only applies during a terminating BCSM for a 2-party call segment, where the call
processing has been suspended at T-busy T_DP. For this reason, a natural service
application would be for an incoming call to a busy destination to be joined into the
connection, after the call had been offered to (and accepted by) the target (busy)
party.
MoveLeg – issued by the SCF to move a leg from one CS to another with which
it is associated.
Reconnect – is used to re-establish communication between the controlling leg
and the (held) passive leg(s) of a call with two or more parties, when the controlling
leg has disconnected.
ReleaseCall – can be used for forward clearing of a B-party (after encountering a
‘Disconnect’ detection point, for instance).
SplitLeg – used to separate one call party from its CS. In the case of a multi-
party CS, SplitLeg will place it in a new associated CS. It interrupts the speech
connection between the leg to be split and the legs remaining in the original CS.
This information flow is the reverse of MoveCallSegments (above).

5.3 Call party handling – A renewal of IN?

In many ways CPH in ITU-T IN CS-2 restores some of the ‘intelligence’ which was
available in the network 50 years ago when the ‘intelligent’ component of networks
was the large inter-connected web of manual-board telephone operators (see Figure
5.8). This network represented a distributed intelligence system that communicated
on a separate plane from the bearer connections. The bearer connections were only
switched through when the necessary transactions had been carried out to ensure
that an end-to-end connection was available.
Service logic and data were available wherever it was needed for a wide range of
services. Switchboard operators in those days could manipulate the legs involved in
a connection. They could provide diversion services, they could put subscribers on
hold whilst they offered the call to other subscribers, they could connect several
subscribers together at once, and they could split a multi-party call into individual
 Call party handling 117

Figure 5.8 An early implementation of call party handling

calls. The versatility of the ‘sleeve control’ [2] manual board system allowed the
operator to provide many IN services, such as connecting several parties together at
once or transferring a call from one party to another. The operator could then drop
out of the call speech path altogether and return to monitor and supervise the call at
any time. In fact they could do all the operations such as split, join, delete and
separate that were subsequently specified in the IN CS-2 recommendations, in
addition to providing most of the regular CS-1 services such as diversion on no
reply and time-of-day routing to Freephone numbers.
 Chapter 6
Distributed intelligence

For the strength of the Pack is the Wolf, and the strength of the Wolf is the Pack.
Rudyard Kipling (1865–1936)

Up to this point in the book the controlling service logic has been presented as
being located (at least logically) at a single point and this has been represented by
the IN SCP. The concept of remote data has been introduced with the IN SDP, but it
has generally been assumed that the controlling logic that runs the service is hosted
by the network operator on a single computing facility. Thus we have service-
switching points, and service control points, such that an SSP might communicate
with a ‘home’, or pre-assigned, SCP for its service requests. For many services,
such as those described earlier, this model has proved to be sufficient. Early
intelligent networking was to do with a set of discrete database nodes serving large
numbers of telephone exchanges, with a reliable and intelligent signalling system in
between. This ensured that the required message reached the intended destination
database processor, and that the reply would be correctly returned to the originating
network node.
We now look at architectures where the intelligence is no longer necessarily
fixed in a single location but is allowed to be distributed across a network. The use
of flexible distribution of computing functions across processor clusters, and the
consequent dynamic location of service logic within a single service control point is
not new. However, extending this distribution of remote intelligence across a
network comprising different vendors’ equipment is still, in practice, relatively new
ground for IN implementations, although theoretical modelling of the associated
issues has been wide-ranging.
Broadly, there are two categories of argument in favour of distributing the
intelligence in a network. Firstly, there are the practical operational reasons for
distribution that are to do with improving network efficiency, and secondly there
are the strategic drivers for open interfaces between network domains and service
provider domains. We now discuss these categories separately.
120 Intelligent networks – principles and applications

6.1 Distribution of intelligence for network efficiency

Initially IN database processors tended to be large. They were points of focus for
heavy real-time traffic demand. As we discussed in Chapter 2, the database
computers themselves therefore needed to be highly reliable and fault-tolerant. This
was because the IN translation data held there was critical for successful
completion of a large proportion of the calls handled by the entire network,
representing large amounts of revenue. Because of this criticality, local node
resilience mechanisms were usually also provided to ensure that back-up
processing and alternative call handling measures were available as expedients in
case of network or equipment malfunction.
Because of the high cost of these early large fault-tolerant and highly secure
processors, an alternative architectural approach has now become popular. This
involves distributing the service control functions across larger numbers of general
purpose, cheaper (but less resilient) servers. The required reliability levels are
achieved by using redundancy-sparing techniques. This allows significant cost
savings to be made because a cluster of general purpose Unix workstation-type
platforms can be provided more cheaply than the earlier high-grade transaction
processors. Similarly, at the network level, greater versatility becomes possible by
distributing the service control function across a larger number of cheaper control
points. This means that, as well as being able to ‘freeze out’ malfunctioning
processors, or processor nodes, the network management systems are able to move
traffic automatically from one part of a network to another as traffic patterns change
through the day. At times of high load at one part of the network, IN service
requests can be routed to alternative host computers by altering the SCCP global
title translation data tables (explained in Chapter 3). Similarly if one processor
cluster requires to be taken off-line for maintenance work, its load can be
reallocated to peer processing nodes beforehand.
It has to be remembered, however, that a disadvantage of using distributed
intelligence for increased flexibility and hardware cost-savings in this manner is
that the supporting data management update mechanisms (as well as the SCCP
global title routing management systems) obviously become more complex. This is
because up-to-date data for IN translations must be made available wherever it is
needed by the service control logic. Up-to-date versions of frequently used data
may need to be held locally at all SCP sites, but it is obviously not efficient to
distribute local copies of rarely used data to every control point in a large network
just in case it is needed. In these cases it is more efficient for the requesting service
logic to obtain the data from the source database when it is needed.
For example, a universal mobility service requires personal data profiles to be
provided for each user. These customer profiles hold details on current location
addresses, time-of-day routing plans, screening lists, etc. When a call attempt is
made to one of these customers, the caller’s local SSP will normally originate an IN
trigger to a controlling SCP, which is unlikely to also be the host for the called
customer’s personal profile.
 Distributed intelligence 121

Therefore, either all customer profiles need to be distributed and updated at all
points which could receive IN requests for this service, or a data request facility16 is
implemented so that local SCFs can query remote SDFs where the required data is
hosted. In a large majority of cases the latter option is likely to be the most sensible,
although some local data caching for heavily used destinations may be employed.
Another aspect of distributed intelligence for mobility services is illustrated in
the CAMEL call flow examples of Chapter 7. In this case the controlling SSP
makes an IN request to the home SCP for a roaming customer in order to access the
required data.

6.2 Distribution of intelligence across open network interfaces

Since the late 1990s there have been strong commercial and legislative pressures
for opening up public network infrastructures for external service providers. The
reasoning for this is that the natural consequence would be an expansion of the
telecommunications industry. Potentially lucrative markets have been identified for
service provision companies (sometimes called ‘soft-telcos’), whose main business
is the provision of new services for their clients through existing public networks.
These independent service providers do not invest in their own information
transport network infrastructures, but instead they negotiate resource usage with
established network providers.
The advantage to the network operator of these open network interfaces is that
the stimulation of service usage results in increased network usage and higher
revenues for the transport of the associated network traffic. The network operators
are able then to offer a service provision interface to external wholesale customers.
This motivation has inspired strong interest in the availability of a standard
interface between the computing and telecommunications worlds. This standard
interface is an example of an application programming interface (API). For
instance, providing customers with direct access to the network to obtain billing
information has potential for reducing the large overhead imposed by internal
billing systems. This leads to a desire for operating companies to evolve their
network structures to be able to interwork with external computer intelligence.
At the same time, the use of object-oriented programming and distributed
processing environments (DPEs) has generated great interest, and this has been a
prime focus of the TINA (telecommunications information network architecture)
project. Since the definitions of the IN SSM (switching state model) in CS-1 and
CS-2 the ITU-T IN standards have tended to define ‘objects’ such as legs and
connections, but essentially IN architectures have been described in a functional

16
This aspect of intelligence distribution is known as ‘location transparency’, because the geographical
location of the target customer profile is irrelevant to the requesting service logic. A roaming user’s
personal profile could be anywhere in the world, so data access procedures must cope with this. The
middleware techniques described later in this chapter are designed to provide solutions for this type of
scenario.
122 Intelligent networks – principles and applications

manner rather than an object-oriented one. Initiatives such as TINA and Parlay are
concerned with architectural structures that incorporate the notion of decoupling
the network resources that are associated with the transport of information (speech
calls or data) from the software that controls the logic of the services themselves.

6.3 A model for distributed IN

This section introduces some general concepts for modelling distributed

intelligence as a preliminary to describing the relevant industrial standards
initiatives, such as the Parlay API.
Figure 6.1 shows a model for the ‘decoupling’ of the service provision resources
from the underlying network resources that control the transport of information
(speech calls or multimedia data). The user terminals, which could be fixed or
mobile phones or multi-media equipment, are at the bottom of the picture and the
services are at the top.
Two different APIs,17 the network API and the resource API, are illustrated here.
The network API is the set of functional interfaces presented to users, third party
service providers, applications and their developers and the Resource API is the
generic interface to the transport networks.
The essence of distributed IN is the separation of these two domains by the
‘middleware’, which is a distributed computing environment responsible for
providing transport layer resources with access to the required application services.
The middleware is a generic distribution facility, providing its users with
connections to a resource of the specified type irrespective of the physical details
of location or routing. The middleware contains some background components (not
shown in Figure 6.1) to carry out the coupling between the transport network user
and the required application service. These components include brokering, generic
call control and service logic execution environment (SLEE) facilities. Middleware
provides the underlying seamless connectivity between the network API and the
various resources and transport networks. It ensures that real-time network
complexities, and others such as security, load balancing, guaranteed message
delivery, etc., are generally taken care of without involving the network API users.
Figure 6.1 shows the middleware interfacing to ‘users’ (such as application
services or transport networks) through dedicated ‘plug-INs’. A ‘plug-IN’ (an
intelligent pun) is a modelling representation of an adaptation function that can be
thought of as containing the service control, or transport, intelligence particular to

17
API is a term that has transferred to the telecoms world from the computing industry. It describes a
concisely defined set of computing methods and rules which allows a computing application to use
external resources, or the facilities of other applications. For instance, a mail-handling program on a PC
has an API to talk to the PC’s modem. It provides the elementary function calls and programming rules.
In the broader sense an API specification consists of a set of interface definitions and this, in principle,
allows application writers and network operators a free choice of implementation technologies.
 Distributed intelligence 123

Figure 6.1 A model for distributed IN

the target service or transport network. The plug-IN18 contains the functions that
form the bridge between the service-specific, or transport-specific, software and the
generic procedures, so the middleware has no notion of the transport technology
details. It is equivalent to a ‘device driver’, where the devices can be signalling
interfaces such as H.323, IETF SIP, INAP, CAP, ISUP or ATM Forum UNI 3.1 for
accessing the various transport network. For connection to the various application
clients and servers across the Network API, plug-INs will ensure the mapping to
DPE technologies such as CORBA, DCOM or Java.
Plug-INs can act as clients or servers. Figure 6.2 shows the DPE middleware
transporting a request from a client plug-IN to a server plug-IN.
As an example, the network of Figure 6.1 could be a PSTN, in which case the
plug-IN that straddles the resource API would be a PSTN-specific plug-IN. This
plug-IN then interfaces between the generic transport layer, represented to the
middleware as the common resource API, and the PSTN-specific API. The PSTN-
specific API could for instance be based on CS-1 INAP. If an IN request (an I_DP)
arrives at the PSTN-specific plug-IN then the plug-IN communicates with the
middleware brokering function to arrange a connection with an appropriate service

18
The term ‘plug-IN’ should not be confused with the familiar computing term ‘plug-in’, which refers to
software components used to extend the features of an Internet browser by providing bypass solutions for
new browser functions.
124 Intelligent networks – principles and applications

Figure 6.2 Plug-INs communicating via DPE middleware

plug-IN. A DPE method call (shown in Figure 6.2) is made to the selected server,
which the middleware has chosen to host this service invocation.
The model of Figure 6.1 demonstrates the main characteristics that form the basis
of standards such as Parlay for the use of APIs and middleware in
telecommunications. This type of model is a consequence of the convergence of
the computing and telecommunications industries, where we have seen the
emergence of strong drivers to present simple functional interfaces to telecom
network application developers and the run-time applications. These interfaces
should be independent of hardware platform, operating system or choice of
programming language.
Figure 6.3 shows the distributed intelligence model of Figure 6.1 extended to
include plug-INs for particular transport network types. Because of the common
resource API, which accesses common middleware features and services, the plug-
INs for both network types are able to access the same services. Obviously this
depends on the network capabilities – a video conferencing service would be
available to the broadband multimedia network, but would only be available from
the PSTN plug-IN for an ISDN call with suitable terminal equipment connected.
Similarly, a voice call diversion service would be available to the PSTN plug-IN as
well as to voice calls served by the broadband network.
The broadband network itself could be an IP (Internet protocol) network
providing multi-media services including voice, or it might be an ATM
(asynchronous transfer mode) network providing ITU-T B-ISDN services.
For the users of the network API, the middleware creates the illusion of the
network as a single large, efficient, ‘virtual computer’, seeming to unify the
different transport networks and their corresponding lower level network services.
An important element, already mentioned, is the ‘broker’, which finds target
server objects, of the required type, on request from client objects. Another
important entity is a ‘naming’ service, which assists the broker in finding the
correct object. For this purpose, the middleware could typically be based or
 Distributed intelligence 125

Figure 6.3 Distributed intelligence and multiple transport types

modelled on CORBA, although there have been several other candidate

technologies for middleware, such as DCOM (Microsoft), RMI (Java) and
DTPM (distributed transaction processing middleware).
The broker function is illustrated in Figure 6.4. As can be seen, each of the plug-
INs has a ‘broker’ interface as well as one or more ‘transaction’ interfaces. The
client plug-IN initially uses the broker interface in order to obtain a reference to an
appropriate server plug-IN. When the reference has been returned, the plug-INs
communicate using the transaction interface.
The notion of the distribution network that has been illustrated in diagrams in
this chapter is similar to the model that is commonly used for networking clients
and servers in a business environment. The concept of ‘network centric computing’
allows specialist client terminals to access different services on network servers
without needing to be aware of the physical location of the servers. A relevant
development in this context is the use of the Java programming language in
network computing. Java programs can be quickly downloaded from the servers
and run on any client running the Java virtual machine. This type of solution hides
the low-level complexities of different computing environments, providing the
‘software glue’ to allow communication between applications hosted on different
equipment.
126 Intelligent networks – principles and applications

Figure 6.4 The broker function

6.3.1 Infrastructure APIs

Many APIs are used, at varying detailed levels, in a computing system. Many will
be internal and unavailable to third party application developers. An example of a
higher level ‘infrastructure’ API that is likely to be internal to a network operating
company’s domain is that serving a company’s operational support system (OSS),
shown in Figure 6.5. Amongst other functions this interface allows call charging to
take place.

6.4 Object request brokers for network intelligence

The object management architecture (OMA), which was defined by the OMG (the
object management group) is based on the object request broker (ORB) component.
This allows other elements in the architecture to exchange request and response
messages. The principle of the ORB is that it represents (‘encapsulates’) the entire
underlying operating system and network. This is the origin of CORBA (common
object request broker architecture).
Most current longer-term strategies for intelligent network architectures
incorporate, as a cornerstone, a distributed processing model based on CORBA.
The current practical issues with CORBA are scalability and performance. Whilst
the OMG has begun work on a real-time version of CORBA, to date a large-scale
real-time implementation of CORBA for a telecommunications network has not
been built. In fact the first applications of CORBA in telecommunications networks
are for off-line, or ‘near real-time’, management support functions required by the
real-time systems.
 Distributed intelligence 127

Figure 6.5 Infrastructure APIs

6.5 The Parlay API

‘‘Parlay, to bet (original stake plus winnings) in a later venture: to succeed in

converting an asset into something more valuable.’’
Chambers Twentieth Century Dictionary.

6.5.1 Background

A leading standard for the network API that has been mentioned more than once so
far in this chapter is the Parlay Consortium API. Parlay achieved a high degree of
publicity after a revolutionary announcement in 1998 from BT, which was part of
the Parlay Group. In this announcement BT indicated its intention to open up its
network to third party application developers, using the new Parlay standard. This
commitment was made at a time when network growth was seen as critical to the
major operators’ strategies because competition for network provision was
becoming intense.
For a long time there had long been demands from independent enterprise
organisations for the ability to offer their own services to their customers in this
manner. However, the prospect of laying bare the inner workings of these networks
128 Intelligent networks – principles and applications

was viewed with trepidation by established operators, who were obviously

concerned about integrity breaches. The BT announcement that the network would
be opened for outsiders, including commercial rivals, was therefore a bold step, and
obviously was not taken before adequate provision had been made for the security
and control management issues to be included.
As discussed previously, from a network operator’s perspective the commercial
driver was that a common standard for third party service provision would
stimulate market expansion in telecommunications services, thereby leading to
greater revenues from network usage. An easy-to-use, freely available API standard
would encourage growth in enterprise start-up companies to the benefit of the
telecommunications industry as a whole. Independent specialist service-providers
would be able to make their new service offerings available to their own customers
using the network operator’s underlying voice and data networks and support
systems.
The details of the API were published on the Internet (http:/ /www.parlay.org),
and so were available for anyone to write a service application that would interwork
with a public network. The Parlay Group originally comprised British
Telecommunications, Microsoft, Nortel Networks, Siemens and Ulticom (formerly
known as DGM&S Telecom). These were joined in 1999 by AT&T, Cisco
Systems, Ericsson, IBM, Lucent Technologies and Cegetel.

6.5.2 What is the Parlay API?

The Parlay API effectively presents the underlying networks as a programmable

entity for independent service developers. The particular network technology is
hidden from the developer. Parlay was originally based on object technology
interfaces from the OMG and so is independent of lower-level messaging
technology. For example it is possible to use CORBA, DCOM, JAVA/RMI or any
other suitable technology for the distribution middleware.
Parlay comprises a set of APIs that is open, extensible and network-independent.
These are intended for use by developers of enterprise-based client applications
who wish to write software for services that can use the facilities of participating
telecommunications networks. The APIs are object-oriented, and are organised
around the functional structure shown in Figure 6.6, which shows the main
categories of the Parlay interfaces. Parlay ‘service interfaces’ provide client
applications with access to network capabilities and Parlay ‘framework interfaces’
provide the backup capabilities such as authentication, load management and fault
handling.
The interface classes are specified using UML (unified modelling language), so
the Parlay specification is documented in terms of framework interfaces, individual
service interfaces, data definitions, class diagrams, sequence diagrams and IDL
files. The interfaces are defined in program statements using the OMG IDL
(interface definition language). These program statements can be generated
automatically in output files from the UML tool.
 Distributed intelligence 129

Figure 6.6 The main Parlay interfaces

Initially, the prime focus of attention for the Parlay interface definition has been
the framework and the services interfaces. These are marked as (1) and (2) in
Figure 6.6. In addition to these, work has proceeded on the definition of other
public interfaces. These include interfaces introduced to support administrative
functions within an enterprise and interfaces (between Parlay framework and Parlay
services in the diagram) that allow third party vendors to provide Parlay services.
Discussion here is confined to the interfaces shown in Figure 6.6.

^ The framework interfaces include the underlying capabilities needed to

provide secure and resilient controlled access to external users. They include
the management, authentication and billing capabilities needed as well as the
generic features for call and connection services, messaging and user
interaction.
^ The service interfaces more specifically detail the functions such as call
routing which are needed to build up the services themselves, including access
to network capabilities and information.

6.5.3 Parlay specifications

The Parlay specification that is in force at the time of writing is Parlay 2.1, although
draft specifications for Parlay 3.0 are publicly available.
Parlay 2.1 API specifications are available (from the Parlay Group Web
homepage, http:/ /www.parlay.org) in the following document groupings:

^ call processing APIs

130 Intelligent networks – principles and applications

^ connection manager APIs

^ framework APIs
^ messaging APIs
^ mobility APIs
^ common data & IDL.

The APIs that have a bearing on the IN principles discussed in this book are
contained in the ‘call processing APIs’ document set, which groups the call control
service interfaces into the following categories:

^ generic service
^ generic call control service
^ multi-party call control service
^ multi-media call control service
^ conference call control service.

Each interface group is described in terms of its interfaces and methods. For
example, the Generic Call Control Service [36] has four defined interfaces:

^ the call control manager interface (has six methods, such as ‘createCall()’ and
‘setCallLoadControl()’)
^ the call control manager application interface (also has six methods, including
‘callAborted()’ and ‘callOverloadEncountered()’)
^ the call interface (has eight methods – examples are ‘release()’ and
‘getMoreDigitsReq()’)
^ the call application interface (has ten methods and is used for call request
responses).

The Generic Call Control Service allows applications to carry out call control
functions in communications networks, including management of IN services. It is
also expected to be adaptable to any required call control technology, including
ITU-T H.323, ISUP, Q.931, Q.2931, ATM Forum UNI 3.1 and IETF SIP. The
objects employed by the Parlay 2.1 call control interfaces are:

^ call (a relationship between a number of parties)

^ call leg (an association between a call and an address)
^ address (a logical representation of a call party)
^ terminal (an end-point).

These objects will be familiar to the reader from the discussions on ITU-T IN CS-2
in Chapter 5. In fact the call model described for Parlay uses similar notation and
manipulation mechanisms as the ITU-T CS-2 call model. The Parlay call party
handling procedures are extended to multi-media calls with the Multi Media Call
Control Service interface set.
 Distributed intelligence 131

Parlay 2.1 was updated in conjunction with ETSI [37] and JAIN2 Community
member companies to form the Parlay 3.0 specifications. These are also now part of
the Open Service Access (OSA) standards, which was adopted by the 3GPP (3rd
Generation Partnership Project) group for UMTS for 3G mobile [38].

