Demystifying ATM/ADSL

(Publisher: Wordware Publishing, Inc.)

Author: Michael Busby
ISBN: 155622592X

Chapter 1What are ATM and ADSL?

Why Do We Need to Communicate?
What is ATM?
Why ATM?
What is ADSL?
Why ATM and ADSL?
Universal ATM Benefits
Business Drivers
Who Will Use ATM and ADSL?
ATM Service Today
Application for Internet Services
Application for Video
Applications for LANs
Broadband Services
Terms and Definitions
And the Winner Is
Summary

Chapter 2Basic CommunicationTechnology

The Basics
AC/DC
Analog Signals
Digital Signals
Signal Conversion
Pulse Modulation/Demodulation
Signal Multiplexing
TDM Carrier Standards
Bandwidth
Narrowband ISDN
Broadband ISDN
Summary

Chapter 3Transmission and SwitchingSystems

Transmission Systems
A Short History of Communication TransmissionSystems
Synchronous
Asychronous

Multiplexing
ISDN Basics
Narrowband ISDN
Broadband ISDN
Copper Wire
RF Propagation
Light Wave (SONET/SDH)
Switching Systems
Circuit Switching
Connectionless Switching
Packet Switching
Multirate Circuit Switching
Frame Relay
Cell Relay
Summary

Chapter 4Communication Protocols

Introduction
Deep Look
Physical Layer
Data Link Layer
Network Layer
Transport Layer
Session Layer
Presentation Layer
Application Layer
Upper Layer Protocol Examples
TCP/IP
IPX/SPX
Frames
Frame Structures
But What Does It All Mean?

Chapter 5Communication Systems

Communication Systems
LAN/WAN/MAN/GAN
Network Topologies
Point-to-Point
Star
Hierarchical
Mesh
Bus
Ring
Whats Up?
Loosen the Grip

Chapter 6ATM, The Basics

OSI Protocol Layers
ATM Protocol Layers
ATM Physical Layer
ATM Layer
AAL Layer
ATM Planes
User Plane
Control Plane
Management Plane
Layer Management
Plane Management
The ATM Cell
ATM Interfaces
ATM Cell Header
Virtual Channels and Virtual Paths
Header Error Check
Provisioning and Signaling
Sustainable Cell Rate and Peak Cell Rate
Minimum Cell Rate
Maximum Burst Size
Cell Loss Priority
Maximum Cell Delay
Cell Delay Variation
Cell Loss Probability
Traffic Shaping
Service Classes
Latency
Port Speeds
ATM Traffic Management and Control
Connection-Oriented Connection
Connectionless-Oriented Connection
PVCs
SVCs
Usage Parameter Control
Multicasting
ATM Network Interfaces
A Cell Family Vacation
ATM and Legacy Systems

Chapter 7ATM, The Gravy

Peak Cell Rate
Sustainable Cell Rate
Low-Speed Data
High-Speed Data
Audio

Full Motion Color Video

Multimedia
The Real Audio and Video World
ATM Switches
Inverse Multiplexing Over ATM (IMA)
Network Node Interface (NNI)
Service Provider Backbone Network
User Network Interface (UNI)
Corporate Access
Consumer Access
Network Management
ATM Interworking
Quality of Service
Switched Virtual Circuits
Switched Virtual Path Connection
Circuit Emulation
Shared Access
Transmission Time
ATM Competition

Chapter 8ATM and the Real World

Introduction
Explosions and the Growth of ATM
ATM Benefits and Network Evolution
U.S. ATM Players
ATM Equipment Suppliers
ATM Customers
ATM Ports By Industry
ATM Port Sales By Type
ATM Services
ATM Traffic
ATM Rate Structure
Non-Recurring Charges
Recurring Charges
ATM PVC Rates
One-stop Shopping
Managed Customer Premises Equipment
Port Oversubscription
ATM Service Contracts
Hurdles
Network Evolution
Crystal Balls

Chapter 9ADSL
What is ADSL?
DSL Technology

DSL Types
DSL
ADSL
HDSL
VDSL
SDSL
Discrete Multi-Tone
FDM
Echo Cancellation
High-Level Online Adaptation
Application Architecture
System Architecture
POTS Splitter
Transport Capacity
Framing
ATM Over ADSL
Hurdles
Market Strategies
Crystal Balls

Appendix A
Appendix B
Appendix C
Appendix D
Glossary
Bibliography
Index

Chapter 1
What are ATM and ADSL?
Questions Answered in This Chapter
Why do we need to communicate?
How does communication unite people?
What is a communication system?
What is information encoding?
What are ATM and ADSL?
Where do ATM and ADSL come from?
Why do we need ATM and ADSL?
Who will use ATM and ADSL?

Why Do We Need to Communicate?

Human communication technologies have evolved over eons from the primitive to the complex,
keeping pace with the evolution of human development. Communication technologies are so
important to the success of day-to-day human interaction that perhaps communication
technologies have provided the pace of human development and may be the foundation upon
which rest the achievements of humankind.
In the beginning, simple hand gestures, body movements, and grunts were sufficient to convey
every possible expression of thought from one cave-dwelling humanoid to another. As cavemen
and cavewomen emerged from dark, dank, smoke-filled caves and began dwelling in open spaces
populated with many unfriendly, and hungry, creatures, new social structures and new methods of
communicating were necessary for humankind to individually and collectively survive the new
challenges. If humans were to thrive in new social structures faced with new threats, more
complex communication skills and reliable transmission systems were essential.
Ideas had to be expressed in their correct context or else misunderstandings might lead to
unfortunate consequences. Information had to be reliably conveyed and shared to ensure the
survival of the group. Body language, hand signals, and grunts gave way to more sophisticated
vocalization as human society evolved from small, simple, family-based, cave-dwelling groups to
more complex and competing tribal groups populating the savannas. As human society began the
conquest of the land, personal language skills necessary for reliable face-to-face communications
became necessary.

As human social groupings further increased in size and split into semiautonomous villages, the
need to communicate long distance between these clan-based groups of similar persuasions
became necessary for their mutual survival. While we cannot be certain, the first transmission
system used for communicating between villages was probably a fleet-footed human messenger.
Messengers have always been a part of our communications systems, even used frequently today.
While runners were more or less reliable, depending upon their individual skill in avoiding
enemies and evading hungry carnivores, some societies learned long ago that sound could travel
faster than the fleetest runners. A system of communicating based upon the sound emitted when a
hollow tree trunk was forcibly struck with a club was developed and perfected, giving us the
basic drum. Where would humans get the idea for drumming to send messages?

Figure 1-1. A modern

Interestingly enough, there is a large beetle that inhabits the African plains that is very adept at
drumming, annoyingly loud, to call a mate. The beetle, in the still of the night, perches upon an
acacia tree and periodically beats the end of his hardened exoskeleton against the tree trunk. A
specific sound pattern is beat each time the beetle pounds his tail against the tree. Perhaps some
individual listening to the sound of the beetle overcame his fear of the night, and roaming
carnivores, and investigated the origins of the drumming. After observing how the beetle made
the interesting sound by striking the tree trunk, this curious individual could have easily picked
up a stick and imitated the beetle by striking a trunk. The drum (and drummer) is born!
Random beating on drums, besides sounding awful (if you do not believe, occupy the same room
as a 4-year-old beating on a set of drums), cannot convey much intelligent information. A method
of encoding information was necessary that the drumbeats might have some meaning to the
listeners. The drummers collectively agreed upon a set of rules that defined the meanings of the
beats.
The drums and drummers, spread out across the land in the tribal villages, represent a
communications transmission system. A system of drummers could keep the citizens informed of
important events faster and much more reliably than a fleet of runners or messengers could.
Drummers were not as likely to be eaten by wild animals or waylaid by hostile neighbors as
runners or messengers were.
Communications technology and transmission systems were keeping pace with the evolving
social and political environments. Imagine intertribal and intratribal warfare based on poorly
communicated ideas and information. The communication channels among and between groups

of tribes had to be clear and the transmission systems had to be reliable, else misunderstandings
could lead to death and destruction! What a parallel with todays society. Saddam Hussein claims
just such a communication error between himself and the United States ambassador to Iraq led to
the 1991 Gulf War.
Other means of communicating between societies were developed, all to keep pace with changing
and evolving human needs. From light, smoke, and fire signals to written messages carried by
mounted messengers and passenger pigeons, humans sought faster and more reliable ways of
communicating in a society growing more and more complex. Advances in natural philosophy,
pacing the headlong rush into the Industrial Revolution, led to the development of the telegraph,
then the telephone.

Figure 1-2. We've come a long way, baby, in just 100 years.
Telephone companies set about furiously wiring the world with copper while they and other
companies began developing more complex communications and transmission systems built
upon radio wave and coaxial cable technologies. Soon, reliable radio transmission was developed
and communications transmission systems were now able to leap previously insurmountable
obstacles, such as terrain and distance, economically. Radio transmission continued to evolve,
ultimately yielding satellite transmission technology. Communications satellites give us
an unprecedented ability to communicate around the globe. While satellites could leap tall
mountains with a single bound, ground transmission systems were still mired in coaxial cable and
microwave transmission systems until the perfection of light wave technology. Light wave
technology gives us the opportunity to communicate locally and around the country at data rates
that we could not even imagine just a few years previously.

As the communications and transmission systems evolved, the manner in which intelligent
information was communicated (encoded) advanced also. From a series of grunts to a meaningful
staccato beat on a tree trunk, written messages, a set of dashes and dots, electronic reproduction

of the human voice, and binary ones and zeroes, state-of-the-art information encoding techniques
have kept pace with the communications and transmission systems evolution. Drums gave way to
Morse code, Morse code gave way to the telephone, and now we have high-speed
communications systems that can transport digital data representing intelligent information at
rates of gigabits per second around the globe.
Modern society has come to depend upon communications and transmission technology for a
wide range of business, personal, and governmental services. Low-speed and high-speed data
communications, audio, video, wireless, personal communications, and Plain Old Telephone
Service (POTS) are some of the communications services that are important to the daily conduct
of human affairs. Additional teleservices such as virtual cinema, virtual theater, video
conferencing, videotelephony, video library, home education, full service multimedia (audio, text,
video, images), and high-definition TV (HDTV) will be economically available to every
business, home, and apartment in the near future.
A quick survey of some recent developments in communications reveals that usage of
communications systems is limited only by the communications networks themselves, and not
our ability to envision more productive uses of the networks.
Figure 1-3. A communications-hungry society
Currently, medical personnel in remote locations in Australia use electronic sensors to capture the
condition of ill patients. Such practices are extremely common and normally would not elicit any
passing comment. However, the patient data garnered, including detailed video images of patients
external and internal ailments, are transmitted via high-speed data communication networks to
centralized medical facilities where specialists are available around the clock to evaluate and
recommend treatment. For difficult operations, health care professionals can even participate
directly in the specific treatment. For budget stricken, small health care facilities, regardless of
location, such usage of the existing communication networks provides a much higher quality
health care service than could otherwise be economically achieved. Imagine a doctor in a studio
operating on a patient thousands of miles away by manipulating robot arms using virtual reality
technology.
High-tech movies today consist of many live action shots intermingled with computer-simulated
characters and scenes. Several high-tech companies located around the country are the creative
technical energy behind many of the recent blockbuster hits. These companies generate digital
data by the bucket full (I am referring to a big bucket). Using expensive supercomputers, they
simulate all manner of things. Connected to dedicated Integrated Services Digital Network
(ISDN) telephone lines, they ship digital data back and forth between their offices and the movie
studios. Imagine participating in a movie as one of the cast while remaining in your home.
Due to the unique requirements of different communication services, multiple information coding
and transmission techniques were developed to deal effectively with the various needs of a
communications- hungry populace. The Public Switched Telephone Network (PSTN),
Community Antenna Television (CATV), Direct Broadcast Satellite (DBS), and private local area
networks (LANs) are some of the incompatible communication systems utilized to transfer
service-specific information from one location to another. These various communication

networks were not designed to transport information in realms other than the specific service they
were designed for. While one can modify a specific communication system to adapt and transport
information derived from another system, such adaptation of the information and system is
expensive and slows the transfer rate. Multiple service-specific communication systems, layered
one upon the other, are expensive to build and maintain.
As our desire to communicate across service-specific barriers grew, driven largely by a desire for
new revenue streams by telecommunications companies and our individual desire to have access
to more and more information, a standardized system of coding and transporting information
transparently and economically from source to destination at data rates sufficient for all types of
service in demand, including unforeseen future needs, had become a global objective of various
communication governing bodies. ATM was chosen as the most viable method to standardize
end-to-end information transport into a single, reliable entity that can accommodate all the needs
of the various communication technologies.

Figure 1-4. Varied communications

What is ATM?
ATM: Automatic Teller Machine? Another Technological Melon? Another Telecom Mess? How
about: Another Telecom Method? The last choice is close.
Actually, it is an acronym for Asynchronous Transmission Mode. ATM is a switching and
transmission technology that can process and transmit any type of information including data,

text, audio, and video at high bit rates with a high degree of accuracy. ATM is the culmination of
all the developments in switching and transmission technology in the last 20 years. This includes
frame relay and packet switching and the change from coaxial to optical fiber transmission
mediums.

Why ATM?
Many types of information services and information transport are present in the various
communication networks around the world. Information services are becoming woven in the very
fabric of our existence. Witness the explosive growth of the Internet. Local area networks (LANs)
and wide area networks (WANs) are expanding into metropolitan area networks (MANs) and
global area networks (GANs). Video related services such as broadcast, CATV, DBS,
videotelephony, and video conferencing are commonly available for business and consumer use.
On the horizon we can see multimedia, video-on-demand (VOD), cyber-cinema, virtual theater,
and other interesting and useful remote data and video related services. What all of these uses
have in common is the need to transport information, or data, transparently through the commoncarrier telephone company transmission systems at speeds faster than can presently be achieved
economically. Transparency refers to the passage of data with little or no processing through all
switching and transmission equipment from source to destination. Source to destination is
frequently referred to as end-to-end.

Figure 1-5. ATM and ADSL public switched network

What is ADSL?

DSL is an acronym for Digital Subscriber Line. Presently, there are at least four types of DSL.
While each type of DSL is given some treatment here, we are primarily concerned with ADSL, or
Asymmetric Digital Subscriber Line. ADSL is a high-speed digital signal processing method for
encoding information that is well suited for transmission over copper lines.
ADSL technology allows the simultaneous transmission and reception of data at speeds of up to
10 Mbps downstream and 650 Kbps upstream over 18,000 feet of copper wire without additional
signal conditioning. Since 80 percent of U.S. homes and businesses are located within 18,000 feet
of a central office, most of us will be able to use ADSL technology without any additional signal
conditioning. Unless otherwise specified, the following discussions will refer to ADSL
technology.

Why ATM and ADSL?

Why? Copper wire and bandwidth, faster faxes, faster Internet access, video-on-demand,
multimedia, teleconferencing, telemedicine, virtual theaters, and cyber-cinemas. Virtually every
homeowner and most businesses, from now on referred to as users, have four copper wires
running from their home/business to a local telephone company switching office called a central
office (CO). An inexpensive technology that can utilize the current infrastructure with minimum
modification and deliver good quality data at the speeds necessary for the applications in demand
is much needed. ATM/ADSL is that inexpensive technology. Competition between
telecommunications service providers, Internet Service Providers (ISPs), and cable TV
companies drives the race to add ATM and ADSL technology to telecommunications service
providers infrastructure.
It is estimated that 650 million dial-up telephone lines are currently installed around the world.
Of the 650 million, 50 million are fiber and 600 million are copper. By the year 2000, there will
be an estimated 700 million dial-up lines. Each one of these dial-up connections represents a
potential customer for Internet, multimedia, and video services using ATM and ADSL
technologies.
Previously, for a user to have multimedia services, expensive technologies were required to get
the information from the service provider to the users facility. ATM and ADSL equipment
manufacturers are predicting that users will be able to have multimedia services for little more
than the cost of a cable TV hook-up without running any additional cabling to the home or
business. All that the user will need is a set-top box, similar to an analog modem.

Universal ATM Benefits

ATM and ADSL may just be the driving force that unifies the world into the much talked about
global community. A global community requires a global information repository, a global
language, and a global information transmission system. English is the default language of the
Internet. The Internet is the default global library. Through Internet usage, the English language is
spread to every corner of the world and will become the default global language. Now, a global
information transmission system is required. ATM and ADSL will become that global
information transmission system.

ATM and ADSL have the ability to effortlessly perform the tasks necessary to become a universal
and global information transmission system. ATM and ADSL will provide one global backbone
network for all types of traffic. A backbone network is a network used by long-haul carriers to
transport data from source to destination. A backbone network is composed of various
transmission equipment and network management functions.
ATM can:

Integrate voice, video, and data

Interwork with all other data types

Extend LANs to MANs, WANs, and GANs

Enable new applications

Simplify network management

Support multiple access platforms/protocols

Provide compatibility with existing cable facilities

Provide long architectural lifetime

Make the information superhighway a reality

ADSL can:

Transport all the worlds data to your home or business over the existing copper wire local
loop

ATM provides for incremental migration capability of legacy (already existing) networks,
assuring economical allocation of scarce resources while allowing users to incorporate ATM
technology in todays networks. ATM is scaleable, meaning it comes in various speeds, from small
(64 Kbps) to large (gigabits/second).

Business Drivers
For ATM to survive and even thrive, there must be economic incentives for OEMs (original
equipment manufacturers) to design and manufacture ATM equipment, for service providers to
install ATM backbone networks, and for users to migrate from legacy networks to ATM.
ATM/ADSL for business is a many-faceted diamond. ATM empowers numerous desirable
network characteristics:
1. Simpler/less costly data networks
2. Replaces private line networks
3. Extension of LAN to MAN, WAN, and GAN

 connect remote corporate entities easily

4. New business markets
data and video services
continental and global alliances
5. Telecommuting
lower enterprise overhead
more competitive enterprise
6. Multimedia interactive conferencing
reduced/eliminated travel expense
7. Faster access to the information superhighway
no speed limit on the superhighway
ATM/ADSL, besides offering exciting, new opportunities for businesses, gives consumers access
to the information superhighway at speeds difficult to even imagine a few short years ago. And
the worlds audio and video library will be at our fingertips. Let us rejoice that we live in an age of
such monumental change in the very way we will evolve as social creatures. Why ATM/ADSL
for the consumer? High-speed Internet access, video, video, and more video.

Who Will Use ATM and ADSL?

Who will use ATM and ADSL now?
Government

defense
national archives

public information and services

distance learning

Health Industry

patient records
patient billing

video based consultation

Figure 1-6. Internet usage by number of dial-ups per day

distance medicine
residential home care

remote teaching

continuing education

Finance

trading rooms
network consolidation

critical site backup

Manufacturing

distributed database management

supplier/vendor database access sharing

just-in-time ordering

personnel training

product training and maintenance

Education

distance learning
virtual classrooms

multimedia classrooms

digital libraries

Home and Consumer Markets

video-on-demand

Internet service

home shopping

personalized entertainment

work at home

continuing education

ATM Service Today

ATM is currently available in the continental U.S., Canada, Western Europe, and Japan. Fifteen
U.S. service providers are ramping up ATM services as fast as they can install switches and
program back office systems to manage the backbone networks. ATM service today has these
service characteristics:
1. permanent virtual circuits
virtual path connections
virtual channel connections
2. switched virtual circuits
3. structured (DS-1) circuit emulation

Figure 1-7. continental perspective of global networks

4. multiple platform/protocol interworking

5. multiple service classes
6. multiple interface rates
7. multiple access interface speeds
8. multiple service providers
9. contract based pricing
10. usage based billing

Application for Internet Services

The explosion in Internet growth and usage fuels the desire for higher data rates. Everyone tires
of sluggish downloads, resulting in interminable waiting while images are slowly painted onto
the screen. Additionally, the demand for Internet multimedia services including video services is
a major driver.
Currently, several technologies exist to provide users with high-speed data access to the Internet.
ISPs are frantically adding 56 Kbps analog modems to their access lines, assuming users will
purchase the expensive analog modems also. Some users are connected to ISDN (Integrated
Services Digital Network) and for an initial outlay of several thousands of dollars and a princely
monthly sum can download data at either 64 Kbps or 128 Kbps, depending upon the class of
service they have subscribed to (paid for). But 128 Kbps is still not fast enough for audio and
video applications. Other users download Internet data via satellite at 400 Kbps for about $550
per month plus a capital equipment outlay of approximately $540. As fast as 400 Kbps seems,
especially compared to the ubiquitous 9.6 Kbps and 14.4 Kbps analog modems, it is still not fast
enough for audio and video applications even with the latest audio and video compression
techniques. Audio in the context of this book means high quality, stereophonic audio.
ISPs currently depend upon POTS (Plain Old Telephone Service) to provide access to most
homeowners and businesses. While some businesses and a few homeowners can afford ISDN and
the attendant fiber cost, most Internet service is provided over copper wire. Internet service is
limited by the speed of the analog modems hooked on each end of the transmission path between
the user and ISP. The analog modems are limited by the 4 KHz audio bandwidth of the copper
connection. Current analog modem technology data rates reach 56 Kbps and most industry
engineers agree that 56 Kbps is the upper limit of analog modem technology. 56 Kbps does not
come close to providing the speed required for audio, video, and multimedia services.

Figure 1-8 Evolution of information systems technology

Application for Video

CATV companies have coaxial cables that have a bandwidth up to 900 MHz running into homes.
But TV channels consume bandwidth at a fierce rate, allowing cable companies to carry a
maximum of 141 program channels downstream (to the user). All CATV programming is
provided over community coaxial delivery systems. In 1996, major CATV companies laid off
thousands of employees and fought for rate increases in many of the communities they serve. The
driver behind the layoffs and rate increases was the rising cost of maintaining the expensive, and
aging, coaxial based delivery system.
ATM and ADSL will allow video services companies to provide more channels of video
programming and still have bandwidth for other purposes, such as Internet access and upstream
(from user to the service provider) communications without laying a single coaxial cable. All they
need do is transmit their programming over Granny Bells humble copper wire.

Applications for LANs

Legacy local area networks (LANs) such as token ring, Ethernet, and token bus are speed limited
to 10 Mbps. Due to the data rate restriction, applications requiring faster rates must utilize other
technologies, limiting LANs to primarily data applications in networks that are not easily
"networked." Multimedia applications typically are a combination of voice, video, and data. Due

to the bandwidth needs of multimedia, they require higher data rates than legacy LANs can
deliver. ATM meets the bandwidth needs of multimedia and can be implemented in LANs, wide
area networks (WANs), metropolitan area networks (MANs), and global area networks (GANs).

Broadband Services
Broadband services is defined by ITU-T (formerly the CCITT) as a service requiring
transmission channels capable of supporting rates greater than 1.5 Mbps or primary rate in ISDN,
T1, or DS-1 in digital technology. Bandwidth is related to data rates. The higher the data rate,
the more bandwidth that is needed.
Analog voice traffic, the historical (legacy) method of communicating over telephone network
systems, gave way to digital data traffic with the advent of computers. With the proliferation of
the home personal computer, the demand for high-speed data services has become a roar. In
addition, the demand for multimedia and video services adds to the cacophony of calls for faster
data rates. To support the demand for the high-speed data services, bandwidth must be increased.
ATM provides for an increased bandwidth compatible with all types of service. Additionally,
ADSL provides the ability to use the increased bandwidth over the readily available copper wires
present in the vast majority of homes and businesses.
ISDN has long been considered the medium for transmitting high-speed digital data. With a
current transmission rate of 64 Kbps or 128 Kbps, ISDN can hardly be considered broadband, or
high speed. Since ATM provides the encoding method to expand ISDN bandwidth, two types of
ISDN must now be distinguished. N-ISDN refers to the Narrowband Integrated Services Digital
Network of 64/128 Kbps transmission speeds, and B-ISDN refers to the Broadband Integrated
Services Digital Network of 1.5 Mbps transmission speed or greater.

Terms and Definitions

One must learn the language of the trade if one hopes to have any chance of understanding what
goes on. When the number of terms in any particular field is small, many people can understand
the terms and therefore understand the field. However, when a particular field has many terms,
few people, relatively speaking, will learn and understand the language, therefore impeding their
ability to grasp all the technicalities of the field. Modern communications is full of terms, old and
new. A presentation of the full communications vocabulary is outside the scope of this book.
There are several good communications/telecom dictionaries available.
When a field, such as telecommunications, uses terms that are used by other fields, some
confusion is sure to creep into the picture. Terms used in this book are defined in context of their
usage in the field of communications, particularly as they apply to ATM and ADSL.
Data and information are used interchangeably throughout the text just as they are used
interchangeably in the industry. Service provider is also known as a carrier and long-haul carrier.
The service providers backbone refers to the service providers network which is available for
users to connect into, usually for a fee.

Communications, networks, and systems are terms that are used loosely to describe seemingly
innumerable, and usually unrelated, ideas and objects. Communications may describe any one of
various methods for the transport of information from source to destination. There is RF
communications which encompasses the field of radio wave propagation, telecommunications
which encompasses the field of communications via the Public Switched Telephone Network
(PSTN), data communications which encompasses the art of transporting information using
computer (originally) technology, and optical communications which concerns the usage of light
wave technology for the purpose of transporting information. There are other communications
technologies that are used today and there will most likely be more in the future.
Networks refers to the connection of equipment, usually computers, to allow the users to
exchange information. There can be a network of computers, a network of communication
switches, or a combination of switches and computers. Network really refers to any loose
association of multiple quantities of anything electronic.
Fields overlap, terms are born and die continuously, and the industry water can get real muddy.
But this book uses all terms in a straightforward manner that is understandable in the context
used.

And the Winner Is

Our communication systems evolved over the years as a function of the technology available at
the time. We designed and built switching and transport systems capable of meeting the
immediate needs of a growing populace. Yet, throughout all the networks and transmission
systems and interconnected mediums, there was not a single standard governing all types of
communications. The result is a hodgepodge of incompatible technologies including
management, encoding, switching, and transmission technologies that are targeted for very
specific communication services. Figure 1-8 details the evolution of primary information systems
from early childhood to adulthood. Notice that all information systems, in their dotage, converge
on ATM and ADSL. ATM is the technology that will combine all the various systems into a
seamless network of high-speed data communications that truly removes the speed limit from the
information superhighway. And ADSL is the transport technology to get the information
superhighway to your home or business.
The telecommunications industry is quickly converging now on ATM. Five major U.S. telecom
companies have ATM networks in place that can provide some degree of access from virtually
any location in the continental United States. Additionally, service providers in other countries
have ATM networks already providing limited service to Canada, Europe, and Asia.
ADSL technology is not as advanced as ATM. By way of comparison, you could say both
technologies are still in their infancy. ATM is just beginning to take tentative steps in the
progression from crawling to walking and is by no means potty trained, while ADSL is just a
newborn, still in the hospital nursery, waiting to come home to loving (we hope) parents.

Summary

Communication is necessary for us to survive in an inhospitable world. Without communication,

the evolution of humankind might have stopped near the same branch of the family tree occupied
by chimpanzees. Because we developed more complex communication methods, we rose higher
on the family tree.
Communication requires a transmission system, such as a drum, to get the intelligent information
from one place to another (source to destination). Communication also requires the encoding,
such as a specific pattern of drumbeats, of information by the source for transmission to the
destination. And the destination must reliably decode the information received.
Any new communication technology that dramatically leaps ahead of all the previous
technologies and is accepted for general use by the populace always results in interesting and
dramatic changes to the way we live, play, and conduct business. Examples are the advent of the
telegraph and, of course, the ever-so-humble telephone set. The telegraphs nearest competitor
when it was invented was the printing press and the nearest competitor to the telephone was the
telegraph. Each succeeding communication technology resulted in dramatic and fundamental
changes to how we live and to us. Now, ATM and ADSL will offer unprecedented access to
virtually all the worlds information and to virtually all the worlds inhabitants. ATM and ADSL
enable global unification of the human tribe. Imagine the world tomorrow.

Chapter 2
Basic Communication
Technology
Questions Answered in This Chapter
What is an analog signal?
What is a digital signal?
What is analog-to-digital signal conversion?
How are digital signals multiplexed?
What is bandwidth?
What is Narrowband ISDN?
What is Broadband ISDN?
What does all this have to do with ATM and ADSL?

The Basics

To understand why ATM and ADSL are necessary for the continued evolution of communication
systems (and the human creature) and how ATM and ADSL fit into the current communication
technology, an understanding of some fundamental communication concepts and a review of the
developments in transmission and switching systems since the advent of the telephone is
necessary. Toward that end, this chapter presents those fundamental concepts in as
straightforward and simplified a manner as possible without reliance on a difficult mathematical
treatment of any topic. The presentation is of an introductory nature beginning with very basic
electronic theory building blocks and only covering those topics necessary to understand ATM
and ADSL. If the reader is interested in a more in-depth treatment of any topic, the bibliography
provides sources rich in material.
The material is presented in a logical sequence. Unless you are already familiar with a topic, the
best course is to read each of the following sections in the order presented, as each lays the
foundation for the next. Upon completion you will be familiar with the concepts, approaches, and
terminology used in modern communications technology.

AC/DC
Generally, there are two basic types of electrical current, alternating current (AC) and direct
current (DC). Typically, AC current produces an AC voltage while a DC current produces a DC
voltage. AC voltage/ current is alternating between a maximum positive level and a maximum
negative level over some period of time. That is, AC voltage/ current is not static, but always
changing. DC voltage/current never changes, always remaining the same level within some
specified tolerance. See Figure 2-1. In communications equipment and systems, DC voltage is
used mainly to bias, or set, the proper operating conditions for semiconductor and integrated
circuit (IC) devices. AC voltage is used primarily to transmit signals over the various
transmission systems, such as coaxial or microwave links. AC voltages are also used in electronic
equipment to generate the DC voltages used to bias the equipment.

Figure 2-1. Basic electrical signals

Analog Signals
An analog signal is characterized as continuous and changing with respect to time. An analog
signal possesses level and periodic characteristics. An AC signal is always an analog signal. Yet,

an analog signal is not necessarily an AC signal. The signal represented in Figure 2-2 is referred
to as a sinusoidal waveform, as the value of the waveform, at any point in time, varies in
accordance with the sine of the function. It is both an analog signal and an AC signal.

Figure 2-2. Graphical representation of an analog signal

AC and analog signal levels are typically measured in volts either from peak to peak (pk-pk
volts), just peak (pk volts), average (avg volts), or RMS (rms volts). Peak to peak levels are
measured from the maximum (peak) positive value to the maximum (peak) negative level. Peak
level is measured from the zero reference level (the point where the waveform goes from concave
up to concave down) to either the positive or negative peak. Average level is the average DC level
(zero in this case since there is as much "negative voltage" as there is "positive"). The RMS level
is .707 times the maximum or minimum peak level.
An analog signal's time period is the time the waveform takes to go from any value through a
positive peak and a negative peak and return to that same value. Typically, the period is
determined by taking the time for the waveform to go from a positive peak to a negative peak
back to a positive peak. One complete cycle represents one period of the waveform. Frequency is
the number of periods of a signal in one second measured in hertz (abbreviated as Hz). See
Figure 2-3. So, 1.5 MHz represents a signal that varies through its complete cycle (positive peaknegative peak-positive peak) 1,500,000 times per second.

Figure 2-3. Signal period

Only a changing signal such as an analog or digital signal can convey intelligent information. A
signal that never changes (DC) cannot be used to convey much useful information. Two analog
signals, one representing the varying information such as voice and called the modulating
frequency, and a constant frequency analog signal called the carrier, can be combined in a process
called modulation (transmit side) and separated in a reverse process called demodulation (receive
side) to transmit the information long distances.

Figure2-4. Graphical representation of binary bit stream

Digital Signals
Digital pertains to data in the form of digits. A digit is the unique numeric value used as a
placeholder in some numbering system. As an example, 0, 1, 2, 3, 4, 5, 6, 7, 8, and 9 are all digits
used in our base 10 numbering system. In communication systems, digital refers to the use of
"digital" devices that use a binary numbering method to represent data. Confused? Binary refers
to the possibility of having only one of two possible values, or states. In this instance, the value is
either a zero or a one. So the terms digital signals, digital computers, digital transmission, etc.,
always refer to using binary signals (zeroes or ones) to represent the data, or information, that is
being processed. "Digital" and "binary" are sometimes used interchangeably.

Figure 2-5. Representation of binary quantities

Binary signals are signals that assume one of two possible states, with each state characterized by
its amplitude. See Figure 2-4. A one (1) is a high level, typically 2.5 to 5 volts, and a zero (0) is 0
to 1.5 volts. Of course, different families of logic types, such as Transistor-Transistor Logic
(TTL) and Complementary Metal Oxide Semiconductor (CMOS), have different levels for ones
values with 0 volts the usual zero value. But for learning purposes, assume a logic one of 5 volts
and a logic zero of 0 volts, as appropriate.
A single binary value is called a bit, also known as a pulse. A bit may be either a one or a zero.
Figure 2-5 shows a bit as a binary value of one. A single bit by itself is only useful for
representing one of two possible states (one or zero). To represent addi0tional states, additional
bits must be included in some manner. For the math hounds, you can represent (x^n) states
(things) with x^n digits. That is, four digits (or bits, same thing) can represent 16 values. That is,
the decimal numbers 0-15 can be represented with four digits.
Typically, bits are grouped as follows: four bits make a nibble, eight bits make a byte
(pronounced bite), 16 bits make a word, and 32 bits make a double word. When two or more
words are grouped together, a data stream is formed. Sometimes, a byte is called an octet.
The rate at which binary signals are passed along from device to device is called the period
(measured in seconds). Usually, the period is measured from the leading edge to the trailing edge
of the pulse. The frequency of the bit stream is the number of periods in any given second.

Figure 2-6. Converting an analog signal

2-Input AND Gate
2-Input NAND Gate
Possible Inputs Possible Outputs Possible Inputs Possible Outputs
Input A Input B Output C
Input A Input B Output C
0
0
0
0
0
1
0
1
0
0
1
1
1
0
0
1
0
1
1
1
1
1
1
0
Table 2-1. Digital AND and NAND gates (digital "state machines")
The use of digital integrated circuits and the continued reduction in circuit geometry is
responsible for the breathtaking evolution in technology in the last 20 years. A common digital
circuit is the AND device, also called an AND gate. The AND gate "ands" the inputs to give a
particular output in accordance with the specific combination of inputs. An AND gate may have
more than two inputs but will always have one output. From Table 2-1 you may see that the
output is always zero (low) until all the inputs are high, and then the output switches to a one
(high). The NAND function is just the opposite of the AND function.

Figure 2-7. Simplified pulse amplitude modulation

2-Input OR Gate
2-Input NOR Gate
Possible Inputs Possible Outputs Possible Inputs Possible Outputs
Input A Input B Output C
Input A Input B Output C
0
0
0
0
0
1
0
1
1
0
1
0
1
0
1
1
0
0
1
1
1
1
1
0

Figure 2-8 Single polarity PAM

Table 2-2. Digital OR and NOR gates
The OR gate output is always high if any input is high. See Table 2-2. The NOR function is just
the opposite of the OR function. There are many types of digital devices, but almost all of them
are composed of a few fundamental gate types such as the AND and OR and their logical
inverses, the NAND and NOR gates. Using the AND and OR gates, one can construct a
wonderful array of very useful digital devices.

Signal Conversion
An analog signal can be represented by a series of digital pulses. The particular circuit that
performs an analog-to-digital signal conversion is called, appropriately enough, an analog-todigital converter, or ADC. An ADC circuit samples the analog signal to be converted at periodic
intervals and generates a binary value representing the value of the sampled voltage. See Figure
2-6. Periodically, at the points in time designated by A, B, C, etc., the analog signal level is
measured. The measured signal level is converted to a digital number representing that level. See
the table in Figure 2-6. As an example, the voltage measured in time indicated by point E is 4
volts. Our ADC converter will output the digital bits corresponding to 0100 to represent the 4volt level at point E.
Converting a digital signal back to an analog signal is the reverse of the analog-to-digital signal
conversion process. The circuit that performs the digital-to-analog signal conversion is called,
appropriately enough, a digital-to-analog converter, or DAC. The DAC outputs a voltage
corresponding to the value of the set of bits received. In our previous example, the ADC
converted 4 volts to the digital bit stream 0100. Now, the DAC will convert the digital bit stream
0100 to 4 volts.

Pulse Modulation/Demodulation
To convert an analog signal to a digital signal, you must start with a uniform periodic reference
called the pulse train. See Figure 2-7. A theorem states that for complete recovery of intelligence,
the minimum sampling rate (the "clock" rate or periodicity of the uniform pulse train) is twice the
maximum frequency of the analog signal. This minimum rate is called the Nyquist rate.
During the sampling period, the analog signal level is "measured." The measured analog signal
level is then represented by a discrete, digital signal whose level has some relationship to the
measured analog signal level. This type of digital sampling is called pulse amplitude modulation
(PAM). The output of a PAM modulator is a series of pulses whose amplitudes are related to the
amplitudes of the analog signal at the point in time the samples were taken.
If the analog signal is a constant frequency and constant level signal, PAM produces a bit stream
with no variation in its level. Not very useful for conveying intelligence. However, voice is not a
constant frequency source. Voice levels and frequencies are changing as we talk. The PAM digital
pulse train that represents analog voice signals is therefore constantly changing its amplitude in
accordance with the voice level changes.
Single polarity PAM can easily be generated by offsetting the AC analog voltage with a DC offset
voltage so that the minimum AC voltage level is above 0 volts. See Figure 2-8. To recover the
original intelligence, the PAM pulse train is applied to a low pass filter. PAM signals are easily
distorted, so the technique is rarely used to transmit information directly. But PAM is usually an
intermediate step in other forms of pulse modulation.

Note that only the level of the sampling pulse train changes in PAM. In pulse time modulation
(PTM), the pulse train amplitude remains constant while the timing of the individual pulses varies
with the level of the analog signal at the moment in time the pulse occurs.
In pulse width modulation (PWM), the pulse train amplitude remains constant while the width of
each individual pulse varies according to the analog level at the moment in time the pulse occurs.
This is also called pulse duration modulation (PDM) and pulse length modulation (PLM).

Signal Multiplexing
For any given transmission type and for most users, bandwidth is wasted if a communications
channel sits idly waiting for a user to transmit something. How often is a homeowner's telephone
offhook, thereby engaging the line for use and compelling the telephone company to assign a
switched channel? (Assume for the moment that there are no teenagers in the household.) How
often does a client computer on a network demand host services? For much of the world, the
answer is, not often. So, data transmission can be characterized as "bursty"; that is, occurring in a
random manner. Dedicating communication channels to users would then result in a huge waste
of the commodity that is most in demand in communicationsavailable bandwidth.
The operative word is "available." The available bandwidth for communicating between your
home and the local access office until now was 4 KHz. Other bandwidth restrictions are present
throughout the public and private communication networks. Governing bodies throughout the
world define the available bandwidth for the various types of communication services.
Math wizards were able to model the bursty nature of human communications with some degree
of accuracy. Using the models for templates, modern communication systems are now designed
to provide reliable service while minimizing the use of available bandwidth and minimizing the
number of copper wires/fiber links necessary to provide the required quality of service. The
ability to perform these feats of magic is embedded in the technique of multiplexing digital and
analog signals.
Bandwidth is the limiting factor in communications. Within the confines of communication
technology, bandwidth is finite and there is just not enough to go around. As we demand more
and more communication services, bandwidth will become even more precious. To accommodate
hundreds of millions of users in the public switched communication networks with an individual
dedicated channel for their sole use would be impossible due to the need to wire the country and
world with sufficient copper and/or fiber to accommodate that many users. The cost alone would
be astronomical. Assuming not everyone desires access simultaneously to the switched networks,
thereby lessening the demand on the number of available channels at any given moment, even to
wire the globe for such service would still be prohibitive. The most economical approach is to
place in service as few physical wires or fiber links as possible, and maximize the utilization of
the available channels within the parameters of some defined Quality of Service (QoS). Both the
public switched telephone network and private networks seek to maximize the use of their
available bandwidth. ATM and ADSL help them do just that.

Initially, Plain Old Telephone Service (POTS) was the only customer of the public switched
telephone network (PSTN). Before long, teletype machines were developed that allowed the
transmission of data faster than a Morse code key operator could tap out messages on a manual
keyset. For many years, voice and teletype data transmission were virtually the only two uses of
the PSTN. After World War II, developments in electronics paved the way for uses of the PSTN
that the original network planners never dreamed of. Faster and faster data rates were the
objective as the race for new services, and revenue streams, was under way, especially after
deregulation of the telephone industry. Numerous types of data acquisition and transmission
technologies came into existence, each demanding its own special treatment by PSTN operating
companies. Telephone companies offered the new data transmission services in conjunction with
POTS as government agencies insisted that all new services had to be compatible with the
extensive in-place POTS network.

Figure 2-9. Transmission channel utilization by PSTN user characteristics

Today, many different services requiring PSTN access are gobbling up the system resources.
Typically, the bit rate of any particular service is constant for the period of time the service
actually transmits its data. But most services do not transmit continuously, 24 hours per day,
seven days per week. From the perspective of PSTN operating companies, these variable and
individually unpredictable demands on system resources give the appearance of users with an
extremely variable bit rate. While any one particular service or company may have a very
constant bit rate, the sum of all services results in a variable one. What we are left with is a batch
of user services with variable bit rates that all operate totally without regard for any other service
or user.
Bit Rate (Kbps)

Call Duration (secs) Burstiness (%)

Min

Max

Min

Max

Min

Max

telemetry

0.01

10

10

10

videotex

64

100

1000

10

10

video surveillance

10

30,100

1,000

10,000

10

100

teletex

10

10

100

100

100

100

electronic mail

10

10

100

100

50

50

voice mail

16

64

10

100

30

50

video telephony

64

70,100

100

100

50

100

telephony (POTS)

64

64

100

100

30

100

facsimile

64

64

10

100

100

100

hifi audio

768

768

1,000

10,000

100

100

video conference

1,000

70100

1,000

10,000

50

100

broadband videotex 1,000

70,100

100

1,000

10

100

video mail

1,000

70,100

10

1,000

10

50

TV

30,100

70,100

1,000

10,000

100

100

HDTV

140,100

565,100

1,000

10,000

100

100

North American
Digital Signal
Number
DS-1
2.048
DS-1C
8.448
DS-2
34.368

CCITT (International)

No. of Voice Data Rate Level No. of Voice

Channels
(Mbps) No. Channels
24
1.544
1
30
48

3.152

120

96

6.312

480

Data Rate
(Mbps)

Table 2-3. Characteristics of PSTN users

Inherently, the PSTN is a constant bit rate medium. The PSTN data streams, such as DS-1, are
fixed bit rate streams. In the case of DS-1, is fixed at 1.544 Mbps by regulatory bodies. Each of
the 24 voice channels comprising the DS-1 pipe have a fixed bit rate of 64 Kbps, regardless of the
particular service that might be using the individual channels moment to moment. However, a
glance at Table 2-3 reveals that most user services operate at bit rates other than 64 Kbps.
Additionally, each user does not transmit continuously, even when switched to the destination.
The "burstiness" is a measure of the amount of time the service actually transmits while
connected to the PSTN. For example, some telemetry source may be switched (connected) to a
destination for 1 to 10 seconds, yet during the time period the source is connected, it is actually
transmitting data 1 to 10 percent of the time. For the remaining 90 to 99 percent of the time, the
switched circuit is actually idle, waiting for the telemetry source to do something.
Channel A in Figure 2-9 shows the channel bandwidth usage of a data service that operates with a
burstiness of 10 percent, such as videotex. Channel B shows the usage by a service that utilizes
the available channel bandwidth with a burstiness of 30 percent, such as POTS. Channel C shows
a service that utilizes the bandwidth with a burstiness of 50 percent. Obviously, regardless of the
call duration, much bandwidth is wasted if a single channel is reserved for the exclusive use of
most data services.
PSTN operating company engineering wizards realized that if they could utilize the idle time the
switched channel was not doing anything for some other purpose, they could dramatically
improve system capacity with little additional investment in transmission infrastructure. So was

born the concept of multiplexing signals onto channels. Electronic circuits at the physical layer of
the OSI model can combine diverse digital signals into a single signal for transmission on a
single channel. Using statistical probability theory, network engineers determine how much
physical capacity any particular network must have to process all requests for service within
some predefined Quality of Service parameter.
Signal multiplexing refers to the technique of combining two or more electrical signals into a
signal that still contains all the data of the two original signals. There are two basic methods of
multiplexing communications signals: frequency division multiplexing (FDM) and time division
multiplexing (TDM).
FDM is an analog signal multiplexing technique where several analog signals are assigned
adjacent frequency bands and transmitted simultaneously over the transmission medium.
TDM is the technique of dividing time into increments and assigning each user a particular
increment or time slot for transmitting their data. The sequence is repetitive, with each user
always using the exact same time slot. Time slots are referenced to some specific starting time.
TDM Carrier Standards

Regulating bodies have assigned the following data rates and channel designations, as shown in
Table 2-4, for TDM multiplexed signals. A DS-1 signal is composed of 24 voice channels, each
specified to operate at 64 Kbps. Astute math wizards will immediately perceive that 24 times 64
Kbps does not equal 1.544 Mbps. It is 1.536 Kbps. We are short 8 Kbps. The 8 Kbps is used for
system maintenance and monitoring purposes.
Table 2-4. TDM carrier standards
Telephone companies and network engineers are accustomed to discussing their networks in
terms of DS signal capacity. Remember, 64 Kbps (a single voice channel) is the standard signal
for ISDN, and as stated previously, ISDN is not broadband at all.

Bandwidth
A spectrum is all the total possible values of some entity. There is a spectrum of electromagnetic
waves ranging from 1 Hz to 10^24 Hz. The electromagnetic spectrum is divided into sections that
are called x-rays, gamma rays, ultraviolet, visible, infrared, microwaves, radio frequency (RF)
waves, and long waves. Each section has unique characteristics useful for human exploitation that
justify the particular groupings. We are primarily concerned with RF electromagnetic waves,
although SONET/SDH is intimately concerned with the visible light portion of the
electromagnetic spectrum.
Each section of the electromagnetic spectrum has a finite number of values. The total possible
values of each section define its total bandwidth. Bandwidth, in the general sense, is the specified
values of some entity of interest that occupy a portion of the spectrum. Bandwidth is then the
range from smallest to largest of some portion of the electromagnetic spectrum. Take RF waves

as an example. RF waves begin at about 1 MHz and go up to 1,000 GHz. While there is a range
of six orders of magnitude from the low end to the high end, there are still only a finite number of
possible values. And in todays communications-hungry world, six orders of magnitude is getting
very small. So, bandwidth becomes a limiting factor that determines how much information can
be transported.
Data rate and bandwidth are related. Take an AC analog signal sampled at some defined sampling
rate. For this example, lets assume the sampling rate is the Nyquist rate, or two samples per
hertz. The analog signal has a positive half-cycle where the signal rises from zero level to some
maximum positive level, then decreases back to a zero level. Then the signal enters the negative
half-cycle, increasing from zero level to some maximum negative level before decreasing back to
a zero level. The combination of the two half-cycles, positive and negative, constitute one
complete cycle, known as a hertz. The cycle is then repeated at some frequency determined by
circuit elements. Since we are sampling at the Nyquist rate, we must divide each cycle into two
periods and sample once during each period. With two half-cycles comprising the full cycle, it
seems reasonable to divide the cycle into its half-cycle components and sample during each halfcycle. So, for each cycle, we get two samples, each corresponding to a half-cycle. Once cycle,
two samples; one hertz, two bits. So, at the Nyquist rate, the data rate is twice the frequency.
Generally, though, sampling is conducted at rates greater than the Nyquist rate. This
oversampling is necessary in audio, voice, and video circuits to ensure true reproduction of the
original information. An oversampling rate of eight times is common, yielding 8 bits per hertz.
POTS has a bandwidth of approximately 4 KHz. The lower sideband cutoff is 300 Hz and the
upper cutoff frequency is 4 KHz. This means that all voice signals varying between 300 Hz and 4
KHz, called the passband, are easily passed by the customer premises equipment (CPE, aka
telephone), central office (CO) equipment, and transmission systems. The highest frequency in
the passband is called the upper cutoff frequency and the lowest frequency in the passband is
called the lower cutoff frequency. Everything in between is in the passband of the circuit.
Frequencies above and below the passband frequencies are attenuated (suppressed). The voice
passband is 300 Hz to 4 KHz and the bandwidth is 4 Khz (technically 3.6 KHz). Audiophiles
know that high fidelity sound transmission requires a much broader passband, as true
stereophonic reproduction requires the presence of signals as low as 60 Hz and as high as 20
KHz. But Ma Bell decided long ago, in the interest of economics, that a 4KHz voice bandwidth
was sufficient to convey enough voice frequencies that voice communications would be
intelligible. Grandma may not sound like she is standing right in front of you, but you can still
recognize her voice over the telephone. That is, there will be few people who confuse Grandma
with Grandpa when talking with them over the phone, although neither sounds like they are
sitting next to you.
The voice passband of 4 KHz has worked well for many years. Now, we wish to pass data over
the system that was originally designed for analog voice communications. Analog signals,
characterized by a continuous change in magnitude with respect to time, require a lot of time to
convey intelligent information. Some method of "speeding up" the transmission of information
was required. Digitization of analog signals and information was the answer.
Now we have a series of digital bits representing an analog voice signal. These bits can be
compressed in time, allowing other bits from another voice signal to be included in the overall bit

stream. This technique of "marrying" several different signals into a combined signal is called
multiplexing. Additional compression techniques can be used, allowing many different signals to
be multiplexed together and transmitted simultaneously in some specified passband.