6.6 TINA

TINA was a relatively long-term, large-scale collaborative project that set out to
define a software architecture for future service provision across all types of
telecommunications networks. It had its original roots in the US, with INA
(information networking architecture), which Bellcore (now Telcordia) instigated
in 1990. INA was formulated as a work-program to surpass the IN standards of the
day, and it was to incorporate newly emerging distributed processing and object-
oriented computing principles.
An important limitation of INA was that it sought to integrate IN with the new
regime. Unfortunately this ambitious aim turned out to be impossible to sustain,
and INA was overtaken by TINA. TINA, on the other hand, was unhindered by any
obligation to be backward compatible with legacy networks, or even intelligent
networks. It was driven by the convergence between telecommunications and
computing and the aim was to produce an over-arching architecture that would
embrace all telecommunications applications.
TINA is rooted in the principles of object-oriented design, autonomous software
components and DPE using technologies that had been defined by the OMG for the
computing industry. CORBA was a favoured technology for the DPE ‘broking’
function for TINA. Output from TINA was in turn taken up by the OMG’s own
telecommunication domain task force.
TINA-C (the TINA Consortium) was formed in 1993 and was disbanded at the
end of 2000. It comprised 40 network operators and equipment vendors with a core
team working in New Jersey, USA. The aim of the consortium was to produce a
document set for the TINA architecture and to work towards easing the path for the
uptake of TINA-based products. The TINA-C approach represented a radical
review of existing ideas about the evolution of the telecommunications industry. It
derived from a general concern that traditional strategic directions were limited and
that wider strategic vision would be needed for the new services to run on
converged data and voice networks. TINA-C set out to re-use existing standards
and other publicly available results, and to produce specifications that could in turn
be standardised by the recognised standards bodies.

6.6.1 TINA roles and reference points

A logical starting point for understanding the TINA approach is the specification
for the TINA business model [39], which is described in terms of actual
132 Intelligent networks – principles and applications

commercial interactions in a business environment. It describes the separate roles

involved. Typical roles described are:

^ consumer – represents the end-user of the service.

^ retailer – concerned with service presentation and interfacing to customers.
^ broker – enables one ‘stakeholder’19 to locate another stakeholder of suitable
type. It therefore requires management of an addressing facility.
^ third-party service provider – concerned with production and maintenance of
the service elements. If a video shop is an example of the retailer, then the third
party service provider (wholesaler) supplies the retailer with his new video-
tapes (the service ‘content’), and is concerned with issues such as production
and supply–maintenance aspects rather than the end-customers.
^ connectivity provider – for setting up and controlling communication paths for
content transmission wherever they are needed.

A set of reference points (RPs) is defined in [39], against which TINA conformance
can be measured. The roles and reference points are illustrated in Figure 6.7.
As an example, a company in the communications industry will typically be
represented by several TINA roles. A video-on-demand retailer, for instance, could
also be a third party service provider. TINA therefore defines business
administrative domains. The use of the standard RPs within a company’s sphere
of business is not important, but the standard RPs must be used between
communicating domains, as shown in Figure 6.8.
Generally, reference point descriptions are broken down into standard
component parts. The ‘business part’ is to do with the overall scope of the
interface, its business purpose and its limitations. The ‘information part’ describes
shared information, such as naming and addressing conventions. The ‘computa-
tional part’ defines the computing object interfaces and the ‘engineering part’ is
concerned with the structure of the DPE nodes, their separation and communication
details.
RP specifications operate at several levels, with the interactions broadly grouped
into two divisions, as indicated in Figure 6.9. The application-related RP part is
concerned with the business, informational and computing viewpoints, whereas the
DPE-related part is for the underlying engineering and technology interactions.

6.6.2 The TINA architecture

The TINA architecture is made up of three separate sub-architectures. These are:

^ the service architecture

^ the network architecture
^ the computing architecture.

19
A stakeholder is an agent that adopts one or more of the TINA roles.
 Distributed intelligence 133

Figure 6.7 TINA roles and reference points

The reference points illustrated in the diagram are:

ret retailer business relationship

RtR retailer to retailer business relationship
Bkr broker business relationship
Tcon terminal connection business relationship
ConS connectivity service business relationship
3pty third party business relationship
LNFed layer network federation business relationship
CSLN client server layer network relationship.

Following on from the TINA business model described earlier, an objective of the
TINA service architecture is to enable the different administrative domains to
interact in pursuance of their commercial aims.
The service architecture definition describes how services are designed and used
in a TINA system. It defines the software components needed, and explains how the
software should be structured in order for it to plug into the overall TINA service
management environment. It provides developers with the necessary programming
instructions to enable an operational TINA service to be controlled and managed by
the TINA environment. The service architecture introduces the concept of
134 Intelligent networks – principles and applications

Figure 6.8 Interaction between business administration domains

Figure 6.9 Grouping of RP interaction types

 Distributed intelligence 135

‘sessions’, which comprise the related activities in a service invocation. There are
four types of sessions:

^ service session (single activation on a TINA service)

^ user session (a user’s relationship with a service session)
^ communications session (the assembly of connections required for the service
session)
^ access session (a user’s connection to a service session).

As an introduction to TINA service architecture concepts, the ‘functional entities’

that comprise the TINA model are shown in Figure 6.10 grouped into three
domains – those of the ‘user’, the ‘service provider’ and the ‘network
provider’.
Associated with the user domain are ‘terminal agents’, through which a user
accesses services. A terminal agent will exist for every terminal connected to the
network, and will be responsible for holding information about the terminal’s
characteristics such as processing capacity, memory and communications
capabilities. This information will be available to the service provider domain.
Associated with the service provider domain are the TINA user agents, service
sessions and user sessions. The user agent represents the user, holding information

Figure 6.10 TINA agents and sessions

136 Intelligent networks – principles and applications

such as preferences, subscription records, usage information, password and

location, thereby controlling their access to services. There would be a default
user agent for unknown users.
When a user invokes a service, a service session is started and adapted according
to information held in the terminal and user agents. A user session is also started for
each user involved in the session. Amongst the data collected and recorded in a user
session is that for subsequent billing. When the user leaves the service session, this
billing information would be passed to their user agent for aggregation with other,
similar, data.
One of the basic TINA separations in the service architecture is that between
‘access’ and ‘usage’, where the former encompasses the interactions for
discovering and requesting services and the latter includes the interactions for
controlling service behaviour or delivering service content. Access interactions
(covering authentication, usage selection, etc.) are needed before the communicat-
ing domains can establish the required usage relationships. Usage relationships are
for controlling service behaviour as well as maintaining stream connections
amongst the service components in order to deliver content information. Service
control is to do with establishing the required quality of service, multi-media
stream binding, connection control and related topics.
Associated with the network domain are ‘communication session managers’,
‘connection co-ordinators’ and ‘connection performers’. A communication session
manager is invoked by a service session from the service provider domain. With the
invocation, the service session also sends terminal capability information, and user
requirements for the session. For example, the terminal may be restricted to 64 kbit/s
voice transmissions and the user may require a specific quality of service. The
communication session maintains the network resources, or modifies them under
the direction of the service session (if, for example, an additional user joins the
session). The connection co-ordinator arranges for the connection performers to
effect the necessary connections.
The network architecture is a technology-independent model and it operates at
three layers. These are the communication session layer, the connectivity session
layer and the network layer.
The computing architecture deals with the DPE, which makes the underlying
distributed computing equipment behave as if there were a single flexible system
across which the applications reside and communicate. The DPE is based on
CORBA. The computing sub-architecture also provides the various models, which
enable the complexity of the software design to be broken up and viewed from
different standpoints.

6.6.3 TINA specifications

The final set of TINA specifications and supporting documents is publicly available
at the TINA-C Web server (www.tinac.com). The document set includes the
following specifications, as well as some supporting texts.
 Distributed intelligence 137

TINA service architecture and specifications

TINA service architecture Version: 5.0 TINA-C June 1997
deliverable
TINA retailer reference point Version: 1.1 April 1999
TINA service component specification – Part B (computational model and dynamics)
Version: 1.0 Jan. 1998

TINA network resource architecture and specifications

TINA network resource architecture
Version: 3.0 Feb. 1997
TINA network resource information model
Version 2.2 Nov. 1997
draft baseline
TINA connectivity service reference point
Version 1.0 Feb. 1997
TINA terminal connection reference point
Version: 1.0 Nov. 1996
TINA network component specifications
Version: 5.0 June 1997

TINA overall model

TINA business model and reference points
Version: 4.0 May 1997

6.6.4 Relationship between IN and TINA

IN has been essentially limited to voice telephony, whereas TINA’s ambitions were
always wider, being concerned with broadband multi-media, multi-party and
information services. TINA is not particularly concerned with transmission details,
and is certainly not constrained to services that are restricted to circuit connections.
TINA is just as concerned with information and multi-media services using
connectionless data streams.
Whereas IN service embellishments start from the ‘here and now’ of telephone
calls, connected via existing telephone exchanges, TINA operates at a higher plane,
considering the new services as software-based applications operating on a
distributed computing platform. This approach is more concerned with the use of
portable code in achieving the required session interactions for the advanced
services than with the access practicalities and network technologies, or even the
distribution details.
Before the current BICC (bearer-independent call control) standardisation
initiative in ITU-T, TINA was recommending a functional split between bearer
control and call control, whereas traditional IN places the main functional split
between call control and service control (i.e. it preserves call control into the
network domain). These differences are illustrated in Figure 6.11.
TINA itself is principally concerned with the service layer, and it defines an
object framework for distributed-object implementations of telecommunications
138 Intelligent networks – principles and applications

Figure 6.11 IN and TINA separations

services of all varieties. Early on it was decided that a sensible migration approach
between the different worlds of IN and TINA would be to introduce ‘islands’ of
TINA within existing IN architectures.
 Chapter 7
Service examples

The concept is interesting and well-formed, but in order to earn better than a ‘C,’
the idea must be feasible.
—A Yale University management professor in response to Fred Smith’s paper
proposing a reliable overnight delivery service. (Smith went on to found Federal
Express Corporation)

This chapter explores some commonly used examples of the application of IN

principles to telecommunications services. As we have indicated previously, IN
was primarily intended for voice services, so most of the examples described here
are therefore, unsurprisingly, voice features. Many of the examples are variants of
the number translation capabilities of IN, but IN number translation has many
interesting and revenue-producing guises, as we shall demonstrate in the next few
pages.

7.1 Simple number translations – Freephone service

The simplest use for an intelligent network is to provide a straightforward number

translation service. This really was the way IN started, as we have discussed in
earlier chapters. With conventional telephony networking the service intelligence is
distributed across the exchanges and the exchange processors pass service
information to each other using the common control signalling system. Service
and customer data are therefore distributed across the telephony network and
conventional (proprietary) exchange management systems are used for data
maintenance. The key advantage of using IN technology, however, is that the
service logic and customer data is logically held in one place. The management of
this data is therefore decoupled from traditional telephony data input systems, and
state-of-the-art computing technology can be used for efficient control and fast
introduction of new services. Number translation services, where strings of dialled
140 Intelligent networks – principles and applications

digits are transformed into network routing numbers, are therefore ideal
applications for IN.
The success of the Freephone service, which combines these straightforward
number translations with the ability for a caller to make a ‘free’ call (at the
recipient’s expense), has been world-wide. Freephone dialling codes (e.g. ‘180x’ in
the US or ‘080x’ in the UK) bear no relation to the geographical location of the
recipient but are convenient to use. In fact, countries that had the foresight to keep
letters as well as numbers on their telephone instruments found this factor to be a
powerful catalyst in the growth of their Freephone services. This encouraged a
source of inventive ingenuity in the development of the services (easy-to-remember
‘numbers’, such as 180x-AIRWAYS, became powerful advertising aids to market
growth).
We looked at the triggering aspect of a Freephone call in Chapter 2. Figure 7.1
summarises the essential network features of the Freephone service.
This is an example of a call to a company which has rented the Freephone
number 08081 570980. When callers dial into the company with this number the
SSP triggers at the 0808 digits (for example). Then, knowing that the number
length for addresses starting ‘0808’ is 11 digits, the SSP waits for a further seven
digits before sending a query (a message containing an INAP I_DP operation) to
the SCP. The SCP consults its translation database and returns a network number
(01632960001 in this case) and instructs the SSP to route the call towards this
number in an INAP ‘Connect’ message. The SSP now recognises ‘01632’ as
corresponding to a physical routing, and so is able to complete the set-up towards
the call’s destination using its normal routing and translation functions, in the same
way as it would if the caller had dialled the digits 01632960001 in the first place.
However, an important difference is that the call is not charged to the caller, and the
resulting call record indicates that a charge needs to be raised against the recipient
of the call.

Figure 7.1 A Freephone call

 Service examples 141

In Figure 7.1 we show the scenario where the caller’s local exchange has been
upgraded to SSP status, but we could equally have shown the second exchange,
which is likely to be a trunk level exchange, as the triggering SSP.
We can see from this example that the function of the SCP is to enhance the
switch’s number translation capabilities. Whilst number-to-address translation is a
normal switch function, it would be an expensive and inefficient overhead to load
the switch’s real-time call-processing software with the functions and data needed
for these extra (non-geographic) translations. Conversely, it would be inefficient to
distribute the Freephone translation data to hundreds of network switches. This is
because these sorts of large-scale network data update operations are slow, labour-
intensive and potentially error-prone procedures. Also, larger Freephone customers
often have a widely-dispersed and dynamic profile of call reception points, and,
through the working day, they may well need immediate changes to their incoming
routing translations in order to adjust their response to changing calling patterns.
So, having a single point of database update becomes an important advantage for
these customers.

7.1.1 Inter-network Freephone

The previous example shows the operation of a Freephone service within a single
network operator’s domain. This next example considers the case of a caller
dialling a Freephone number that belongs to another network operator, in the same
or a different country. There may, or may not, be co-operation between the
intelligent network controls in these two domains.
Figure 7.2 shows the case of ‘non-co-operating’, independent INs, where SCP
processing is invoked in the first network to obtain a translation that can be used to

Figure 7.2 Inter-domain Freephone – non-co-operating INs

142 Intelligent networks – principles and applications

route the call towards the second network. This example shows that the translation
process does not alter the Freephone number (which is controlled by the second
network operator) but adds a prefix (represented by þ41 in this example) that
allows the underlying transport network to route the call forward in the right
direction. The current example shows ISUP being used as the telephony-level
interconnect protocol.
When the call request arrives at the destination network the second network
operator’s SCP is separately used to find the translation of the Freephone number to
the required network termination. The call attempt is then routed towards this
required destination.
As an alternative implementation, Figure 7.3 shows the case of ‘co-operating’
INs, where the first SCP interrogates the second SCP to find the destination routing
number directly before sending instructions to the first SSP to route the call.
Overall the processing in this mode of working is likely to be more efficient, as
can be seen from a quick glance at the diagrams – a single IN access operation is
needed rather than two separate transactions. Furthermore, the ‘co-operating
networks’ approach allows the second operator’s SCP to be aware of traffic loading
in advance of the incoming calls being offered. An alternative interconnect route,
maybe via a third network if necessary, could be specified if the traffic control
algorithms in the second SCP are aware of congestion or other service-affecting
problems, on the first choice routing to the target destination.
SCF–SCF interface details are included in the ITU-T IN CS-2 recommendations
but because there are other commercial issues concerning interconnect agreements,
the former scenario (Figure 7.2) is still the more commonly used arrangement in
fixed networks. From a standards viewpoint, the diagrams illustrate that the
network interconnect could be between SCFs, between SDFs, or between an SCF
and a foreign SDF.

Figure 7.3 Inter-domain Freephone – co-operating INs

 Service examples 143

7.1.2 More complex number translations – Geographical routing

Once a basic Freephone service is operating, a natural development is to introduce

further applications on the number translation theme, making more use of the
inherent capabilities of the computer on which the central database is based.
Examples are routings based on time of day, day of week or the geographical
location of the calling line.
An example of geographical routing would be a national windscreen repair
company that offers a single Freephone contact number nationally, with incoming
calls always being connected to the geographically closest depot. Another example
is the well-known pizza service, where a chain of pizza restaurants advertises a
single Freephone number for a home-delivery pizza service and the call is routed to
the nearest restaurant to the caller. In the USA this is known as an ‘area number
calling’ (ANC) service.
As before, the SSP recognises the Freephone access code (080X in this case, but
could be 1800, or 0500, depending on the local regional numbering plan) and routes
a query to the remote database. This time the query includes the CLI (calling line
identity). The SCP database recognises, from the Freephone number, that this is to
be geographical routing service and so it then obtains the caller’s location from the
CLI to determine a suitable network destination address for a local agent. The SCP
then returns this local agent’s network address to the SSP for onward routing of the
call.
There can be situations where the triggering SSP does not have a full CLI to send
in the IN request message to the SCP. This could arise for instance if the triggering
SSP is at the trunk level and is several switches away from the caller’s local
exchange, which might be an analogue switch that does not support SS7. In this
case the normal procedure is for the SS7 network to provide a ‘partial CLI’, which
identifies the digital (SS7-enabled) switch nearest to the calling user. This partial
CLI is then delivered to the SCP by the SSP. This will give the service logic some
indication of the caller’s local area and this can be used instead of the full CLI to
find a suitable local agent.

7.2 Personal numbering

Personal numbering (PN) is a service that allows users to have overall control of the
handling of their incoming calls. The PN subscriber is allocated a ‘personal’
number, which is likely to be part of a reserved number range so that it can be
easily identified by callers as a personal number from the initial digit string.
When callers dial the user’s personal number, the service will divert incoming
calls to wherever the user wants the calls to terminate. This could for instance be a
mobile number, a home number, work number or answering machine. The service
maintains a diary with default numbers that can be over-ridden by the user. For
instance, a user might want calls directed mostly to his office phone during normal
144 Intelligent networks – principles and applications

working hours, although whilst travelling to a meeting he might want calls to go to

a mobile phone and then to an answering service whilst the meeting is in progress.
These frequent changes to the network destination for calls dialled to the same
customer number suggest that an IN architecture with central control of the routing
numbers is likely to provide the most suitable implementation. In practice, variants
of this service usually have a pre-set background diary profile, with a mechanism
provided by the network operator to enable the user to update his profile at short
notice.
One automatic and commonly used method is to provide the means for users to
enter into dialogues with automatic voice response units prompting the user to enter
DTMF digits on the telephone key-pad. In the network this would use intelligent
peripheral (IP) equipment and the IN CS-1 ‘prompt and collect’ procedures,
described in Chapter 2. In some implementations these updates are more
conveniently carried out via an Internet browser and a secure communication
with a server that is associated with the controlling SCP.

7.3 Incoming call screening

A popular IN application is the screening of incoming calls, where customers need

to be able to control which calls are offered to their telephones. Such a service
would be attractive for instance for a customer who suffers from a large number of
unwanted and unsolicited marketing telesales calls. A ‘blacklist’ of calling numbers
is held in the database associated with the controlling service logic and a
terminating trigger (DP 12 on the BSCM in Figure 2.8, p. 29) is set against the
subscribing customer’s line. Every incoming call to that line causes an I_DP to be
sent to the SCF. The I_DP will include the CLI as a parameter, and this will be
checked against the barred CLI list in the SCP. If the caller’s number appears on the
list, a ReleaseCall message is returned to the SSP and the call is cleared. Statistics
can be maintained at the SCF, so the subscribing customer can be charged, if
necessary, at a rate that is based on the number of rejected calls.
This screening scenario is illustrated in Figure 7.4. The service, or its variants, is
sometimes marketed under names such as ‘do-not-disturb’ or ‘choose to refuse’.
Figure 7.4 shows a version of the call screening service where the SCP checks
the CLI and instructs the SSP to release the call if the CLI is on the screening list.
In practice, the SCP would probably precede the ReleaseCall operation with a PA
(Play Announcement) operation so that a suitable rejection tone or automatic
announcement could be provided from the SSP’s internal SRF before the SSP
releases the call.
A more sophisticated variant of the screening service could involve an external
SRF provided in an IN IP node. This would allow the SCP to instruct the SSP to
extend the caller’s connection to the external SRF for more complex screening
treatment. For instance, the different levels of screening might apply at different
times of the day and customised announcements and mailbox facilities may be
required, depending on the screening conditions.
 Service examples 145

Figure 7.4 Incoming call screening

Another variant, shown in Figure 7.5, could be offered where the customer can
invoke a ‘do-not-disturb’ feature, which would result in all incoming calls that are
not rejected outright being connected to the IP. The caller can then be invited to
enter a 3 or 4 digit ‘pass-code’ to achieve connection or be connected to the
customer’s mailbox. If a pass-code is entered by the caller, it could be validated by
SRF software, or reported back to the SCF for validation using the Prompt and
Collect User Information response message. If the validation is positive, the SCF

Figure 7.5 More complex call screening

146 Intelligent networks – principles and applications

would instruct the SSP to break the IP connection and connect the caller to the
customer’s line.
The message flow for a successful connection after the screening process just
described is similar to that of Figure 2.15 (p. 42). The pass-code entry occurs at the
‘user interaction’ phase, when the caller is in communication with the SRF at the
IP.

7.4 Least cost routing

In an environment where there is strong competition amongst network carriers for

transit traffic, IN techniques can provide an effective solution for a ‘least cost’
routing service. With this service, trunk calls can be routed over one of several
transit networks depending on cost. Figure 7.6 illustrates the operation of the
service for a call between customers on two separate local operator networks.
There is competition at the transit level, and so the diagram shows a choice of
three different transit carriers. Each is likely to have advantages for different call
scenarios, and so the service provided at the SCP needs to be aware of the details of
the tariff arrangements for each transit network. There could, for instance, be
variations depending on time of day, call destination and call origin.
The service triggers to an SCP at the customer’s local exchange, which is the
SSP shown in Figure 7.6. This could be a conditional line-based trigger, with a

Figure 7.6 Least cost routing service

 Service examples 147

condition that the call triggers if a call is recognised to be a long distance call (e.g.
on the condition that there is a leading ‘0’ in the dialled digit string).
The SCP returns a routing prefix, or an outgoing route identifier, which enables
the SSP to route the call onward through the required transit network, which relays
the call set-up messages to the destination operator’s network. The call is then
connected to the destination customer in the normal manner.
Providing a fallback option to cover the event that the preferred route is
unavailable could enhance this service. An E_DP for a ‘route-select fail’ (DP-4)
event could be set. If this triggers, the call could be connected to an SRF in an IP
node and the caller offered a connection over a second choice route. This second
option may not be the ‘least cost’ alternative, but at least the call would be
completed – if this is what the customer wants.
This is another example of a service that is dependent on specialist and up-to-
date data. In this case tariffs may fluctuate and special price offers may be available
at short notice for limited periods. It may therefore be practical for the network
operator to out-source the data maintenance function to a specialist service
provider. Figure 7.7 shows a scenario where a network operator might offer its own
basic service, but may also provide access to a more advanced specialist service
operated by an independent service provider. Figure 7.7 shows an internal SDF for
the operator’s service and an external SCF–SDF interface to the service provider’s
own SDF, shown in a separate SDP node.
The decision on which data to access could be made on the basis of the service
key sent from the SSP in the I_DP message. Alternatively, the SSF could access a
customer profile database, which would indicate which service variant has the
user’s subscription. An advantage of this latter procedure is that the underlying
network management need only be concerned with the triggering criteria, and can
be kept free from details of the service provisioning.
The SCP application therefore either accesses its own database for least cost
routing information or it generates a query to the external database (SDP) which

Figure 7.7 Service data provided externally

148 Intelligent networks – principles and applications

will return the identity of the preferred carrier. The SCP will convert this to routing
information and convey it to the originating SSP.
In this scenario the service provider could charge the network operator a per-
transaction fee, and the network operator could recover the call costs through
normal billing procedures. A different approach would be for the service provider
to provide the least cost routing service directly to the customer, making its own
provision for billing, using the CLI or PIN verification for identification on a call-
by-call basis. In this case the network operator’s SCP would just operate as a
gateway, providing access, control of SRF resources and an onward routing service.