Narrowband ISDN
Long ago in the Dark Ages, circa 1970, and long before anyone really imagined the proliferation
of desktop computers, Integrated Services Digital Network (ISDN) was conceived as a means of
transmitting digital data at a whopping 64 Kbps speed. In terms of digital speeds available then,
64 Kbps was blazingly fast. Cold molasses, quietly gripping the sides of the bottle and refusing to
budge no matter how hard you beat on the side of the bottle, seems speedier than 64 Kbps today.
Yet, some visionaries spoke of a future broadband, or high-speed, ISDN. Only now, with ATM
technology, can we really begin to differentiate between two ISDNs: narrowband and broadband.
Narrowband ISDN (N-ISDN) now refers to the original 64 Kbps ISDN service. Many ISDN
connections were made in the U.S. and service providers are not in any hurry to abandon
profitable legacy systems. So we give it another name while making room for an additional ISDN
technology.

Broadband ISDN
Broadband ISDN (B-ISDN) refers to integrated digital services with a bandwidth from l.5 Mbps
up to hundreds of Mbps. ATM technology is the father and fiber optic cable is the mother that
will make possible the high data speeds necessary to give birth to B-ISDN. The birth of B-ISDN
will have a profound impact on every living creature on this planet. Our ability to communicate
determines how we evolve, and B-ISDN will revolutionize our ability to communicate.

Summary
Human intelligence is conveyed from person to person through the transfer of information.
Information must be properly encoded by the source and decoded by the destination. And
information must be converted to the proper type of signal for the transmission medium used
from source to destination.
Human perception is inherently analog in nature. Our speech and our hearing, two important
ingredients for sending and receiving information between two humans, are analog functions.
Yet, digitally encoded information is more efficient in terms of the bandwidth used. So, while
humans must necessarily receive and transmit information in an analog fashion, modern
communication systems must change the basic human-originated analog information into digital
data to accommodate all the information requiring transmission. To preserve precious bandwidth
even more, communication systems multiplex and demultiplex the digital data. Still, digital
signals are not useful for long-distance communication. The digitally encoded information must
be superimposed on an analog signal for efficient long-distance transport. ATM is basically a
technique for encoding and multiplexing digital data for transmission and ADSL is basically a
technique for multiplexing digitally encoded analog data for transmission.

Chapter 3
Transmission and Switching
Systems
Questions Answered in This Chapter
What is a transmission system?
What is a switching system?
What do transmission and switching systems have to do with ATM and ADSL?

Transmission Systems
Why does someone working to understand ATM and ADSL require a dose of mundane and
flightless transmission and switching systems? Because ATM is a transport and switching
technology. Because ADSL is a transport and switching technology. Because ATM and ADSL
must exist and work with legacy transmission and switching systems. Because ATM and ADSL
must cope with the multitude of transmission and switching seams. Because ATM and ADSL
are the wings of a now flight-enabled global communications system that can soar with no known
bounds. A believer must have a fundamental idea of the legacy transmission and switching
systems empowered by ATM and ADSL to fully appreciate the power of emerging ATM and
ADSL technologies.
Transmission systems are necessary to convey intelligent information from source to destination.
Knowing all the information in the world does not have much practicality if one cannot transmit
some portions to a recipient and therefore make some use of all that information. Whether the
transmission system is the voice and ear working together and jointly using sound waves to
transmit the information from source to destination (also called end-to-end), or an expensive
government network of computers and digital radios spread around the globe, transmission
systems are an integral element of our communications networks. Without the ability to transmit
our knowledge, the possession of knowledge itself would dramatically lessen in import and there
would be an attendant decrease in our standard of living and even our ability to evolve. Without
information transmission systems, we might be little better off than chimpanzees.
Given that transmission systems are so important to our ability to communicate, which dictates
our ability to evolve, it is no surprise that humans utilize numerous transmission systems. We are
concerned with transmission systems that have found particular usage in the world of
telecommunications. Since the topic of this book is ATM and ADSL, we will confine the
discussion to those systems specifically useful for the transmission of ATM and ADSL
information. But first, we need to know something of the past and how we came to where we are
today.

A Short History of Communication Transmission

Systems

In the late 1800s, the dominant communications technology was analog in nature. Analog signals
are characterized as being continuous in time with respect to some constantly changing value
such as amplitude and/or frequency. Mathematically speaking, an analog signal has at least a first
derivative. The advent of the telephone gave rise to analog voice communications using twisted
copper pairs for single channel communications transport. That is, one pair of copper wires could
carry one-half of a two-way (half duplex transmission) conversation mechanically switched in the
central office from the source of the call to the destination.
Development of the vacuum tube in 1907 led to analog systems utilizing FDM (frequency
division multiplexing) in 1925. FDM is the process of simultaneously transmitting two or more
signals representing intelligent information over a common path by using a different frequency
band for each signal. That is, a frequency domain is divided into discrete bandwidths, called
channels, with each channel used to transmit a unique signal representing some particular
information. Once multiplexed, FDM signals could be transmitted at low cost over long
distances. Multiplexing was expensive, so the technology was reserved for long-distance and
high-volume service where the multiplexing cost was more than offset by the reduction in
transmission cost. FDM technology resulted in the consolidation of the long-distance network
into fewer routes with much more capacity while microwave transmission technology,
characterized by its analog nature to date, progressed from 4 GHz through 11 GHz frequency
regions.
The invention of transistors in 1954 was to have a profound impact on every aspect of our lives
including communications systems. Electronic equipment became smaller, faster, less power
hungry, easier to maintain, easier to operate, easier to build, and less expensive. Also, the
transistor opened the door to a technology-based world full of opportunities that previously was a
window that only the interesting characters among us could occasionally peer into.
In 1969, transistor-based digital wireless systems were developed. Wireless systems is another
name for transmission systems that exclude wire-based transmission systems, such as copper and
coaxial cable, and encompasses RF transmission systems including microwave and satellite.
The concept of PCM (pulse code modulation) was developed in the late '60s. PCM is the process
where an analog signal, like voice, is digitally encoded. To generate a PCM signal, an analog
signal is sampled at discrete periodic intervals. The level of the analog signal as measured during
the sample interval is equated to a binary value. There can be an almost infinite range of binary
values to represent the continuous analog signal if exactitude is a requirement, which becomes
impractical. Why? Well, our digital world is pretty much based upon either 8-bit, 16-bit, 32-bit,
or 64-bit representations of data. And one can represent only so many things with 8, 16, 32, or 64
bits. Practically, discrete binary values of some limited range are chosen to represent the analog
signal. Typically, the number of values are closely related to the number of bits available in the
digital equipment. An 8-bit machine can represent values from 0 to 255 (0 - 2^8-1). Obviously,
every sample must be made to fit one of the available digital values. The process of fitting the
sample value to an available digital value is called quantizing and the loss of resolution is called
quantizing error, a source of noise generation in digital systems including ATM and ADSL.
An analog signal is sampled at the same repetitious rate known as the sample rate which must be
at least twice the maximum frequency of the analog signal for full recovery of the signal. The

minimum sampling rate is called the Nyquist rate. A signal sampled at less than the Nyquist rate
will result in less than full reproduction of the original intelligence. Since telephone systems limit
the audio (analog) signal to a maximum of 4 KHz, the sampling rate must be 2 times 4 KHz, or 8
KHz. Granny Bell does not really care about the higher order frequencies that add the warmth to
our voices.
By 1972, a digital channel bank combining digital wireless transmission systems and PCM was
developed that offered toll quality service. Toll quality refers to the minimum measure of service
quality sufficient to charge for the service. No one wants to pay for less than toll quality service.
A technique called time division multiplexing (TDM) was developed to maximize the number of
voice channels that could be transmitted over a single transmission line. An analysis of
communications revealed that during a connection much of the time was empty space where no
information was sent or received, especially if the call was between two lovelorn individuals.
Determining optimum rates of multiplexing based upon statistical methods, diverse signals were
combined on the same channel at the local exchange in such a manner that the daw signal still
contained all the intelligence of the original signals.
Soon after the appearance of digital wireless and PCM, a high bit rate digital system employing
TDM was developed. TDM is the process of dividing time into slots and assigning specific
channels to particular slots. After the analog signal was pulse code modulated at 8 KHz, the
resultant bit stream could be assigned a slot (channel) in a much faster transmission system. The
number of digital channels available depended upon the speed of the transmission system.
TDM was an expensive technology. Only upon the deployment of optical transmission systems
was TDM economically employed as a coding method. Optical transmission is accomplished by
combining the intelligent information with an unintelligent carrier (modulating) signal. The
modulated electrical signal that now includes the unintelligent carrier and the intelligent
information is converted to a light wave signal by a laser diode or light emitting diode (LED).
The modulated signal is carried over optical fiber from source to destination. The intelligent
information is then demodulated (carrier is removed). The intelligent information is reconstructed
electrically from the light wave by the action of pin diodes or avalanche photo diodes.
PCM and TDM encoding methods are used in optical communications systems which have
increased in data rates from 45 Mbps to 1.2 Gbps, 1.7 Gbps, and 2.4 Gbps. These
communications systems are identified in the United States and Canada as Synchronous Optical
Network, or SONET. Elsewhere, particularly in Europe, fiber optic systems became known as
Synchronous Digital Hierarchy (SDH). Now, we commonly see the two acronyms combined as
SONET/SDH. SDH is the ITU-TSS international standard for transmitting information over
optical fiber. SONET is the American National Standards Institute standard for optical fiber
transmission. A SONET system deviates slightly from an SDH system.
Before SONET/SDH, the most common digital encoding scheme in use, in the mid-80s, was
Plesciosynchronous Digital Hierarchy (PDH). PDH is still used in the form of DS-1/2/3/4 signals.
PDH is called plescio (meaning almost) because the transmission method is neither wholly
synchronous nor asynchronous. Synchronous in this context refers to the use of a specific time
relationship between two or more signals that must be maintained for the successful transmission

of the data and its eventual recovery by the end user. Asynchronous in this context refers to the
lack of a time relationship. PDH was replaced by SDH. SDH multiplexed the PDH signals into a
synchronous time division multiplexer (STM) of n = STM-n signal where n represents some
integer value.
There are two common transmission methods utilized by the majority of the communications
networks today. One method is called synchronous transmission and the other is asynchronous
transmission. Some confusion may result from the usage of these two terms as they mean two
very different things in RF communications and in telecommunications. The following discussion
endeavors to explain what the terms mean in telecommunications.
Synchronous

Synchronous in RF communications, and in digital systems, refers to the relationship of the data
to some time reference that is used to either generate and/or regulate the transmission of the
information between the sender and the receiver. A synchronous signal is clocked by some
repetitious reference electrical signal. The clock is like a drummer keeping the beat for a group
of professional dancers. See Figure 3-1. Like the dancers moving simultaneously to and fro to the
beat of the drum, clocked signals do not do anything until the proper clock signal provides the
go-ahead. This is because many electrical circuits are gated. Think of it as a complex maze
with a locked gate at the end of each section with all manner of people trying to get to the finish,
except that no one can enter a new section of the maze without a gatekeeper opening the gate. But
it is a fixed race, as the gatekeepers have specific instructions concerning who gets through first.

Figure 3-1. A synchronous bit stream

In telecommunications, especially in ADSL terminology, synchronous refers to the simultaneous
transmission and reception of data over the same physical medium (wires, coaxial cable, fiber,
etc.). See Figure 3-2.

Figure 3-2. Another synchronous bit stream

A user can receive data from a source while transmitting data to the source at the same time.
However, in RF communications, the simultaneous transmission and reception of information
over the same physical medium is called duplex operation. Usually, one can tell by the context
which synchronous is referred to. But not always.
Asychronous

Asynchronous in RF communications, and in digital systems, refers to the lack of a timing

relationship between the sender and the receiver of the data. Serial communication through the
voice frequency analog modem is a good example of asynchronous transmission. See Figure 3-3.

FIGURE 3-3. An asynchronous bit stream

In telecommunications, and particularly in ADSL terminology, asynchronous refers to the
differing data rates sent from and received by an end point of the transmission path, such as
customer premises equipment (CPE). If the CPE receives data at one frequency and either does
not transmit any data back to the source or transmits back at a much slower rate, then the link
between the sending and receiving points is considered an asynchronous link. As an example,
ADSL sending equipment can transmit 1.5 Mbps to a user who can return up to 640 Kbps.

Multiplexing
Multiplexing is the framework of the ATM and ADSL foundations. It allows multiple data signals
to be combined for transport. Multiplexing promotes efficient use of bandwidth as the process
maximizes the number of data signals combined into a single signal suitable for transmission.
Current telephone installations do not use multiplexing from the home to the central office (CO).
Multiplexing occurs at the CO where multiple voice circuits are combined into large bundles of
circuits.

Multiplexing is the business of placing data signals from multiple sources into a known sequence
on the same transmission medium for transport from source to destination. Two or more data
signals are combined in the transmission equipment in such a manner that neither signal disturbs
the fundamental relationships (level and timing) of the other signals and each independent data
signal is recovered intact by the receiving equipment. The technique of multiplexing signals is
made possible due to increased integrated circuit speeds and increased transmission speeds.
Take voice circuits as an example. To digitize a voice circuit, or channel as they are sometimes
called, the voice signal itself is periodically sampled. The Nyquist rate is two samples per hertz of
frequency. Telephone circuit voice bandwidth is about 4 KHz. So, the Nyquist sampling rate is 8
KHz x 8 bits / Hz = 64 Kbps. Now the transmission rate for ISDN is 1.544 Mbps. Somehow,
multiple voice circuits, 24 of them to be exact, must be combined to fully utilize the 1.544 Mbps
ISDN rate. Well, multiplexing is the technique that gets the job done.

Figure 3-4. A lightly loaded (multiplexed) and very inefficient train

Figure 3-5. A fully loaded (multiplexed) train

In a simplified explanation, each voice circuit is continuously digitized at the 8 Kbps sampling
rate. Each group of 8 bits from each circuit must wait its turn to board the ISDN train. Think of
each group of 8 bits as a tour group occupying exactly one passenger car in the train. If the train
leaves the station with only a single voice circuit occupying one car of the train, as shown in
Figure 3-4, the train does not operate very efficiently (wastes space or bandwidth). But, if the
train is fully loaded, as shown in Figure 3-5, then the train operates at maximum efficiency
(utilizes all the bandwidth to get the data to the destination). If you want to understand modern
communications technology, and particularly ATM and ADSL, you must understand the concept
of bandwidth and come to grasp the idea that bandwidth is gold, green gold, the color of money.
Service providers charge for the amount of bandwidth used by networks. And it is not cheap.
Now, each voice circuit has a specific order, like a seating arrangement, to load its bits into its car
in the train. The order the cars are in and the seating arrangement of each car is known to both the
departing station (source multiplexer) and the arriving station (destination demultiplexer). If any
circuit gets out of order, either loading onto the wrong car or not placing its bits into the correct
seats when loading, then the destination will unload (demultiplex) the signals incorrectly.
Kind of like multiple tour groups milling about a train station, some members will certainly
become separated from their tour group. Yet every tour group must immediately exit the station in
groups of eight people (bits). Another train is immediately behind and there is no time to waste.
So the mean stationmaster forces every group to immediately vacate the premises, whether or not
the original group member structure is intact.

Figure 3-6. Frequency division multiplexing

There are two flavors of multiplexing of interest in ATM and ADSL. One is called frequency
division multiplexing (FDM) and the other is time division multiplexing (TDM). The difference
between the two flavors is indicated by their names. One multiplexes signals in the frequency
spectrum, also called frequency domain, and the other multiplexes signals in the time domain.
Frequency multiplexing combines signals of differing frequencies into a single signal that is
composed of the differing signals. As an example, two original voice signals, channel one of 4
KHz bandwidth and channel two of 4 KHz bandwidth, are combined into a single signal that is
composed of the two original signals plus some carrier signal. See Figure 3-6. The technique
involves adding (modulating) the two signals onto the carrier at differing frequencies so the two
original signals do not interfere with each other. In the figure, 2,140 KHz is chosen for one voice
channel and 2,150 KHz is chosen for the other voice channel. The two resulting signals are
applied to a multiplexer circuit which is essentially a filter combiner that just combines the two
signals onto the same transmission path. Note if the two original voice channels are combined
onto the same transmission path while still original 4 KHz signals, the resultant combined signal
is gibberish.

Figure 3-7. Frequency division demultiplexing

Figure 3-8. Time division multiplexing

To restore the two original voice signals, the multiplexed signal must undergo a reverse
multiplexing process called demultiplexing. See Figure 3-7. Figure 3-7 is essentially the same as
Figure 3-6 except the arrows are turned around as the demultiplexer signal flow is the reverse of a
multiplexer. Also, instead of an adder circuit to mix the carriers with the original voice signals,
now there is a subtractor circuit to remove the original voice signals from the carriers.
Time division multiplexing (TDM) is a simple concept. Divide some increment of time, say one
second, into multiple slots and assign specific slots to individual voice circuits. If the sending
multiplexer and receiving demultiplexer agree upon the assignment of slots, then the recovered
data is identical to the transmitted data.
Figure 3-8 shows an increment of time divided into 24 slots. If the time period shown is one
second, then each slot occupies 41.67 milliseconds of time. Each bit of each slot then occupies
5.2 milliseconds.
How does this relate to telephone circuits? A voice circuit is sampled at 8 KHz, or every 125
microseconds (1/8 KHz). The voice sampling data rate is 64 Kbps (8 Khz / sample times 8 bits /
sample). The DS-1 transmission rate is 1.544 Mbps. There are 25 voice slots, each 64 Kbps, in
the DS-1 rate. So, if we fully loaded a 1.544 Mbps data stream with voice circuits, we could carry
25 phone conversations simultaneously. Simultaneously in this sense does not mean at the exact
same instant of time, but together as a data stream. In actual practice, there are 24 voice circuits
carried over a DS-1 line. The 25th slot is used by the telephone companies (telecos) for system
management. Have you ever clearly heard a strange, one-sided conversation of short duration
over the telephone? You guessed it. The mean stationmaster unloaded the wrong car on your
circuit and while you could hear the other party, they could not hear you. This situation was fairly
common in the '70s when multiplexing was in its infancy.
What does all this have to do with ATM and ADSL? The previous TDM example used voice
circuits as the model for explaining TDM. However, any digital data signal can be time
multiplexed. The very existence of ATM and ADSL is dependent upon the ability to multiplex
signals. ATM is commonly called statistical multiplexing because the telecos gamble with the
ATM TDM time slots. Sometimes you win and sometimes you lose, and if you want to beat the
house percentage, do not skip over the remaining chapters. ADSL uses FDM techniques to get the
message to the home.

ISDN Basics

Integrated Services Digital Network (ISDN) was originally conceived as an implementation of

digital technology that integrated the separate analog switching, transmission, and user premises
equipment and interfaces into a seamless network. To understand seamless, perhaps we
should first understand what a seam is.
In the old days, CO switching equipment, transmission facilities, and user equipment were analog
and considered as separate entities. For signals, especially digital signals as we entered the Age of
Gates, to transit the telephone company networks from home (source) to CO to long-haul carrier
(long-distance teleco) to CO to home (destination), the signal had to pass through the various
equipment seams. Each of these seams represents a beginning/ending in a stage of the journey
of the signal from source to destination. So a seam is the interconnection, or interface, between
each stage of the signals journey. Each of these seams might require the equipment to operate on
the signal in some manner that changes fundamental signal relationships such as level, timing,
and/or frequency, or even the type of signal (light wave, electrical, electromagnetic wave, etc.).
The idea of seams and stages in the transmission path is illustrated with a person setting out on a
journey around the country using various modes of transportation. The first stage of the trip
requires the person to walk from the home out to an automobile. The first seam occurs when the
person opens the car door and gets inside the auto. The second stage of the journey occurs when
the person drives to the airport. The second seam occurs when he exits the car at the airport. The
third stage occurs when he walks into the airport. Another seam occurs when the fellow gets on
the airplane. And the next stage of the journey occurs when the airplane takes off and then lands
at the destination. And the journey continues with stages and seams for each change in
transportation mode. Yet, a trip by automobile from source to destination would only require two
seams, getting into and out of the automobile at the source and destination, respectively.
Slow analog voice circuits could deal with mechanical switch gear and analog transmission
techniques relatively easily. But digital signals were appalled at the antiquated methods and
equipment used for analog signals. Since digital signals could not deal very well with the seams
in the transmission systems, the desire to develop a digital signal standard that would provide
seamless telephony service was born. That is where the Integrated comes from in ISDN. It was
originally IDN, or Integrated Digital Network, meaning a seamless digital network. As folks
talked about the concept, the idea of multiservices (meaning voice, video, data) was conceived.
So, we finally wound up with the Services part of ISDN.
ISDN is a transmission and switching concept that provides a seamless digital signal path from
source to destination. Assuming, of course, that all the telecos in the transmission path have
ISDN compatible equipment installed, seamless connectivity is end-to-end connectivity. Not all
telecos have ISDN equipment installed, especially in the former Soviet Union and third world
countries. (From this point forward the term seamless will be used without the quotes.)
An important enabling legacy technology for ISDN is light wave transmission. Previously,
without light wave technology, only very expensive coaxial delivery systems could transport
ISDN data rates. See Table 3-2 for SONET/SDH transmission rates. Light wave technology, with
the attendant fiber-based transmission facilities, provided a sufficiently fast transmission medium
for the ISDN data rates. Now, ATM and ADSL enable ISDN services over copper wire.

Narrowband ISDN

Narrowband ISDN is the term applied to ISDN services with a 64 Kbps data transmission rate.
When ISDN was originally conceived, 64 Kbps seemed terribly fast..... Please stop
laughing..... :)..... OK, thats enoughyou are starting to dribble. In those caveman, or cave-PC
days, modems operated at a blindingly fast 200 bps, making 64 Kbps seem like a rocket. Just be
glad you did not live in those dark ages and have to communicate with a 50 bps teletype.

Figure 3-9 A bit bucket too small

In those long ago days, all of the countrys information databases were text based. If you kept the
message short and sweet, 50 bps got the job done, more or less. Ever see a teletype (playboy)
picture? My, how the machines could labor trying to reproduce a graphic image.

Figure 3-10. A bit barrel just right(for now)

But why 64 Kbps ISDN? Commercially deployable digital technology was based on 16-bit
processors. How high can a 16-bit processor count? How about 64,000, rounded off? The actual
number is 65,536 but it is customary to write the number as 64 Kbps. This original 64 Kbps
ISDN is now referred to as Narrowband ISDN (N-ISDN) to distinguish it from Broadband ISDN
(B-ISDN).
Why is ISDN important to ATM and ADSL? Ever hear of legacy systems? Well, ISDN is the
legacy system that ATM and ADSL must deal with in todays teleco systems.
Broadband ISDN

ISDN at 64 Kbps was okay for the text-based traffic and the digitally encoded voice circuits that
originally used the ISDN networks. But faster speed processors (whoever thought of desktop 300
MHz processors 20 years ago?), 32-bit processors, the proliferation of personal computers, and
the Internet virtually made (N-) ISDN obsolete in a couple of years.

The original ISDN is a (bit) bucket too small for the Age of Gates with the attendant multimegabit data needs. See Figure 3-9. Now, we want data in really big (byte) buckets. One
economical solution that does not trash our legacy ISDN systems is to use multiple ISDN bit
buckets for the byte-hungry, image-based applications. If 25 ISDN buckets are combined, then
the resultant data barrel is 1.536 Mbps (64 Kpbs/ bucket x 24 buckets) deep. See Figure 3-10.
Telecos take the 1.536 Mbps barrel, stretch it by adding 8 Kbps for traffic management purposes,
and call it Broadband ISDN, also known as DS-1 at 1.544 Mbps.

Copper Wire
Copper wire was the original electrical transmission medium for the industrial revolution. As
revolutions go, the industrial could not have had a better partner. Copper has desirable
characteristics that make it very useful for industrial revolutions. Copper is cheap (well, it used to
be), has good electrical conductivity, weathers very well, and is easy to work with. Seems like the
industrialists were waiting with bated breath for someone to invent something that could use
copper as a major component of the product. Because the telephone and copper wire went
together like bread and butter, bacon and eggs, Elvis and 7-Elevens...you get the idea. As soon as
the telephone was invented, the race to literally (copper) wire the world was under way.
Copper wire is everywhere. There are 600 million homes and businesses worldwide wired with
copper. It is estimated 80 percent of wired premises are within 1,800 feet of a CO or access node.
The 80 percent/1,800 feet is a U.S. estimate, but let us extend the figure around the globe, not an
unreasonable feat. Then let us assume the average distance from CO to home is half of 1,800 feet,
or 900 feet, or approximately 0.2 miles. Ball park, 120 million miles of copper are snaking
around underground and overhead. That is a lot of copper wire. All this wire and all these
premises are candidates for ADSL modems.
Wire Type

Max Frequency

STP

155 Mbps

UTP-3

25.6 Mbps/51.84 Mbps

UTP-5

155 Mbps

Coaxial

45 Mbps/155 Mbps

Table 3-1. Maximum copper transmission frequencies

Copper wire has various characteristics that influence electrical signals including ADSL signals.
Telephone wire from the access to the premises, as I have mentioned once or twice (and will
mention once or twice more), is a twisted pair of wires. The twist is to reduce signal coupling
between the two wires and the additional surrounding wires when the twisted pair is placed in

large bundles. The twisted pair can be shielded, known as shielded twisted pair (STP), or
unshielded, known as unshielded twisted pair (UTP).
Shielding, or the lack of it, length, size, and age are important considerations for determining
what flavor of DSL technology is best for any particular application and location. All copper wire
is not made equal. Table 3-1 shows the maximum practical frequency for data transmission by
typical telephony wire type. Coaxial cable, of the flavor used in CATV premise wiring, is tossed
in for comparison. Most homes have the UTP-3 flavor. Well, how can twisted pair at 25.6 Mbps
compete with CATVs 155 Mbps? Read on, until the end (of the book) is near.

RF Propagation
While copper wire and telephony have had a 100-year marriage made in heaven, the marriage has
not been totally without discord. Sometimes, after the honeymoon glow was a faded memory,
Granny Bell wanted to go places Papa Copper could not take her. But Granny had a secret suitor
waiting to do her bidding.
Radio frequency (RF) wave propagation technology was under development at the turn of the
century. By the end of World War I radio amateurs were demonstrating the practicality of longdistance communications using radio frequency wave propagation through the atmosphere. After
launching these things called radio waves into space and receiving the signal many miles away,
as if by magic, the usefulness of radio frequency wave propagation in long-distance
communications systems was self-evident.
RF communications technology was initially relatively low frequency, voice-based transmission
systems. RF transmission links were particularly useful for over-the-horizon (out of sight)
communications with remote locations. Today, aircraft flights from Alaska to the Orient must still
rely upon such antiquated technology for a safe ride. But RF transmission technology has kept
pace with the rest of the developments in communication systems. RF transmission rates range
from the very slow (16 Kbps) to the very fast (Gbps).

Figure 3-11. Traveling light waves

Currently, Direct Broadcast Satellite (DBS), microwave, personal communications (paging
systems), and wireless systems (cellular) use RF propagation to get data from source to
destination. The upper bounds of RF propagation keep moving up in frequency as new enabling
technologies are discovered and refinements in current processes are made. Continued advances
in RF technology may yield 300 GHz transmission systems. With current digital encoding
schemes, the total amount of bits that can be transmitted without interference from 0 to 300 GHz
is 9.6 tera bps (9,600,000 Mbps). That is the equivalent bandwidth of 3,858 OC-48 fiber
connections. While RF has an upper bound limiting the amount of bits that can be transported,

fiber based systems are only limited by the demand for additional bandwidth, as additional fibers
can be connected as demand warrants.
RF wave propagation is highly regulated by all the worlds governments. Some places are very
serious about throwing a person in jail for unlawful, that is non-government approved, radiation.
RF spectrum is finite with fierce competition for bandwidth. And RF wave generation,
propagation, and reception requires expensive facilities, equipment, and maintenance. Just ask the
CATV service providers.
While there is a role for RF wave propagation in our communications systems of the future, ATM
and ADSL will make some forms of RF communications, such as CATV, obsolete. Dont need that
expensive coaxial delivery system anymore, and with ATM multicasting capability, satellite
delivery systems, with their limited delivery capabilities for video products, are a technology
dinosaur. And we dont need the expensive microwave hops used for relaying CATV video
products from source to destination anymore.

Light Wave (SONET/SDH)

Light wave technology is really a simple idea. Take a light wave, make it pulsate in relation to the
information to be conveyed, and ship it over some medium from source to destination. The
implementation of that simple idea in the United States is called Synchronous Optical Network
(SONET). In Europe, light wave technology is called Synchronous Digital Hierarchy (SDH).
While pushing light waves down a tube is a simple idea, the underlying technology to accomplish
the simple is very complex. But, simply put, SONET/SDH is light waves pulsating with
passionate electrical pectorals, passing through fiber cable from source to destination, albeit at a
very fast speed. Why SONET/SDH? Speed, speed, and more speed. Speed is synonymous with
bandwidth. And remember, I said bandwidth was green gold. See Table 3-2 for the various
SONET/SDH transmission rates.
STS Level

OC Level

Frequency (Mbps)

STS-1

OC-1

51.84

STS-3

OC-3

155.52

STS-9

OC-9

466.52

STS-12

OC-12

622.08

STS-18

OC-18

933.12

STS-24

OC-24

1244.16

STS-36

OC-36

1866.24

STS-48

OC-48

2488.32

Note: STS is the acronym for Synchronous Transport Signal and OC is the acronym for Optical
Carrier.
Table 3-2. SONET/SDH transmission rates

The light waves are contained within and guided on their journey from source to destination by a
special type of material called an optical fiber. See Figure 3-11.

Figure 3-12. Simplified switching scheme

Figure 3-13. Typical digital switching element

The fiber is made from silicon and resembles a narrow glass tube. The light wave travels from
one end to the other by reflecting off the inside wall of the fiber tube. The light wave can be
stimulated to carry signals with a bandwidth up to 2.4 Gbps (OC-48).

Figure 3-14. Four-stage Delta-2 switching matrix

Switching Systems
As long as there were just two telephones in the world comprising the network, there was no need
to provide any switching capability. As soon as a third telephone was added to the network
(perhaps within the week after discovering the technology?), switching became an issue. Three
telephones give any one user the choice of two possible connections. Some specialized circuitry
was required to provide the correct path depending upon the choice made by the caller. So, the
birth of the switch followed soon after the discovery of the telephone.
Originally, switches located in a central office (CO) were mechanical contraptions with the circuit
connection provided by a telephone switchboard operator. A switchboard had rows of plugs
arranged in pairs, called cord pairs, over rows of jacks. The plugs were used to mechanically
connect two jacks together to form a completed connection, or circuit. A user rang the operator
and advised her of the party to be called. The operator would then use the cord pair to connect the
two parties.
In 1891, Almon B. Strowger patented his mechanical switch that formed the building blocks of an
automatic switching system, eventually replacing the plug and jack mechanical switchboard and
putting many operators out of work, a familiar story. A Strowger switch, connected to each
incoming line, responded to dialed digits by rotating a certain amount with each digit. After
completing the final rotation when the last digit was received, the incoming line would be
connected to the correct outgoing line. The Strowger switch was so successful that in 1978, 53

percent of the Bell System exchanges in service, with 23,000,000 subscribers, still used Strowger
switching.
Due to certain significant limitations of Strowger switching, common control of
electromechanical switch matrices called crossbar switching was introduced. Crossbar switches
were slow, complex electromechanical switching devices that required significant amounts of
current to activate.
Common control refers to the designation of some device that receives the dialed information,
interprets the number, then actuates the proper circuit elements to make the physical connection.
Also, the common control breaks the connection when one party hangs up. When the phone is
picked up by the calling party the common control is assigned to listen for an incoming call to the
CO. The common control provides the dial tone to the calling party, then accepts the dialed digits.
After receiving the dialed digits, the common control determines the path through the switch
matrix by either hard-wired or stored program (computer) rules. The rules determine which two
points to close to complete the path. This particular method of identifying a connection is very
simple and handy. It is the most common way of detecting when and what key is pressed on a
computer keyboard.
Reed relays were developed to reduce the large currents required to activate the crossbar switch
and to increase the reliability of switches. Reed relays are small, glass encapsulated
electromechanical switching devices. The common control selects which relays to close to
complete the physical path.
Step-by-step (Strowger switching), crossbar, and reed relay switching are called space division
circuit switching, or just circuit switching, because every single telephone call is assigned a
unique physical path, or connection, through the telephone system, from source to destination
(end-to-end). Circuit switching is also commonly referred to as connection switching.
Figure 3-12 depicts a simple mechanical switching scheme where the Input can be switched to
Output A, Output B, Output C, or Output D, depending upon the desires of the circuit operator.
The diagram could represent any one of the mechanical switching types just discussed. Either an
operator can physically place a cord in a jack and complete the circuit or a mechanical switch can
be positioned electrically using a coded signal from the source telephone, as the Strowger switch.
Figure 3-13 shows a logical diagram of a digital switching element. As you may readily tell, any
input, 0 or 1, may be switched to any output, 0 or 1. Digital switching matrices are composed of
many of these types of elements.
Figure 3-14 shows a four-stage (A, B, C, D) Delta-2 Switching Matrix. The switching matrix is
composed of 32 basic digital switching elements, labeled A1-A8, B1-B8, C1-C8, and D1-D8.
There are 16 inputs and 16 outputs. Any input can be switched to any output. In the real world,
fast 32-bit microprocessors, under program control, switch the matrices. As a packet/cell comes
into the switch, the header information is read and decoded while the packet/cell is held in a
buffer (waiting room). After the microprocessor decodes the header (address) information and
determines that a correct switch path is available, the cell/packet is moved from the buffer into

the switching matrix, to follow the path from input to output through the individual switch
elements.
TDM multiplexing of channels gave us a new way to switch signals from source to destination. It
is commonly called connectionless switching. Every channel in a TDM multiplexed signal is
uniquely identified by its very position in the bit stream. All a particular device must do is read
the bit stream during the time period of interest. The device reading the bit stream must know
when to start reading and when to stop. A method of identifying the proper location that did not
depend entirely upon some reference clock signal was used with PDH multiplexing. With
synchronous systems, such as SONET/SDH, devices and bit streams must be operating from or
have some relationship to the same fundamental timing source. Granny Bell provided that single
timing source in the nationwide distribution of a master clock reference signal located at
Hillsboro, Missouri. Connectionless switching development progressed from plesciosynchronous
multiplexing used in PDH for conversion to a higher bit rate to synchronous multiplexing used in
SONET/SDH.
Circuit Switching

Circuit switching was the original type of transfer mechanism used. A dedicated line between
source and destination is allocated to the parties involved in the transfer of data over the network
path. Channel capacity must be available and reserved to make the connection between each pair
of nodes on the path, and each node must have a switching capability. For different applications,
utilization of the line may vary enormously from no load to overload. However, there is little
signal delay and effective transparency for the user. Effective transparency means the
transmission path from source to destination appears to the source and destination as one
homogeneous highway with no speed bumps, traffic lights, or left turns. Circuit switching is very
efficient for constant bit rate (CBR) applications.
Connectionless Switching

Connectionless switching is more suited to data services that require access of a bursty nature.
Rather than tie up a physical channel that may remain idle for long periods of time, data services
that transmit data in a pseudo random or bursty manner use connectionless switching. Their data
is multiplexed with other bursty services. Network engineers, using connection probability
theory, calculate the optimum number of connectionless switches/multiplexers required to
support a network within the limits of some defined Quality of Service (QoS) parameters. To
manage a bursty data stream through the network, various methods of grouping the data bits of
each channel have been devised. Data stream management is necessary to ensure that the
specified QoS is maintained. That is, all the bits, within acceptable limits, must get successfully
from the source to the destination.
Packet Switching

Packet switching hacks the data into segments. The data to be transmitted is divided into chunks
or packets. Each packet contains two groups of bits. One group is the data and the other group of
bits is the header information used for routing and quality control. At each node of the network

the packet is received, stored briefly, then passed on to the next node. At each node the packets
may be put on a queue to wait for an open channel before moving further into the network. There
can be several nodes in the transmission path. There are two approaches used to transport the
packets. A datagram is where each packet can take any path through the network as long as all the
packets reach the destination, eventually and not necessarily in the order transmitted. Virtual
circuit (VC) is where all the packets are routed through the same path without having the path
dedicated. Datagram allows for dynamic handling of congestion (sends packets around the
bottleneckyou take the high road, Ill take the low road kind of thing and we will meet in the
morn) and no call setup is necessary. VC channels allow for sequencing (you get each packet in
the same order I sent it), and error (bad puppy...uh packet...send it to me again) and flow control
(slow down, I cannot read as fast as you can e-mail me).
Multirate Circuit Switching

Multirate circuit switching (MRCS) is an enhancement of the synchronous TDM approach used
initially in circuit switching. In circuit switching, a node must operate at a fixed data rate which
must be used regardless of application. In multirate switching, multiplexing is introduced. A node
attaches to the network by means of a single physical link which carries multiple, fixed data rate
channels called B-channels at 64 Kbps. Traffic on each channel can be switched independently
through the network to various destinations. This is used for simple ISDN. The user has a limited
number of data rate choices but they are fixed so variable bit rate (VBR) is difficult to
accommodate efficiently. A user that desires to transmit MCS data at rates other than the CBRs
must use a higher rate than the maximum predicted.
Since VBR has been introduced, a few comments about it is in order. VBR is a useful tool for
controlling the transmission system loading, ensuring that during peak traffic periods the system
resources, such as transmission facilities, do not get overloaded (you can only shove so many
bytes down the throat at any one moment regardless of the ability to swallow), resulting in lost
data.

Frame Relay
Frame relay is essentially the same as packet switching. Frame relay was developed as a result of
the high data rates desired combined with the need for low transmission error rates in modern
high-speed data communications systems. In packet switching, there was considerable overhead
involved in error recovery (leave me alone, I can fix it myself), redundancy enhancement (play it
again, Sam, I didnt get it right the first time), and routing information (where, oh where, has my
little dog gone? Where, oh where, can he be?). Overhead is synonymous with inefficiency. With
frame relay, the packets are variable length, not fixed length, meaning that they are designed to
operate at up to 2 Mbps. The variable length frame relay packets provide a means for
implementing VBR.

Cell Relay
Cell relay is an evolutionary descendant of frame relay and multirate circuit switching. Cell relay
uses small, fixed-sized data packets called cells. The cell relay switching approach allows for the

definition of virtual channels (multirate circuit switching has fixed channels) with data rates
dynamically defined. A small cell size allows an almost constant data rate even though a packet is
the transport structure. Cell relay error control is much improved over frame relay, and allows
more errors to be handled at a higher logical level (more play it again, Sam ...but now the
request comes from Louie). The small, fixed-size cells reduce overhead even more and thus allow
rates of tens to hundreds of Mbps. Cell relay is equally adapted for use with variable bit rate
(VBR) and constant bit rate (CBR) switching technologies.

Summary
We are now in possession of a fast switching technology, cell relay, and a fast transmission
technology, SONET/SDH, over optical fiber. Cell relay supports VBR, CBR, and varied
applications. SONET/SDH allows large amounts of data to be transmitted efficiently over a
network. The combination of these two technologies, cell relay and SONET/SDH, provides the
foundation for ATM.

Chapter 4
Communication Protocols
Questions Answered in This Chapter
What is a communication protocol?
What is the OSI communication protocol model?
How do ATM and ADSL relate to the OSI model?
What are some communication protocol examples?
What is a frame and why is it important to ATM/ADSL?

Introduction
What is a protocol? Diplomatic protocols are the established ceremonial forms and courtesies
used by diplomats in the exercise of diplomacy. Deviations in the established protocol are
considered serious breaches of convention and can lead to unfortunate consequences. Similarly,
communication protocols define acceptable queries and responses between users of a particular
network type. The objective of course, is to provide for the orderly and predictable behavior of
hardware and software used in the network.
While governing bodies have established communication protocols to be machine independent,
setting down rules and guidelines for the general cases without regard to the specific platforms
performing the communication tasks, protocols, in a sense, are very environment oriented.
Designers must consider the specific equipment and their peculiarities, or uniqueness, used in
assembling network components. A DOS-based machine will not behave identically to a UNIX-

based machine in a network unless such behavior differences are accounted for in the design of
the network. Currently, the most common method of accounting for machine dependencies in
network communications is to purchase third-party software and hardware. Protocols are
concerned with every aspect of communications, from equipment physical interfaces, data
formats, and communication speeds to high-level software interfaces. The objective is to provide
a communications service to users that is reliable.
A protocol is an agreed-upon set of rules relative to specific objectives that are meaningful to the
parties that agreed upon the rules in the first place. Any particular protocol, diplomatic or
communication, may not necessarily be understood by others only indirectly related to the
process. Communication protocols are rules that attempt to define specific events which are
intended to control, in some acceptable and understandable manner relative to the objective to be
accomplished, the devices that are involved in a communications network. The events of interest
and the sequence or manner in which they occur define the protocol. There are numerous
communications protocols existing today. Some such protocols are IEEE 488, RS-232, and IEEE
803.
Why does the world need communications protocols? Variety. Someone said variety is the spice
of life. I do not know the context of the statement but I can certainly see how it may be applicable
in some settings and totally inappropriate in others. Variety is certainly the norm in
communications. There are many different ways to transport information from end to end and
there are many different ways of encoding information for transport. For any two similar
transport or encoding systems to work, a defined set of rules must specify all the particulars of
any importance. Two similar transport or encoding systems will provide gibberish, i.e., garbage,
results unless all the rules are followed within any specified tolerances. Just as humans need
varied behavioral protocols to define acceptable behavior in varied social settings so that human
actions may be interpreted with a certain degree of accuracy, so do communication systems need
protocols to define acceptable behavior, that all communications between end users of a similar
system will be understandable.
What if a rebel, with or without a cause, decides to flaunt convention and attempt to
communicate outside the established protocol? A communication outside the established protocol
may go unrecognized or result in a message to the sender, and perhaps to the intended recipient,
that an unacceptable or unrecognizable communiqu was attempted. Totally unexpected results
may occur or, in poorly designed systems, equipment may crash.

Figure 4-1. Open Systems Interconnection (OSI) model

A protocol template to use for modeling communication systems is beneficial that we might more
easily visualize the underlying concepts. ITU-T (formerly CCITT) developed the Open Systems
Interconnection (OSI) model that serves the purpose of a protocol template very well. See Figure
4-1. The OSI model represents a stack of seven layers that can be considered as gradations from
simpler to more complex functionality as one goes up the stack. This does not mean that one
layer is simpler than another. In fact, each layer in its own right is a complex interworking of
rules, hardware, software, and human endeavor.
The OSI model, especially above the physical layer, is really a hierarchical software model that
provides for coherent communications between groups of software procedures and/or programs.
Each layer can be represented by a software program that communicates only with the layer
above and below it. Typically, each layer is comprised of software routines that are called by the
next higher layer, when appropriate. The software routines are written for specific
communication applications. The interface between higher and lower layer routines is performed
by procedure calls known as application programming interface (API) calls.
The first three layers, the Physical Layer, the Data Link Layer, and the Network Layer, are
closely linked together. Kind of a you cant have one without the other sort of thing.
The three layers comprise the minimum structure required to establish and maintain
communications between two entities (nodes). These three layers pass from node to node in the
transmission path and contain, besides the data, address and error-detection information

necessary to ensure error-free routing to the correct destination. The control and routing
information is called Operations Administration and Management (OAM) data and is added to
the user data stream in quantities called headers and trailers. Each layer adds it own headers and
trailers to the data packet.
The control and routing information is interpreted by switches, routers, and bridges along the
transmission path. The control information and data are bound together until received by the
destination, where the control information and data are separated. Generally, the first three layers
represented by the OSI model comprise the fundamental ISDN/ BISDN software package
running on a host processor responsible for the physical switching duties, with the user providing
the application specific software package, comprising the remaining OSI layers, that typically
runs on a remote host computer.

Figure 4-2. OSI Physical Layer

Deep Look
The layers can be viewed as a trickle down hierarchy of software procedures but must also
include a trickle up hierarchical view as well.
While a detailed description and analysis of each OSI model layer is beyond the scope of this
book, a knowledge of the basic purpose of each layer is useful. In the ATM and ADSL literature,
there is often reference to the OSI model and the interworking relationships of the layers. The

following discussion serves only as an introduction to the OSI model. The interested reader is
encouraged to seek additional information and understanding.
Physical Layer

The Physical Layer interface of the OSI model is concerned with the various physical interfaces
of the equipment. Some of the issues this layer is concerned with are: voltages, electrical
currents, frequencies, connectors, and transmission media, such as fiber, coaxial, or twisted pair.
This layer is responsible for the physical generation and transmission of information and control
signals. The Physical Layer interacts with the layer immediately above, the Data Link Layer.
Control and data information are passed between layers through a software and hardware
combination called a low level driver (LLD). The LLD is an electrical circuit that responds to low
level (basic) software commands.
The Physical Layer controls the physical link between nodes of the communication path. The
connections are called physical service access points (PSAPs). It also supervises the specific
mediumcoaxial, fiber, RF wave, twisted pairof the transmission path. And it transmits the
bits between nodes. The controlling aspect of the Physical Layers job includes such mundane
activities as turning things off and on, as appropriate. The supervisory aspect includes monitoring
the path and the data to detect conditions conducive to error generation. The transmission portion
of its job is to cause the transmission medium to physically emit or receive the signal, either light
wave, electrical, or RF wave, that represents the data.

Figure 4-3. OSI Data Link Layer

The Physical Layer responsibilities include the following functions:

Providing the correct electrical signals for the proper transmission medium
which can be:
fiber optic cable
coaxial cable
unshielded twisted pair wire (UTP)
shielded twisted pair (STP)
RF (satellite, microwave)

Managing the medium before, during, and after the transmission including:
data (bit) transmission
physical link control

Some of the specifications that define Physical Link Layers for specific types of networks
include:

DS-1 basic rate interface (BRI)

DS-1 primary rate interface (PRI)

DS-3

SONET

RS-422A

RS-423

RS-232D

RS-449

B-ISDN

N-ISDN

V.xx (21, 22, 26, 27, 29, 32, 42, etc.)

The Physical Layer interface may attach a trailer, header, or both to the data element. The
trailer/header includes control information about the source and destination address, and error
control information.