7.5 A virtual private network (VPN) service example

To further illustrate triggering principles we now examine a simple VPN service,

which is illustrated in Figure 7.8. The features offered by a VPN service typically
include:

& abbreviated dialling

& reduced call charges for ‘on-net’ calls
& a call forwarding service
& a mailbox service.

In this example the triggering criteria is arranged such that when a user dials the
digit ‘7’ the SSP triggers but waits for three more digits before sending the request
to the SCP. ‘7’ is the prefix for the VPN numbering scheme and users have three

Figure 7.8 A VPN service

 Service examples 149

digit private addresses. These are in addition to their public addresses, as shown in
Figure 7.8. This trigger is equivalent to the ‘customised dialling plan trigger’ in the
US AIN specifications. The service operates as follows:
User A dials user B’s VPN number, which is ‘124’, preceded by a ‘7’, which is
the pre-assigned identification digit for VPN users calling each other. The SSF
assembles an I_DP message, ensuring that it contains the digits that user A has just
dialled, as well as A’s own public number (01154 960004). The I_DP also contains
the service key (SK) that is associated with the particular set of triggering
conditions provided in the SSP for this VPN service.
The I_DP is despatched to the SCP. The ensuing transaction is shown in Figure
7.9. The request is initially handled by the portal functions of the SLEE (service
logic execution environment, which was described in Chapter 2) at the SCP. The
SCP as illustrated has four different service logic programs (SLPs) running, which
are accessible through the SLEE. The SLEE uses the SK parameter of the I_DP to
identify the appropriate SLP for handling the request.
The network operator will be providing VPN services for many customers, and
each one is a separate and independent virtual network. The particular VPN
associated with a new call is identified by the network CLI (i.e. the advertised
public number in this case) of the user who has called. The VPN access digit (‘7’)
indicates that the following three digits (‘124’) will be an address for another user

Figure 7.9 Service logic selection

150 Intelligent networks – principles and applications

on the VPN. Having established the context, the SLP will consult a translation
database in order to find a translated number for the onward routing. The database
consulted can be local to the SCP or remote, or even under the customer’s control.
When the service logic has located the database entry and has found a
corresponding network destination number (‘01632 960007’ in this case) the SCP
returns an INAP ‘connect’ operation to the SSP. The Connect operation normally
contains the required public network address (‘01632 960007’) that is used by the
SSP to continue with the call set-up.
However, if the called user has a call-forwarding (unconditional) feature active,
then the IN service logic would have returned the network number for the call-
forwarded destination. So, if user B is currently forwarding incoming calls to user
C, the address string returned in the Connect message would have been ‘01174
960008’. If the conditional call-forwarding service is active, such that calls are
forwarded if the user is not available (‘no reply’ or ‘busy’), the IN service logic will
precede the earlier Connect operation with E_DP-arming operations. The SCF can
arm for the network events ‘busy’, ‘no reply’ or ‘route unavailable’, using ‘request
report BCSM’ operations with ‘event’ parameters set to 4, 5 or 6. The message
sequence diagram of Figure 7.10 illustrates the action taken when one of these
E_DPs is encountered.
In order to specify a special charging rate for the VPN service, the SCF also
sends an FCI (‘furnish charge information’) operation. This allows service-related
information to be temporarily stored in a working record in SSP so that, at the end
of the call, an appropriate charging event record is output to the off-line billing
system for consequent pricing and billing.
Figure 7.10 illustrates a typical SSF–SCF message interchange for the example
IN VPN service activation. The three messages from the SCF (request report
BCSM, FCI and Connect) could be sent as separate SS7 messages, as shown in

Figure 7.10 Message flow for a VPN call

 Service examples 151

Figure 7.10, but alternatively a single message could be used for all three, leading
to greater messaging efficiency.
If the called customer answers, there is no further SCP involvement, and the
INAP transaction is closed. However, if the destination is busy, as is the case in the
example of Figure 7.10, the SCP will receive an event report BCSM operation, in
response to the request report BCSM, and further SCP processing is needed. The
SCP will then instruct the SSP to route the call to an alternative destination, as
shown in the message sequence.
This service example illustrates some of the principles. In most cases there will
be a richer service specification, requiring a more complex interchange of
messages. For example, it is likely that E_DPs for other call-failure conditions (e.g.
‘no reply’, or faults or congestion) will be armed, requiring appropriate SCF logic
to be executed when these are triggered. Also a mailbox destination would
normally be included as an alternative destination for calls that fail to reach their
intended destinations.

7.6 A directory enquiry call completion service

This is an example of an IN-enabled directory enquiry service that offers automatic

connection to the required destination number when the enquiry operator has found
the required number.
The example illustrates ITU-T CS-1 procedures and messages for the SCP-
controlled automatic connection of a call to IN IP devices. This is in order for the
SCF service logic to collect further information from a caller in order to progress a
call. Typically the extra information is collected by the IP, using voice guidance.
The caller is invited to enter DTMF tones from key-pad sequences, or to respond
verbally to the automated voice unit.
As with the other service examples, the procedure illustrated here is just one of a
number of possible ways of arranging the service logic support for a network
service. For example, for clarity of explanation a single SCP is shown as the host
for the required service logic for this service. In practice, the constituents of this
SCP may well be provided in a geographically distributed computing system, but
the network control principles illustrated here remain fundamentally similar.

7.6.1 IN distribution of operator calls

In preparation for describing the ensuing stages in the call completion scenario,
Figure 7.11 shows the basic arrangement for accessing an enquiry operator from a
common access number. Typically a short well-known digit string is used to access
services such as directory enquires – this illustration uses the digits ‘123’. Enquiry
operators will typically be geographically distributed over a wide area and the
controlling service logic needs to arrange for a new call to be routed to a suitable
free assistance operator.
152 Intelligent networks – principles and applications

Figure 7.11 A directory enquiry call

Commonly, dedicated distribution and local queuing systems are employed for
operator assistance centres. If an IN solution is used, these systems must either be
provided within the IN service logic in their entirety, or else there must be inter-
working between the overall IN logic and the local management servers. This is to
ensure that the IN systems have up-to-date information available on termination
availability, and to update the SCF with results of individual searches. The latter is
the scenario chosen for the current example, and the scenario is illustrated in Figure
7.11. These local management servers control local call centres (and possibly
home-working groups).
As well as performing normal housekeeping and supervisory functions, the
management servers also continually update the service data function (SDF) at the
controlling SCP with information on availability and preferred choice of selection
of the network numbers corresponding to the enquiry operators. This provides the
SCF service logic with continual real-time information to enable it to provide fast
responses to incoming requests.
Triggers are provided at SSPs, which can be at the local or trunk level in the
network, so when a customer dials ‘123’ to access the service, an IN request
message (I_DP) is sent to the SCF to obtain a routing. The SCF’s choice of network
routing will depend on the CLI (sent from the SSP in the I_DP) and on information
available in the associated SDF. In reply to an I_DP it responds with a ‘connect’
message instructing the SSP to route the call to a suitable enquiry operator’s
network number, in the customer’s local area if necessary.

7.6.2 Automated response and call completion service

As human labour costs rise relative to costs of equipment and software, the business
imperative with enquiry operator procedures is the trimming down of the overall
 Service examples 153

call-handling time in the enquiry operator procedures. One or two seconds’ saving
here and there on enquiry operator services using automation can make a big
financial difference to a network operating company’s balance sheet. One way to do
this is for the enquiry operator to pre-record a salutation (such as ‘Good morning.
Which name and area do you require?’). This is then played to the caller by a voice
response unit (VRU) at the start of every call, so the enquiry operator is then able to
join in the call when the caller is giving the details.
Another commonly used way of saving time is, after the trace has been
completed, for the required number to be given by an automatic announcement,
freeing the operator to move on to another enquiry after finding the number. The
call completion service can be offered at this point by another VRU, and if the
caller agrees, the call can be established automatically by the SCF.
The illustrative sequence here (Figure 7.12) is for a call that passes through each
of these stages. The overall process therefore requires a series of connections to a
VRU, then to an enquiry operator, then another VRU and finally to the destination
number. The VRUs are numbered 1 and 2 in Figure 7.13, but obviously they could
be the same unit. They are shown separately here to illustrate the principle of serial
interconnections that is characteristic of IN CS-1 INAP procedures – each
connection is cleared as the next is established.
Connections are numbered in Figure 7.12 as follows. Connection 1 plays the
initial salutation and leads into the main interaction between the caller and
the enquiry operator in connection 2. Connection 3 to the second VRU is to tell the

Figure 7.12 A 4-stage directory enquiry service

154 Intelligent networks – principles and applications

Figure 7.13 Call completion call flow, part 1

caller the number that has been found and to offer the call completion service, and
connection 4 is the actual call to the destination number.
A message flow that will achieve the network control for the call as far as the
enquiry operator interaction (connection 2) is shown in Figure 7.13.
When the caller dials the access code for directory assistance triggering occurs as
previously described, but this time the SCF instructs the SSP to route the call
temporarily to a VRU using the ETC procedures for IN IPs which were described in
Chapter 2.
The call then arrives at the VRU and is connected using normal ISUP
procedures. Meanwhile the VRU’s SRF uses the SCF Id. parameter to find the
signalling location of the controlling SCP. The SCP had sent the SCF Id. parameter,
with the correlation Id. (CI on the diagram), to the SSP in the ETC operation, and
these parameters were then relayed to the VRU in the network signalling address
message.
The VRU is therefore able to send a query message (ARI operation) to the SCF,
including the correlation Id. parameter. The SCF uses the CI parameter to bind this
request to the earlier transaction that it had opened with the originating SSP. The
SCF returns the ‘Play Announcement’ instruction to the SRF. This includes details
of the announcement that should be played to the caller. The VRU equipment duly
plays the requested announcement and reports completion of this task back to the
SCF using the SRR (‘specialised resource report’ operation).
 Service examples 155

The SCF now instructs the SSP to clear down the speech path to the first VRU
and to set up a new connection to a free enquiry operator. This uses the DFC and
connect operations. This call scenario assumes the existence of a background
relationship between the controlling IN SCF and the management servers for the
enquiry operator call centres, as shown in Figure 7.11 above. This keeps the SCF up
to date with information on the network addresses of available enquiry operators.
The SSP now releases the speech path to VRU 1 and sets up a new path to the
nominated enquiry operator just in time to hear the caller respond to VRU 1’s
earlier salutation, enabling the enquiry operator to start the number trace.
Having located the required number the operator presses a keyboard button
which launches an information message, containing the required number, from the
management server to the SCF. This uses the (non-IN) transaction interface in
Figure 7.11 to send a ‘number found’ message to the SCP. The operator will then
be free to drop out of this enquiry to move on to other work. The operator’s server
locates the correct target SCF, using the SCF identifier and CI, which had been sent
forward in the network set-up message (e.g. ISUP IAM), which had been received
in turn from the SCP in the connect message.
The message sequence continues in Figure 7.14. On receipt of the ‘number
found’ message, the SCF releases the operator connection, using the DFC
(disconnect forward connection) message and instructs the SSP to establish a new
connection to the second VRU. Again, the SCF uses the ETC operation for this,

Figure 7.14 Call completion call flow, part 2

156 Intelligent networks – principles and applications

sending the SCF ID and CI parameters for the same transaction binding and SCP
identity reasons as before.
VRU 2’s role here is to announce to the caller the number that had been found in
the preceding interaction by the enquiry operator and to offer the call completion
service. This invites the caller to respond verbally, or to enter DTMF digits to
indicate acceptance or otherwise. Because more is required of the SRF in VRU 2
than was required of VRU 1, the SCF this time replies to the VRU’s ARI with a
‘prompt & collect’ (P&C) instruction rather than the passive PA request as before.
This time the SCF asks the SRF to collect a response from the caller, either via
DTMF digits or voice recognition equipment.
The caller’s response is returned in the P&C Response operation. If the caller has
accepted the offer of automatic call completion, the SCF once more instructs the
SSF to arrange for the disconnection of the temporary connection to the VRU and
sends a Connect operation for the SSP to make the final network connection to the
required destination number.
From this point on, the call is routed and connected to the required destination
number in the normal way. The SSF–SCF transaction could be terminated at this
point, or it could be kept open and a mid-call trigger armed, and the call could be
re-presented to the operator if necessary.
In practice there are most likely to be other INAP messages used; we have just
shown the ones needed to demonstrate a skeletal service. For instance, charging has
not been considered; an FCI could be sent with each Connect and ETC message,
this would cause the SSP to generate a billing record, if required, for each of the
three separate connections used by the caller.

7.6.3 Call completion service summary

This example has aimed to illustrate the ‘behind-the-scenes’ workings of the

intelligent network using various CS-1 INAP operations to effect the series of
connections needed to make up a complex service of this nature. The actual
connection methods used in physical implementations vary widely – we have
chosen to illustrate the use of standard Core INAP operations for all interfaces
except the SCF to operator management service. However it is common for
proprietary optimisations to be included to avoid messaging delays.
Obviously there are stringent messaging performance targets to meet because the
various connection changeovers need to appear instantaneous to the user, and
delays of more than a few hundred milliseconds at any stage will become
noticeable. To achieve the serial coupling of the different service logic functions
throughout, the ‘correlation Id.’ (CI) parameter is essential, as is the SCF Id.
parameter, which serves to complete the triangular relationships between SCF, SSF
and SRF. This CI parameter appears as an optional parameter in several INAP
operations, particularly the Connect, ETC and ARI, which are used here. However,
some telephony signalling systems in force today (e.g. TUP derivations and early
ISUP versions) do not have standard mechanisms implemented for conveying CI
 Service examples 157

and SCF ID between the SSP and the IP, and so ingenuity is often needed in
specific implementations to find ways of re-using existing signalling fields.
Because the majority of today’s IN implementations are still based on the CS-1
core INAP procedures, we have illustrated these CS-1 procedures for this service.
In fact, this type of service is likely to be more suited to the CS-2 leg-control
procedures, because of the numbers of connections and legs involved and the time
constraints. Call party handling (Chapter 5) techniques have a big advantage with
services where these sorts of time-critical call-leg manipulations are not necessarily
serial, and so are handled more efficiently.

7.7 Call gapping

Independently of IN, call gapping is a commonly used flow control technique in

telephony networks. It is therefore naturally adopted as a mechanism for load
control protection at IN service control points and other vulnerable IN nodes that
might be at particular risk where they are centrally provided resources.
CS-1 specifies a flow control procedure where an SCP is able to send call-gap
request messages to an SSP in the event of resource overload risk in the
components of the SCP. The core INAP operation that is sent from the SCF to the
SSF in order to reduce the rate of service requests is the ‘CallGap’ operation.
CallGap will instruct the SSF to gap on destination number or service key, or both,
and to gap calls for a specified duration with a specified minimum inter-call
interval. The INAP CallGap operation is sent to an SSP in a response message for a
currently open transaction.
An example is shown in Figure 7.15. In this illustration a call triggers and the
SCF returns a translated number for routing in a ‘connect’ operation, and an FCI
operation to convey IN charging information via the SSP to the off-line billing
centre. However, the SCF has detected a high level of calls to this particular
service, which is identified by the service key value, and so it returns a call gap
instruction for this service key value. The gapping does not apply to the current
request – it comes into force after this request has been dealt with.

Figure 7.15 Use of the INAP CallGap operation

158 Intelligent networks – principles and applications

Figure 7.16 Activate service filtering example

7.8 Activate service filtering

The ActivateServiceFiltering (ASF) operation was designed to ease the load on the
SCP in the event of high-volume televoting services. This procedure devolves a
large part of the mechanics of the service handling to the SSP by instructing it to
only trigger to the SCP for this service every nth call. The value of n could be any
value within the implementation limits, but might commonly be between 1 and
1000 depending on the dynamics of the service.
Details of the votes are maintained in the SSP in counters, which are defined in
the ASF operation. The ASF also specifies the action to be taken by the SSP in
terminating the televote calls, in terms of playing announcements or tones to the
callers.
Figure 7.16 shows an example where n is set to 20. The results are reported after
the next n calls in a ServiceFilteringReport (SFR) operation, which can accompany
another I_DP operation.
The issue of charging for the service must of course be settled. The normal
procedures allow the SCP to return an FCI operation (e.g. with a connect operation)
in response to an I_DP. However, in the case where the SCP has instructed the SCP
to take autonomous action, such as connecting the caller to a standard
announcement, there may be a need to generate a billing record to the SCP’s
specification for each of these individual calls. Alternatively, the billing could be
left to normal procedures, where typically a call record will be generated by a local
exchange for every (apparently) successful call.

7.9 Simple CTI (computer–telephony integration)

CTI is what it stands for – the integration of computer technology with telephony.
A typical application is that employed in a call centre, where the CLI is used to
present customer information on an agent’s computer screens whilst simultaneously
presenting the associated call through the same agent’s headset.
 Service examples 159

Figure 7.17 CTI call pick-up

Figure 7.17 shows an example of a simple CTI-group function, where waiting

calls can be directed to alternative agents. The terminals represent four users of a
CTI call pick-up group, each having a telephone and a computer terminal. A call is
routed through the network destined for telephone D and a terminating trigger (‘DP
12 – T_Attempt_Authorised’) is encountered.
The SCF processes the I_DP, which includes the dialled number, the CLI, the DP
number and the service key value, which enables the SCP to identify the SCF
application logic that handles the call pick-up service. The SCF service logic sends
a message on the notification channel to the CTI server containing the CLI. The
CTI server is able to use the CLI to access records about the caller, and these can be
displayed on user D’s computer screen while the caller is waiting for an answer.
The SCF returns an instruction (containing connect and requestReportBCSM
operations) to the SSP to attempt a call connection to D and to arm event triggers
(E_DPs 13 & 14) to cover ‘no reply’ or ‘busy’ conditions.
In this example the call is offered to D but is not answered, so after the pre-set
time period the SSF activates the ‘no answer’ event trigger and notifies the SCF,
using the Event Report BCSM operation. The SCF then accesses information in a
local or remote SDF (which could be the CTI server) about other destinations that
could accept the call. Having found an alternative destination number, say B on the
diagram, the SCF instructs the CTI server to transfer the CLI information from D’s
computer to B’s computer. At the same time the SCF instructs the SSP to clear the
call attempt to D and to make another call attempt to B.
The communication between the SCP and the CTI server (the ‘notification
channel’) could be a proprietary interface, or it could be based on IN CS-1 and CS-
2 SCF-SCF or SCF-SDF protocols. A key benefit of using IN-based solutions here
is not so much the call deflection capability, which many PBXs possess anyway,
but that the group members could be geographically remote from each other spread
across a wide-area CTI network.
160 Intelligent networks – principles and applications

7.10 CAMEL calls

7.10.1 An outgoing CAMEL call

Figure 7.18 illustrates the call set-up messages for an outgoing mobile call, using a
CAMEL IN request, to a PSTN user. This example call flow is for a number
translation service. It is concerned only with the call control signals; it is not
concerned with the sequence of preliminary signalling needed for the mobile
station to obtain a channel from the BSS, or with the authentication dialogue that
takes place with the HLR and VLR.
When a mobile station originates a call, and a set-up message arrives at the MSC
from the BSC (not shown in the diagram), the MSC interrogates its associated
VLR, using a MAP ‘send info for outgoing call’ message. This message contains
the following key parameters:

. the requested call destination, in the form of an E.164 address

. the requested service, which in this example is just voice.

On receipt of the ‘send info for outgoing call’ message the VLR will check the
caller’s profile information, which would already be lodged in this VLR as a result
of a previous location update. In particular it checks that this caller is authorised to
make calls and that call barring does not apply to the requested destination address.

Figure 7.18 An outgoing CAMEL call

 Service examples 161

When the VLR checks the profile it will also discover that the CAMEL
subscription information (CSI) mark is present, because the caller is a CAMEL
customer. The CSI in the subscriber’s data profile in the VLR will have been
updated from the HLR. The VLR then returns a ‘complete call’ message that
includes an O_CSI (originating CAMEL subscription information) indicator.
O_CSI will identify the particular CAMEL service environment (CSE) for this
subscriber.
The home CSE could now be in a foreign mobile network – it makes no
difference because CAP (CAMEL application protocol) is an international standard
and the gsmSSF does not have a particular relationship with any gsmSCF.
As well as the address of the target CSE, the CSI also contains a service key so
the destination gsmSCF knows which services are to be triggered. This is an
interesting extension of the CS-1 service key (Chapter 2), which is chosen as a
result of the triggering criteria in the SSF. The gsmSSF, however, has no
knowledge of the trigger ‘conditions’, so it does not know which services are to be
triggered in individual service invocations.
If the CSI had not been present in the caller’s profile information, the VLR would
have checked whether conditional barring was set. Because this is a CAMEL
customer, the ultimate destination may be changed as a result of the consequent
CAMEL service request. The VLR is now only concerned at this stage if barring of
all calls applies – in which case the call is immediately rejected.
When the gsmSSF receives the ‘complete call’ message with the CSI, it
assembles an SS7 CAP message (to be carried by TCAP and SCCP) containing the
CAP I_DP operation. The address (of the home CSE) for the message is obtained
from the CSI. The CAP I_DP parameter set would typically include:

^ service key
^ calling party number
^ calling party category
^ bearer capability
^ called party number
^ BCSM event (DP number)
^ IP SSP capability
^ caller’s IMSI (international mobile subscriber identity)
^ call reference number
^ MSC address.20

When the gsmSCF at the user’s home CSE has processed the I_DP contents, it
returns a CAP ‘connect’ message to the MSC (containing the gsmSSF). This
contains a new call destination address to be used by the MSC for this call.