Data Link Layer

The Data Link Layer is responsible for establishing connection between networks, framing the
data and control bits, and ensuring data integrity. The data integrity function provides flow and
error control of the data transmitted over the physical link. The Data Link Layer operates on the
physical devices involved in the electrical transmission of the data.
A primitive is a term used to describe a request for circuit connection, request for circuit
deactivation, or a request to transfer data. Primitives are software routines of the lowest kind, in a
manner of speaking. The Data Link Layer uses a defined set of primitives (basic software
routines) that control the physical devices. An LLD interprets the primitives and manipulates the
appropriate control circuits, typically called registers, that activate the circuit devices involved in
determining the actual physical transmission path.
The Data Link Layer includes the Media Access Control (MAC). The MAC is just what it says it
is. It is responsible for loading data into the transmission medium and unloading data received by
the transmission medium. The MAC includes buffers (bit buckets, or more appropriately
nowadays, byte barrels) to hold data on the transmit and receive paths. The MAC is able to
accommodate differences in speed between applications and mediums to some degree. Also, the
MAC frames the bits into ATM cells, both coming and going. And the MAC can request
retransmission of data, if required.
The Data Link Layer also includes the Logical Link Control (LLC) Sub-layer. The LLC is the
logical element that controls acknowledged, unacknowledged, connection-oriented, and
connectionless-oriented services for the network node. The LLC manages and controls the flow
of information into and out of the node based on the type of network connection bought and paid
for.
The Data Link Layer responsibilities include:

Establishing a physical connection (called building a connection) between

nodes
Deactivating a physical connection (called tearing down a connection)
between nodes

Framing the individual bits before transmission

Framing the individual bits after reception

Retransmission service for protocol data units (PDUs)

Detecting transmission errors

Ensuring proper addressing

Controlling the flow (speed) of data

Some Data Link Layer specifications include:

ATM Adaptation Layer (AAL)

Figure 4-4. OSI Network Layer

ATM Layer
IEEE 802.2

IEEE 802.3

IEEE 802.4

IEEE 802.5

HDLC

X.25

ISDN

The Data Link Layer attaches a trailer, header, or both to the data element. The trailer/header
includes control information about the source and destination address, and error control
information.
Network Layer

The Network Layer is intended to provide the upper layers a high degree of freedom from
specific network connection protocols such as voice analog modem protocols V.xx. The Network
Layer is therefore involved in the setup and teardown of connections. Also, the Network Layer

identifies the connection as connection oriented or connectionless oriented. Sometimes, it is also

used in data transfer. The Network Layer provides routing and addressing information to its
adjacent layers. It is the uppermost layer of the three chained layers. The Network Layer also
provides for the orderly interconnecting of both similar and dissimilar networks.
The Network Layer, which includes a predefined set of route tables, will calculate the openshortest-path-first route used to identify the route the data will take from source to destination. In
order to minimize the number of blocked transmissions (and therefore maximize revenue), the
open-shortest-path-first routine examines the current connections and determines an appropriate
transmission path that is not currently in use. As the name states, the routine looks for the shortest
path available. In many instances, particularly in LAN applications, there may be only one choice
and the calculation really does not exist. However, in the public sector, there could be thousands
of choices available. (Do you want the Russian operator or the Chinese operator to listen to your
Washington, D.C.-Bombay conversation?)
This aspect of communications has extremely important implications for video applications. And
ATM/ADSL are very much concerned with video applications. Video that is sent to a user
destination, stored, then played back at the users leisure is not too worrisome. But live video
must traverse the transmission path in the exact order sent and the transmission delay from source
to destination must be negligible, else the quality of the picture suffers dramatically. While the
shortest path available may be okay for voice and bursty data communications, it is not
necessarily any good for live video.

Figure 4-5. OSI Transport Layer

Here is $10,000 worth of consulting advice. The Network Layer includes the Management
Information Base (MIB). The MIB is a difficult to read (and understand) software database that
includes all of the known information regarding the network node. If you want the network to
provide you any particular information, reports, etc., concerning the status of individual
connections or group of connections (by location, SVC, PVC, port, switch, etc.), ask someone
(engineering? sometimes they do not understand the MIB) if it is in the MIB. If the answer is no,
dont waste your time trying to get network information that does not exist.
The Network Layer responsibilities include:

Determining the transmission path

Establishing the connection

Releasing the connection

Transmitting and receiving data

Acknowledging data reception

Requesting data retransmission

Some Network Layer specifications include:

IP
IPX

Transport Layer

The Transport Layer is responsible for the delivery of data between origination and destination
within the bounds of established reliability levels. There are five defined levels of reliability. The
reliability level is established by the type of service requested from the service provider. This
layer is also responsible for data multiplexing and demultiplexing. ATM does not utilize the
Transport Layer functionality. ADSL does.
The five reliability levels are simple, multiplexing, basic error recovery, error recovery and
multiplexing, and error detection and recovery class. For simple reliability, flow control and
connection release are provided by the underlying network layers. Multiplexing utilizes flow
control but does not specifically utilize error control which is provided by the underlying layers.
Basic error recovery does not utilize flow control but can detect errors and provide some error
control functionality. Error recovery and multiplexing utilizes flow control, error detection, and
correction. Error detection and recovery utilizes flow control, error detection and correction
including retransmission, routing around network path failures, and detecting and reacting to link
inactivity.

Figure 4-6. OSI Session Layer

The Transport Layer responsibilities include:

Ordering the establishment of connections

Ordering the release of connections

Figure 4-7. OSI Presentation Layer

Notifying the source of errors

Establishing the priority order of multiple users

Multiplexing and demultiplexing data

Some Transport Layer specifications include:

TCP
UDP

SPX

TP0, TP1, TP2, TP3, TP4

SPP

SEP

ADSP

VIPC

VSPP

Session Layer

The Session Layer establishes and maintains the exchange of data between origination and
destination. Also, the Session Layer must provide for an orderly recovery from failures caused by
any number of predictable and unpredictable events. This layer is the lowest layer of an
application-oriented communications software program.
The Session Layer is responsible for:

Data transfer to/from the lower layers

Establishing the connection

Releasing the connection

Re-establishing a broken connection

Enforcing protocols between applications

Some Session Layer specifications include:

DAP
RPC

SAP

DNS

SCP

ASP

PAP

Figure 4-8. OSI Application Layer

Presentation Layer

The Presentation Layer is responsible for manipulating the data such that the application host will
understand it. As an example, this layer residing on a UNIX host will interpret DOS formatted
data so that the UNIX machine will understand the data correctly. A common interpretation issue
addressed by this layer is the different method of using the carriage return/line feed in files
between UNIX- and DOS-based machines.
Some Presentation Layer responsibilities are:

Establishing the connection

Releasing the connection

Negotiating and formatting platform independent data syntax (i.e., DOS vs.
UNIX)

Encrypting and decrypting the data

Some Presentation Layer specifications are:

LPP
NCP

NetBIOS

X.25 PAD (Packet Assembler/Disassembler)

Application Layer

The Application Layer is responsible for providing the interface between lower layers and the
users application programs. It is rich with application programming interface (API) function
calls. The Application Layer utilizes API function calls to pass data and control information to
and from the lower layers.
Some Application Layer responsibilities include:

Providing an interface between the network and user applications

Requesting the execution of an operation (i.e., file transfer to a printer)

Reporting the results of operation execution (file sent to printer)

Reporting the status of an operation (printer out of paper, cannot print)

Aborting an operation

Error and flow control

Some Application Layer specifications include:

NFT (Network File Transfer)

RFA (Remote File Access)

NTP (Network Time Protocol)

TFTP (Trivial File Transfer Protocol)

NFS (Network File System)

SNA/FS File Services

FTAM (File Transfer and Access Management)

VT (Virtual Terminal)

PostScript

Upper Layer Protocol Examples

There are data communications protocols that exist to provide the orderly exchange of
information in a network. One such protocol is the Transmission Control Protocol/Internet
Protocol (TCP/IP). TCP/IP was originally developed by the Department of Defense to connect, or
network, Department of Defense computers with university computers. TCP/IP is a set of rules

used by software programmers who write networking code. There are other network protocols,
usually proprietary.
Every protocol must specify how the network components will identify data and control
information. A fundamental component of network protocols is the grouping of data and control
information into clearly defined and therefore manageable buckets called frames. The most basic
component of a frame is the simple binary bit. Bits are grouped into bytes which consist of eight
bits. The bytes are then grouped together into frames. Frames are grouped together to form
packets.
The position of the bits in a byte and the position of the bytes in the frame determine if the
network components will interpret the bits as data or as control information. The bits are
transmitted serially, that is, one after the other. Depending upon the network software, the first bit
in a byte is either interpreted as the most significant bit (big endian) or as the least significant bit
(little endian). And the network software also must interpret each transmitted or received byte as
either a big endian or little endian. The choice of big endian or little endian is not significant as
long as all the network components interpret the bits and bytes in the same fashion.
Why do we care about upper layer protocols when ATM is itself a lower layer (layers 1 to 3)
protocol? Because ATM/ADSL must interwork with all protocols for ATM/ADSL to realize their
full potential as unifying global technologies. Too much capital is invested in legacy systems for
them to be replaced solely for migration to an ATM network platform. We must understand the
interworking relationships between ATM/ADSL and the legacy protocols that we might
maximize the opportunities ATM/ADSL provide. A beginning to the understanding follows.
TCP/IP

Transmission Control Protocol/Internet Protocol (TCP/IP) is a software protocol for the

configuration of networks. Included in the protocol is the ability to provide remote logon
services, remote file transfers, and electronic e-mail. In reference to our OSI layering model,
TCP/IP protocol makes use of the Network Layer, Transport Layer, and Application Layer.
IPX/SPX

IPX/SPX is Novells proprietary network software. IPX/SPX protocol provides for the ability to
set the framing type used by network adapter drivers. The following framing specifications are
supported by the IPX/SPX protocol: IEEE 802.2, IEE 802.3, Ethernet II, Ethernet SNAP, Token
Ring, and Token SNAP. Specifically, a network administrator has the ability to choose the
particular framing used by the network.

Frames
One of the key components of a network protocol is the ability to group electrical signals into
precise, meaningful units. Meaningful for whom or what? Well, eventually, meaningful to the end
user which may or may not be a human (how about automatic feeders for livestock?), and
certainly meaningful to the hardware and software that acts upon or reacts to the individual

signals. Electrical signals that are without a precise, known ordering based upon some defined
relationship are not much more useful than noise. The typical bit arrangement of modern
networks is called a frame. Each protocol must define the bit framing used to pass information
from source to destination. The following is a more in-depth look at bit framing.
The objective of a communications network is to reliably transfer information from source to
destination within some specified performance criteria, such as speed and bandwidth utilization.
In order to accomplish the purpose of communication networks, communications protocols
specify how a system will segment and package the data, called a Protocol Data Unit (PDU).
Segmentation and packaging of the user data is necessary to maximize the use of the available
bandwidth due to the bursty nature of data communications. Each data package, or PDU, is called
a frame. Contained within the PDU is not only the user data but also routing and frame control
information. The routing and frame control data is called the PDU header. The data is called,
interestingly enough, the data unit. In terms of the OSI communications model, a frame is a group
of data at the Data Link Layer while a group of frames forms a packet at the Network Layer.
A LAN connecting several computers in an office setting represents a communications network
that does not need to be switched to perform its intended function. Each computer is hard-wired
to the server. However, a server connected to another computer through the public switched
telephone network in a dial-up mode is switched. An example of such a network is an office LAN
whose server provides Internet access to the office client computers by dialing up the Internet.
Communication networks may or may not be switched between source and destination,
depending upon the application and the geographical location of the system elements. WANs,
GANs, and the Internet are examples of networks that must pass data through switched
intermediate networks. Also, any particular user may have the ability to connect to more than one
user. If a network connection is not switched, it is referred to as a permanent virtual circuit
(PVC). When a network connection is switched, it is referred to as a switched virtual circuit
(SVC).
In order to route data from the source to the proper destination, whether the network is switched
or not, the PDU contains the address of the sender and the address of the receiver of the data.
Regardless of the network topologytoken ring, Ethernet, or some other topology including
the address of both the sender and intended receiver in the PDU provides some measure of
confidence that the data will reach its correct destination. And the correct destination will
recognize who sent the data.
The use of sender and receiver addresses in the PDU gives a network the freedom to route PDUs
through the network in the most efficient manner possible, allowing the network to maximize
bandwidth usage. The result is one PDU may follow a specific path through the network while
the next PDU may follow a different path. Also, PDUs may be buffered, or held temporarily, by
network elements due to the network link status (busy), the intended receiver not ready to
receive, or supercession by higher priority traffic. Yes, Uncle Sam can preempt your Internet
session and so can companies that have paid telecommunications companies for a higher class of
service. The result? Not all PDUs arrive at the destination in order of transmission, and the delay
between each PDUs arrival at the destination can vary significantly. Such routing of the data is
anathema to audio- and video-based applications.

To determine if the PDU is corrupted when received, the PDU contains an eight-bit byte (octet)
that represents a magical number called a CRC (cyclic redundancy check). The CRC is calculated
by the sender using a polynomial and the number of ones (or zeroes) bits stuffed into the PDU.
Upon receipt, the receiver calculates the CRC based upon the number of ones (or zeroes) bits
received and compares the number calculated to the CRC received. If the two numbers are equal,
the receiver can assume, with a high degree of accuracy (better than 1 in 10^12), that the data
received was actually the data sent.
However, if the CRC calculation by the receiver does not match the number transmitted, the
receiver can perform some sleight of hand and may be able to reconstruct the correct data, using
the transmitted CRC. But sometimes the tricks do not work and the receiver must ask the sender
to retransmit the corrupted PDU. Such retransmission of the data is anathema to audio- and
video-based applications.
So, we are shipping PDUs all around the countryside and we do not have good control over when
they reach their destination and in what order. This situation is okay for data communications
applications that do not have an intimate relationship with time. But if the data is time sensitive,
such as video and multimedia, then data processing issues such as time lapses and delays result in
unacceptable system performance.
To provide the throughput necessary to support process-sensitive communication applications a
better method is needed. Fortunately, just when we were beginning to need it most, ATM was
discovered! Actually the real story is our transmission systems (SONET) have become
sufficiently reliable that we can now do away with much of the header overhead, giving us a slim,
trim, speedy PDU.

Frame Structures
To really appreciate the benefits of ATM, and to understand why ATM is much faster than legacy
systems, a look at the framing used by some legacy systems is worthwhile. When one
understands where one has been, then it is much clearer to see where one is going.
Header Type

FiEld

Size

802.3

destination address

6 bytes

802.3

source address

6 bytes

802.3

length

2 bytes

802.2

destination SAP

1 byte

802.2

source SAP

1 byte

802.2

control data

1 byte

Table 4-1. IEEE 802.3 framing with 802.2 headers

Table 4-1 shows the framing organization for Ethernet IEEE 802.3 with IEEE 802.2 headers.
Novell Netware networks use this type of framing by default. It is the framing type automatically

selected when the network software for a Novell Netware adapter driver is installed. Novell sets
the Service Advertising Protocol (SAP) field to 0xe0, specifying that the upper layer protocol is
IPX.
Header Type

Field

Size

802.3

destination address

6 bytes

802.3

source address

6 bytes

802.3

length

2 bytes

802.3

0xffff

2 bytes

data

Table 4-2. IEEE 802.3 framing

IEEE 802.3 framing is used most often in Novell networks that use Netware 2.x/3.x servers.
Since Novell developed this framing while IEEE 802.3 was still being developed, it is not 100
percent IEEE 802.3 compliant. The data size is unlimited as indicated by the x.
Header Type

Field

Size

Ethernet II

destination address

6 bytes

Ethernet II

source address

6 bytes

Ethernet II

type

2 bytes

data

Table 4-3. Ethernet II framing

Ethernet II framing is an attempt to simplify framing and header overhead.
Header Type

Field

Size

802.3

destination address

6 bytes

802.3

source address

6 bytes

802.3

length

2 bytes

802.2

0xaa

1 byte

802.2

0xaa

1 byte

802.2

UI

1 byte

SNAP

protocol ID

1 byte

SNAP

type

1 byte

data

Table 4-4. Ethernet SNAP framing

Ethernet SNAP framing allows networks to use Ethernet II frames on IEEE compliant networks
without any modification to the network. Notice the use of three protocol headers in the frame:
IEEE 802.2, IEEE 802.3, and SNAP.
Header Type

Field

Size

802.5

AC

1 byte

802.5

FC

1 byte

802.5

destination address

6 bytes

802.5

source address

6 bytes

802.5

routing data

0-18 bytes

802.2

destination SAP

1 byte

802.2

source SAP

1 byte

802.2

control

1 byte

data

Table 4-5. Token Ring framing

Token Ring framing includes the SAP field which Novell sets to 0xe0 to indicate that the upper
layer protocol is IPX. Token Ring framing is specified by IEEE 802.5 and IEEE 802.2
Header Type

Field

Size

802.5

AC

1 byte

802.5

FC

1 byte

802.5

destination address

6 bytes

802.5

source address

6 bytes

802.5

routing data

0-18 bytes

802.2

0xaa

1 byte

802.2

0xaa

1 byte

802.2

UI

1 byte

SNAP

protocol ID

1 byte

SNAP

type

1 byte

data

Table 4-6. Token Ring SNAP framing

Token Ring SNAP framing allows networks to use Ethernet II frames on IEEE compliant
networks without any modification to the network. Notice the use of three protocol headers in the
frame: IEEE 802.2, IEEE 802.5, and SNAP.

But What Does It All Mean?

Framing is really an ATM issue and not an ADSL issue. The ATM protocol fits in with the first
three layers of the OSI model. The OSI model features and constraints of the first three layers are
well represented in the features offered by ATM as a transmission and switching technology.
ATM framing is similar to the legacy framing structures from which it is derived.
ADSL, though, is a multiplexing/demultiplexing technology that does fit in as a higher level
implementation of the OSI model. ATM fits in with the first three layers of the OSI model, and
ADSL fits in with the higher layer protocols.

Chapter 5
Communication Systems
Questions Answered in This Chapter:
What is a communication system?
What are LANs, MANs, WANS, and GANS?
What do ATM and ADSL have to do with LANs, MANs, WANS,and GANs?
What are network topologies?

Communication Systems
Why do we need to know anything about communication systems to understand ATM and
ADSL? Mainly because ATM and ADSL are communication technologies. Because ATM and
ADSL will replace, work with, modify, and/or unify our existing communication systems. If there
is no understanding of what a communication system is, there is no understanding of the import
of the technologies. Opportunity comes with understanding the impact ATM and ADSL have on
our communication systems.
A communication system is a collection of sometimes diverse elements such as computers,
routers, hubs, bridges, gateways, switches, radios, modems, software, and infrastructure working
together to achieve a common purpose: getting information to and from the systems users. There
are public communication systems, such as the Public Switched Telephone Network, there are
Department of Defense communication systems, and there are private communication systems,
such as local area networks. LANs may be connected through public or private transmission
mediums to form MANs, WANs, and GANs. Regardless of the ownership (public, Department of
Defense, private, etc.) of a communication system, there exist some fundamental building blocks
that all communication systems have in common.

Figure 5-1 Typical digital data terminal communications network

In Figure 5-1, a digital data terminal communication network is shown. The figure shows two
Digital Terminal Interfaces (DTIs) arranged in duplex communications mode. Duplex, in this
instance, refers to the ability to simultaneously transmit and receive data. Usage determines if the
two DTIs actually transmit and receive simultaneously. However, the ability to transmit while
receiving, whether or not the users choose to, defines duplex operation. A cursory glance at the
figure will reveal each DTI possesses both a receive path and a separate transmit path.
Lets look at the transmit path of the left DTI. Digital data, or information already in digital
format, is applied directly to the DTI. Audio frequency information, also called analog
information, must be converted first to a digital signal, hence the analog input is processed by an
A/D converter before being applied to the DTI. The DTI is responsible for converting the
information into whatever particular format is applicable based upon the network specifications.
The digital and/or audio information is then conveyed through the transmission medium to the
DTI on the right side. The right side DTI receives and processes the digital data. The DTI then
transmits the data out to the D/A converter and the digital data channel. The D/A converter
converts the digital data signal into an analog signal that closely matches the original analog
signal. If we followed the transmit path of the right DTI, it would be exactly the same as that
described for the left DTI.
The DTI may be a computer and all its component parts, such as the Universal Asynchronous
Receiver Transmitter (UART) and analog modem and any additional components such as hubs,
routers, and gateways to form the particular network of interest, or it may be a special device
designed and built for the express purpose of transmitting and receiving digital data directly to
and from another device of the exact type. While the figure shows just two DTIs interconnected,
the number of DTIs that can be interconnected is virtually unlimited. Notice the transmission
medium physically connects the DTIs.
The figure represents a general digital data terminal representing any one of several types of
protocols. The digital data can be encoded in various ways such as Ethernet, frame relay, or ATM.
Lets assume the encoding scheme is ATM. The DTI is responsible for framing the data bits into
ATM cells and adding the control information to the headers. Also, the DTI includes the ADSL
modem functions, if ADSL is utilized.

Figure 5-2 Typical digital radio data communications network

Figure 5-2 represents a digital radio data communication network. The digital radio network
works identically to the data terminal interface network with the exception that the digital radio
must modulate for transmission and demodulate for reception the digital data signal before
presenting the signal to an antenna for propagation into the atmosphere. Modulation consists of
combining the digital data signal with an analog carrier suitable for transmission at the frequency
desired. Demodulation consists of removing the transmitted carrier from the received signal and
recovering the original digital data.
The digital radio is responsible for framing the data bits into ATM cells and adding the control
information to the headers. The transmission medium between radios is the atmosphere. Since
ADSL is intended as a twisted wire pair technology, the digital radios do not convert the ATM
cells to ADSL. However, the input to the digital radio could be an ADSL input (after all, the
ADSL signal is an analog signal) that is transmitted to another digital radio which provides the
appropriate ADSL output. Such an ATM/ADSL application exists for remote video and remote
Internet access.
A modern communication system typically consists of many hardware and software elements
working together to achieve reliable communications from end to end (between users). The
Public Switched Telephone Network includes all of the previously mentioned elements in various
flavors and in a multitude of locations interconnected in seemingly incomprehensible complexity.
An example of a Public Switched Telephone Network inter-regional communication system is
shown in Figure 5-3. The Local Access and Transport Area (LATA) is the geographical area the
local telephone company is confined by regulation to operate and offer local and long-distance
telecommunications services in.
Local Exchange Carriers (LECs) are typically the Regional Bell Operating Companies (RBOCs
or Baby Bells) and other suppliers of local service such as GTE. The Interexchange Carriers
(IECs) are typically AT&T (Granny Bell), Sprint, MCI, etc.

Figure 5-3 Typical LATA Network

The lines connecting the various elements pictured in Figure 5-3 seem simple enough. But,
thousands of miles of twisted pair wires, hundreds of miles of coaxial cable, tens of thousands of
miles of optical fiber, hundreds of microwave transmission facilities, and several satellite links
may be required to provide the geographical coverage shown in Figure 5-3 by the Interexchange
Carriers and Local Exchange Carriers.
Why do we care about LECs and Interexchange Carriers? Because one provides the ATM
permanent virtual circuit (PVC) that hauls the ATM data from source to destination, and the other
provides the local ADSL access. One charges for the ATM PVC and Quality of Service (QoS) and
the other charges for the local and ADSL access. The two, or more in some cases, carriers must
work together to bring ATM and ADSL technology to the marketplace.
Currently, there typically is more than one Interexchange Carrier providing ATM service to or
from a particular geographical region. Not all Interexchange Carriers service all geographical
regions. And their service rates differ dramatically. The ATM Interexchange Carriers, their service
areas, and typical rates are given in a later chapter.
LECs install and maintain the local access switches that must be provisioned with ADSL modems
for ATM/ADSL to be a viable communication system. The LEC receives the ATM cells from the
Interexchange Carrier ATM network and converts the cells to an ADSL signal, then transmits the
ADSL signal over the local access phone lines to the customer (consumers and small businesses)
via the ADSL modem.

LAN/WAN/MAN/GAN
Local area networks are typically building and/or campus based. Generally, campus based means
the server and the remote desktop computers served are physically located in a small
geographical area, such as a cluster of several colocated buildings. When two or more LANs are
interconnected through their respective servers in the same general geographical area, the
network becomes known as a metropolitan area network. A MAN could be two separate
corporate locations within the same metropolitan area that have their LANs networked together.
A LAN can be interconnected to its network elements with wiring confined to the customer
premises. However, a MAN generally requires access to the public communications network.

Figure 5-4 ATM network with LAN topology

Network Topologies
Topology means a way of doing something. Network topologies are various ways of connecting
the network elements, also called nodes, such as hubs, bridges, routers, switches, gateways, and
users, to accomplish the intended function, whatever that intended function is. The common
network topologies are point-to-point, star, hierarchical, mesh (full and partial), bus, and ring.
Each network topology has specific attributes that make it especially useful for accomplishing
specific functions. The following discussion centers around the topologies utilized by LAN
networks. Why do we need to look at LAN topologies? ATM and ADSL will find a large market
in the consumer video segment, but they will also find a huge market empowering LANs as
MANs, WANs, and GANs. Sort of like a goodly dose of castor oil mixed with growth hormone
for LANs. Take them from childhood to adulthood. Give them some muscle. Allow them to
mature. Let us pray they do not overdose and become the technological monsters of yesterdays B
movies. By the way, where is Hal nowadays?

Figure 5-5 Point-to-point network topology

Point-to-Point

Point-to-point is the most simple of the various network topologies. Only two nodes are involved
in direct communications in the point-to-point network, so each node always knows who it is
talking to. There is no need for any sophisticated addressing and switching schemes. In Figure 55 Node A can only talk to Node B and Node B can only talk to Node C. There are no other
choices. There is little delay in the signal path due to routing issues (is this my packet or do I
send it on to the next node?). Just fire it up and go! A point-to-point network example can be
represented by your desktop computer and mouse. The mouse is hardwired to the computer.
There is no place else for the signals to go, except from the mouse to the computer. A good
example of an RF point-to-point network is microwave links.

Figure 5-6 Star network topology

Star

Star network topology is characterized as having one boss and many workers. In Figure 5-6,
the boss node is represented by Node A and the worker nodes are Nodes B, C, D, and E. All
the worker nodes must communicate directly with the boss node. As long as the worker nodes
only want to discuss something with the boss, network efficiency remains reasonable, unless all
of the workers want to talk at the same time. Then line up outside the door and wait your turn.
If any worker wants to talk to another worker, then the message must be relayed through the boss.
Slow, slow, slow. Imagine if you must pass all workplace communications through your boss.
Your boss knows everything you say and do. Scary. And everyone works at something less than
peak efficiency, especially the boss, if there are many workers involved in the network.

Figure 5-7 Hierarchical topology

Figure 5-8 A networking hierarchical topology

A good example of the star topology is the workplace LAN. All the users are tied to a server
which acts as the boss of the LAN. All the LAN users must go through the boss to get to the other
users, which may be desktop computers, printers, etc. An e-mail server is a prime example of this
type of topology.
Hierarchical

In hierarchical topologies, each node communicates only with the node below and above it.
Notice that the top node does not loop around to the bottom node. The top node can only
communicate with the bottom node through the intervening nodes, just like a chain of
command. And just like a chain of command, sometimes it isnt easy getting the attention of the
top dog. Sorry, I mean top node.
A good example of a hierarchical network is the OSI protocol model. Each layer can only pass
data to and from the layers above and below. From a networking perspective, a hierarchical
topology is represented by MAN, WAN, and GAN gateways. For a user of a LAN to get access to
another user on another LAN, located perhaps around the globe, the user must gain access to the
intervening gateways. See Figure 5-8.
While a hierarchical topology might keep you from getting to the top (dog) node as fast as you
might want, at least there is a path that does allow access. It is probably a practical impossibility
(that is a legal defense in the technical world) to wire every LAN user directly to every other
LAN user in the world. The amount of wiring alone would probably sink our ship. If users are

grouped together by some meaningful criteria, such as geographical location, and each group has
access to other groups via single entry points, then indirectly we can give every user in the world
access to every other user. Hierarchical networks give us the ability to do just that in the form of
MANs, WANs, and GANs. But, there are barriers regarding network commonality and
interworking, transmission mediums, and access speed. ATM and ADSL are the enabling
technologies that remove the barriers. We can fly, we can fly, we can...okay, we get the picture.

Figure 5-9 Full mesh topology

Mesh

Figure 5-10 Bus topology

Mesh topology is the direct connection of every network node to every other network node. Mesh
technology is fast since every user can talk directly to every other user. But mesh topology is
expensive due to the amount of wiring necessary to connect every node to every other node and
the need to provide sophisticated address schemes (equals complex circuitry and software) so
each user knows who they are talking to.

Figure 5-11 Ring topology

Figure 5-9 depicts a full mesh network. There is also a partial mesh topology that is the same as a
full mesh except that not all the nodes are connected together. One or more nodes do not connect
to another node. Mesh topology finds a use in connecting servers together, such as the Internet.
Bus

Bus topology is just like getting on a real bus through the front door. As you walk down the aisle,
you look for an empty seat with your seat number that was previously assigned. When you find
your empty seat, you sit down. But if you were assigned a seat number that does not exist, or you
failed to recognize your seat when you passed by, you continue down the aisle and pass out
through the rear door.
Bus topology was common in the early days of networking but has been largely replaced by the
ring topology.
Ring

The ring topology is a bus topology, except there are no front or rear doors. Any node can put
data on the ring bus and the data will circulate around the bus until it finds the right seat. There
are some issues regarding who puts data on the bus when (don&39t want two passengers up and
walking about the bus at the same time), but all in all, it is an efficient topology for
communicating.
The ring topology has found use as a metropolitan backbone network for interconnecting LANs
into MANs. The service providers are installing metropolitan fiber-based rings around all the
major cities of the U.S., Europe, and Asia. Not only does ATM provide an efficient transmission
technology for the metropolitan rings, but it is also the enabling technology for connecting the
MANs into WANs and GANs. Soon, the world will be one. One what? One communications
network. One global community.

Whats Up?
There are at least nine different, and for the most part proprietary, communication protocols
interconnecting computers and LANs today. The protocols listed in Table 5-1 are only a sample
of the different protocols in use today. There are many more. Each protocol is intended to work
on specific vendor equipment. When a company decides to purchase proprietary networking
equipment from a vendor, the vendor will take the company in hand and provide all of that
companys networking needs. And we all know what happens when vendors have you by
the...hand. These protocols are not intended to communicate directly with the other protocols.
Even the one vendor (not really a vendor in this instance but rather a vendor specification) listed
that is not a proprietary product (TCP/IP) has many vendors producing it with their own little
quirks that affect the ability to efficiently network their flavors together.
Vendor

Protocol

TCP/IP

TFTP

Novell NetWare

PCLANP

ISO

FTAM

IBM

SNA

3Com

3-Open

DecNet

DAP

Zerox

XNS

AppleTalk

AFP

Banyan Vines

StreetTalk

Hewlett-Packard

RFA

Sun Network

NFS

Table 5-1. Some communication protocols

AFP - Apple-Talk Filing System; DAP - Data Access Protocol; FTAM - File Transfer and
Management; NFS - Network File System; PCLANP - PC Local Area Network Program; RFA Remote File Access; SNA - Systems Network Architecture; TFTP - Trivial File Transfer Protocol;
XNS - Xerox Network Systems

Loosen the Grip

Where does that leave us? Well, if you have a particular networked LAN implementation in
South Dallas and the same networked configuration in North Dallas (or East St. Louis and West
St. Louis, about the same difference) and you dont mind having your...hand held, you can easily
connect the two networks. However, if you have a network in Dallas and one in London, well, let
me say that ATM loosens the grip on the...hand considerably. Read on and liberate yourself from
the grip.

Chapter 6
ATM, The Basics
Questions Answered in This Chapter
What is the ATM protocol?

What are ATM function layers?

What is a cell?
What are UNIs and NNIs?
What are virtual paths, virtual channels, PVCs, and cell headers?
What are the ATM service classes?
What is Quality of Service?

OSI Protocol Layers

Now is a good time to revisit the OSI protocol model. See Figure 6-1. The Open Systems
Interconnection (OSI) model was developed in 1984 by the International Standardization
Organization (ISO). The OSI model specifies a standard telecommunications frame of reference
for developers, operating companies, and other interested parties. Today, descriptions of
architectural protocols are generally espoused in OSI model terms. The seven OSI layers are
arranged in a functional manner. That is, all functionality associated with related activities is
defined by a particular layer. The model is an idealized conception of how the world would be, if
we lived in a perfect world. While most companies follow the model fairly closeafter all it does
lend itself adequately to the taskthe demarcation from one layer to another frequently is not a
cut-and-dried issue. However, we are not concerned with splitting hairs and debating gray areas.
The OSI model will be used just as it was intended, as a model to present appropriate material in
a more understandable form.

Figure 6-1. Open Systems Interconnection layering model

At the heart of the OSI model is the idea of peer-to-peer communications. Peer-to-peer
communications means only those layers on the same level will exchange data. That is, the layers
of Application A will only exchange data with the same layers of Application B. Put in terms of
the everyday work environment, vice presidents only talk to vice presidents, directors only talk to
directors, and senior managers only talk to senior managers. Ever known any such workplaces?
See Figure 6-2. Remember, the OSI model is an idealized representation and in the real world
layers usually must exchange information with the layers at least immediately above and below
them. More like the real world of the workplace, too. While the OSI model is comprised of seven
layers, the ATM model is only three layers whose functionality closely follows the bottom three
OSI model layers.

Figure 6-2. OSI layering model peer-to=peer communications

ATM Protocol Layers

A graphical representation of the relationships of ATM protocol components is shown in Figure
6-3. Notice in the figure the components nearer the bottom front are closer to the fundamental
transmission elements. The functions of the three ATM protocol layers are shown in Figure 6-4.
The three ATM protocol layers are depicted in Figure 6-4 with each layers task and sublayers
identified. The AAL layer is responsible for adding/removing the ATM header from the payload
and consists of five AAL layer types: AAL1, AAL2, AAL3/4, and AAL5.

Figure 6-3. ATM protocol cube

ATM Physical Layer

The ATM physical layer converts the data into the appropriate electrical or optical signal format
and the corresponding voltage levels necessary to drive the electrical or optical transmission
circuits.
ATM Layer

The ATM layer performs switching, multiplexing, transmission, and control functions. The ATM
layer is responsible for congestion control including tagging cells that exceed the users
maximum allowable transmission rate as Discard Enable (DE) by setting the Cell Loss Priority
(CLP) bit to a binary 1. The ATM layer congestion control mechanism includes management of
incoming data buffers. The ATM layer adds the cell header at the source and removes the cell
header at the destination. To get the cell from source to destination, the ATM layer translates the
cell address. Also, this layer is responsible for placing the cells onto the transmission media in the
sequence in which they arrived.

Figure 6-4. Functions of ATM protocol layers

AAL Layer
AAL Type

User Services Connection Type

Example

AAL 1

CBR

Connectionless
Oriented

voice services at DS-0, DS-1, and

DS-3 rates, video

AAL 2

VBR-rt

Connectionless
Oriented

compressed packet mode video

AAL 3/4

VBR-nrt

Connection or
Connectionless
Oriented

delay tolerant data with

sequencing and error detection
support

AAL5

VBR

Connection Oriented delay tolerant data requiring

minimal sequencing and error
detection support

Table 6-1. ATM AAL layers

The Adaptation Layer (AAL) inserts and extracts data into the 48-byte payload and adds and
removes data into the 5-byte header. There are four types of AAL layers as depicted in Table 6-1.
Each AAL layer type is designed to cope with a specific user service type.

AAL Type 1

AAL Type 1 functions are used for constant bit rate (CBR) services. CBR services are time
dependent which include real-time audio and video.

AAL Type 2
AAL Type 2 functions are used for variable bit rate video transmission. AAL Type 2 is not fully
defined at this writing.

AAL Type 3/4

Originally two separate types, AAL 3 (connection oriented) and AAL 4 (connectionless oriented),
AAL Type 3/4 functions are used to support variable bit rate, delay-tolerant services requiring
sequencing and error detection functionality. Data traffic in this category includes such activities
as large file transfers.

AAL Type 5
AAL Type 5 was intended for variable bit rate, delay-tolerant data traffic that does not require
sequencing and error detection to the extent of AAL Type 3/4. File transfers that are on high
quality links might fall into this category.
AAL Type 5 has less overhead bits than AAL Type 3/4. Each cell length field is not used and
neither is there a cell CRC. AAL Type 5 allows up to 64 Kbps of user data to be encapsulated in
an ATM super-cell. The super-cell is identified by the Payload Type Indicator (PTI) portion of the
ATM cell header.

ATM Planes
Boss, da planes! Da planes! (Sorry, could not resist the urge.) In the literature, ATM planes are
usually defined in very general terms that make it difficult for an individual with little or no
previous networking experience to understand just exactly what the planes are and what they do.
I suppose the reason for such oblique attempts to define the planes has to do with the complexity
of the underlying functionality.
Fundamentally, ATM planes are higher level software procedures that operate upon the ATM
hardware, its associated software procedures, and the ATM data to provide a controlled and
orderly process of establishing a communications link between two or more ATM users and
transferring data between the users. Figure 6-5 shows the relationship among and between the
planes. Well, trade one general explanation for another and you are still left with only a vague
notion of the purpose of ATM planes. While each ATM plane can be a book-length topic itself,
hopefully the following explanation will provide the reader with more substance and less fluff.

Figure 6-5. ATM planes

User Plane

The User Plane is responsible for transferring user application information to network elements.
Such information includes Quality of Service (QoS) parameters. This plane identifies to the
network the particular ATM Adaptation Layer type used in the transmission. User Plane
functionality permeates the physical layer, ATM layer, and AAL layer.
Control Plane

The Control Plane is responsible for establishing and releasing the connection in connectionoriented services, and providing other signaling information necessary to perform a switched
service. The Control Plane functionality permeates the physical layer, ATM layer, and AAL layer.
Management Plane

The Management Plane is responsible for managing the User Plane and Control Plane. Through
the Management Plane, the User Plane and the Control Plane are able to exchange information.
The Management Plane is composed of two subplanes, the Plane Management and Layer
Management subplanes.
Layer Management

The Layer Management subplane is responsible for managing each of the ATM layers, the
physical layer, the ATM layer, and the AAL layer. Layer-specific management functions are
grouped under the Layer Management umbrella.

Plane Management

The Plane Management subplane is responsible for managing and coordinating the functions of
the complete ATM product.

The ATM Cell

ATM cell length is fixed at 53 bytes (octets) with 48 bytes (octets) for payload (user information)
and 5 bytes (octets) reserved for the cell header. Cells can transport burst, voice, video, and image
data. ATM cell format is simpler and requires less overhead than other standardized "packet" type
technologies. Figure 6-7 depicts a frame relay frame format. Contrast Figure 6-7 with the ATM
cell format of Figure 6-6.

Figure 6-6. ATM cell format

Figure 6-7. Frame relay frame format

Frame relay frame lengths are variable. Frame length can be up to 4,096 bytes. Frame relay was
designed for bursty data with no intention of ever shipping real time-dependent traffic, such as
audio and video, across the frame networks. Currently, several bodies are attempting to develop a
Voice Over Frame Relay protocol.
ATM provides for four different flavors of ATM cells. The particular cell flavor of individual
cells depends upon the usage of the link at the moment of interest. ATM cell types are: idle cells,
unassigned cells, VP/VC traffic cells, and VP OAM traffic cells.
ATM Interfaces

There are two network interfaces encountered in ATM networks. There is the Network Node
Interface (NNI) and the User Network Interface (UNI). An NNI is the interface between two
network nodes and the UNI is the interface between the user and the network. ATM interfaces are
discussed in detail on page 135.

Figure 6-8. NNI-to-UNI ATM header format

Figure 6-9. UNI-to-NNI ATM header format

ATM Cell Header

The ATM cell header is composed of 5 bytes (octets) of information that is used for signaling,
routing, error detection, and Quality of Service purposes. The cell header is somewhat different
depending upon the ATM interface under scrutiny. The NNI-to-UNI interface includes a Generic
Flow Control field that is replaced with a Virtual Path Identifier in the UNI-to-NNI interface.
Figures 6-8 and 6-9 depict the NNI-to-UNI and UNI-to-NNI header cell formats, respectively.
The NNI-to-UNI ATM header format includes the Virtual Path Identifier (VPI), Virtual Channel
Identifier (VCI), Generic Flow Control (GFC), Payload Type Indicator (PTI), Cell Loss Priority
(CLP), Header Error Control (HEC), and Reserved fields. The value of the bits, either a zero or a
one, in each of the fields determines the destiny of the ATM cell.

Figure 6-10. An ATM cable bundle

The UNI-to-NNI ATM header format includes the Virtual Path Identifier (VPI), Virtual Channel
Identifier (VCI), Payload Type Indicator (PTI), Cell Loss Priority (CLP), Header Error Control
(HEC), and Reserved fields. Of course, the value of the bits, either a zero or a one, in each of the
fields determines the destiny of the ATM cell.

Figure 6-11. ATM VPI and VCI signaling

Virtual Channels and Virtual Paths

Notice that the ATM cell header is comprised primarily of Virtual Path Identifiers (VPIs) and
Virtual Channel Identifiers (VCIs).

Virtual channels (VCs) form bundles that comprise the virtual paths (VPs). The relationship
between VCs and VPs is clearly depicted in Figure 6-10. Just imagine Figure 6-10 is a graphical
representation of an actual transmission cable bundle. The individual wire pairs form the VCs and
the bundles of VCs compose the VPs.
The virtual path identifier of a virtual path entering the ATM switch is different from the virtual
path identifier of the virtual path exiting the switch. Likewise, the virtual channel identifier of a
virtual channel entering the ATM switch is not the same as the virtual channel identifier of the
virtual channel exiting the ATM switch. Figure 6-11 shows two VPs, called VPI 1 and VPI 5,
each composed of three VCIs (VCI 1, VCI 2, VCI 3), entering an ATM switch. VPI 5 is passed
through the switch unchanged. VPI 1 is unbundled in the switch and the individual VCIs are
switched to other VPI bundles. Figure 6-11 is a simplified representation of the inner working of
the ATM switch. VPI 2, VPI 3, and VPI 4 include other switched VCIs from other VPIs in
addition to the ones pictured. The ATM switch software includes routing tables that interpret the
VPI/VCIs and switch the cells to the correct output, ensuring the cells go on to the correct
destination.

Figure 6-12. Three-port bidirectional ATM switch

The VP path portion of the ATM switch will switch whole VPs from switch ingress to switch
egress. The VC path portion of the ATM switch will break down the VP bundles and switch the
individual VCs.
The VPI/VCI value has only "local significance," meaning it is useful only to the switch that the
value is currently associated with. Each switch in the transmit path changes the VPI/VCI
according to its own needs. Figure 6-12 depicts a three-port, bidirectional ATM switch with the
ports identified as Port 1, 2, and 3. Port 1 has two cells with VPI/VCI set to 15 and 32. Port 2 has
a cell identified with VPI/VCI of 27. Port 3 has a cell identified as VPI/VCI 48. Since the switch
depicted is bidirectional, cells can enter and leave each port in either direction. So, cells entering
Port 1 are assigned VPI/VCI values of 15 and 32. Likewise, cells leaving Port 1 are assigned
VPI/VCI values of 15 and 32. Cells entering or leaving Port 2 are assigned a VPI/VCI value of 27

and cells entering and exiting Port 3 are assigned a VPI/VCI of 48. The table included with
Figure 6-12 shows all the possible ingress and egress combinations the three-port switch can
manage. It is important to understand the VPI/VCIs assigned to cells coming into and exiting the
switch are unique to this switch. The next switch in the transmission path will assign its own
VPI/VCIs necessary for it to manage the cells.
There are several VPI/VCI combinations that are of special interest. Signaling cells are indicated
when the VPI equals 0 and the VCI equals any positive integer (positive whole number). Empty
cells are indicated when both VPI and VCI equal zero.
Header Error Check

The ATM Header Error Check (HEC) code is derived from the polynomial 1 + x + x^2 + x^8. In
the HEC polynomial, x refers to the number of ones bits in the cell. The source counts the ones
bits, calculates the value of the polynomial, and places the value in the header. The destination
calculates a new HEC based on the number of ones bits received. If the value transmitted agrees
with the calculated value by the destination, then all is well. If the values do not agree, the
algorithm used allows the destination to correct single bit errors or to detect multiple bit errors.

Provisioning and Signaling

An ATM circuit can be either connection oriented or connectionless. A connection-oriented
circuit is one that is established between two or more network entities using signaling between
the entities. A connectionless-oriented circuit is a circuit that is "nailed up" between two or more
network entities. "Nailed up" is telecom lingo meaning the entities of concern are always
connected to each other and no switching or signaling is required for one entity to communicate
with another.
Provisioning in ATM parlance means a circuit is "nailed up." That is, a provisioned circuit is a
circuit that is hard wired along a transmission path from source to destination. There is no
switching involved to get from source to destination.
Signaling is synonymous with switching in ATM. The switching function is performed by the
ATM switches using the signaling information provided by the ATM cell header.

Sustainable Cell Rate and Peak Cell Rate

Sustainable cell rate (SCR) is a transmission rate defined as the average cell rate that may be
transmitted over the backbone. The SCR is measured in megabits per second. The operative word
in this definition is "average." The average is taken over some time period. The average time
period depends upon the ATM switches deployed by the carrier in their network and the switches
specific capabilities. Typically, the average time period is taken over either 30-minute periods or
15-minute periods. Cells exceeding the SCR are tagged Discard Enable (DE).
Peak cell rate (PCR) is a transmission rate defined as the maximum allowable cell transmission
rate. PCR is measured in megabits per second. In a perfect world, all cells transmitted above PCR

are discarded. Guaranteed. However, we do not live in a perfect world and carriers are attempting
to attract ATM business to fatten the lightly loaded ATM backbones. So, most carriers are
allowing users to burst above PCR, up to the access speed, without penalty, as long as network
congestion remains under control. Now is the time for all good men and women to get maximum
bang for the buck.
Technically, SCR and PCR are applicable to VBR services, which are bursty by nature. Most
carriers, however, are going to impose an SCR and PCR of some value on all classes of service to
manage system capacity. The only sane way the carrier can manage the carrier backbone is to
have a known SCR and PCR for every user. That way, the carriers capacity planning
departments can, hopefully, actually plan the capacity of the backbone network. Granted, the
PCR may be the access speed of the connection but whatever it is, it must be defined for every
user, else havoc will reign, eventually.
The challenge for the user is to select the proper service class for the users specific network
requirements. Selecting the proper service class can still result in poor system performance and/or
wasted resources if the correct SCR and PCR are not chosen. The network manager who knows
his network data transmission rates can make wise choices.

Minimum Cell Rate

The minimum cell rate (MCR) applies only to the available bit rate (ABR) class of service. The
MCR is the transmit rate in cells/second that the source is always allowed to send. It is the
minimum bandwidth required by the end application.

Maximum Burst Size

The maximum burst size (MBS) is defined as the maximum number of cells that can be
transmitted per burst at PCR. A burst is defined as the consecutive transmission of two or more
cells. A typical maximum burst size is 256 cells. However, certain access technologies have a
specific MBS. See Table 6-2.
Access Network AAL5 (cells)
Ethernet

32

FDDI

91

IP over ATM

192

Table 6-2. MBS of several access technologies

Cell Loss Priority

Cell loss priority (CLP) is a bit field of the ATM header. The CLP field is used by the ATM
network to determine the eligibility of a particular cell for discard. Normally, the CLP bit is set to
binary 0. When the user exceeds, or bursts above, the contracted SCR, the CLP bit of each cell
exceeding the SCR is tagged Discard Enable (DE) by setting the CLP bit to 1. If the carrier

backbone becomes congested, then the carrier can discard the DE tagged cells. When the network
becomes congested, the DE tagged cells can be discarded, even if the transmission rate is still less
than the PCR.

Maximum Cell Delay

Maximum cell delay is the time required for a cell to travel from the first ATM ingress port to the
last ATM egress port. Put another way, maximum cell delay is the time required for a cell to
travel from the source ATM port to the destination ATM port. Maximum cell delay is measured in
milliseconds and is measured in one direction only. The transmit path from point A to point B
may be different from the transmit path from point B to point A, according to the carriers
network topology and cell routing strategy. So, the maximum cell delay in a duplex
communication mode can vary considerably in the transmit direction between two locations.

Cell Delay Variation

Cell delay variation is a measurement of the maximum variance from the average time interval
between successive cells transmitted from source to destination. Like the maximum cell delay
descriptor, it is measured from ingress ATM port to egress ATM port, and it is measured in
milliseconds. Cell delay variance describes the maximum delay any cell has experienced as the
cell transisted the carriers ATM network. This descriptor is extremely important to real-time
audio and video applications where a delay of microseconds can distort the signal, resulting in
poor quality of the received signal.

Cell Loss Probability

Cell loss probability is a measurement of the possibility that a particular cell will be discarded by
the carriers network. Cell loss probability is different for each Quality of Service which defines
specific user service categories such as constant bit rate (CBR).

Traffic Shaping
Traffic shaping is the ability to change the rate of cell transmission on the fly at the customers
premises in order to comply with the carriers traffic contract requirements such as SCR, PCR,
and MBS. Traffic shaping in this context is performed by CPE. However, not all CPE can
perform traffic shaping. Lets say the user network is pumping data into the ATM CPE at a rate
far above the SCR of a CBR service class. If there is no traffic shaping capability in the ATM
CPE, all the cells exceeding the SCR are tagged Discard Enable and may or may not be
transmitted to the destination, depending upon how the carriers network is loaded at the moment
of attempted transmission and the carriers attitude toward allowing customers to burst above the
service contract limits. However, if ATM CPE with traffic shaping capability is used, the excess
cells are held in abeyance and allowed to enter the ATM port at the SCR rate. Think of traffic
shaping as an ATM cell bucket with a funnel of varying diameter, as set by the users service
contract, attached to the bottom of the bucket that all the cells must pass through upon egress. Do
big buffers come to mind?