20
It can be seen that, apart from these three last GSM-specific parameters, there are strong similarities
between the CAP I_DP and the INAP I_DP. More information on CAMEL standards is given in Chapter
3, and an overview comparison between mobile and PSTN signalling is in Chapter 3.
162 Intelligent networks – principles and applications

At this stage a second check with the VLR is carried out, using the MAP ‘send
info for outgoing call’ to ensure that this user is not barred from calling the new
destination number. If this check passes, a second ‘complete call’ instruction
(without CSI) returns to the MSC, which sends an ISUP IAM across the network
towards the required destination. Figure 7.18 assumes that the next MSC is in fact a
gateway MSC, and that the call is destined for a PSTN user. The call set-up and
clear-down sequences then proceed in the same way as for a fixed network SS7 call.

7.10.2 An incoming CAMEL call

Figure 7.19 shows an incoming call from a fixed network user to a CAMEL mobile
subscriber. In this example a CAMEL IN request is made before connection. This
allows the CSE to run CAMEL-based call screening services (for example) before
authorising the gsmSSF to connect the call. This illustration shows the SCE
instructing the gsmSSF to connect the call to the subscriber.
The call enters the mobile network via a Gateway MSC, which queries the HLR
in order to find out which MSC to direct the call towards. If there is more than one
HLR, the G-MSC, which (in this example) is also the gsmSSF, will identify the

Figure 7.19 An incoming CAMEL call

 Service examples 163

correct one from the called address. This query uses MAP, and, as shown in Figure
7.19, the message used is a ‘send routing info.’ message.
The HLR will first check that calls for this user are not being forwarded, and will
then send a MAP ‘provide subscriber info.’ request to the last-used MSC/VLR for
this mobile subscriber. For non-CAMEL subscribers, the HLR would use the MAP
‘provide roaming number’ request. However, a different request message is now
appropriate because the gsmSSF needs the CAMEL subscriber’s location and status
details, which need to be sent in the CAP I_DP message to the home SCE at the
gsmSCF.
The VLR returns the required details and allocates a temporary mobile station
roaming number (MSRN). This information is returned to the G-MSC via the HLR
in the acknowledgment messages. The MSRN is subsequently used as part of the
destination address in routing the call to the MSC currently serving the mobile. The
MSC associates the temporary MSRN with the mobile station, after which it frees
up the MSRN for re-use.
The G-MSC (the gsmSSF) next sends an SS7 CAP I_DP request to the CAMEL
subscriber’s home CSE. The destination address (for the home CSE) for the
message is obtained from the information returned in the previous MAP response
from the VLR.
After the gsmSCF has processed the I_DP contents and checked the screening
conditions it returns a CAP ‘continue’ message to the gsmSSF, instructing it to
carry on processing the call from the point it reached when it sent the I_DP to the
gsmSCF. The call then proceeds as a normal incoming GSM call. The G-MSC
sends an ISUP IAM, using the previously allocated MSRN, and the target MSC
sends a paging signal for the mobile in its last-known location area, then waits for
the response from the mobile station. When this is received the call can be
connected.

7.10.3 Pre-pay mobile service

The ‘Pre-pay’ mobile service is an example of how IN features can be translated

directly into results in the marketplace. Pre-pay was introduced originally as the
poor man’s mobile phone service – a second class service for users who didn’t have
the financial basis, or credit worthiness, to enter into a fixed minimum period
contract and receive monthly bills. In fact Pre-pay has now turned into a
mainstream service with big advantages for users and operators. The overheads of
contracts, credit checks, bills, minimum contract periods and monthly network
access charges become unnecessary. Bad debt is not possible because airtime is
paid for, simply and electronically, in advance. As well as reducing the mobile
operators’ management overheads, these features expand the market footprint
because many potential customers had been deterred by the prospect of contracts
and direct debit arrangements. As a result, mobile phones have become affordable
commodity items, available from supermarket and other popular outlets.
164 Intelligent networks – principles and applications

Figure 7.20 An outgoing pre-pay CAMEL call

The CAMEL system has provided the Pre-pay service with an inspired technical
solution. With CAMEL, all Pre-pay calls involve a request to the CSE, which
controls the credit checking and call monitoring. The CSE is able to instruct the
gsmSSF if the caller is running out of credit so that the caller can be informed using
an automatic announcement. Figure 7.20 shows the messaging involved in a typical
Pre-pay call.
As in Figure 7.18, the MSC–VLR interchange results in an I_DP being sent to
the home CSE, which recognises the user as a Pre-pay subscriber and returns a
request report BCSM event message, arming the gsmSSF for the two events
T_Answer and T_Disconnect. This is followed by a continue message, assuming
there are no number translation or call barring services applicable, and that the
caller has sufficient pre-paid credit in his account. When an ANS (answer) message
is returned from the far end network, the CSE is informed, via an event report
BCSM message, and it can start timing for the call. At the end of the call another
event report BCSM message is sent to the CSE, which can stop the call duration
timer and reduce the user’s available credit appropriately.
Had the user run out of credit during the call, the gsmSCF could have sent a
‘release call’ instruction to the gsmSSF and the MSC would have terminated the
call.
 Service examples 165

7.11 IN control for Internet dial-up access

An important application for IN today is in the management of the routing of dial-

up access to Internet service providers (ISPs). Where the basic access to the IP
network is provided by dial-up connections, IN control can substantially streamline
the delivery of dial-up data calls to appropriate modem banks in one of the ISP’s
points of presence (PoPs).
An example of this is shown in Figure 7.21, where an ISP has two PoPs
containing modem banks that can handle incoming dial-up calls for Internet access.
The caller dials an Internet access number, which would typically be via a
Freephone or premium rate number controlled by an IN platform. The IN SCP
therefore needs to provide a translation to a network address that will deliver the
call to a suitable modem bank.
In the example shown, the call is offered to PoP1, but all ports on PoP1 are busy
and so it is unable to handle the call. The IN platform is able to intercept this call
failure, by arming E_DPs (in the way we have demonstrated in previous examples
in this chapter) and re-routing the call to a secondary PoP (PoP2) which has spare
ports. In this way, a call that would have otherwise failed has been connected to the
IP network.
As the IN platform is routing the call to the PoP, it is able to apply conventional
voice network control features such as proportional routing, call gapping, etc. An
obvious enhancement would be to provide a data link between the SCP and the
ISP’s own management system so that SCP routing decisions can be moderated by
information or instructions from ISP-controlled service logic. An example of a
standard ITU-T IN component that could be used for this type of application would
be the CS-2 intelligent access function (IAF), which is designed to provide for
linkage between non-IN and IN network structures.

Figure 7.21 IN control of Internet dial-up access

 Chapter 8
Concluding summary

Everything that can be invented has been invented.

—Charles H. Duell, Commissioner, U.S. Office of Patents, 1899.

8.1 IN today – How did we get here?

Today, intelligent network principles are well established in most large fixed public
networks, and are controlling many popular high-volume services providing
backbone revenue streams for the large network operators.
Most intelligent networks in operation today have their basis in the ITU-T’s
highly successful CS-1 standard, or the (broadly equivalent) AIN standard in the
USA. In order to illustrate the potential for the evolution of the current
implementations, we have, in this book, mapped out the range of intelligent
network recommendations that have been produced by the ITU-T, which is the
main international standards body for telecommunications. Regional variants of IN,
such as AIN in the USA and ETSI’s Core INAP in Europe have been based on the
ITU-T’s recommendations.
ITU-T CS-1 was a major early success for IN, but the next standards landmark,
CS-2, represented a considerable expansion of detail that has so far proved to be too
much for many network operators, although this applies less to newer operating
companies who do not have such large legacy overheads. CS-3 consolidated on
some of the CS-2 recommendations and started to venture towards broadband
networks. We have seen that the recently released CS-4 documents reassess the role
of IN against a background of converging PSTN and IP networks. CS-4 seeks to
reconcile traditional IN with the IETF initiated activities such as PINT and SPIRIT
for supporting services such as click-to-dial and Internet call waiting. CS-4 also
tracks the API Open Service Access initiative for interfacing IN with external
service provider applications.
However, this book has been primarily concerned with implementations based
on the ETSI Core INAP standard because this has become a widely popular variant,
168 Intelligent networks – principles and applications

and was in fact fed back into the ITU-T to be adopted as a refinement (CS-1R) to
the early CS-1 ITU-T IN capability set. For this reason this book has concentrated
primarily on the practicalities of the ‘here and now’ of IN as it is currently
operating, rather than indulge in speculation of how evolving IN principles might
be implemented in future mainstream networks.
We have also provided some coverage on the origins of IN, from the early
pioneering Bellcore work in the 1980s (and before). This is because it is important
to understand the context into which IN grew in order to understand the
possibilities, as well as the limitations, of the technology.
During the 1990s the intelligent network was the prevailing vision for the
telecommunications industry. It replaced the previous vision of the 1980s, which
had been the exploitation of computer technology to achieve a fully digital
telephone network. As in most branches of engineering, yesterday’s visions tend to
become today’s ‘business as usual’. The use of IN principles has now become
commonplace, in the same way that earlier visions, such as the ‘new’ digital
networks of the 1980s and fibre optic transmission, also became commonplace.
However, whilst many of the original IN aims have been met, several factors
conspired to make the introduction of IN a long process. These included issues such
as:

^ need for short-term financial justification

^ vendor control of switching equipment
^ competition with legacy switch-based features.

We now expand on the background to these issues because an appreciation of how

practical realities intrude effortlessly into the implementation of a visionary
concept such as IN is crucial to understanding how we can successfully evolve our
networks for the future.

8.1.1 Financial justification

In the 1990s, the aims of intelligent networking seemed perfectly clear, and they
were well articulated. The large operators needed to re-arrange their networks to
enable them to introduce new service ideas quickly and cheaply and the ability to
do this became the ‘holy grail’ in their strategic plans. Words and phrases like
‘vendor-independence’, ubiquity’, ‘open interfaces’, ‘re-usable platforms’, ‘service
independent functionality’ became the currency of the newly emerging IN
community, whose enthusiasm was possibly as unbounded as that of today’s
Internet community!
As time moved on, however, there was a realisation that good ideas needed to be
turned into practical business cases, because network upgrades actually cost a lot of
money. This inevitably resulted in a tempering of ideals, and a measure of
pragmatism and compromise started to mellow the intelligent network expecta-
tions.
 Concluding summary 169

Another complication that caused delay in the arrival of IN in public networks

was that it tended to fall on the first candidate IN services to justify the expense of
the necessary network upgrades for the entire new IN infrastructure! Hence, the
search for the ‘silver bullet’ service started – a single service which would make it
all worthwhile, and whose backers were prepared to bear the initial brunt of the
required network upgrades to move the IN vision into reality. This often proved
difficult. When whole-life costs of services are analysed it quickly becomes clear
that as well as the real-time network mechanisms one has to include the
surrounding legacy operational support environments – the systems for billing,
order handling, maintenance, etc. In fact, as is the case with most major network
rearrangements, these latter costs are liable to dwarf the real-time network upgrade
costs.

8.1.2 Vendor control

Another factor that tended to slow the speed of IN rollouts was the understandable
reluctance of domestic equipment suppliers to share the operators’ high degrees of
enthusiasm for the idea of ‘vendor-independent’ networks. Operators traditionally
used to buy switching equipment from a small number of vendors, who had in turn
become used to the idea that much of their business was derived from selling new
services through costly network upgrades, which, because of their scale, involved
long lead-times and lengthy roll-out procedures.

8.1.3 Legacy switch-based features

Before IN, new telephony services had generally been introduced by software
modifications to the switch call control software, and this usually involved a
negotiated contract with the switch manufacturer. Whilst this was just the sort of
lengthy process and expense that IN was intended to eliminate, ironically the IN
upgrades themselves were subject to the same (or even worse) delays and expenses.
Most network operators’ sales teams are under continuous pressure to introduce
new services regularly in order to ‘delight’ – and indeed to keep – their customers
in the face of the strong competition. Therefore, when a demand for a new service
arose, a working implementation for it was needed as quickly as possible. In these
circumstances, the network operator could be faced with a choice between a service
independent IN solution that might be available in 12 months, or a vendor-tested,
service-specific network upgrade providing all the functions needed, available
maybe in just 6 months. With these options the obvious temptation is to choose the
latter to meet existing targets.
Obviously the consequence of this ‘short-term’ view is of course that switch
designs become more and more cluttered with disjoint functions, added in an ad
hoc manner, and the risk of the features interacting with adverse consequences
grows steadily and exponentially. Meanwhile the task of introducing IN-based
170 Intelligent networks – principles and applications

features on these switches became more difficult because of increasingly complex

call control software and erosion of potential IN business cases.

8.2 Tomorrow’s intelligent network

At the start of the book we reported on an early authority on IN [1] who, in 1993,
predicted that IN would be history by the year 2000. Accordingly, at the turn of the
century we were asking whether IN was the shooting star or the discarded catalyst!
We can now see that IN has in fact grown into an essential cornerstone technology
in voice networks. In many ways IN has now reached a plateau of stability, and
there are fundamental questions about how IN will now assist with the evolution of
future networks.
An important role for IN has been the part it has played as a ‘spawning-ground’
for the advancement of important technologies that are now shaping the future of
the industry.
An obvious example of this is the TINA project which, whilst rooted in early IN
developments, quickly dissociated itself from the original ‘bottom-up’ IN approach
in order to place new telecommunications service requirements in a wider business
context. In doing so it introduced the OMG’s concepts of object-orientation and
distributed computing into the telecommunications arena, which have now become
important concepts for next-generation distributed intelligence networks.
Another example of an IN-spawned technology is the concept of an open
network API, for which pioneering work was carried out by the Parlay team. Such
an API is able to effectively transform a telecommunications network into a
programmable entity, allowing new service development opportunities for third
party service providers. The pay-back for the network operator is that expanded
service activity will lead to greater network infrastructure use, hence increasing
network operator revenues. The Parlay output has now become an important
standard under the auspices of the ETSI OSA.

8.2.1 Future services

The operation of IN within a network should obviously be entirely transparent to

the users of IN services. They need know nothing about IN. IN is essentially a
network re-arrangement, internal to the network operator’s business, which enables
advanced telephony services to be offered in an efficient and cost effective manner.
Examples have been given of the type of popular IN services such as Freephone,
premium rate, VPN, personal number, call screening, Calling Card, etc. In addition,
the use of IN implementations of services that result from regulatory obligations,
such as Number Portability and Carrier Preselection, are commonly used. This is
particularly the case in the USA, as a result of which it is common for the majority
of PSTN calls to use IN features. The immediate future for IN in telephony
 Concluding summary 171

networks will undoubtedly see continued expansion and adaptation of these

familiar services.
Whilst IN services contribute significantly and steadily to PSTN revenue streams
in fixed networks, the fastest growth in the use of IN has recently been in mobile
networks, and the popularity of the ‘Pre-pay’ and inter-network roaming services
have been largely responsible for this. In this book we have limited our treatment of
IN and mobile networks to overview explanations and some examples that illustrate
the use of CAMEL in GSM in various scenarios. However, the currently most
exciting prospects for the expansion of the use of IN principles is almost certainly
in this field of mobile communications. For a more thorough treatment of
intelligence in current and future mobile networks the reader is referred to
Reference 45.
Another field where IN promises to make a significant impact is that of Internet
convergence. We have described (in Chapter 6) the use of IN for managing the
traffic routing for PSTN dial-up Internet access, so that it is directed to the service
provider’s equipment at the most appropriate point of presence. However, the
biggest single issue facing the future of IN evolution is likely to be the role IN has
to play in the new voice over IP (VoIP) networks. In considering the evolution of
IN we have to continually bear in mind that today’s IN has primarily been designed
for new services on voice networks. However, the success of the Internet protocol
(IP) has meant that the traditional circuit-switched networks for voice telephony
may now start to give way to VoIP technology. This raises the question of the new
role of IN technologies and how they transfer to the new order. In-depth coverage
of the impact of network intelligence on IP networks is provided in Reference 52.
 References

Several of the references cited are for ITU, ETSI and IETF output documents that
were valid at the time of publication. However, before drawing on their content it is
advisable to check on their currency. This can be done at the following Internet web
sites:

ITU: http/www.itu.int/publications/itut.html
ETSI: http/www.etsi.org/getastandard/home.htm
IETF: http/www.rfc.net/std1.html.

1 THORNER, J.: ‘Intelligent networks’ (Artech House Incorporated, 1994)

2 ATKINSON, J.: ‘Telephony – a detailed exposition of the telephone exchange
systems of the British Post Office, vols I & II’ (Sir Isaac Pitman & Sons Ltd,
1950)
3 ‘CCITT signalling system No. 7’, Journal of the Institution of British Telecom
Engineers (special issue on C7), 1988, 7, (1) Oliver, G.P. et alia
4 HILLS, T., and ABLETT, S.: ‘Advanced intelligent networks’. Analsys
Briefing Report Series No. 7, 1993
5 BROOKS, F.K.: ‘The mythical man-month’ (Addison Wesley, 1979)
6 ITU Recommendation Q.1200: General series, ‘Intelligent network recommen-
dation structure’, Revision 1, September 1997
7 ITU Recommendation Q.763: ‘Signalling system no. 7 – ISDN user part
formats and codes’, December 1999
8 ITU Recommendation Q.1901: ‘Bearer independent call control protocol’, June
2000
9 ITU Recommendation Q.1214: ‘Distributed functional plane for intelligent
network CS-1’, October 1995
10 ETSI Standard: ‘ETS 300 374-1; IN CS-1 Core INAP protocol specification,
version 8’ (ETSI Publications, July 1994)
11 ITU Recommendation Q.764: ‘Signalling System No. 7 – ISDN user part
signalling procedures’, September 1997
12 ETSI Standard: ‘ETS 300 403-1: DSS1 specification’ (ETSI Publications,
November 1998)
174 Intelligent networks – principles and applications

13 ETSI Standard: ‘ETS 300 374-5: IN capability set 1 (CS-1); Core INAP; Part 5:
Protocol specification at the SCF-SDF interface’
14 ITU Recommendation Q.773: ‘Transaction capabilities formats and encoding’,
June 1997
15 WELCH, S.: ‘Signalling in telecommunications networks’, IEE Telecommu-
nications Series 6 (Peter Peregrinus Ltd, 1981)
16 FLOOD, J.E. (Ed.): ‘Telecommunication networks’ (IEE Publications, 1997,
2nd edn)
17 HILLS, M.T.: ‘Telecommunications switching principles’ (George Allen &
Unwin, 1984)
18 MANTERFIELD, R.: ‘Common channel signalling’ (Peter Peregrinus Ltd,
1991)
19 ITU Recommendation: X.200: ‘Open system interconnection: the basic
reference model’, July 1994
20 LAW, B., and WADSWORTH, C.A.: ‘SS7 signalling system no. 7: message
transfer part’ Journal of the Institution of British Telecom Engineers, April
1988
21 ITU Recommendation Q.721: ‘Functional description of the signalling system
no. 7: telephone user part (TUP), Blue Book, 1988
22 ETSI Standard: EN 300 195-1 v2.1.1 (2001–06) ISDN supplementary service
interactions; DSS no. 1 protocol
23 CLARKE, P.G., and WADSWORTH, C.A.: ‘CCITT signalling system no. 7:
signalling connection control part’, British Telecommunication Engineering
Journal, 1988, 7(1)
24 ITU Recommendation Q.711: ‘Functional description of the signalling
connection control part’, March 2001
25 BALE, M.C.: ‘Signalling in the intelligent network’, in Dufour, I. (Ed.):
‘Network intelligence’, Chapman & Hall, 1997
26 JOHNSON, T.W., LAW, B., and ANIUS, P.: ‘CCITT signalling system no. 7:
transaction capabilities’, British Telecommunication Engineering Journal,
1988, 7(1).
27 ITU Recommendation: Q.1211: ‘Introduction to intelligent network capability
set 1’, March 1993
28 ITU Recommendation: Q.1213: ‘Global functional plane for intelligent network
capability set 1’, October 1995
29 TURNER, G.B.: ‘Service creation’, in Dufour, I. (Ed.): ‘Network intelligence’
Chapman & Hall, 1997
30 ITU Recommendation: Q.1215: ‘Physical plane for intelligent network
capability set 1’, October 1995
31 ETSI/3GPP Standard: ETSI TS 101441 V7.7.0 (2001–12): Digital cellular
telecommunications system (phase 2þ); CAMEL phase 2, stage 2 (3GPP TS
03.78 version 7.7.0 Release 1998).
32 ITU Recommendation: Q.1224: ‘Distributed functional plane for intelligent
network capability set 2’, September 1997
33 ITU Recommendation: Q.1222: ‘Service plane for intelligent network
capability set 2’, September 1997
34 ITU Recommendation: Q.1223: ‘Global functional plane intelligent network
capability set 2’, September 1997
 References 175

35 UK Office of Telecommunications: ‘Network Interoperability Consultation

Committee: PNO-ISC Specification Number 007 (ISUP). 2001
36 The Parlay Technical Team: ‘Parlay APIs 2.1 Generic call control service
interfaces’, March 2000. http/www.parlay.org
37 ETSI Standard: ES 201 915-1 V1.1.1 (2001–12) ETSI Standard – Open Service
Access; Application programming interface (The Parlay Group & ETSI): DES/
SPAN-120070-1. 2001
38 ETSI Standard: TR 121 905: ‘Universal mobile telecommunications system
(UMTS). Vocabulary for 3GPP specifications (3GPP TR 21.905)’ 1999
39 TINA business model and reference points. Version 4.0 TINA-C Deliverable,
May 1997. http/www/tinac.com
40 ITU Recommendation: Q.1204: ‘Intelligent network distributed functional
plane’, March 1993
41 ITU Recommendation: Q.1241: ‘Introduction to intelligent network capability
set 4’, July 2001
42 ITU Recommendation: Q.1244: ‘Distributed functional plane for intelligent
network capability set 4’, July 2001
43 PETRACK, S., and CONROY, L.: ‘IETF RFC 2848. The PINT service
protocol: Extensions to SIP and SDP for IP access to telephone call services’
2000
44 FAYNBERG, I., GATO, J., LU, H., and SLUTMAN, L.: ‘SPIRITS protocol
requirements’. IETF Internet Draft, expires August 2002. Note that this is only
referenced as work in progress at the time of publication. It is necessary to
check with www.ietf.org/internet-drafts or http/www.rfc.net/std1.html to
ascertain the current situation with SPIRITS work.
45 CLAPTON, A. (Ed.): ‘Future mobile networks: 3G and beyond’. (IEE Books,
2001).
46 ITU Recommendation: Q.1248: ‘Interface recommendation for intelligent
network capability set 4’ (a series of seven documents), July 2001 (pre-
publication version)
47 ETSI/3GPP Standard: ETSI TS 123 078 V4.3.0 (2001–12): ‘Digital cellular
telecommunications system (Phase 2þ) (GSM); Universal mobile telecommu-
nications system (UMTS); Customised applications for mobile network
enhanced logic (CAMEL) Phase 3 – Stage 2 (3GPP TS 23.078 version 4.3.0
Release 4)’
48 HOLMES, P.E., and STREET, N.J.: ‘Operational support systems for an
intelligent network’ BT Technology Journal, 1995, 13(2)
49 ADAMS, E., and WILLETS, K.: ‘The lean communications provider: surviving
the shakeout through service management excellence’ (McGraw-Hill, July 1996)
50 ITU Recommendation: M.3010: ‘Principles for a telecommunications manage-
ment network (TMN)’, February 2000
51 ITU Recommendation: M.3040 ‘TMN management functions’, February 2000
52 SWALE, R. (Ed.): ‘Voice Over IP: Systems and Solutions’. (IEE Books, 2001)
 Appendix 1: List of ITU-T
IN documents