Carriers may be lax right now, due to very lightly loaded networks, in policing their ATM
networks for contract compliance by users. As the lightly loaded networks begin to experience
capacity ratios of 25 percent or better, users can expect carriers to rigorously enforce contractual
obligations. If the CPE equipment in the user location is not equipped to perform traffic shaping,
plan on either upgrading or changing CPE or experiencing some degree of network performance
less than optimum.
Here is a traffic cost analysis that may be fruitful for users to perform: ATM traffic volume is not
constant all through the day and night. Traffic waxes and wanes just like the moon. The period of
least traffic is midnight to 6 a.m. CST. Sending traffic that exceeds the contractual rates during
lightly loaded periods instead of the busy part of the day may not get the users cells discarded.
Remember, ATM is the statistical multiplexing of transmission and signaling theory. Some
combination of traffic shaping and transmitting during off peak hours may yield higher network
throughput while allowing the user to purchase lower rate, and less expensive, PVCs. The
difference can be several hundreds of thousands of dollars per year for the heavier hitters.
Statistical multiplexing. Probability. If the user understands the concepts as used in ATM
networks, significant cost savings can be realized.

Service Classes
Service classes are really nothing more than a hierarchical set of user categories that determine
the access priority to the switch and transmission facilities. Service classes are tied directly to
carrier revenue potential and are therefore of great significance to carriers. From the users
perspective, choosing the proper service class for the network application is important for two
reasons. If the user is paying for a service class that results in circuit underutilization, the unused
bandwidth is money wasted. However, if the service class chosen is not adequate for the volume
and nature of data transmitted, dropped cells mean data must be retransmitted, perhaps
innumerable times. Of course, for real-time audio and video applications, dropped cells are not
good. The user should give careful consideration to the application and the various service classes
offered by the carriers before choosing the service class for the application.
ITU-T QoS ATM Forum QoS

Timing Between
Source/Destination

Bit Rate

Connection Mode

Necessary

Constant

Connection

Necessary

Variable/Constant

Connection

Not Necessary

Variable

Connection

Not Necessary

Variable

Connectionless

Table 6-3. Adaptation service classes

Table 6-3 lists the four service classes defined by ITU-T and the ATM Forum. Class A and B
require a timing relationship between the source and destination, while class C and D do not.
Table 6-4 provides more specific details concerning the attributes of each service class. Notice
Class B will be used for real-time audio and video and packetized audio and video.

ATM Forum
QoS

QoS Category

Cell Rates

ABR Available Bit Rate

SCR guaranteed Intelligent network control,

SCR<PCR
LAN emulation

CBR Constant Bit Rate

PCR=SCR

Real-time audio and video,

private line emulation

C,D

VBR-nrt Variable Bit Rate not real time

SCR<PCR

Burst LAN

VBR-rt Variable Bit Rate - real SCR<PCR

time

Packet audio and packet video

UBR Unspecified Bit Rate

Any application not requiring

time dependencies between
source and destination

SCR,PCR

Uses

Table 6-4 Quality of Service classes

The ATM service class categories, listed in order of decreasing cost to the user, are:
ABR Available Bit Rate
ABR class of service specifies a minimum bandwidth that is guaranteed by the service
provider. Each service provider has specific bandwidths that are sold ABR. The
bandwidths available from each carrier are dependent upon the ATM network backbone
capacities of the carrier and the carriers order entry, provisioning, and billing systems.
ABR is ideally suited for intelligent network congestion control and LAN Emulation
(LANE). The user can burst above the SCR to the PCR.
CBR Constant Bit Rate
CBR is just what the name says. The peak cell rate always equals the sustainable cell rate.
Any burst above the SCR results in dropped cells and the need to either retransmit the
data or accept lost data. CBR is used primarily for private line emulation and real-time
video and voice applications.
VBR-nrt Variable Bit Rate - not real time
VBR-nrt is used for connection-oriented data applications. The timing relationship
between source and destination is not important. The transfer of data does not rely upon
any timing dependencies between source and destination. The bit rate is not constant and
may burst up to some predetermined PCR. Bursty LAN traffic is well suited to this class
of service. PCR and SCR are not equal for this class of service.
VBR-rt Variable Bit Rate - real time
VBR-rt is used for connection-oriented audio and video applications. Another name for
connection-oriented audio and video is packet audio and packet video. The timing
relationship between source and destination is very important to the successful transfer of
data. The bit rate is not constant and may burst up to some predetermined peak cell rate
(PCR). PCR and SCR are not equal for this class of service.
UBR Unspecified Bit Rate
UBR is just like flying standby. You get on only when there is an empty seat. There is no

guarantee by the service provider that your data will get from source to destination. This
service class is the least expensive way to transmit ATM cells. (PCR and SCR can be
anything negotiated with the carrier providing the carrier order entry, provisioning, and
billing systems can accommodate the desired data rates.)
The service classes are subsets of the ATM Adaptation Layer. A CBR can be either AAL1 or
AAL5.

Latency
Latency is a fancy way of saying "wait your turn." Cells are multiplexed onto some physical
medium for transport around the country or globe. Multiplexing requires all good cells (and bad
ones too at this stage of the transmission path) to queue up and wait for their turn to be
multiplexed. How long is the wait? Assuming (and we know about assume, dont we?) all other
things are perfect, the latency of a cell is 53 bytes times 8 bits/byte divided by the transmission
rate. More formally, latency = (53 bytes x 8 bits / byte) / (transmission rate). Assuming (there we
go again) the transmission rate is DS-1, then latency = (53x8)/(1.544 x 10^6), or latency = 27
milliseconds.
ITU-T QoS

ATM Forum
QoS

Service Class
Category

Cell Delay

Cell Loss

ABR

Very High
(~150msec)

Very Low (10^-12)

CBR

Very Low (~5msec) Very Low (10^-12)

C,D

3,4

VBR-nrt

High (~50msec)

Low (10^-6)

VBR-rt

Low (~10msec)

Low (10^-6)

UBR

Very High
(~150msec)

Very High (10^-3)

Table 6-5. Latency by Class of Service

Latency is extremely important to real-time applications such as video and audio. Too much
latency, and the picture or audio begins to become distorted until the picture or sound is
unrecognizable.

Port Speeds
Port speeds are governed by the equipment the service provider has installed in the ATM
backbone and the physical transmission medium speed accessible from the customers premises.

If the service provider has done an adequate job of engineering its ATM network, the switch
speed is much greater than the users access speed.
ATM UNI

Port Speed

DS-0

0.064
Mbps

DS-1

1.544
Mbps

2xDS-1

3.0 Mbps

3xDS-1

4.50 Mbps

4xDS-1

6.0 Mbps

5xDS-1

7.50 Mbps

6xDS-1

9.0 Mbps

7xDS-1

10.50
Mbps

8xDS-1

12.0 Mbps

DS-2

6.312
Mbps

DS-3

44.736
Mbps

OC-1

38 Mbps

OC-3

155 Mbps

OC-12

622 Mbps

OC-48

2.4 Gbps

OC-192

10 Gbps

Table 6-6. Typical ATM port speeds

ATM Traffic Management and Control

ATM traffic management and control is responsible for the orderly flow of data from source to
destination. Traffic management and control issues include the type of connection established
between the source and destination and traffic policing.
Connection-Oriented Connection

Communication between parties is established through a direct physical path that is expected to
exist for the duration of the communication. Data is sent sequentially from source and arrives
sequentially at destination. The PSTN, CATV, and Frame Relay networks are connection
oriented.
Connectionless-Oriented Connection

Communication between parties does not have a direct physical path from source to destination.
While data is sent sequentially from the source, the data can arrive in any order at the destination
due to differing transmission paths the data elements may take. The Internet and LANs are
considered connectionless oriented.
PVCs

PVCs use switch connections to route data from source to destination. The cells contain the
routing information in the header VCI and VPI fields necessary for the switch to direct the cells
to the correct destination path. PVCs may exist for any time duration desired such as weeks,
months, or years. The telecom industry refers to a PVC connection as a "nailed up" connection. A
PVC does not require any call establishment procedures as everything necessary to get the data
from source to destination exists in the 5-byte header.
SVCs

SVCs use ATM signaling and ATM addressing to direct cells from source to destination. SVCs
exist for the duration of the data transfer only, which may be milliseconds or seconds. SVCs
require call establishment procedures including address screening and source address
authentication. The source transmits a signaling request that is routed through the network,
establishing the link from source to destination. The destination will accept or reject the sources
signaling request. If the destination accepts the sources request to establish a path, each switch
along the transmit path will update its VC translation table. After all the switches complete the
VC translation, data is then allowed to transit the network.

Figure 6-13. Multicasting

Usage Parameter Control

Usage parameter control (UPC) is a high-tech way of saying "traffic cop." Usage parameter
control is high-tech traffic policing of the ATM "superhighway." Usage parameter control detects
and reports user contract violations and can prevent users from exceeding their contract specified
traffic limitations. The intent is to ensure the network QoS is maintained. The ATM switch
Message Information Base (MIB) stores individual user traffic usage parameters which are used
by network control software to compare to contract parameters. The service provider has the
options to ignore contract violations, notify the user of violations, and/or prevent user traffic from
exceeding contract imposed limitations, thereby preventing violations from even occurring. The
specific action the UPC performs after detecting a violation is to discard an illegal cell or tag as
Discard Enable, meaning it will be discarded if the network becomes congested while the cell is
in transit.

Figure 6-14. NNI-to-UNI/UNI-to-NNI interface

Multicasting

Multicasting is nothing more than a single source, referred to as the root, sending ATM cells to
two or more destinations, called a leaf. Sometimes, multicasting is referred to as a point-tomultipoint connection. Multicasting allows one location in the network to broadcast data to as

many other locations as desirable. A typical application would be the corporate headquarters
broadcasting new product information to a sales force located in many cities across the country,
continent, or globe.

ATM Network Interfaces

ATM classifies three kinds of physical interfaces. The physical interfaces are, in the real world,
ATM switches, CSU/DSUs and/or routers. The interface between the customer premises and the
service providers backbone network is called the User Network Interface (UNI) to Network
Node Interface (NNI), or UNI-to-NNI. See Figure 6-14.

Figure 6-15. Off to Grandma's house

The customer premises interface has two types of UNI interfaces. They are:

Public UNI
Private UNI

The service providers backbone network interface has two types of interfaces. They are:

Public NNI
Private NNI (P-NNI)

Data Exchange Interface (DXI) is the interface between an ATM CSU/DSU and another
topologys router. A CSU/DSU is a Channel Service Unit/Data Service Unit. CSU/DSUs are
devices that terminate PVCs/ports on customer premises and provide signal conditioning. DXI is
responsible for converting variable length, frame-based data to fixed length ATM cells.

A Cell Family Vacation

Perhaps the interrelationships of the three ATM layers still escape the casual reader. Lets take a
Cell Family Vacation to Grandmas and see for ourselves how the three layers interwork. Like all
journeys, this one will begin at the beginning as depicted in Figure 6-15.
The members of our ATM Cell family, named John Audio, Marsha Video, and Baby Data, arrive
at the source AAL. Here they get into the same vehicle (ATM cell payload) in the AAL parking
lot, open the road map, and plan their route to their destination (assemble VPI/VCI). The source
ATM layer directs the Cell family vehicle to the correct parking lot exit. Now, the source physical
layer is the type of road that is the route between source and public ATM network. The route the
family must take might be a dirt road (twisted copper pairs), an improved road (coaxial), or a
freeway (optical). Today, they are lucky to find themselves speeding down a superhighway at 155
Mbps (OC-3). Now, the members of the ATM Cell family are seasoned travelers who know that a
good rest is needed after traveling so far so fast. The ATM public network acts as a rest stop (cell
buffers) in the middle of a major interchange (backbone ATM switch) for the family. When the
family reaches the Public ATM network (physical layer), the family receives a much needed rest
as they wait their turn to enter the busy interchange (switch). The family finally reaches the
interchange (ATM public network switch) and must decide which direction to go in. One wrong
turn and they are lost forever. John finds an exit and quickly drives off the crowded expressway
(OC-3 light wave) onto a dusty dirt road (twisted copper pairs). Slowly driving along (64 Kbps)
trying to miss all the chuckholes, John almost misses Grandmas driveway (destination ATM
layer). Marsha quickly shouts out to make a turn and the Cell family enters Grandmas driveway
(destination physical layer). Turning off the engine, all the family members exit (destination AAL
layer) the rental car (ATM cell payload).

ATM and Legacy Systems

ATM is fundamentally a transmission technology responsible for assembling data received from a
variety of sources into payloads. After attaching the proper addressing and error control
information into the cell header, ATM is responsible for placing the bit stream representing the
cell-based data onto the transmission media that may be any one of the various media types, such
as copper wire, coaxial, or fiber. To perform the tasks required, ATM must be compatible with a
wide range of legacy technologies.

Chapter 7
ATM, The Gravy
Questions Answered in This Chapter
What are SCR and PCR?
How do audio, video, and multimedia signals fit in with ATM transmission?
What is an ATM switch, NNI, and UNI?
What does a service provider ATM backbone look like?

What are corporate and consumer ATM access costs?

What is interworking and what technologies are ATM compatible?
What is shared access?
What transmission technologies does ATM compete with?
This chapter introduces some new ATM concepts and terminology and explains their function and
importance. Also, since ATM is the transmission technology that can seamlessly transmit audio,
video, and multimedia, the issues regarding ATM transmission of these applications are given
treatment in some detail.

Peak Cell Rate

Peak cell rate (PCR) is the maximum allowable cell rate for any given PVC or SVC. PCR is the
upper boundary on the rate that data can be loaded onto the ATM connection without potential
loss of cells. Cells exceeding PCR are tagged Discard Enable (DE) and are eligible to be dropped.
Whether or not the cells are actually discarded depends upon the volume of traffic transiting the
switch at the same time, the switch capacity, and the service providers attitude toward PCR and
the user. Currently, ATM backbones are lightly loaded and service providers are not discarding
DEs unless switch capacity is maxed out. Since switch capacity is seldom maxed, some service
providers are giving users a free ride, for now. PCR is expressed in megabits per second. Some
service providers are setting the PCR to the port connection speed, allowing users to burst to the
maximum possible. Setting PCR to port speed is another freebie from the service providers to
encourage migration to the lightly loaded ATM backbones. Do not expect such generous
handouts as the backbones begin to exceed 25 percent of capacity.
Peak cell ratesounds simple enough. The maximum number of cells transmitted. Yes, but
when? Over the highest one second of traffic of the last one minute? The highest one minute of
the last hour? Or the highest one hour of the past one day? Confusing? Well, ATM switch
manufacturers have pretty much removed the confusion. Most ATM switches have their
databases polled every 15 minutes with polling times set to coincide with the quarter intervals of
the clock. The switch polls on the hour and every quarter hour thereafter. The data reporting
systems take either 15 or 30 minutes, depending upon switch manufacturer, to present the data to
the user. So every 15 or 30 minutes you can view the switch statistics, at least those statistics the
service provider deems appropriate for user consumption, of the previous 15-minute time slot.
This means the peak cell rate that is visible to the user and to the service provider is going to be
an average cell rate taken over at least a 15-minute interval. So, a clever user can burst above the
PCR any portion of a 15-minute time period, as long as the average cell transfer rate remains at or
below the PCR without incurring any monetary penalties from the service provider.
PCR is a traffic parameter that is most often used in the VBR class of service. PCR is intended to
allow the user to burst above some sustainable cell rate (SCR) for short durations. Applications
requiring PCR are real-time based, such as audio and video.

Sustainable Cell Rate

Sustainable cell rate (SCR) is the cell rate the user is allowed, by contract, to transmit over the
ATM network without incurring any monetary penalties or resulting in dropped cells. SCR is
measured in megabits per second (Mbps). SCR is an average, like PCR, as service providers poll
their switches Management Information Base (MIB) periodically. The typical polling time is 15
minutes, although some service providers poll at 30-minute intervals. So, if the number of bits
transmitted over the polled period are equal to or less than the contracted SCR, there is no
monetary penalty. However, during some portion of the period, the actual transmit rate might be
much greater than the SCR.

Low-Speed Data
Low-speed data requires a bit rate of 64 Kbps or less. Examples of low-speed data are Morse
code communications (still used in military and maritime applications), telemetry, remote
monitoring, and data gathering applications. Low-speed data does not present any technical
challenges to ATM and/or ADSL technologies.

High-Speed Data
Bandwidth is the commodity ATM (and Frame Relay) service providers sell. Like most
commodities, the more bandwidth you want, the more you have to pay. High-speed data is the
same thing as large bandwidth. The higher the speed, the larger the bandwidth required to
transmit the data and the higher the price to place the data on the backbone of some service
provider.
Some applications, such as real-time video- and audio-based applications, require, for acceptable
performance, the transmission of data at very high speeds. Some applications do not necessarily
need to transmit data at high speeds for acceptable performance, but the sheer volume of data
may necessitate a large bandwidth. Such applications include banks and financial institutions that
need to ship huge volumes of data concerning financial transactions in short time periods.
Another example is a government bureaucracy such as the Social Security Administration. Unless
these large entities use a large bandwidth transmission medium, they cannot get all their data
transported in the time required. Unlike audio and video applications, if they do not succeed in
getting the data from Point A to Point B, the quality of the end-user application does not suffer.
While our bank account might not be debited or credited in a timely fashion if the financial
institution does not transmit data at a sufficiently high speed, the end result is still the same.
Eventually the account is updated, although the account holder might suffer some financial loss
due to the delay. These large, commercial enterprises and government agencies are the groundbreaking leaders in adopting ATM service. The amount of data they need to ship is sufficient that
the economies of scale justify the current expense of ATM.
Audio

Real-time audio applications are not as complicated as video, but they do have their own set of
unique characteristics. The bandwidth for concert hall stereo sound is 20 KHz. For full life-like

stereo reproduction of sound, many reproduction machines sample the original analog signal at 8
or 16 times the peak frequency. If we have a machine that oversamples at x8, the bandwidth
required to transport the digitized signal to and fro is 20 KHz x 8 samples / Hz x 8 bits / sample =
1.28 Mbps. For a machine that oversamples the audio signal x16, the bandwidth required for
transport is 20 KHz x 16 samples / Hz x 8 bits / sample = 2.560 Mbps. Both of the preceding
examples utilized an 8-bit / sample sampling machine. If the digitized signal were created on a
16-bit / sample machine, the results would be double.
There is a trade-off between how well we can hear and how much bandwidth is necessary to
carry the digitized audio signal. The more bandwidth we use, to a point, the better the reproduced
signal matches the original signal. However, reproducing the signal to a greater accuracy than we
are capable of distinguishing with the limitations of our hearing abilities is not sound economics
and wastes bandwidth. So, some folks argue that oversampling x4 and limiting the upper
bandpass to 10 KHz (bandwidth = 10 KHz x 4 samples / Hz x 8 bits / byte = 320 Kbps) is
sufficient for virtually all audio applications.
However, audiophiles want x8 or even x16 oversampling and 16 bits / sample. Typically, 1.28
Mbps is sufficient for most audiophiles and 320 Kbps is sufficient for the rest of the world.
For telephone quality audio connections, the bandwidth required is 4 KHz x 2 samples / Hz x 8
bits / sample = 64 Kbps or the basic T1 rate. Of course, anyone listening to the sound reproduced
with this set of parameters would not mistake the sound as originating from Carnegie Hall.
ATM can transport the bit rates required for audio without any effort. However, if the sound is to
arrive at the destination in a live transmission and, after conversion to an analog signal, be a
faithful reproduction of the original sound, the transmission latency (delay) must be acceptable. A
latency greater than 100 microseconds begins to become noticeable at the higher frequencies. To
keep latency from noticeably degrading the audio signal, a transmission rate of 4.2 Mbps (3xDS1 UNI) or greater is required.

Figure 7-1. Simplified composite video test signal

Full Motion Color Video

Real-time video applications have unique characteristics that require special consideration for
encoding and transporting the data. The key words are "real time." For an audio and/or video

signal to arrive at its destination (your monitor/TV set) with an acceptable level of distortion (less
than that required for the human eyes and ears to perceive), the time relationship between the data
bits must be preserved within the acceptable limits.
Figure 7-1 shows a composite baseband video test signal. The signal varies from 40 IRE to +120
IRE over a 62-microsecond period. The timing relationship between the various components of
the composite signal, which vary over a 62-microsecond period, must be maintained within strict
limitations, or else the signal received, and viewed, will contain a varying degree of distortion.
The amount of distortion is dependent upon the shift in the relationships, but it usually doesnt
take much to muck up a color signal sufficiently to say the picture is not viewable.
A closer examination of Figure 7-1 in the 0- to 15-microsecond time period reveals the waveform
in Figure 7-2. The timing values given have typical 0.1 microsecond tolerances. With 0.1
microsecond tolerance, you can readily ascertain there are critical timing relationships involved.
The particular video waveform segment shown in Figure 7-2 is the video blanking pulse which
includes the chrominance subcarrier. Picture effects of improper timing relationships include: no
picture, picture too light, picture too dark, apparent color saturation, picture breakup, incorrect
color saturation, color smearing or bleeding, poor reproduction of sharp luminance transitions,
fuzzy vertical edges, brightness variations between the left and right side of the screen, horizontal
streaking and smearing, top-to-bottom brightness inaccuracies, picture flicker, changes in hue,
incorrect reproduction of picture colors, loss of detail, transient brightness effects, grainy or
snowy picture, color sparkles, picture blurriness, and an audible buzz in the audio channel. For
most of the video problems just listed, it does not take much change in the composite signal
relationships to produce the undesired picture effects. What causes changes in signal relationships
in ATM transmission? Primarily, signal delay as cells transit the network. There is the finite time
a cell takes to travel from point A to Point B. Also, there is a queueing time, called latency,
associated with every network node the cell must transit. Again, all cells must wait their turn
before they are processed by the equipment at every node.

Figure 7-2. Simplified video pulse width requirements

Figure 7-3. Composite color video bandwidth diagram

Another issue related to video signals is the amount of bandwidth (BW) required to transmit the
video signal. Figure 7-3 depicts a standard composite NTSC video signal that has not been
compressed by any compression scheme. The video signal bandwidth is 5.75 MHz with an
additional 0.25 MHz guard band on the upper sideband for a total bandwidth of 6 MHz. To
transmit just one frame of this signal requires 192 MHz. The Nyquist rate specifies 2 x signal BW
(or 2 x 6 MHz) x 16 bits for digital signals and 256 color resolution of bandwidth. 192 MHz is a
whopping amount of bandwidth. An OC-3 port speed at 155 MHz and $10,000 per month access
charge is not sufficient to transmit standard, full motion, NTSC video images with only 256
colors. To transmit 65K colors, a 384 MHz BW is required. Well, fortunately, there are
compression schemes that allow the video signal to be reduced in BW when it is digitally
encoded. The current digital compression scheme giving the most compression for the buck is
Motion Picture Equipment Group-2, or MPEG-2. MPEG-2 compresses the video signal down to
1.2 Mbps to 7.5 Mbps, depending upon the quality of signal preferred with the lower bit rate
yielding the poorest quality picture and typically used for applications that do not require high
quality video, such as teleconferencing. 7.5 Mbps full motion color video has a bandwidth that is
a good fit with ATM and ADSL technologies. A 5xDS-1 port costs about $4,000 per month and
provides 7.5 MHz bandwidth. A metropolitan area the size of Dallas has approximately 450,000
CATV subscribers who pay approximately $40 per month for subscriber basic rate access. With a
port charge of $4,000 per month, it does not take too many of those $40 per month CATV
subscribers to pay the expense of shipping video over ATM (and ADSL) to the home. Sure, the
expense of an ATM-based video network headend will be about the same as a CATV headend.
However, the largest expense in the CATV industry is installing and maintaining the coaxial
delivery system. The ATM video network eliminates the coaxial delivery system burden and there
is essentially no replacement burden. If ATM does not put the CATV industry out of business
within five years, I will eat my hat.
Multimedia

At the beginning of 1996, it was estimated there were 34 million PCs in American households
with an additional 16 million in households around the rest of the world. Some estimates of the
total number of PCs in the world equal 100 million in the year 2000. It is anticipated that the
growth in PCs will occur primarily in Asia and Europe. These 100 million households are the

more affluent households that are likely to subscribe to data access services for pleasure and work
at home.
Computer games and other forms of digital entertainment are washing over the gunwales of our
cultural ark. While the entertaining virtual worlds can certainly claim to be multimedia products
with their video images and audio bytes, it is really the Internet that provides the greatest
motivation for developing multimedia applications.
Internet-based electronic commerce and educational applications will absorb an amazing quantity
of audio and video bytes over the next five years.
We are familiar with the interminable wait while images slowly flicker to life when accessing
web pages. Modem companies delivered 56 Kbps modems in short order and if you are lucky
enough to have one in your computer, assuming your Internet Service Provider (ISP) has installed
56 Kbps modems in its facility, then the "web wait" portion of the "www" is not as wearing on
the patience as 9.6 Kbps access.
When the web first came into existence, everything available on the web was text based and 9.6
Kbps seemed adequate for viewing the information. Then, folks started getting creative and
graphics, images, video, and sound were soon added to web sites in a grab bag of multimedia
potpourri. As the graphics and images become more detailed and varied in color and tone, more
bits are required to display the images. Table 7-1 lists the number of bits required to display some
typical monitor images. The figures given in the table are "raw" numbers that represent images
without compression. Compression techniques for video reduce the number of bits significantly
while also reducing the quality of the video image somewhat. A 65K color, full motion video
image can be compressed as much as 100 fold, giving a 1.5 Mbps (157 Mbps/100) transmit
speed.
Monitor
Display

640x480
640x480 full
still images motion video
(bits)
(bits/second)

audio stereo
Total video+sound (bits)
quality (x8
sampling) (bits)

text (80x25)

16,000

N/A

N/A

16,000 (text only)

black & white 2,457,600

image

78,643,200

160,000

78,803,200

16-color
image

2,457,600

78,643,200

160,000

78,803,200

256-color
image

2,457,600

78,643,200

160,000

78,803,200

65,536-color
image

4,915,200

157,286,400

160,000

157,446,400

Table 7-1. Computer monitor display bandwidth

The Real Audio and Video World

Audio and video signals come in two basic flavorsReal Time (rt) and Not Real Time (nrt). A
real-time audio and/or video signal is a signal that has a fixed relationship to the clock as the
seconds are ticking by. An example of a real-time application is the transmission of live sports
events or the transmission of a video signal over a coaxial cable from the local CATV service
provider. Not real-time video applications include those video movies rented from certain
locations on Friday night and played in the VCR. Real-time applications require the audio and/or
video signal to be transmitted and displayed virtually simultaneously in real time. Not real-time
audio and video applications do not require the signal to be transmitted and displayed virtually
simultaneously.
In other words, a not real-time application can receive the audio and video signal in some known
sequence, then store the information for any service provider and/or user for a specified period of
time before playback. With this approach, the order of the cells as they arrive at their destination
is unimportant since they are assembled/stored in their original sequence by the equipment at the
destination.
So, a video services provider can ship movies over the ATM network to homeowners through
copper wire to the ADSL set-top boxes. If there is some storage device either at the customer
premises or at some intermediate point (perhaps the CO), the audio and/or video signal can be
saved for playback at the convenience of the customer.