The ITU-T Q-series of recommendations covers switching and signalling in

telecommunications networks. Within the Q-series, recommendations numbered
Q.12. . .. are for intelligent network standards, and the documents between Q.13. . ..
and Q.19. . .. are also IN-related. The I-series covers ISDN, so the dual Q and I
numbers on some documents indicate collaborative work between the IN and ISDN
working groups in ITU-T. The list below gives the dates of the documents that are
in force at the time of publication of this book.
Q.1200 (09/97) General series on intelligent network recommendation structure
Q.1201 (10/92) Principles of IN architecture
Q.1202/I.328 (09/97) IN – service plane architecture
Q.1203/I.329 (09/97) IN – global functional plane architecture
Q.1204 (03/93) IN – distributed functional plane architecture
Q.1205 (03/93) IN – physical plane architecture
Q.1208 (09/97) General aspects of the IN application protocol
Q.1210 (10/95) Q.1210-series IN recommendation structure
Q.1211 (03/93) Introduction to IN CS-1
Q.1213 (10/95) Global functional plane for IN CS-1
Q.1214 (10/95) Distributed functional plane for IN CS-1
Q.1215 (10/95) Physical plane for IN CS-1
Q.1218 (10/95) Interface recommendation for IN CS-1
Q.1218 Add. (09/97) Definition for two new contexts in the SDF data model
Q.1219 (04/94) IN user’s guide for CS-1
Q.1220 (09/97) Q.1220-series IN CS-2 recommendation structure

Q.1221 (09/97) Introduction to IN CS-2

Q.1222 (09/97) Service plane for IN CS-2
Q.1223 (09/97) Global functional plane for IN CS-2
Q.1224 (09/97) Distributed functional plane for IN CS-2
Q.1225 (09/97) Physical plane for IN CS-2
Q.1228 (09/97) Interface recommendation for IN CS-2
Q.1229 (03/99) IN user’s guide for CS-2
178 Intelligent networks – principles and applications

Q.1231 (12/99) Introduction to IN CS-3

Q.1236 (12/99) IN CS-3 – management information model, requirements and
methodology
Q.1237 (06/00) Extensions to IN CS-3 in support of B-ISDN
Q.1238 (to be published) Interface recommendation for IN CS-3
Q.1238.1 (06/00) Common aspects
Q.1238.2 (06/00) Interface recommendation for IN CS-3: SCF-SSF interface
Q.1238.3 (06/00) Interface recommendation for IN CS-3: SCF-SRF interface
Q.1238.4 (06/00) Interface recommendation for IN CS-3: SCF-SDF interface
Q.1238.5 (06/00) Interface recommendation for IN CS-3: SDF-SDF interface
Q.1238.6 (06/00) Interface recommendation for IN CS-3: SCF-SCF interface
Q.1238.7 (06/00) Interface recommendation for IN CS-3: SCF-CUSF interface

Q.1241 (08/01)* Introduction to IN CS-4

Q.1244 (08/01)* Distributed functional plane for IN CS-4
Q.1248.1 (08/01)* Interface recommendation for IN CS-4 – Common aspects
Q.1248.2 (08/01)* Interface recommendation for IN CS-4: SCF-SSF interface
Q.1248.3 (08/01)* Interface recommendation for IN CS-4: SCF-SRF interface
Q.1248.4 (08/01)* Interface recommendation for IN CS-4: SCF-SDF interface
Q.1248.5 (08/01)* Interface recommendation for IN CS-4: SDF-SDF interface
Q.1248.6 (08/01)* Interface recommendation for IN CS-4: SCF-SCF interface
Q.1248.7 (08/01)* Interface recommendation for IN CS-4: SCF-CUSF interface
Q.1290 (05/98) Glossary of terms used in the definition of intelligent networks

Q.1300 (10/95) TASC Telecommunication applications for switches and computers.

General overview
Q.1301 (10/95) TASC architecture
Q.1302 (10/95) TASC functional services
Q.1303 (10/95) TASC management: architecture, methodology and requirements
Q.1400 (02/95) Architecture framework for the development of signalling and
OA&M protocols using OSI concepts

Q.1521 (06/00) Requirements on underlying networks and signalling protocols to

support UPT
Q.1531 (06/00) UPT security requirements for Service Set 1
Q.1551 (06/97) Application of CS-1 INAP UPT Service Set 1
Q.1600 (09/97) SS7 – Interaction between ISUP and INAP
Q.1600bis (12/99) SS7 – Interaction between ISUP and INAP: test suite structure
Q.1601 (12/99) SS7 – Interaction between N-ISDN and INAP CS2
Q.1701 (03/99) Framework for IMT-2000 networks
Q.1711 (03/99) Network functional model for IMT-2000
Q.1721 (06/00) Information flows for IMT-2000 capability set 1
Q.1731 (06/00) Radio-technology-independent requirements for IMT-2000 layer 2
radio interface
Q.1751 (06/00) Inter-network signalling requirements for IMT-2000 capability set 1
Q.1901 (06/00) Bearer independent call control protocol.

* These are new recommendations, approved pre-published versions available (July 2001)
Information in this Appendix is derived from the ITU-T web-site (http://www.itu.int) on 18
December 2001.
 Appendix 2: Additional CS-1
INAP operations

Additional INAP operations specified in ITU-T CS-1

The following is a list of the 26 operations included in ITU-T CS-1, but excluded
from ETSI Core INAP and ITU-T CS-1R:

IN operation

analysedInformation
analyseInformation
CollectedInformation
CancelStatusReportRequest
HoldCallInNetwork
OAnswer
OCalledPartyBusy
ODisconnect
OMidCall
ONoAnswer
OriginationAttemptAuthorized
RequestCurrentStatusReport
RequestEveryStatusChangeReport
RequestFirstStatusMatchReport
RouteSelectFailure
SelectFacility
SelectRoute
StatusReport
TAnswer
TCalledPartyBusy
TDisconnect
TMidCall
TNoAnswer
Query
SdfResponse
UpdateData
 Appendix 3: Example SS7 message
sequence coding

SS7 SCCP and TCAP support the non-connection-oriented messaging transactions

required by IN. The operation of SCCP and TCAP were outlined in Chapter 3 and
we now illustrate the message coding details for a typical IN transaction. In fact the
particular message flow illustrated here is for a mobile call involving a CAP
(CAMEL application protocol) request. As we discussed earlier, there is strong
equivalence between the CAP and the INAP protocols. There are some application
parameter differences, but the behaviour of SCCP and TCAP in support of the
messaging is identical in both cases and so the illustration here would be similar for
an INAP transaction scenario.
We show the messaging required for a call where the IN service logic requests
the SSF (the gsmSSF in this case) for end of call notification. This might be used,
for instance, where the SCF needs to log the call’s duration, perhaps in order to
decrement the customer’s prepaid account by the cost of the call.
Figure A3.1 shows the seven messages in the example sequence, and transcripts
of the individual messages themselves are shown on the following pages.
The first event shown is the receipt by the gsmSCF of the CAP I_DP in a TCAP
Begin message. The accompanying parameters transferred are:

^ service key
^ called party number
^ calling party number
^ calling party’s category
^ IP SSP capabilities
^ bearer capability
^ event type BCSM
^ IMSI (international mobile subscriber identity)
^ Call reference number
^ MSC address.
182 Intelligent networks – principles and applications

Figure A3.1 TCAP messages illustrated

The first seven parameters are taken directly from Core INAP (details in Chapter 2)
and the last three are CAMEL-specific.
About 500 ms after the gsmSCF receives the I_DP it returns a Request Report
BCSM Operation (500 ms is a typical duration for the processing on an IN request
in an SCF). This is the second message of the sequence and it is a TCAP Continue
message. It includes a single IN parameter that gives details of the particular events
for which the service logic requires the SSP to monitor. Shortly following this is a
third message, in the same direction as the first, and this is another TCAP Continue,
carrying an IN Continue operation, requiring no parameters. The purpose of this
message is to instruct the gsmSSP to continue with normal call processing. In this
case the gsmSSP will use the digits dialled by the caller to instigate the routing of
the call towards the required destination.
Eight seconds later an ‘answer’ notification is sent from the gsmSSF to the
gsmSCF in another Continue message and the call enters the ‘speech’ phase. After
about 38 s a ‘disconnect’ is generated by the first leg of the call and the gsmSSF
consequently sends another TCAP Continue message carrying another Event
Report BCSM operation.
The 6th message carries a Release Call operation from the gsmSCF, and this is in
another TCAP Continue, so the transaction remains open until the 7th message,
where the gsmSCF sends an explicit TCAP End message (with no CAP operation
this time).
The following transcripts for the seven messages of this example call sequence
illustrate the overheads involved at each SS7 layer. We show the actual
 Appendix 3: Example SS7 message sequence coding 183

hexadecimal code for each message followed by a decomposition and interpretation

for this code. Each message has MTP and SCCP header information as well as
TCAP transaction, dialogue and component headers as a prelude to the actual IN
operations and parameters.
MTP headers contain the sequence numbers, the point codes and the service
information octet, which holds the service indicator (value ‘3’ for SCCP in these
messages) and the network indicator (value ‘2’ for national in these messages).
SCCP message headers include the called and calling SCCP addresses with
indicators giving information on the particular assemblies of global titles,
subsystem numbers and point code addressing used for each message. The use of
these addressing options was outlined in Chapter 3. The message sequences also
show how the TCAP headers are used. Details of interpretation for TCAP are in
Reference 14.

Message No. 1:
CAMEL I_DP operation þ parameters in a TCAP BEGIN message
gsmSSF ? gsmSCF
Time: 18:47:53,853

HEX 0 1 2 3 4 5 6 7 8 9 A B C D E F

0 cd 97 3f 83 20 00 0c b0 09 81 03 0e 12 0b 52 05
10 00 12 04 21 43 65 87 19 32 04 43 30 00 05 7b 62
20 79 48 04 2b 09 03 00 6b 1e 28 1c 06 07 00 11 86
30 05 01 01 01 a0 11 60 0f 80 02 07 80 a1 09 06 07
40 04 00 00 01 00 32 00 6c 51 a1 4f 02 01 01 02 01
50 00 30 47 80 02 00 de 82 08 04 90 21 43 65 11 11
60 11 83 07 83 13 88 88 11 11 01 85 01 0a 88 01 04
70 bb 05 80 03 90 90 a3 9c 01 0c 9f 32 08 72 02 03
80 01 00 68 90 f3 9f 36 08 03 01 0c 3e 00 00 01 fe
90 9f 37 07 22 22 22 22 22 22 22

Translation for Message No. 1:

BITMASK ID name Comment or value

MTP header information:

-1001101 Backward Sequence Number 77
1------- Backward Indicator Bit 1
-0010111 Forward Sequence Number 23
1------- Forward Indicator Bit 1
--111111 Length Indicator 63
00------ Spare 0
----0011 Service Indicator SCCP
--00---- Sub-service: Priority Spare/priority 0 (USA
only)
10------ Sub-service: Network Ind. National message
**b14*** Destination Point Code 32
184 Intelligent networks – principles and applications

BITMASK ID name Comment or value

**b14*** Originating Point Code 48

SCCP message header:
1011---- Signalling Link Selection 11
00001001 SCCP Message Type 9 (Unitdata)
----0001 Protocol Class Class 1
1000---- Message Handling Return message on
error
00000011 Pointer to parameter 3
00001110 Pointer to parameter 14
00010010 Pointer to parameter 18
Called address parameter:
00001011 Parameter Length 11
-------0 Point Code Indicator PC absent
------1- Sub-system No. Indicator SSN present
--0100-- Global Title Indicator Has translation,n-
plan,encoding
scheme, nature of
address
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
00000101 Sub-system number CAP
00000000 Translation Type Not used
----0010 Encoding Scheme BCD, even number of
digits
0001---- Numbering Plan ISDN/Telephony
(E.164/E.163)
-0000100 Nat. of Address Indicator International number
0------- Spare 0
***B6*** Called Address Signals 123456789123
Calling address parameter:
00000100 Parameter Length 4
-------1 Point Code Indicator PC present
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Calling Party SPC 48
00------ Spare 0
00000101 Sub-system number CAP
Data parameter
01111011 Parameter Length 123
**B123** Data 62 79 48 04 2b 09 03 00
6b 1e28
TCAP transaction portion
Begin:
01100010 Tag
01111001 Length 121
 Appendix 3: Example SS7 message sequence coding 185

BITMASK ID name Comment or value

1 Origination Transaction ID
01001000 Tag
00000100 Length 4
***B4*** Orig Trans ID 722010880
2 Dialogue Portion
01101011 Tag
00011110 Length 30
2.1 Dialogue External
00101000 Tag
00011100 Length 28
2.1.1 Dialogue Object ID
00000110 Tag
00000111 Length 7
00000000 Authority CCITT Recommendation
00010001 Name Form q
10000110 Rec Number 7
00000101 Rec Number 73
00000001 AS 1
00000001 Dialog-AS Dialogue PDU
00000001 Version 1
2.1.2 Dialogue Single ASN1
10100000 Tag
00010001 Length 17
2.1.2.1 Dialogue Request
01100000 Tag
00001111 Length 15
2.1.2.1.1 Protocol Version
10000000 Tag
00000010 Length 2
00000111 UnusedBits 7
1------- Version 1 Yes
-0000000 Filler 0
2.1.2.1.2 Application Context Name (ACN)
10100001 Tag
00001001 Length 9
2.1.2.1.2.1 ACN Object ID
00000110 Tag
00000111 Length 7
0000---- ObjId CCITT
----0100 Organisation Identified-organisa-
tion
00000000 ETSI
00000000 Domain Mobile Domain
00000001 Mobile Sub-domain GSM-Network
00000000 Common Component ID AC-ID
00110010 Application Context cap-gsmSSF-to-
gsmSCF
00000000 Version Version1
3 TCAP Component Portion
01101100 Tag
01010001 Length 81
186 Intelligent networks – principles and applications

BITMASK ID name Comment or value

3.1 Invoke
10100001 Tag
01001111 Length 79
3.1.1 Invoke ID
00000010 Tag
00000001 Length 1
00000001 Invoke ID Value 1
3.1.2 CAP Local Operation
00000010 Tag
00000001 Length 1
00000000 Operation Code Initial DP
3.1.3 Parameter Sequence
00110000 Tag
01000111 Length 71
3.1.3.1 Service Key
10000000 Tag])
00000010 Length 2
***B2*** Service Key 222
3.1.3.2 Called Party Number
10000010 Tag
00001000 Length 8
0------- Odd/Even Indicator Even number of
address signals
-0000100 Nature of Addr. Indicator International number
1------- Internal Netw. No. Indic. rout.2 internal net.-
not allwd
-001---- Numbering Plan Indicator ISDN numbering plan
----0000 Spare Spare
***B6*** cld.pty.no. Address Signal 123456111111
3.1.3.3 Calling Party Number
10000011 Tag
00000111 Length 7
1------- Odd/Even Indicator Odd number of address
signals
-0000011 Nature of Addr. Indicator National [signifi-
cant] number
0------- Number Incomplete Indic. Complete
-001---- Numbering Plan Indicator ISDN numbering plan
----00-- Addr. Present. Restr. Ind. Presentation allowed
------11 Screening Indicator Network provided
**b36*** clg.pty.no. Address Signal 888811111
3.1.3.4 Calling Party’s Category
10000101 Tag
00000001 Length 1
00001010 CallingPartysCategory Ordinary calling sub-
scriber
3.1.3.5 IPSSP Capabilities
10001000 Tag
00000001 Length 1
0------- End of Standard Part End of standard part
-00----- Reserved 0
 Appendix 3: Example SS7 message sequence coding 187

BITMASK ID name Comment or value

---0---- Voice Announ.Gen.from Text Not supported

----0--- Voice Inf.via Voice R. Not supported
-----1-- Voice Inf.via Speech R. Supported
------0- Voice Back Y/N Not supported
-------0 IP Routing Address Y/N Not supported
3.1.3.6 Bearer Capability Constr
10111011 Tag
00000101 Length 5
3.1.3.6.1 Bearer Cap
10000000 Tag
00000011 Length 3
1------- Extension Indicator Fixed to last octet
-00----- Coding Standard CCITT standardised
coding
---10000 Info. Trans. Cap 3.1 kHz audio
1------- Extension Indicator Last octet
-00----- Transfer Mode Circuit mode
---10000 Info. Transfer Rate 64 kbit/s
1------- Extension Indicator Last octet
-01----- Layer User Info. Layer 1
---00011 User Info. L1 Protocol Rec. G711, A-Law
3.1.3.7 Event Type BCSM
10011100 Tag
00000001 Length 1
00001100 Event Type Bcsm termAttemptAuthor-
ised
3.1.3.8 IMSI (International Mobile Sub-
scriber Identity)
***B2*** Tag
00001000 Length 8
**b60*** MCC þ MNC þ MSIN 272030100086093
1111---- FILLER 15
3.1.3.9 Call Reference Number
***B2*** Tag
00001000 Length 8
***B8*** Call Reference 03 01 0c 3e 00 00 01 fe
3.1.3.10 MSC Address
***B2*** Tag (CONT P [55])
00000111 Length 7
***B7*** MSC Address 22 22 22 22 22 22 22
188 Intelligent networks – principles and applications

Message No. 2:
CAMEL Request Report BCSM operation þ parameters in a TCAP
Continue message
gsmSSF / gsmSCF
Time: 18:47:54,409

HEX 0 1 2 3 4 5 6 7 8 9 A B C D E F

0 98 d0 3f 83 30 00 08 90 09 01 03 07 0b 04 43 30
10 00 05 04 43 20 00 05 6f 65 6d 48 04 00 6a b6 e3
20 49 04 2b 09 03 00 6b 2a 28 28 06 07 00 11 86 05
30 01 01 01 a0 1d 61 1b 80 02 07 80 a1 09 06 07 04
40 00 00 01 00 32 00 a2 03 02 01 00 a3 05 a1 03 02
50 01 00 6c 33 a1 31 02 01 02 02 01 17 30 29 a0 27
60 30 0b 80 01 0f 81 01 01 a2 03 80 01 02 30 0b 80
70 01 11 81 01 00 a2 03 80 01 01 30 0b 80 01 11 81
80 01 00 a2 03 80 01 02

Translation for Message No. 2:

BITMASK ID name Comment or value

MTP header information:

-0011000 Backward Sequence Number 24
1------- Backward Indicator Bit 1
-1010000 Forward Sequence Number 80
1------- Forward Indicator Bit 1
--111111 Length Indicator 63
00------ Spare 0
----0011 Service Indicator SCCP
--00---- Sub-service: Priority Spare/priority 0 (USA
only)
10------ Sub-service: Network Ind National message
**b14*** Destination Point Code 48
**b14*** Originating Point Code 32
SCCP message header:
1001---- Signalling Link Selection 9
00001001 SCCP Message Type 9 (Unitdata)
----0001 Protocol Class Class 1
0000---- Message Handling No special options
00000011 Pointer to parameter 3
00000111 Pointer to parameter 7
00001011 Pointer to parameter 11
Called address parameter:
00000100 Parameter Length 4
-------1 Point Code Indicator PC present
------1- Subsystem No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
 Appendix 3: Example SS7 message sequence coding 189

BITMASK ID name Comment or value

-1------ Routing Indicator Route on DPC þ Sub-

system No.
0------- For national use 0
**b14*** Called Party SPC 48
00------ Spare 0
00000101 Sub-system number CAP
Calling address parameter:
00000100 Parameter Length 4
-------1 Point Code Indicator PC present
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Calling Party SPC 32
00------ Spare 0
00000101 Sub-system number CAP
Data parameter
01101111 Parameter length 111
**B111** Data 65 6d 48 04 00 6a b6 e3
49 042b
TCAP transaction portion
Continue:
01100101 Tag
01101101 Length 109
1 Origination Transaction ID
01001000 Tag
00000100 Length 4
***B4*** Orig Trans ID 6993635
2 Destination Transaction ID
01001001 Tag
00000100 Length 4
***B4*** Dest Trans ID 722010880
3 Dialogue Portion
01101011 Tag
00101010 Length 42
3.1 Dialogue External
00101000 Tag
00101000 Length 40
3.1.1 Dialogue Object ID
00000110 Tag
00000111 Length 7
00000000 Authority CCITT Recommendation
00010001 Name Form q
10000110 Rec Number 7
00000101 Rec Number 73
00000001 AS 1
00000001 Dialog-AS Dialogue PDU
00000001 Version 1
190 Intelligent networks – principles and applications

BITMASK ID name Comment or value

3.1.2 Dialogue single ASN1

10100000 Tag
00011101 Length 29
3.1.2.1 Dialogue Response
01100001 Tag
00011011 Length 27
3.1.2.1.1 Protocol Version
10000000 Tag
00000010 Length 2
00000111 UnusedBits 7
1------- Version 1 Yes
-0000000 Filler 0
3.1.2.1.2 Application Context Name
10100001 Tag
00001001 Length 9
3.1.2.1.2.1 ACN Object ID
00000110 Tag (UNIV P Obj Identi-
fier)
00000111 Length 7
0000---- ObjId CCITT
----0100 Organisation Identified-organisa-
tion
00000000 ETSI
00000000 Domain Mobile Domain
00000001 Mobile Sub-domain GSM-Network
00000000 Common Component ID AC-ID
00110010 Application Context cap-gsmSSF-to-
gsmSCF
00000000 Version Version1
3.1.2.1.3 Result
10100010 Tag
00000011 Length 3
3.1.2.1.3.1 Associate Result
00000010 Tag
00000001 Length 1
00000000 Associate Result Accepted
3.1.2.1.4 Result Source Diagnostic
10100011 Tag
00000101 Length 5
3.1.2.1.4.1 Dialogue Service User
10100001 Tag
00000011 Length 3
3.1.2.1.4.1.1 Dialogue Service User Value
00000010 Tag
00000001 Length 1
00000000 Dialogue Service User Null
4 TCAP Component Portion
01101100 Tag
00110011 Length 51
4.1 Invoke
10100001 Tag
 Appendix 3: Example SS7 message sequence coding 191

BITMASK ID name Comment or value

00110001 Length 49
4.1.1 Invoke ID
00000010 Tag
00000001 Length 1
00000010 Invoke ID value 2
4.1.2 CAP Local Operation
00000010 Tag
00000001 Length 1
00010111 Operation Code Request Report BCSME-
vent
4.1.3 Parameter Sequence
00110000 Tag
00101001 Length 41
4.1.3.1 BCSM Events
10100000 Tag
00100111 Length 39
4.1.3.1.1 BCSM Event
00110000 Tag
00001011 Length 11
4.1.3.1.1.1 Event Type BCSM
10000000 Tag
00000001 Length 1
00001111 Event Type Bcsm tAnswer
4.1.3.1.1.2 Monitor Mode
10000001 Tag
00000001 Length 1
00000001 Monitor Mode Notify and continue
4.1.3.1.1.3 Leg Id Constr
10100010 Tag
00000011 Length 3
4.1.3.1.1.3.1 Sending Side ID
10000000 Tag
00000001 Length 1
00000010 Sending Side Id Leg2
4.1.3.1.2 BCSM Event
00110000 Tag
00001011 Length 11
4.1.3.1.2.1 Event Type BCSM
10000000 Tag
00000001 Length 1
00010001 Event Type Bcsm tDisconnect
4.1.3.1.2.2 Monitor Mode
10000001 Tag
00000001 Length 1
00000000 Monitor Mode Interrupted
4.1.3.1.2.3 Leg Id Constr
10100010 Tag
00000011 Length 3
4.1.3.1.2.3.1 Sending Side ID
10000000 Tag
00000001 Length 1
192 Intelligent networks – principles and applications

BITMASK ID name Comment or value

00000001 Sending Side Id Leg1

4.1.3.1.3 BCSM Event
00110000 Tag
00001011 Length 11
4.1.3.1.3.1 Event Type BCSM
10000000 Tag
00000001 Length 1
00010001 Event Type Bcsm tDisconnect
4.1.3.1.3.2 Monitor Mode
10000001 Tag
00000001 Length 1
00000000 Monitor Mode Interrupted
4.1.3.1.3.3 Leg Id Constr
10100010 Tag
00000011 Length 3
4.1.3.1.3.3.1 Sending Side ID
10000000 Tag
00000001 Length 1
00000010 Sending Side Id Leg2
 Appendix 3: Example SS7 message sequence coding 193

Message No. 3:
CAMEL Continue operation in a TCAP Continue message
gsmSSF / gsmSCF
Time: 18:47:54,416

HEX 0 1 2 3 4 5 6 7 8 9 A B C D E F

0 98 d1 2d 83 30 00 08 90 09 01 03 07 0b 04 43 30
10 00 05 04 43 20 00 05 18 65 16 48 04 00 6a b6 e3
20 49 04 2b 09 03 00 6c 08 a1 06 02 01 03 02 01 1f

Translation for Message No. 3:

BITMASK ID name Comment or value

MTP header information:

-0011000 Backward Sequence Number 24
1------- Backward Indicator Bit 1
-1010001 Forward Sequence Number 81
1------- Forward Indicator Bit 1
--101101 Length Indicator 45
00------ Spare 0
----0011 Service Indicator SCCP
--00---- Sub-service: Priority Spare/priority 0 (USA
only)
10------ Sub-service: Network Ind. National message
**b14*** Destination Point Code 48
**b14*** Originating Point Code 32
SCCP message header:
1001---- Signalling Link Selection 9
00001001 SCCP Message Type 9 (UDT)
----0001 Protocol Class Class 1
0000---- Message Handling No special options
00000011 Pointer to parameter 3
00000111 Pointer to parameter 7
00001011 Pointer to parameter 11
Called address parameter:
00000100 Parameter Length 4
-------1 Point Code Indicator PC present
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Called Party SPC 48
00------ Spare 0
00000101 Sub-system number CAP
Calling address parameter:
00000100 Parameter Length 4
194 Intelligent networks – principles and applications

BITMASK ID name Comment or value

-------1 Point Code Indicator PC present

------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Calling Party SPC 32
00------ Spare 0
00000101 Sub-system number CAP
Data parameter
00011000 Parameter length 24
**B24*** Data 65 16 48 04 00 6a b6 e3
49 042b
TCAP transaction portion
Continue:
01100101 Tag
00010110 Length 22
1 Origination Transaction ID
01001000 Tag
00000100 Length 4
***B4*** Orig Trans ID 6993635
2 Destination Transaction ID
01001001 Tag
00000100 Length 4
***B4*** Dest Trans ID 722010880
3 TCAP Component Portion
01101100 Tag
00001000 Length 8
3.1 Invoke
10100001 Tag
00000110 Length 6
3.1.1 Invoke ID
00000010 Tag
00000001 Length 1
00000011 Invoke ID value 3
3.1.2 CAP Local Operation
00000010 Tag
00000001 Length 1
00011111 Operation Code Continue
 Appendix 3: Example SS7 message sequence coding 195

Message No. 4:
CAMEL Event Report BCSM operation þ parameters in a TCAP
Continue message
gsmSSF ? gsmSCF
time: 18:48:02,562

HEX 0 1 2 3 4 5 6 7 8 9 A B C D E F

0 f2 b2 38 83 20 00 0c b0 09 81 03 05 07 02 42 05
10 02 42 05 27 65 25 48 04 2b 09 03 00 49 04 00 6a
20 b6 e3 6c 17 a1 15 02 01 02 02 01 18 30 0d 80 01
30 0f a3 03 81 01 02 a4 03 80 01 01

Translation for Message No. 4:

BITMASK ID name Comment or value

MTP header information:

-1110010 Backward Sequence Number 114
1------- Backward Indicator Bit 1
-0110010 Forward Sequence Number 50
1------- Forward Indicator Bit 1
--111000 Length Indicator 56
00------ Spare 0
----0011 Service Indicator SCCP
--00---- Sub-service: Priority Spare/priority 0 (USA
only)
10------ Sub-service: Network Ind. National message
**b14*** Destination Point Code 32
**b14*** Originating Point Code 48
SCCP message header:
1011---- Signalling Link Selection 11
00001001 SCCP Message Type 9
----0001 Protocol Class Class 1
1000---- Message Handling Return message on
error
00000011 Pointer to parameter 3
00000101 Pointer to parameter 5
00000111 Pointer to parameter 7
Called address parameter:
00000010 Parameter Length 2
-------0 Point Code Indicator PC absent
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
00000101 Sub-system number CAP
196 Intelligent networks – principles and applications

BITMASK ID name Comment or value

Calling address parameter:

00000010 Parameter Length 2
-------0 Point Code Indicator PC absent
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
00000101 Sub-system number CAP
Data parameter
00100111 Parameter Length 39
**B39*** Data 65 25 48 04 2b 09 03 00
49 0400
TCAP transaction portion
Continue:
01100101 Tag
00100101 Length 37
1 Origination Transaction ID
01001000 Tag
00000100 Length 4
***B4*** Orig Trans ID 722010880
2 Destination Transaction ID
01001001 Tag
00000100 Length 4
***B4*** Dest Trans ID 6993635
3 TCAP Component Portion
01101100 Tag
00010111 Length 23
3.1 Invoke
10100001 Tag
00010101 Length 21
3.1.1 Invoke ID
00000010 Tag
00000001 Length 1
00000010 Invoke ID value 2
3.1.2 Local Operation
00000010 Tag
00000001 Length 1
00011000 Operation Code Event Report BCSM
3.1.3 Parameter Sequence
00110000 Tag (UNIV C sequence
(of))
00001101 Length 13
3.1.3.1 Event Type BCSM
10000000 Tag
00000001 Length 1
00001111 Event Type Bcsm tAnswer
3.1.3.2 Leg Id Constr
10100011 Tag
00000011 Length 3
 Appendix 3: Example SS7 message sequence coding 197

BITMASK ID name Comment or value

3.1.3.2.1 Receiving Side Id

10000001 Tag
00000001 Length 1
00000010 Receiving Side Id Leg2
3.1.3.3 Misc Call Info
10100100 Tag
00000011 Length 3
3.1.3.3.1 Message Type
10000000 Tag
00000001 Length 1
00000001 Message Type Notification
198 Intelligent networks – principles and applications

Message No. 5:
CAP Event Report BCSM operation þ parameters in a TCAP
Continue message
gsmSSF ? gsmSCF
18:48:40,486

HEX 0 1 2 3 4 5 6 7 8 9 A B C D E F

0 85 a9 3f 83 20 00 0c b0 09 81 03 05 07 02 42 05
10 02 42 05 2f 65 2d 48 04 2b 09 03 00 49 04 00 6a
20 b6 e3 6c 1f a1 1d 02 01 03 02 01 18 30 15 80 01
30 11 a2 06 ac 04 80 02 80 90 a3 03 81 01 01 a4 03
40 80 01 00

Translation for Message No. 5:

BITMASK ID name Comment or value

MTP header information:

-0000101 Backward Sequence Number 5
1------- Backward Indicator Bit 1
-0101001 Forward Sequence Number 41
1------- Forward Indicator Bit 1
--111111 Length Indicator 63
00------ Spare 0
----0011 Service Indicator SCCP
--00---- Sub-service: Priority Spare/priority 0 (USA
only)
10------ Sub-service: Network Ind National message
**b14*** Destination Point Code 32
**b14*** Originating Point Code 48
SCCP message header:
1011---- Signalling Link Selection 11
00001001 SCCP Message Type 9
----0001 Protocol Class Class 1
1000---- Message Handling Return message on
error
00000011 Pointer to parameter 3
00000101 Pointer to parameter 5
00000111 Pointer to parameter 7
Called address parameter:
00000010 Parameter Length 2
-------0 Point Code Indicator PC absent
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
00000101 Sub-system number CAP
 Appendix 3: Example SS7 message sequence coding 199

BITMASK ID name Comment or value

Called address parameter:

00000010 Parameter Length 2
-------0 Point Code Indicator PC absent
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
00000101 Sub-system number CAP
Data parameter
00101111 Parameter length 47
**B47*** Data 65 2d 48 04 2b 09 03 00
49 0400
TCAP transaction portion
Continue:
01100101 Tag
00101101 Length 45
1 Origination Transaction ID
01001000 Tag
00000100 Length 4
***B4*** Orig Trans ID 722010880
2 Destination Transaction ID
01001001 Tag
00000100 Length 4
***B4*** Dest Trans ID 6993635
3 Component Portion
01101100 Tag
00011111 Length 31
3.1 Invoke
10100001 Tag
00011101 Length 29
3.1.1 Invoke ID
00000010 Tag
00000001 Length 1
00000011 Invoke ID Value 3
3.1.2 Local Operation
00000010 Tag
00000001 Length 1
00011000 Operation Code Event Report BCSM
3.1.3 Parameter Sequence
00110000 Tag
00010101 Length 21
3.1.3.1 Event Type BCSM
10000000 Tag
00000001 Length 1
00010001 Event Type Bcsm tDisconnect
3.1.3.2 Event Specific Info Bcsm
10100010 Tag
00000110 Length 6
200 Intelligent networks – principles and applications

BITMASK ID name Comment or value

3.1.3.2.1 T Disconnect Specific Info

10101100 Tag
00000100 Length 4
3.1.3.2.1.1 Release Cause
10000000 Tag
00000010 Length 2
1------- Extension Indication 1 Last octet
-00----- Coding Standard CCITT standard
---0---- Spare Spare
----0000 Location User
1------- Extension Indication 2 Last octet
-0010000 Cause Value Normal call clearing
3.1.3.3 Leg Id Constr
10100011 Tag
00000011 Length 3
3.1.3.3.1 Receiving Side Id
10000001 Tag
00000001 Length 1
00000001 Receiving Side Id Leg1
3.1.3.4 Misc Call Info
10100100 Tag
00000011 Length 3
3.1.3.4.1 Message Type
10000000 Tag
00000001 Length 1
00000000 Message Type Request
 Appendix 3: Example SS7 message sequence coding 201

Message No. 6:
CAP Release Call Operation in a TCAP Continue message
gsmSSF / gsmSCF
Time: 18:48:40,735

HEX 0 1 2 3 4 5 6 7 8 9 A B C D E F

0 ac 86 31 83 30 00 08 90 09 01 03 07 0b 04 43 30
10 00 05 04 43 20 00 05 1c 65 1a 48 04 00 6a b6 e3
20 49 04 2b 09 03 00 6c 0c a1 0a 02 01 04 02 01 16
30 04 02 80 90

Translation for Message No. 6:

BITMASK ID name Comment or value

MTP header information:

-0101100 Backward Sequence Number 44
1------- Backward Indicator Bit 1
-0000110 Forward Sequence Number 6
1------- Forward Indicator Bit 1
--110001 Length Indicator 49
00------ Spare 0
----0011 Service Indicator SCCP
--00---- Sub-service: Priority Spare/priority 0 (USA
only)
10------ Sub-service: Network Ind National message
**b14*** Destination Point Code 48
**b14*** Originating Point Code 32
SCCP message header:
1001---- Signalling Link Selection 9
00001001 SCCP Message Type 9
----0001 Protocol Class Class 1
0000---- Message Handling No special options
00000011 Pointer to parameter 3
00000111 Pointer to parameter 7
00001011 Pointer to parameter 11
Called address parameter:
00000100 Parameter Length 4
-------1 Point Code Indicator PC present
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Called Party SPC 48
00------ Spare 0
00000101 Sub-system number CAP
Calling address parameter:
202 Intelligent networks – principles and applications

BITMASK ID name Comment or value

00000100 Parameter Length 4

-------1 Point Code Indicator PC present
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Calling Party SPC 32
00------ Spare 0
00000101 Sub-system number CAP
Data parameter
00011100 Parameter length 28
**B28*** Data 65 1a 48 04 00 6a b6 e3
49 042b
TCAP transaction portion
Continue:
01100101 Tag
00011010 Length 26
1 Origination Transaction ID
01001000 Tag
00000100 Length 4
***B4*** Orig Trans ID 6993635
2 Destination Transaction ID
01001001 Tag
00000100 Length 4
***B4*** Dest Trans ID 722010880
3 Component Portion
01101100 Tag
00001100 Length 12
3.1 Invoke
10100001 Tag
00001010 Length 10
3.1.1 Invoke ID
00000010 Tag
00000001 Length 1
00000100 Invoke ID Value 4
3.1.2 Local Operation
00000010 Tag
00000001 Length 1
00010110 Operation Code Release Call
3.1.3 Release Call Arg
00000100 Tag
00000010 Length 2
1------- Extension Indication 1 Last octet
-00----- Coding Standard CCITT standard
---0---- Spare Spare
----0000 Location User
1------- Extension Indication 2 Last octet
-0010000 Cause Value Normal call clearing
 Appendix 3: Example SS7 message sequence coding 203

Message No. 7
TCAP End message
gsmSSF / gsmSCF
time: 18:48:40,740

HEX 0 1 2 3 4 5 6 7 8 9 A B C D E F

0 ac 87 1d 83 30 00 08 90 09 01 03 07 0b 04 43 30
10 00 05 04 43 20 00 05 08 64 06 49 04 2b 09 03 00

Translation for Message No. 7:

BITMASK ID name Comment or value

MTP header information:

-0101100 Backward Sequence Number 44
1------- Backward Indicator Bit 1
-0000111 Forward Sequence Number 7
1------- Forward Indicator Bit 1
--011101 Length Indicator 29
00------ Spare 0
----0011 Service Indicator SCCP
--00---- Sub-service: Priority Spare/priority 0 (USA
only)
10------ Sub-service: Network Ind National message
**b14*** Destination Point Code 48
**b14*** Originating Point Code 32
SCCP message header:
1001---- Signalling Link Selection 9
00001001 SCCP Message Type 9
----0001 Protocol Class Class 1
0000---- Message Handling No special options
00000011 Pointer to parameter 3
00000111 Pointer to parameter 7
00001011 Pointer to parameter 11
Called address parameter:
00000100 Parameter Length 4
-------1 Point Code Indicator PC present
------1- Sub-system No. Indicator SSN present
--0000-- Global Title Indicator No global title
included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Called Party SPC 48
00------ Spare 0
00000101 Sub-system number CAP
Calling address parameter:
00000100 Parameter Length 4
-------1 Point Code Indicator PC present
------1- Sub-system No. Indicator SSN present
204 Intelligent networks – principles and applications

BITMASK ID name Comment or value

--0000-- Global Title Indicator No global title

included
-1------ Routing Indicator Route on DPC þ Sub-
system No.
0------- For national use 0
**b14*** Calling Party SPC 32
00------ Spare 0
00000101 Sub-system number CAP
Data parameter
00001000 Parameter length 8
***B8*** Data 64 06 49 04 2b 09 03 00
TCAP transaction portion
End:
01100100 Tag
00000110 Length 6
Destination Transaction ID
01001001 Tag
00000100 Length 4
***B4*** Dest Trans ID 722010880
 Glossary

3G Third generation of mobile communications systems, based on

broadband technology. 3G follows 2G, which refers to mobile
systems such as GSM, using digital speech-coding technology,
and 1G (earlier analogue mobile telephony systems). 2.5G refers
to intermediate technologies between 2G and 3G such as GPRS-
based GSM. An example of a 3G system is ETSI’s UMTS.
3GPP 3G Partnership Project. This was originally pioneered by ETSI
and it focuses on GSM evolution to 3G. Earlier ETSI work for
GSM was integrated into 3GPP to achieve a single UMTS
development stream. 3GPP provides the currently prevailing
worldwide standard for the core network architecture and
protocols, although the radio interface standard used is subject
to regional differences because of spectrum allocations and
licensing variations. See also ‘UMTS’ below.
ACM Address Complete Message. This is an ISUP (and TUP) message,
sent by a destination node and relayed back to the call origination
node, in reply to an IAM (or IFAM). It indicates that the address
has been checked and found to be valid.
ACP Action Control Point, the term used for a network switch in the
early DSDC number translation system offered by AT&T in the
USA.
AD Adjunct – this is a CS-1 physical (network) entity that is similar to
an SCP, but is more closely coupled (using faster signalling than
SS7) to the SSP.
AIN Advanced Intelligent Network. This term was coined in the early
1990s by Bellcore (now Telcordia) for a set of widely accepted IN
standards for the USA.
ANS ‘Answer’ message. This is an ISUP (and TUP) message, sent by a
destination node and relayed back to the call origination node,
when the termination device identified by the previously sent
address digits has answered the call. This may be automatic, in the
case of a modem or ISDN device, or not, in the case of a recipient
of a voice call lifting the handset.
206 Intelligent networks – principles and applications

ANSI American National Standards Institute. ANSI facilitates the

development of US domestic standards through consensus
processes and communicates US standards to the international
standards arena.
ANSI-41 An abbreviation of ANSI/EIA/TIA-41, which is the joint ANSI &
EIA (Electronic Industries Alliance) for mobile networking.
ANSI-41 is applicable to non-GSM networks.
APDU Application Protocol Data Unit.
API Application Programming Interface.
ARI AssistRequestInstructions – an INAP operation. INAP operations
are described in Section 2.5.1.
ASF ActivateServiceFiltering – an INAP operation. INAP operations
are described in Section 2.5.1.
ASN/1 Abstract Syntax Notation No. 1. This is the formal message-
description language used by the ITU-T for defining IN
information exchanges. http:/ /asn1.elibel.tm.fr/ refers.
ATM Asynchronous Transfer Mode. ATM is a flexible data transfer
mechanism, based on 53-octet cells, chosen by the ITU-T as the
data transport technology to be used for ITU-T broadband ISDN.
AT&T American Telephone and Telegraph company.
BCM Basic Call Manager. The BCM encompasses the CCF and the IN
triggering part of the SSF in an IN SSP node.
BCP Basic Call Process. BCP is an ITU-T IN SIB that represents the
underlying thread of basic call control in IN service creation
systems. Other SIBs are introduced to model particular service
behaviour patterns and these extra SIBs are superimposed on the
BCP.
BCSM Basic Call State Machine. This is the basic IN call model, which is
fundamental to ITU-T IN architecture.
BCUSM Basic Call Unrelated State Model. This is a separate (from the
BCSM) call model introduced in ITU-T CS-2 to handle non-call-
related features such as location updating. See also CUSF.
Bellcore Bell Communications Research – the previous name for the US
Telcordia Technologies Incorporated, which develops network
solutions, including IN architectures.
BICC Bearer-Independent Call Control.
BSC The Base Station Controller in a mobile network.
BSS Base Station System, in a mobile network. The BSS contains the
BTS and BSC functions.
BSSMAP Base Station System Mobile Application Part of the SS7
signalling system.
BTNR British Telecom Network Requirements. Early ‘de facto’ scheme
of documenting British Telecom’s UK network interworking
standards.
BTS The Base Transceiver Station in a mobile network.
C7 ITU-T (CCITT) Signalling System No. 7 – usually called ‘SS7’
nowadays.
 Glossary 207

CAMEL Customised Application of Mobile networks Enhanced Logic.

CAMEL an intelligent networking protocol, like INAP, but
optimised for GSM networks. CAMEL is the IN standard for
GSM based networks and WIN is the IN standard for non-GSM
mobile networks.
CAP CAMEL Application Protocol. This is the SS7 IN signalling
protocol for CAMEL. It is closely related to INAP.
CCAF Call Control Agent Function. This is an IN functional entity, in the
ITU-T distributed functional plane, which represents a user who
has more sophisticated network interface, such as ISDN, than an
analogue POTS phone.
CCAFþ A CCAF with wireless access functionality. CCAFþ was
introduced in the ITU-T CS-2 standard.
CCBS Completion of Calls to Busy Subscribers. The CCBS service is
sometimes known as ‘ring-back when free’.
CCF Call Control Function. This is an IN construct that models the
fundamental call control software in network switches.
CCITT International Consultative Committee on Telephony and Tele-
graphy. This was the previous (before June 1994) name for ITU-T.
Centrex The term used for providing a customer with a ‘virtual’ PBX from
a public network switch.
CG CallGap – an INAP operation. INAP operations are described in
Section 2.5.1.
CI Correlation Identifier. An ITU-T INAP parameter used for
correlating call-processing requests that occur in separate TCAP
transactions.
CLI Calling Line Identity. This is sometimes abbreviated to CgLI, to
distinguish it from the called line identity (CdLI).
CLIP Calling Line Identity Presentation. This is a telephony network
service that enables customers to display callers’ line identities on
incoming calls.
CLIR Calling Line Identity Restriction. This telephony service allows
the originating end to inhibit the display of calling line identity at
the far end, where it over-rides CLIP.
COLP Connected Line Presentation. A telephony service that enables
callers to see a display of the terminating line identity when a call
is connected.
COLR Connected Line Restriction. A telephony service that allows the
terminating end to inhibit the display of terminating line identity
at the originator’s display, where it over-rides the COLP service if
the caller has it enabled.
CON Connect message – an INAP operation. INAP operations are
described in Section 2.5.1. Note that the same term is also used in
other signalling contexts (besides IN) for different messages.
CORBA Common Object Request Broker Architecture. This is a
distributed object technology for client/server middleware that
was introduced by the Object Management Group (OMG).
Core INAP See ETSI Core INAP.
208 Intelligent networks – principles and applications

CPE Customers’ Premises Equipment. CPE is telecommunications

equipment used in the domestic or business (or other premises) to
originate or terminate telecommunications traffic.
CPH Call Party Handling. This was the most significant extra IN
function package that was included in the ITU-T CS-2 standard,
introduced in 1997. It allows call legs of multi-party calls to be
manipulated individually under IN service logic control. See also
CVS.
CRACF Call-related Radio Access Control Function. This was introduced
in CS-2 to allow IN to be applied to radio and mobility network
control.
CS Call Segment. A CS is the collection of legs and connection points
that represent the transmission and connection resources for an
isolated portion of a call.
CSA Call Segment Association. A CSA is a coupling of two or more
CSs.
CS-1 Capability Set number 1. This is a generic ITU-T term for the first
tranche of a standard. However, in this book CS-1 generally refers
to the ITU-T Intelligent Network CS-1 standard.
CS-1R CS-1 Refined. ETSI had taken a reduced subset of the original CS-
1 release and had included missing detail in order to formulate
ETSI Core INAP. This was then fed back into the ITU-T study
group and it was released as the ITU-T CS-1R recommendation.
CSE CAMEL Service Environment. In a CAMEL GSM system the
CAMEL CSE is equivalent to the SCP in a fixed IN architecture.
CSI CAMEL Subscription Information (mark). This is in a subscri-
ber’s data profile in the HLR and VLR and indicates that the user
is a CAMEL subscriber.
CS-n Capability Set-n. This refers to ITU-T sequential steps of
functional capability. The term is generally applicable to all
standards disciplines in the ITU-T, but in this book we nearly
always use it to refer to the series of releases of IN standards. In
particular, the term CS-1 is used to refer to the initial milestone
series of intelligent networking standards of March 1993.
CTI Computer Telephony Integration.
CTR ConnectToResource – an INAP operation. INAP operations are
described in Section 2.5.1.
CURACF Call-Unrelated Radio Access Control Function. See also CRACF.
CUSF Call-Unrelated Service Function. This is a CS-2 network function
that provides access between a user agent (SCUAF) and an SCF
for call unrelated transactions such as location updating. It runs a
state model called the BCUSM.
CVS Connection View State. This is a term used to describe the
connection and call-leg configurations that can be modelled by the
ITU-T IN CS-2 call party handling (CPH) features.
DDSN Digital Derived Services Network. An early AT&T proprietary
intelligent network system based on AT&T’s earlier number-
translation based Direct Services Dialling Capabilities (DSDC).
DDSN is sometimes abbreviated to DSN.
 Glossary 209

DFC DisconnectForwardConnection – an INAP operation. INAP

operations are described in Section 2.5.1.
DFP Distributed Functional Plane. This is the next plane down from the
Global Functional Plane in the ITU-T IN hierarchy. It effectively
provides a ‘road-map’ showing the relationships between the
various logical network objects (the functional entities) compris-
ing an intelligent network.
DLE Digital Local Exchange. A term that is sometimes used for a
traditional PSTN circuit-switching node that has analogue (POTS)
and digital (e.g. ISDN) lines serving end customers.
DP (IN) Detection Point in the Basic Call State Model. DPs are
points, between the PICs (see below) where IN triggers are
embedded. For example, there are 17 DPs defined for the ETSI
Core INAP BCSM. DPs can be configured as Event or Trigger
detection points (E_DPs or T_DPs).
DPC Destination Point Code in an SS7 signalling network.
DPE Distributed Processing Environment.
DSDC Direct Services Dialling Capabilities. This was an early IN
development from AT&T in the USA. See also DDSN.
DSS1 Digital Subscriber Signalling System no. 1. This is an ISDN user
network interface. It is a common-channel digital access
signalling system. The ITU-T recommendation for DSS1 is
Q.931 and the ETSI standard is EN 300 403–1.
DTAP Direct Transfer Application Part (of SS7), which is carried over
TCAP and SCCP and is for call control messages passing between
the BSC and MSC in mobile networks.
DTID Destination Transaction IDentity, in an SS7 TCAP message.
DTMF Dual Tone Multi-Frequency – a tone signalling system available
to POTS users with a multi-frequency key-pad. Also known as
MF4 signalling.
E_DP Event Detection trigger Point. Typically used to trap call-
processing events such as ‘answer’ or ‘release’. E_DPs are
dynamically armed by the SCF sending RequestReportBCSM
messages to the SSF. E_DPs can be active (E_DP-R for Response)
or passive (E_DP-N for Notify).
EIA Electronics Industry Alliance. This is a body that organises
standards development in the USA.
En Bloc A mode of telephony signalling (ISUP and TUP) where all address
digits are sent in a single call set-up message. The alternative is
‘overlap’ signalling.
ETC EstablishTemporaryConnection – an INAP operation. INAP
operations are described in Section 2.5.1.
ETSI European Telecommunications Standards Institution. ETSI pro-
duces regional standards that are subsequently used in Europe and
elsewhere.
ETSI Core INAP This is the 1994 ETSI-defined subset of the ITU-T IN CS-1
Standard. It was adopted and ratified as an international standard
by the ITU-T in the following year (1995) under the name CS-1R
(R for ‘Refined’). ETSI Core INAP is a widely used standard.
210 Intelligent networks – principles and applications

EventReportBCSM This is an INAP operation that is used by SSPs to respond to

RequestReportBCSM operations. INAP operations are described
in Section 2.5.1.
FAM Final Address Message. This is optionally used for the last address
message in a sequence in the SS7 TUP-derived protocols for
telephony call control. It is used if the sending node is aware of
the number length of the address digit string. See also ‘overlap’
signalling.
FCC (US) Federal Communications Commission.
FCI FurnishChargingInformation – an INAP operation. INAP opera-
tions are described in Section 2.5.1.
FE Functional Entity. Examples of typical intelligent network FEs are
the SSF, SRF and the SCF.
FIM Feature Interaction Manager. Part of the IN SSF.
FISU Fill-In Signalling Units. These are short signalling ‘packets’ sent
by SS7 MTP on an idle signalling link.
GFP Global Functional Plane. This is the high-level theoretical
description of how service requirements are mapped into network
functions in the ITU-T IN standards.
G-MSC Gateway MSC. This is the MSC in a mobile network that
interfaces to other networks, typically to a PSTN.
GPRS General Packet Radio Service. See also 3G.
GSL Global Service Logic. GSL is the abstract service logic that
describes the operation of the SIBs in the IN global functional
plane.
GSM Originally this was ‘Group Speciale Mobile’, but the abbreviation
was subsequently re-allocated in ETSI to the phrase ‘Global
System for Mobile communications’. GSM is the European 2nd
generation mobile communications standard.
gsmSSP GSM Service Switching Point. (The initials ‘gsm’ before any IN
entity denotes the mobile network variant of that entity.)
GT Global Title. This is a form of address used in SS7 SCCP routing.
GTs do not refer to particular nodes but need to undergo
translation (in a GT table – GTT) to resolve to a physically defined
address.
GTT Global Title Translation – see GT.
H. 248 H.248 is the ITU-T standard for the control of media gateways by
media gateway controllers. It is also known as the ‘Megaco’
protocol.
H. 323 An ITU-T suite of protocols designed for controlling multimedia
communications over IP.
H. 323 Gatekeeper The Gatekeeper is a point of ‘central’ control for service control
and provision in H.323 networks. H.323 devices (such as
terminals and gateways) register with gatekeepers before they
can be given permission (by the gatekeeper) to receive or originate
H.323 ‘calls’.
HLR Home Location Register. Every mobile network has at least one
HLR to manage the records and account details for the network’s
customers.
HTML HyperText Mark-up Language.
 Glossary 211

IAF Intelligent Access Function. The IAF is provided in the CS-2

distributed functional plane to allow interconnection between IN
and non-IN structured networks.
IAM Initial Address Message. This is the initial call set-up message
used by the ISUP and TUP protocols in SS7.
i/c ‘Incoming’. Used as an abbreviation to denote call set-up
direction.
ICA InitiateCallAttempt – an INAP operation. INAP operations are
described in Section 2.5.1.
I_DP InitialDP (Detection Point) message – an INAP operation. INAP
operations are described in Section 2.5.1. I_DP is usually the first
message in a CS-1 IN transaction and it is sent from an SSF to an
SCF.
IETF The Internet Engineering Task Force.
IF Information Flow. Intelligent Network functional entities (FEs)
communicate using IFs. Often there is a one-to-one relation
between FEs and INAP operations. INAP operations for CS-1 are
described in Section 2.5.1.
IFAM Initial and Final Address Message. This is an SS7 telephony
signalling message, used in SS7 TUP (but not in ISUP), and TUP-
derivatives, as an alternative to the IAM for ‘en bloc’ signalling.
IMSI International Mobile Subscriber Identity. This is a series of
numerical digits that allows the identification of mobile phones
internationally when they are roaming. IMSI uses the ITU-T
E.212 numbering plan.
IN Intelligent Network. This is a well used anachronism that has been
applied to many things, but it usually refers to the introduction of
separated service logic into public telephony networks in the
manner described in this book. Sometimes the upgraded PSTN is
described as an intelligent network itself; sometimes IN refers just
to the new network elements.
IN/1 (&2) IN/1 and IN/2 were early standards initiatives by Bellcore in the
USA.
INA Information Networking Architecture. INA was the 1990
Bellcore-inspired forerunner to TINA.
INAP Intelligent Network Application Protocol (part of SS7). See also
ETSI Core INAP.
INCM Intelligent Network Conceptual Model. This is the top-down
process that was used as the framework for the production of IN
standards by the ITU-T.
INWATS Inward Wide Area Telecommunications Service. This was a
feature of AT&T exchanges in the 1970s that provided an early
Freephone service. See also DDSN.
IP Internet Protocol or IN Intelligent Peripheral, depending on the
context.
ISDN Integrated Services Digital Network.
ISO International Standards Organisation. See also OSI.
ISP Internet Service Provider.
ISUP ISDN User Part (of SS7) signalling.
212 Intelligent networks – principles and applications

ITU-T International Telecommunications Union – Telecommunications

Division. Before June 1994 the ITU-T was known as CCITT.
IUP Interconnect User Part (of SS7) signalling IUP is the TUP
derivative telephony signalling system that is historically used
between public operators in the UK. The preferred standard
interconnect protocol is ISUP.
IVR Interactive Voice Response.
JAVA API The JAVA API for telephony call control (also known as JTAPI)
is an extensible API designed for various call control scenarios.
JAIN Java for Advanced Intelligent Network. JAIN is a software
industry community that specifies APIs for converged PSTN and
IP (Internet Protocol) networks.
LEC Local Exchange Carrier in the USA.
LNP Local Number Portability
Local Exchange Same as DLE.
LSSU Link Status Signalling Unit. This is used to exchange status
information between each end of an SS7 signalling link.
MAP Mobile Application Part (of SS7). MAP is a well established SS7
user protocol for handling mobility management messages
between HLRs, VLRs and MSCs in mobile networks. It is also
used between BSCs and MSCs.
MCI Malicious Call Interception. A service that is typically imple-
mented using traditional switch-based functions (but nevertheless
does appear on the ITU-T CS-1 list of ‘benchmark services’).
MDF Main Distribution Frame. A termination point for cables and
subscriber lines in a traditional telephone exchange. The MDF
provides a flexibility point for jumpering cables to exchange
equipment.
MF4 Multi-Frequency number 4. This is an ITU-T standard for in-band
dual tone signalling used by push-button phones. Also known as
DTMF and sometimes abbreviated to ‘MF signalling’.
MGC Media Gateway Controller. An MGC is an external controller for
media gateways in IP (Internet Protocol) networks. The MGC
provides call control functions, and commonly communicates
with the media gateways using the H. 248 standard for supervising
call channels.
MSRN Mobile Station Roaming Number.
MSU Message Signalling Unit. The MSU is the ‘payload’ of an SS7
MTP message.
MTP Message Transfer Part of SS7 signalling.
NANP North American Numbering Plan.
NAP Network Access Point. The NAP is a CS-1 physical entity that just
contains CCAF and CCF functions, and so does not communicate
with an SCP.
NCP Network Control Point, the term used for a central database in the
early DSDC system in the USA. See also DDSN.
NICC Network Interoperability Consultative Committee. This is a UK
standards body, part of OFTEL, which oversees the production of
PNO specifications for the interworking of public network
operating companies in the UK.
 Glossary 213

NUP National User Part of SS7 signalling. NUP is derived from the SS7
Telephony User Part (TUP).
O_ Any IN message or trigger point prefixed by ‘O_’ is to do with the
originating (as opposed to terminating) half of the BCSM call
model.
O_BCSM Originating BSCM.
OCCRUI Out-Channel Call-Related User Interaction. This is an ITU-T CS-
2 function introduced to allow direct communication between a
user’s equipment and the controlling service logic in the SCF.
OCUUI Out-Channel Call-Unrelated User Interaction. This is an ITU-T
CS-2 feature that is modelled by the BCUSM.
o/g ‘Outgoing’. Used as an abbreviation to denote call set-up
direction.
O&A Operational and Administration.
OFTEL UK Office of Telecommunications. OFTEL is the regulator for the
UK telecommunications industry.
OMG Object Management Group. The OMG is a powerful software
industry consortium of over 800 companies and is a distributed
object industry standards group. It is noted for the introduction of
CORBA.
OPC Originating Point Code in an SS7 signalling network.
OSA Open Service Access. OSA is a consortium that takes account of
ETSI, JAIN, 3GPP and Parlay Consortium results and is working
to harmonise the production of APIs that can be used with ITU-T
IN functional components.
OSI Open System Interconnect. The OSI Reference Model (7-layer
model) was developed in 1978 by the International Standards
Organisation (ISO).
OSS Operational Support System. This is a generic term encompassing
the background management schemes for administering network
functions such as service provision, billing, network management,
performance monitoring, statistics collection, etc.
OTID Originating Transaction IDentity, in an SS7 TCAP message.
Overlap A mode of telephony signalling (in SS7 ISUP and TUP) where all
address digits are sent in individual set-up messages as digits
become available. The series of set-up messages consists of firstly
an IAM, then subsequent SAMs and (optionally, in TUP) a FAM.
This can lead to faster call set-ups when interworking with slower
signalling systems, but with the penalty of less efficient messaging
in terms of quantity of separate messages. The more efficient
alternative is ‘en bloc’ signalling.
PA Play Announcement – an INAP operation. INAP operations are
described in Section 2.5.1.
Parlay Consortium The Parlay Group was formed to produce technology-independent
APIs for use across multiple network platform environments.
Parlay is now part of the OSA initiative.
P&C Prompt and Collect User Information – an INAP operation. INAP
operations are described in Section 2.5.1.
PBX Private (Branch) Exchange.
PCM Pulse Code Modulation (for digital transmission).
214 Intelligent networks – principles and applications

PE (IN) Physical (network) Entity.

PINT PSTN/Internet Interworking. PINT services are characterised by
requests originating from the Internet and resulting in the
invocation, or the enrichment, of calls in a traditional telephone
network.
PIC ‘Point In Call’. This is a call processing reference point defined for
the IN basic call state model (BCSM).
PIN Personal Identification Number. PINs are used in IN structured
networks to provide users with secure access to their network
services.
PNO Public Network Operator. See also NICC.
PoP Point of Presence (for an Internet service provider).
POTS Plain Old Telephone System.
PSTN Public Switched Telephone Network – sometimes referred to as
POTS.
Q The identifier for the ITU-T series of recommendations that cover
the topics of switching and signalling.
Q.931 A well known ITU-T standard for access signalling between an
ISDN terminal and a narrowband ISDN network.
RBOC Regional Bell Operating Company.
RCF Radio Control Function. See also CRACF.
RCU Remote Concentrator Unit. The RCU houses terminations and
control equipment of a group, typically of several thousand, of
customers’ telephone lines. The RCU is connected to the parent
exchange over 2Meg links.
RequestReportBCSM This is an INAP operation that is used to arm IN E_DPs in the
ITU-T IN call model. INAP operations are described in Section
2.5.1.
RIL 3 GSM Radio Interface Layer 3, which is based on DSS1 and is
carried over SS7 DTAP and SCCP.
ROSE Remote Operations Service Element. This is the basis of the
TCAP component sub-layer and is defined in ITU-T X.229.
RP Reference Point (in TINA).
SAD Send All Digits. This is a TUP message that can be sent in
response to an IAM in order to obtain all remaining address digits.
SAM Subsequent Address Message. This is an ISUP (and TUP)
message and it follows the IAM. It contains extra address digits.
SA-GF Service Application Gateway Function. SA-GF is an IN CS-4
component that provides an open API interface to an IN structured
network.
SC-GF Service Control Gateway Function. SC-GF is an IN CS-4
component that provides an interface between and IP (Internet
Protocol) network node and an IN SCF.
SCCP Signalling Connection Control Part (of SS7). This allows non-
circuit-related signalling (such as TCAP) to be carried over SS7
routes.
SCE(F) Service Creation Environment (Function) in an intelligent
network.
 Glossary 215

SCF Service Control Function. This is most often located in an SCP,

but can also be deployed in other IN nodes such as IP, SSP or SN.
SCF Id. SCF Identifier. An ITU-T INAP parameter.
SCP Service Control Point.
SCUAF Service Control User Agent Function. Introduced in IN CS-2. See
also CUSF.
SDF Service Data Function. This is a database function that may be
located in an IN physical entity such as an SCP, SN or SSP. It can
also reside in a dedicated and separate SDP.
SDP Service Data Point. See also SDF.
SDL Specification Description Language. This is a commonly used
diagrammatic method of illustrating and describing the sort of
state changes that are typical of IN functions. An example of the
use of SDL is the BCSM description in Q.1214.
SFR Service Filtering Response – an INAP operation. INAP operations
are described in Section 2.5.1.
SHP Service Handling Protocol. An SS7 TUP signalling parameter,
carried in the IAM or IFAM that informs the far end about the
nature of the service that is required, e.g. POTS or ISDN.
SIB Service Independent Building block. SIBs are abstract representa-
tions of an IN’s capabilities and they are intended to provide the
fundamentals for a generic capability for building IN-based
services.
SIO Service Information Octet. SIO contains information on the
identity of the user of an SS7 MTP message. SIO is also an
abbreviation for ‘signalling indication ‘out-of-alignment’, which
is transmitted when one end of an SS7 signalling link is
attempting to align with the other end.
SIP Session Initiation Protocol (SIP). SIP is an IETF initiative that
provides an alternative (to the ITU-T H.323 standard, where
Q.931-based call set-up messages are exchanged) method for
setting up and modifying conferences between multiple parties in
a VoIP scenario. SIP is based on HTML-type techniques.
SIP Proxy An SIP Proxy contains both an SIP client and server. It is therefore
able to act as a ‘go-between’ at the functional interface between
SIP-based packet networks and traditional IN-structured networks.
SK Service Key. SK is a parameter carried in an I_DP that allows a
service logic reference to be conveyed from the SSF to the SCF.
SLEE Service Logic Execution Environment. This is traditionally the
term used for the ‘front’ end of an SCF. The SLEE controls the
transaction management aspects of an IN invocation session,
ensuring that IN requests and responses are handled by the correct
SLP.
SLP Service Logic Program (in an SCF). The SLP represents the
service logic for a particular IN service or feature.
SMAF Service Management Access Function. SMAF provides the access
interface to SMF from service users as well as network operator
functions.
SMF Service Management Function (see also TMN).
216 Intelligent networks – principles and applications

SMG Special Mobile experts Group – an ETSI working party that

invented GSM.
SMS (GSM) Short Message Service or (IN) Service Management
System.
SN Service Node. This is a self-contained IN network element that
typically contains SCF, SSF and SRF entities.
SND Send ‘N’ Digits. This is a TUP message that can be used as an
alternative to SAD if the recipient of the preceding IAM knows
how many digits to request. ‘N’ represents the number of digits
requested. The use of SND is more precise and leads to faster call
set-ups because the recipient node can proceed with call set-up
rather than waiting for possible arrival of extra address digits, as is
the case with the SAD procedure.
SP SS7 Signalling Point.
SPC Stored Program Control. This was the term used for early
computer-controlled telephone exchanges.
SPIRITS Services In the PSTN/IN Requesting Internet Services. SPIRITS
services are the converse of PINT services – events in an
intelligent telephony network can instigate actions in a data
network. Internet call waiting is an example of such a service.
SRF Specialised Resource Function. The SRF provides speech path
resources that are additional to those provided in the CCF. SRFs
are typically located in IN IPs or in SSPs.
SRR Specialised Resource Report – an INAP operation. INAP
operations are described in Section 2.5.1.
SSCP Service Switching Control Point. This is a composite node,
encompassing the IN functions typically contained by SSPs and
SCPs.
SSF Service Switching Function. The SSF is distinguished from the
SSP in that the latter term refers to a physical network node
containing an SSF, possibly in addition to other IN functions.
SSN SCCP Sub-system Number. SSN is used in non-circuit-related
SS7 signalling to represent the SCCP user.
SSP Service Switching Point – generally refers to a traditional voice
network switch that has been enhanced with IN functions such as
the ability to ‘trigger’ to external service logic when the trigger
conditions are met.
STP SS7 Signal Transfer Point.
T_ Any IN message or trigger point prefixed by ‘T_’ is to do with the
Terminating (as opposed to originating) half of the BCSM call
model.
T_BCSM Terminating BSCM.
TC Transaction Capabilities. See also TCAP.
TCAP Transaction Capabilities Application Part. TCAP is an SS7 user
part that provides a remote-operations transaction service to
higher level signalling protocols, such as INAP. TCAP uses the
underlying SS7 transport services of SCCP and MTP.
T_DP Trigger Detection Point. T_DPs can be active (T_DP-R for
response) or passive (T_DP-N for Notify).
 Glossary 217

Teletex An ITU-T defined international store and forward data service for
error free communications over the switched telephone network.
TIA Telecommunications Industry Association (a US standards
authority).
TINA Telecommunications Information Network Architecture. TINA
was a 7-year project that was concerned with the introduction of
DPE solutions and products into the telecommunications arena.
TMN Telecommunications Management Network. TMN is an ITU-T
standard for management functions that is used by IN CS-2. The
ITU-T TMN standards are M.3010 and M.3400.
TUP Telephony User Part (of SS7) signalling.
UFM Unified Functional Methodology. This is a CS-2 mechanism for
managing feature interactions.
UMTS Universal Mobile Telecommunications Service. UMTS provides
the European standard for 3G, specified in ETSI GSM Release
‘99. Because of the prominence of ETSI GSM in mobile standards
UMTS is set to become a major worldwide standard for 3G mobile
users (see also ‘3GPP’ above). Whereas GSM uses TDMA (Time
Division Multiple Access), where each device has a fixed time-
slot allocation for the duration of a call, UMTS uses W-CDMA
(Wide-band Code Division Multiple Access). The UMTS
frequency bands are 1920–1980 MHz for the upward direction
and 2110–2170 MHz for the downward direction.
UNI User Network Interface.
Videotex An ISDN service that delivers textual information via the phone
line (or cable) to a TV set. This is a service that did not achieve
any significant popularity.
VLR Visitors’ Location Register. This is associated with an MSC and
holds temporary identity information about mobiles which are
currently in the MSC’s domain.
VoIP Voice over Internet Protocol.
VPN Virtual Private Network. This appears to the users to be a private
network that uses public network resources.
VRU Voice Response Unit. Commonly provided as an IN IP node.
WIN Wireless Intelligent Network. This is the IN standard for ANSI-41
based networks in the same way that CAMEL is the accepted
standard for the introduction of IN features to GSM networks.
 Index

1996 US Telecommunications Act 73 authorisation 28, 72

2-motion selector 2 code 19, 81
3G 131, 205
3GPP 131, 174–5, 205, 213 base station controller (BSC) 68, 206
A-Interface 68 basic call process (BCP) 82, 93, 206
abbreviated dialling service 4, 78, 148 Beatles 11
account card calling: see automatic Bellcore 33, 71–5, 85, 168, 205–11
alternative billing BCSM
ACP 73, 205 instance 31, 108
address complete message (ISUP) 21, 60, see also CS-1 BCSM
205 BCUSM 91, 206, 208, 213
adjunct 75, 85, 205 bearer-independent call control (BICC) 16,
Advanced Intelligent Network (AIN) 73–5, 101, 137
83–90, 149, 167 benchmark services: see CS-1 benchmark
alerting pattern 46, 89 services
alternate billing 1–3, 73–4 billing 7, 148–50, 158
automatic alternative billing 78 and call completion service 156
and operational support environments
implementation 81
168, 213
ANSI-41 89, 206, 217 and PARLAY interfaces 12
answer message (ISUP) 21, 60, 164, 205 and service management 36
as a CAMEL event report 87–8, 182 and TINA service sessions 135–6
as a CS-1 event trigger 22, 159 centre 81
API 14, 99, 121–4, 127–9, 167, 170 information through an API 121
application protocol data unit (APDU) 67 see also call charging and alternate
‘apply charging’ operation 44–5, 89 billing
area number calling 143 Blue Book 63, 174
ASN/1 18, 25, 185, 190, 206 broadband ISDN 76–7, 124
assist procedure 40–6, 83 broker 122–6, 132–3, 207
assistance operator 151–2 BT (British Telecom) 59, 73–4, 127–8
see also switchboard operators business administrative domain 132
assisting SSF 83 business cases for IN 9–11, 23, 104,
AT&T 3, 71–4, 128, 205–11 168–70,
ATM 77, 124, 206
Forum UNI 3.1 23, 130 C7: see common channel signalling
authentication 33, 68, 128–9, 136, 160 call control agent function 82, 91, 94, 207
220 Index

call control function (CCF) and POTS 62–3

and the BCP 82 and terminating call screening 79–80
and co-located service logic 32 presentation (CLIP) service 61–2, 207
in a CAMEL network 160 restriction (CLIR) service 61–2, 207
in a distributed intelligence model 122 see also ISDN CLI services
in GSM networks 69 CAMEL 14, 86–90, 170, 174–5
in the API model 130 application part (CAP) 67–8, 88–9, 123,
in the IN CS-1 DFP 82 161–3, 181–207
in the IN CS-2 DFP 94 call flow examples 67, 160–4, 181–203
in the IN CS-2 call model 106 IN detection points 88
pre-IN CCF 4–7, 49, 53, 169 CAP: see CAMEL application part
representation in an IP network 100–1 capability set 10, 71, 76–7, 83, 168
separation and break-out to IN 14–17 CCITT 7, 53, 63, see also ITU-T
triggering 27
charging see ‘call charging’
call barring 16, 61, 85, 160, 164
‘choose to refuse’ 144
call charging 60, 45–8, 126, 157–8
circuit identity code (CIC) 61
record 21, 36
circuit-related signalling 7, 16, 55, 63,
and ‘alternate billing’ service 73
split between caller and called parties 80 214–16
for pre-pay mobile calls 89 ‘classical’ IN 6
for VPN service 150 CLI services: see ISDN CLI services
for ‘call completion’ service 156 common channel signalling 7, 53–69, 73
call completion to busy subscriber service competitive differentiation 4, 85
(CCBS) 62, 79, 207 component portion (of TCAP message)
call distribution service 48, 78, 151 66–7, 185–202
call diversion service 33, 39, 79, 124 computer-telephony integration 158–9
and presentation number 16 ‘connect’ operation 19–20, 43–5, 50, 105,
by switchboard operators 1–2, 116–17 152–62, 207
on ‘no reply’ in CAMEL networks 89 for a number translation service 140
call forwarding 16–20, 26, 46, 79, 85, 148 for a VPN service 150
call gapping 27, 31, 157, 165 connected line presentation (COLP) service
call model 21 62, 207
see also CS-1 BCSM connected line restriction (COLR) service
call party handling (CPH): see ‘leg control’ 62, 207
call routing 71 conference call 2, 72, 79, 90, 130, 215
function 16, 129 convergence of mobile and fixed networks
service 79 69, 91
call segment 31, 90, 107–16, 208 convergence of IN and IP networks 101,
call segment association 31, 113, 115, 208 171
call transfer 90, 107 convergence of computing and
call screening 2, 79, 144, 145, 162, 170 telecommunications 124, 131
call waiting 4, 16, 90–1, 115 CORBA 99, 123–8, 131, 136, 207, 213
Internet call waiting 98, 101, 167, 223 core INAP: see ETSI core INAP
calling card service 73–4, 170 credit card calling service 78
calling name delivery 1 see also automatic alternative billing
calling line display services 61 CS-1 167, 208
calling line identity (CLI) 16, 27, 61–2, and APIs 123
143–4, 148–9 and GSM 86–8
and call completion services 152 and TINA 121
and CTI service 158–9 basic triggers 18–19
and least-cost routing service 149 BCSM 28–9
 Index 221

compared with CS-2 103–8 dial-up internet access 165, 171

benchmark services 77–8 digital exchange 2–4, 8, 37–8, 49, 63
detection points 29 direct transfer application part (of SS7) 68,
DFP 29–30, 77–8, 82–3 209, 214
documents 77, 177 distributed functional plane 104–6, 173–4,
feature interaction manager 30–1, 50 177–8, 209
GFP 77–8, 81–2 see also CS-1, CS-2 or CS-4 DFP
intelligent peripheral control 151–9 distributed intelligence 58–9, 116, 119–25,
implementation issues 48 170
limited to 2-party calls 72 distributed processing environment 14,
operations 43–8 121–4, 131–6, 217
physical plane 78, 83–5 distributed service logic 39, 99
PICs 22 diversion: see call diversion
service(s) plane 81 ‘do-not-disturb’ 144–5
standard 76–85 DP: see CS-1 detection points
see also ETSI core INAP DPE 14, 123–4, 131–2, 136, 209
CS-1R (refined) 78, 85, 101, 108, 168, Drake, Edwin L. 16
208–9 DSDC 73–4, 209
CS-2 90–8, 103–17, 167, 206–17 DSS1 40, 62, 209
and PARLAY 130 to IN parameter equivalents 25
and TINA 121 DTMF 17–19, 73–5, 90, 209
call model 104–5 and voice guidance 5, 81, 144, 151
core switch capabilities 108 as a user signalling interface 39
DFP 93–5, 109–12 with call completion service 156
feature interactions 50 DTPM 125
GFP 92–5 Duell, Charles H. 167
hybrid approach 109
IAF 165
E.164 number format 64, 160, 184–204
in CAMEL 89
leg control 74 echo-cancellation 8
see also ‘leg control’ Ecole Superiore de Guerre 103
physical plane 95–7, 101 ‘embedded’ service 3, 17, 50, 53
SCF – SCF interface 142, 159 ‘en bloc’ signalling 60, 209, 213
service management 37 ETSI see European Telecommunications
service plane 91–2, 95 Standards Institute
standards 177 Europe 10, 49, 62, 74, 85, 167
switching manager 104, 107 European Telecommunications Standards
CS-3 97 Institute (ETSI) 209
feature interaction manager 97 core INAP 10, 17, 21–4, 43–4, 49–50
CS-4 97 standard 78, 85–6
DFP 97, 101 and CAMEL 87
GFP 101 standards and CLI-based services 62
physical plane 101 external service providers 74, 121, 167
service plane 101
CUSF 91, 94–7, 178, 206, 208 fault handling 92, 128
CVS 103–9, 113, 115, 208 feature interactions 13, 17, 50, 91
manager 30–1, 106
database 32, 63–5, 73, 119–20, 140–50 and standards 91–2
DCOM 123–5, 128 FCC 73
derived services network (DSN) 74, 208–9 FIM: see CS-1 feature interaction manager
dialled-digit trigger 18 firewall 99
222 Index

flow control 38, 55, 157 initial address message (IAM) 21, 26–8, 42,
Foch, Marechal Ferdinand 103 59–61, 211
follow-me diversion service 79 in GSM networks 162–3
follow-on call 91 intelligent network application protocol
Freephone number (INAP) 7, 43–8, 77–8
and presentation number 62 see also ETSI Core INAP
services 2–3, 8, 14, 62–5, 79, 170 intelligent peripheral (IP) 5, 20, 39–43, 46,
and SPIRITS technology 98 73
in US 73–4 for mobile networks 86–9
and number translations 139–43 for out-channel call-related user
for Internet access 165 interaction (OCCRUI) 90
functional entity 77, 82–4, 93–5, 210 ‘leg-control’ functions 105
‘furnish charge information’ operation 44, and personal number service 144
47, 150 and Internet Protocol 100
interactive voice functions 5
generic call control 130, 174 International Telecommunications Union
global functional plane 177, 210 see ITU-T
see also CS-1, CS-2 or CS-4 GFP Internet 13, 98–101, 124, 128, 168, 211–17
global service logic 93, 210 browser 144
global system for mobile communications call waiting 167
(GSM) 14, 85–9, 171, 205–17 dial-up 165
and PSTN signalling differences 68 and IN convergence 171
call flow example 163 documents 175
supplementary services 85 INWATS 73, 211
global title 7, 183–204 ISDN
addressing 64–5 and IN 13
translations 14, 120 and non-ISDN signalling specifications
GPRS 89, 205, 210 63
access and IN CS-2 90–1
H.248 100–1, 210 access indicator 27
H.323 99, 101, 123, 130, 215 bearer capability 18
Gatekeeper 98, 100, 210 CLI-services 61–2
Hartley, L. P. 1 interface to IN IPs 75, 98
help-desk 61 interworking and IN CS-2 97
home location register (HLR) 66–8, 87–91, set-up request message 60
160–3, 208, 210–12 standards 177
supplementary services 62
‘hot line’ service 20
teleservices 27
user part (ISUP) 16, 59
IETF 98, 123, 130, 167, 173–5, 211 indicator 27
IN conceptual model 76–7, 80–2, 92, 95, see also broadband ISDN
211 ISUP: see ISDN user part
IN switching manager 103, 106–8, 205 ITU-T 10, 71–8, 85–116, 177
IN switching state model 107, 121
IN vision 6, 13, 34, 76, 168–9
IN/1 73–5, 211 JAIN 99, 131, 212–13
IN/2 33, 72–5, 211 JAVA 99, 123, 125, 128, 212
INAP: see ETSI core INAP Johnson, Dr. Samuel 71
information element 27, 66, 91
information flow 77, 83, 92–5, 113, 116, Kansas City 2
211 Kipling, Rudyard 119
 Index 223

least cost routing service 72, 146–8 number portability service 73, 97, 170
leg control 72, 90–2, 104–17, 157 number translation 12, 32, 36, 48, 139–43,
local exchange carriers (US) 73 164, 205
location transparency 121 see also Freephone service

malicious call 61 object management architecture 126

indicator (MCI) service 79, 212 object management group 126–31, 170, 213
manual board 2, 117 object-oriented
see also switchboard architecture 14, 122
mass calling service 79 design 38, 128
media gateway 100–11, 210–12 programming 121
controller 100 techniques 107, 113, 131
MEGACO 101, 210 object request broker 126
message transfer part (MTP) 55–9, 64–7, OCCRUI 90, 213
183–204, 210–16 OCUUI 91, 213
MF4 212 off-hook trigger 4, 19–21, 26
see also DTMF operational support system 7, 35–7, 126,
mid-call trigger 75, 108, 156 213
middleware 14, 122–8, 207 originating call screening service 79
mobile application part (MAP) 55, 63–8, OSA 131, 170, 213
160–3, 209–12 API 99
and CAMEL 88 OSI 178, 213
mobile networks 14, 67–9, 85–6, 91, 161–2, model 54
stack 66
171
overlap signalling 60, 209, 213
mobile switching centre (MSC) 66, 86–7,
160–4, 181–7
mobile telephony 9, 85, 205 PARLAY 122–4, 127, 131, 170, 213
API 99, 122, 127–8
mobility management 14, 58, 212
group 128
multi-media 122, 130, 136 interfaces 129–30
multi-party calls 75, 90, 107–16, 130, 137, specifications 129, 174–5
208 payphone 26, 46, 73
multi-vendor environments 9, 68, 82, 92 PBX 98, 159, 207
PE – FE mappings 84
naming 132 physical entity 84, 214
service 124 physical plane 77, 177
NANP 73, 211 see also CS-1, CS-2 or CS-4 physical
NCP 73, 212 plane
NSCX 73 PIN code 78, 81, 148, 214
network access point (NAP) 84, 212 PINT 98–101, 167, 175, 214
network API 122–4, 127, 170 PIZZA service 143
network centric computing 125 plain old telephone service (POTS) 39, 62,
network management 7, 35–7, 63, 120, 147, 83, 207–15
213 see also PSTN
network number 62, 140, 150–2 ‘play announcement’ 20, 45–7, 144, 154
non-circuit-related signalling 7, 63, 181– plug-IN 122–5
204, 214 point codes (OPC & DPC) 56, 64–5, 183–
see also signalling connection control 204
part Point of Presence (PoP) 165, 214
‘non-stop’ transaction processors 38 Pre-pay service 86, 89, 163–4, 171
224 Index

premium rate service 24, 80, 165, 170 remote operations service element (ROSE)
presentation layer (in OSI) 55, 66 67, 214
presentation of called line identity 16, 61 RMI 125, 128
presentation number 62
see also call diversion Sarnoff, David 53
‘prompt and collect’ 19, 41, 45, 144–5, 213 SCUAF 91, 94–5, 208, 215
proprietary security 37–9, 58, 99, 122, 128
extensions to INAP 86 screening service 80
INAP variants 68, 86–8 selective call barring service 61
interfaces 37, 159 selective call forwarding service 79
optimisations to INAP 156 service architecture (TINA) 132–7
systems and solutions 10–11, 14, 50, 64,
service application gateway function 99,
91, 139
214
propogation delay 8
service control gateway function 99–100,
PSTN 170–1, 209–14
and CAMEL 160–2 214
and GSM – signalling differences 68 service control point (SCP) 4, 32–8, 119
intelligence 13 service control function (SCF) 8, 19, 31–2,
services 78 82, 90, 94, 215
and IP 98–9, 101, 167 for mobile networks 69
plug-IN 123–4 distributed 72, 120
service creation 35, 75, 82, 206
Q.931 standard 27, 42, 77, 101, 130, 214 environment (SCE) 7, 12, 34–7
function (SCEF) 34, 82–3
in CAMEL 162–3
radio interface (RIL 3) signalling 68, 214 in CS-2 91–4
Red Book 63 service data point (SDP) 7, 33–4, 65
reference points (TINA) 132–4 service data function (SDF) 33–4, 82, 94,
Regional Bell Operating Company (RBOC) 152, 215
3, 73–5, 214 service filtering 31, 48, 80, 158
regional standards 76, 209 service handling protocol (SHP) 63, 215
register translator 2 service independent building blocks (SIBs)
release 7, 14, 82–3, 206, 215
0 & 1 (AIN specification) 75 for CS-1 77
96 (of GSM) 87 for CS-2 91
call party connection 115, 105 for service logic representations 35
complete message (ISUP) 60 service key 92, 147–9, 159, 215
cycle of builds 12 in call filtering services 43
message (ISUP) 60 in call gapping services 45, 157
method in PARLAY specification 130 in CAMEL calls 161, 181, 186
of a call 47, 87–8, 144, 164, 182 parameter 17–18, 25–6, 33, 47
of CS-1 standard 76 service logic 8–33, 38–41, 65, 72–6, 150–9,
of CS-2 standard 93
181–2
of speech-circuits 16
‘embedded’ and separated 3–5, 49, 218
reason 45
program 13, 25, 32–3, 149, 215
request report BCSM 88, 150–1, 164, 182, execution environment (SLEE) 33, 149,
188–91 222
resource API 122–4 and service creation 35
reverse-charge call 1, 86 distributed 39, 95, 99, 121–2
‘ring-back’ service 62–3 and Televote services 80
see also ‘call completion to busy and CPH 90 103–8
subscribers’ global (abstract) 93, 210
 Index 225

centralised 119–20, 139 and IN CS-1 benchmark services

service management 35–7, 91–2, 133, 175 78–9
access function (SMAF) 83, 92–4, 215 and feature interactions 31, 50
function (SMF) 34–6, 83, 94 information sent to IN SCP 27, 46
system (SMS) 7, 39 in mobile networks 86–7
service node 8–9, 84, 216 in legacy networks 39, 168–70
see also feature interaction
service plane 77–8, 177
see also CS-1, CS-2 or CS-4 service plane switchboard operators 1–2, 116, 123
service profile:
database 12–13 Telcordia Technologies Ltd. 3, 131
details for mobile customers 66 see also Bellcore
service switching control point (SSCP) 8, telephone operator: see switchboard
84, 216 operators
telephony user part (TUP) 42, 55, 59–68,
service switching point (SSP) 4, 9, 16, 61
data management 51 156, 217
for mobile networks 86 televote service 43, 158
service switching function (SSF) 15–20, 82, terminating call screening service 79
94, 216 terminating trigger 20, 144, 159
see also triggering
session 13, 55, 135–7
TIA 89, 206, 217
session layer (in OSI) 66
TINA 14, 121–2, 131–8, 170, 217
signalling connection control part (SCCP)
TMN 92, 95, 97, 175, 217
7, 26–7, 63–5, 120
transaction capabilities 7, 65–9
signalling gateway 100 see also non-circuit-related signalling
signalling point 55–8 transaction capabilities application part
signal transfer point (STP) 55–7, 65 (TCAP) see: transaction capabilities
signalling transfer point network 74 transaction identity 67, 209, 213
silver bullet 169 transaction portion (of TCAP message)
‘single-ended’ control of IN features 71–2, 66–7, 184–204
90 transport layer (in OSI) 66
SIP 99, 101, 123, 130, 215 triggering 8, 11, 15, 29–30, 75, 206
proxy server 98, 100, 215 and Freephone service 140–3
soft-telco 121 and VPN service 148
specialised resource function (SRF) 20, 40, and call completion service 154
82, 94, 216 conditions 47, 51
in CAMEL networks 86 criteria 43, 147
see also intelligent peripheral and in CAMEL calls 161
interactive voice function introduction 17–20
speech recognition 50, 75, 156 multiple triggering 97
to remote service logic 17
SPIRITS 98–101, 167, 175, 216
trigger data 24
split charging 80 trigger data management 36–7, 51,
SS7: see common channel signalling 92
Stored Program Control (SPC) 2, 53, 216 see also trigger table
Strowger, Almon B. 2 trigger detection points 22
subsequent address message 214 see also CS-1 detection points
subsystem number (SSN) 64–5, 184–203, trigger table 17, 24–7, 31, 43–4
216
supplementary services 61–3, 68, 85, 107 UFM 50, 92, 217
switch-based services 4, 13, 53 UK 18, 55, 59, 73, 212
and customer data 16–17, 31–3 UMTS 131, 175, 205, 217
226 Index

universal version control 92

access number service 80 virtual private network (VPN) 3, 12, 37, 80,
integrator 13 148, 217
mobility 9, 120 visitors’ location register (VLR) 66, 88,
personal telecommunications service 160–4, 217
(UMTS) 80, 217 voice announcements 5–6, 75
US 3–4, 49, 71–6, 89–90, 131, voice recognition: see speech recognition
149
user-defined routing 80 Wireless Intelligent Network 89–90, 207,
user-interaction 82, 90–1, 129, 146, 217
213 wireless music box 53