Figure 7-4. ATM switches

Finally, if the ADSL set-top boxes include buffers with sufficient size to hold 50 milliseconds
(approximately 425 Kbits) worth of ATM cells, then all the issues with real-time video signals
and transmission time essentially become non-problems.

ATM Switches
There are two distinct types of ATM switches. For the network backbones, there are the fast core
switches with speeds in the 20 Gbps to 50 Gbps range. The highest speed core switch on the

market in 1998/1999 has a maximum port speed of OC-48. Core switches are physically located
on service providers premises and form the service providers network node to network node
interface (NNI). Typical prices of core switches range from $100,000 to $250,000.
ATM switches located on customer premises are called edge or access switches and form the
User Network Interface (UNI) (see Figure 7-4). These switches are typically utilized to connect
LAN networks to the ATM backbone. These switches also have the ability to concentrate various
network traffic from multiple sources at the customer premises and convert the non-native data
streams into a combined ATM bit stream for egress and disassemble an ATM bit stream into
separate native bit streams on ingress. Typical edge switch speeds range from 12 Gbps to 50
Gbps. The maximum port speed for 1998/1999 is OC-12. The number of ports available for a
typical ATM switch is detailed in Table 7-2. The table shows the maximum number of ports a
12.8 Gbps switch can have for several port speeds.
Typical ATM Switch Capacity (Switch speed 12.8 Gb/s)
Port Type

Maximum Ports per Switch

DS-3 Ports

282

OC-3 Ports

94

OC-12 Ports

15

DS-1/NxDS-1 Ports

752+

Table 7-2. Typical ATM switch capacity

ATM switch prices in Table 7-3 are general. The user must include 6 percent for maintenance if a
maintenance contract is purchased and about $1K per person per day (average 5-day training
period) for training if the equipment is self-maintained.
Switch Type

Application

Speed

Price

Service Provider High capacity backbone transport for

public carriers

OC-3 +

$100K-$2M

Enterprise WAN High capacity backbone transport for

private ATM networks

DS-3 +

$50K-$250K

WAN Access

Multiplexing multiple technology bit

streams onto public service providers
network

NxDS-1 + $10K-$100K

LAN Access

Concentrate multiple technology bit stream NxDS1 + $10K-$135K

for workgroups

ISP

Switch ISP traffic onto public ATM

network

Table 7-3. ATM switch price ranges

DS-3 +

$55K-$500K

Inverse Multiplexing Over ATM (IMA)

DS-1 access is about one-sixth the cost of DS-3. DS-1 is too slow for many applications and DS3 is much too fast. If the service provider only offers those two choices and the application
requires something in between, well, we all have to make tough decisions from time to time. Do
you install the more expensive DS-3 access and buy more bandwidth than necessary and waste
the resources (how about becoming a reseller?), or dont buy enough and cope somehow with the
ensuing traffic problems (oversubscription and 24 hour per day transmission)? Or do you sign up
with a service provider who offers IMA? IMA is the ability to multiplex traffic in multiples of
DS-1, such as 2xDS-1 or 2.088 Mbps. Typically, IMA is offered by service providers in multiples
of DS-1 up to 8xDS-1. With a DS-3 access line six times more expensive than a DS-1, the user
can multiplex up to five DS-1s and still be ahead.
There are some modest price differences between ATM access equipment that have IMA
capability and access equipment lacking the IMA functionality. A financial analysis will
demonstrate the most cost effective solution. DS-3 prices are expected to drop over the next two
years. Determine the price differentials between DS-1 and DS-3 access in the geographical area
of interest before choosing a solution.

Network Node Interface (NNI)

The NNI issues relative to the user are networking monitoring, management, and link reroute.
Network monitoring and management consists of someone or something plugged into the
network and making some kind of intelligent decisions based upon what is happening. Intelligent
network circuits deal with the mundane moment-to-moment, that is, the predictable, situations.
Congestion control comes to mind when network management is mentioned.

Figure 7-5. ATM "Cloud"

It takes about 30 seconds to reroute if a link goes bad. Time to reroute is switch dependent and
also dependent upon the number of switches in network. It is very important to determine time to
reroute if traffic is heavy. Ask your service provider for this information.

Figure 7-6. Typical national ATM commercial backbone network

Service Provider Backbone Network

Typically, the ATM backbone network is referred to as the "ATM Cloud" and is depicted as a
cloud in the literature. The cloud may consist of many ATM switches interconnected by some
particular topology. Various network topologies including point-to-point, star, partial mesh, and
full mesh are utilized by the service providers. The interconnections between switches are
referred to as trunks. Also, the connection between any two switches is called a Network Node
Interface (NNI). Typical trunk speeds are OC-12 and OC-48. Contrast the trunk speeds to the
customer access speeds of T1/T3/NxDS-1/OC-3. Of course, the trunk speed must be greater as
the trunk is carrying the traffic of many customers simultaneously.
Figure 7-6 depicts a typical ATM backbone network. The network shown in some detail in the
figure might also be shown as a network "cloud." Notice the network topology is a partial mesh
topology. The choice of a particular network topology is dependent upon the origin and
destination of customer traffic. Like most businesses, ATM service providers seek to minimize
costs and maximize revenues and profits. Switches are geographically located with the profit
motive clearly in view.
Each service provider has a Network Management Center (NMC) to monitor the health of their
respective network. When there is a network trunk failure, the NMC will detect the failure very
quickly and dispatch a maintenance crew to fix the problem. Such a situation might arise when
someone cuts the fiber interconnecting two switch sites (nodes). The typical NMC mean-time-torepair (MTTR) goal is four hours. Network coverage is maintained 24 hours per day, seven days
per week (7x24).

Figure 7-7. ATM corporate subscriber unit

User Network Interface (UNI)

The User Network Interface (UNI) is the physical point where the user equipment is connected to
the service providers backbone. The typical location of the physical connection is the customers
premises. Currently, corporate customers use equipment such as ATM access switches or
concentrators to connect to the ATM network. In the future, consumers may use set-top boxes as
the UNI for ATM access.
The typical UNI access speeds now are NxDS-1, DS-3, and OC-3.The number of ports per PVC
is most often between 64 and 256. A typical corporate customer has three to five ATM UNI sites.
As more corporate players enter the market, there could be some additional port and PVC choices
available. For consumer access, 7xDS-1 access speed is the most interesting speed due to the
bandwidth requirements for full motion video.

Figure 7-8. ATM consumer subscriber unit

Service providers offer UNI managed services including CPE installation and maintenance, and
training. Network monitoring of CPE is provided by the service providers NMC. Just as in NNI
trunk monitoring and maintenance, the NMC will dispatch a maintenance crew if the equipment
has failed. The user can choose from several maintenance options including 5x8 coverage (five
days per week/eight hours per day), 7x24 coverage (seven days per week, 24 hours per day), or
some combination of coverage such as 5x12 (five days per week/12 hours per day). Or, the user
can elect to maintain the CPE themselves after the warranty period has expired. If the user elects
to maintain the CPE themselves, there are usually several choices available. One maintenance
approach is to purchase spare cards and replace failed cards as required, returning the failed card
to the manufacturer for repair at the users leisure. Another approach does not require the
purchase of spare cards. The user requests a replacement card from the manufacturer who ships
immediately and bills the user after the failed card is returned or 30 days later, whichever occurs
first.
The typical 7x24 CPE maintenance contract specifies a four- or six-hour MTTR. The "M" in
MTTR means the average. While any particular failure might take many hours to repair, the goal
is an average time of four hours. Since these things are specified in the corporate users contract,
a site MTTR log that records CPE downtime is imperative if the user wishes to enforce the
service providers legal obligations. If the actual MTTR is significantly above the stated contract
MTTR, a user can negotiate consideration from a service provider for the perceived, and actual,
loss of user network resources. Also, if any legal dispute arises concerning the service providers
legal obligations, the MTTR log becomes an indispensable document. The importance of keeping
records cannot be overstated.
The typical service provider is not really providing the UNI network monitoring and CPE
maintenance. The service providers do not have internal business systems capable of dealing with
ATM. To provide the UNI monitoring and CPE maintenance, they are subcontracting the work

out to small, third-party vendors. And the ability of the service provider to monitor the
performance of the third-party vendor may not be adequate. If the user does not keep records,
then disputes concerning monitoring and maintenance become the users word against the thirdparty vendors. Such situations could possible give rise to some ugly disputes over the next couple
of years, as several thousands of dollars for each incident are at risk.

Corporate Access
Corporate ATM access is characterized by the volume of diverse, non-native traffic, and the need
to maintain site integrity 7x24. Figure 7-7 depicts a typical corporate UNI. All data traffic from
the user is piped to concentrators, also known as aggregators or access builders. Concentrators
take the non-native upstream bits and convert/combine them into an ATM cell stream. The
downstream ATM cells are desegmented into their native bit streams and routed to the proper port
by the concentrators. The CSU/DSU function may or may not be integrated into the concentrator
functionality.

Figure 7-9. Frame Relay to ATM Interworking

Consumer Access
Consumer access to ATM is enabled by the ADSL middleman. ADSL is necessary to provide the
1.5 Mbps to 10 Mbps bandwidth necessary to provide full motion video over the copper wire pair
into the home (or business, cheap corporate access!). Figure 7-8 depicts a typical consumer ATM
UNI interface, also called a set-top box.
The Line Interface Unit provides signal conditioning. The ADSL modem, besides up/down
converting the signal to/from the analog signal with the desired bandwidth, converts the ATM cell
stream to/from the non-native bit stream. The non-native bit streams include POTS, video, audio,
multimedia, and Internet access applications.

Network Management
Network management requires the service provider to poll the ATM switch, both backbone and
edge switches. The service provider reads the switch MIB variables of interest periodically.
Typical reporting periods are 30-minute intervals with some providers moving toward a 15minute interval in 1998. Trunk and edge access utilization are monitored for service engineering
design compliance to ensure QoS class objectives are achieved. Cell loss, peak and average cell
times, PVC, and port utilization are some of the network variables that are monitored. Since the
data is removed from the switch every 30 minutes, all values are an average of the previous 30minute time period. Therefore, a user can regularly burst above their PCR without incurring any
additional cost, as long as the average PCR over the measurement interval is below the negotiated
PCR rate.

Some service providers offer online reports accessible by users for network monitoring. These
reports are available for selected time periods, in the measurement increments of 15 or 30
minutes, of hourly, daily, weekly, and monthly. The online port and PVC statistics do not match
the billing statistics the user receives due to different business reporting systems utilized for
network monitoring and billing.
PVC test loops trace every path in the network. These loops are used to measure loss, delay, and
delay variation. Currently, typical network availability is around 99.95 percent.
A 99.95 percent network availability means the network is not available to the user for 36
minutes during a 31-day month.

ATM Interworking
ATM Interworking is also called Multiple Services Platform (MSP). ATM Interworking provides
the ability to pass traffic originating from sources utilizing other protocols across the ATM
network. The source data, called non-native since in its original format it is not ATM, must first
be converted to the ATM protocol format before it can transit the ATM network. Customer
premises equipment called aggregators or concentrators can perform the conversion to ATM
protocol.
After the non-native data is converted to ATM format it can be transmitted to any destination
regardless of the native protocol of the destination as long as an aggregator or concentrator
reformats the data in the native protocol if it is not ATM.
As an example, a Frame Relay concentrator can convert the frames to cells and ship them across
the ATM network to an SMDS destination (Switched Multimegabit Data Service), as long as the
SMDS destination has an ATM-to-SMDS concentrator to reformat the ATM cells. Also, for
Frame Relay interworking, frame PVCs can be carried over ATM PVCs. Or, frame PVCs can be
multiplexed onto one ATM PVC.
There exist in the "clouds" many communication technologies, some with great similarity and
many with great disparities between their protocols and technologies. There also exist in the
"clouds" many user networks utilizing the diverse protocols and technologies seeking to connect
to not so similar networks and finding that it can be an acrimonious, complicated, and expensive
proposition that can yield a marriage between technologies made in hell. ATM Interworking is the
magic that makes the marriage between diverse networks a marriage not made in the "clouds" but
one made in the heavens. ATM Interworking provides several benefits worth noting here.
With customer premises concentrators and aggregators providing the ATM-to-whatever interface,
ATM Interworking provides a seamless, global data, video, and voice service. CPE is available
now that can provide interworking for several legacy technologies. And the CPE is not expensive.
Original equipment manufacturers (OEMs) plan to introduce new concentrators and aggregators
for the technologies and protocols that are not yet able to be converted by the end of 1998.

ATM Interworking provides users with a smooth, quick, and very economical migration path to
ATM. ATM Interworking is as easy as it gets in the networking business. Just purchase, install,
turn on, and configure the CPE concentrator or aggregator. Now, you have global access to any
other network connected to ATM. ATM Interworking quickly extends the LAN (LAN Extension,
or LANE), giving access to WANs and GANs. The day of plug-and-play internetworking is not
far off.
ATM Interworking is not without some challenges. Extending LANs to WANs and GANs creates
a significant network management issue. To cope with the countrywide, continental, and global
management issues, service providers are turning up network management services that not only
manage the providers backbone network but can also manage the customer premises. While the
service providers are struggling to define who, what, when, why, where, and how, entrepreneurial
third-party vendors are stepping into the vacuum and offering complete network management
services, including site preparation, purchasing, installing, and maintaining CPE, and providing
around the clock network monitoring. Additionally, these creative companies can provide any
number of network usage and status reports online for the customer.
IP traffic is already heavily multiplexed, which makes ATM overhead a more noticeable burden.
Demultiplexing IP traffic may be a short-term solution; changing to native ATM as equipment
and systems age is a viable long-term solution.
Service Level Guarantees (SLGs) for Frame Relay is stated contractually in terms of frames.
Service Level Guarantees for ATM are stated in terms of cells. Usually, there is not a one-to-one
correspondence between the contents of a Frame Relay frame and the contents of an ATM cell.
Service providers must determine how to interpret SLGs for ATM Interworking network
topologies.
A robust ATM Interworking solution supports all transmission technologies including SMDS,
Frame Relay, X.25, IP, audio, voice and video. Currently, ATM Interworking does not support
SMDS and X.25 data technologies. Voice Over ATM (VOA) is available from several service
providers that use equipment supplied from OEMs each using proprietary solutions. Until the
ATM Forum completes the VOA specification and the subsequent adoption of the specification
by the industry, the proprietary nature of current solutions makes VOA interworking problematic.
The ATM Forum expects to complete a VOA specification by the end of 1998. Vendor compliant
equipment will follow soon afterwards.
ATM Interworking Quality of Service is also an issue. Depending upon a service providers
backbone capabilities, a level of service available in one technology is not necessarily available
in another. Remember, not all networks are created equal. The inequality of the networks is due to
the implementation of ATM before standardized specifications, and service providers and OEMs
stampeding to be "first-to-market" with ATM products and services. Service providers bought
ATM switches several years ago from different OEMs that utilized proprietary designs in their
equipment. Now, some service providers are finding it necessary to trash their previous ATM
backbone to install ATM Forum standards compliant equipment. Anyway, check with each
service provider; after all, there are not that many companies offering ATM servicesyet.
Determine which providers can offer the interworking required. As the ATM Forum standards

effort progresses, technology interworking solutions will become as easy as booting up a

computer.
A term that is occasionally seen in the ATM literature is "tunneling." Tunneling refers to the
encapsulation of Frame Relay frames into ATM cells. The Frame Relay frame, including data and
signaling information, is treated just like any other bit stream received, and completely
encapsulated. The frame signaling information remains embedded in the ATM cells along with
the data. The signaling for the ATM backbone network comes from the CPE aggregator,
concentrator, or ATM switch.
The interface between a variable length frame-based technology and ATM is called a Data
Exchange Interface (DXI). (See page 136.) The DXI is responsible for converting the variable
length frame into the fixed length ATM cell and vice versa. The actual piece of equipment that
performs the conversion is variously called a CSU/DSU, concentrator, or aggregator.

Quality of Service
When a user desires ATM service, the user must contract with an ATM service provider to carry
the users data traffic over the service providers ATM network. One of the several choices the
user must make when contracting ATM service is the level of Quality of Service (QoS) the user
desires. Of course, network design requirements dictate the particular Quality of Service the user
chooses. Remember, however, a contract that is legally binding must be negotiated and signed
before any service provider will haul data over its ATM network. Contract lengths vary
depending upon service provider and user and the amount of traffic expected. Typically, a service
provider requires the user to sign a three-year contract. And we are discussing a contract that
costs the typical user anywhere from several thousand dollars per month to $50,000 plus per
month. It is not a good thing to be stuck in a contract that does not really provide the Quality of
Service the network requires for optimum performance. Renegotiating the contract is always a
possibility but the loser most likely is the user.
Automatic rerouting of links after link failure takes time, from 30 seconds for a small number of
switches in the backbone to several minutes for a large number of switches. Rerouting occurs if
there is a trunk or port failure or if the bit error rate exceeds predetermined thresholds for an
extended period of time, usually two minutes or more. Rerouting is a function of switch
manufacture. Since the service providers have installed ATM switches from various switch
manufacturers, rerouting time varies from service provider to service provider.

Switched Virtual Circuits

Switched virtual circuits (SVCs) is a user on-demand connection. SVCs are only established
when one party desires to transmit to another party, and disconnect occurs when the parties have
completed the transmission. ATM layer service class agreements and negotiation between
network elements require complex handling by the hardware and software interface components.
ATM layer class and traffic descriptors for SVCs are the same as those for PVCs. The SVC
connection is referred to as the switched virtual circuit connection, or SVCC. The SVP
connection is referred to as the switched virtual path connection, or SVPC.

Two extremely important service features are necessary to make SVCs attractive: screening lists
and source address authentication. Without these security features, no network manager worth
their pay could allow an SVC connection to their network. Imagine the havoc if an unauthorized
connection were allowed.
When the network receives a connection request, the source address authorization is checked.
The calling party number in the setup message must match the ATM port address. If the two do
not match, the connection is refused.
SVCs allow connections between two or more UNIs, or DXIs/FUNIs. The connection is made
when the calling party signals the called party, and disconnect occurs when either party signals
the other party of the end of the connection. Signaling is performed in-band and all data transfer
occurs in real time. The in-band signaling guarantees the same circuit control over the integrity of
the transmission as that provided by PVCs.
Switched virtual circuits (SVCs) offer the user PVCs on demand. The user establishes and tears
down circuits using Q.2931 (ATM Forums UNI 3.1) signaling. Traffic descriptors and service
classes prescribed for PVCs also are applicable for SVCs. The calling party can send the called
party, in SETUP, the called party subaddress, calling party subaddress, AAL parameters,
Broadband High Layer Information, and Broadband Low Layer Information. The called party can
send the calling party, in CONNECT, AAL parameters and Broadband Low Layer Information.
SVCs support point-to-point connections, point-to-multipoint connections (also called
multicasting), and multipoint-to-multipoint connections. SVCs also support simplex point-topoint, duplex point-to-point, and simplex point-to-multipoint transmission.
The calling partys source address is authenticated. An added feature of SVCs is the ability to
provide address screening to users. Address screening provides an imperative security feature that
allows users to control who connects to their network. With address screening, only those
connection requests with valid source addresses are accepted. Address screening lists using
allowed or disallowed decisions can be kept. A calling party address list may contain up to 256
allowed/disallowed addresses. Also, a called party address list may contain up to 256
allowed/disallowed addresses. When a call is placed, the calling party switch determines if the
address called is allowed or disallowed and only places the call if the called address is on the
allowed screening list. The call is blocked if the address is on the disallowed list. When the called
party receives the request to connect, the called party compares the calling party address to its
allowed or disallowed screening list and accepts or denies the connection based upon the results
of the screening.

Figure 7-10. World wide ATM revenue

Currently, only two long-haul service providers, GTE and AT&T, offer SVCs but all the ATM
players plan to offer SVCs by the end of 1998.

Switched Virtual Path Connection

Switched virtual path connections (SVPC) are not currently supported by ATM Forum or Q.2931
signaling. SVPCs are anticipated to be available sometime near term as specifications are
completed.

Circuit Emulation
Emulation of time division multiplexing circuits, such as DS-1 and DS-3, is an easy task for the
adroit ATM AAL1 layer. The AAL1 layer receives the bit stream and repackages the bits as cells.
A 193-bit stream with DS-1 framing is converted to 53-bit ATM cells with hardly any effort as
long as an accurate synchronizing clock signal is used in the physical hardware performing the bit
shuffling.
There is a structured, meaning channelized, DS-1/NxDS-0 circuit emulation that provides an
emulated point-to-point DS-0, fractional T1, and DS-1 circuits. Non-channelized DS-1 circuit
emulation provides a point-to-point DS-1 circuit. Non-channelized DS-3 circuit emulation
provides a point-to-point DS-3 circuit. Circuit emulation provides a means for suppressing the
idle, or empty, frames, thus preventing the frames conversion to ATM cells.
Several coding and signaling schemes are available depending upon customer preferences.

Shared Access
A hybrid of permanent virtual path and switched virtual circuit, a user establishes a virtual path
connection (VPC) to the ATM network. A signaling channel within the VPC is used to set up and
tear down virtual channel connections (VCCs). The use of hybrid PVCs allows users with shared
access the ability to have SVC service when the LEC does not support ATM signaling
capabilities. Users with direct access must use this method of connecting instead of SVCs to
communicate with users with shared access.

Transmission Time
Transmission time is important in situations requiring the transfer of huge amounts of data. A
survey of file types and typical sizes is listed in Table 7-4. As you would expect, the larger the
file, the longer it takes to transmit it. Also, two transmission times are listed, 1 Mbps and 1000
Mbps. Again, as you would expect, the faster transmission time takes less time to transmit a file
of any given size.
Example

Data

Transmission Time (at 1

Mbps)

Transmission Time (at

1000 Mbps)

Page of Multimedia Text 10 Mbytes

100 seconds

.1 seconds

CAD File

12 Mbytes

120 seconds

.12 seconds

Digitized X-Ray

15 Mbytes

150 seconds

.15 seconds

Integrated Circuit Design 20 Mbytes

200 seconds

.2 seconds

Newspaper

50 Mbytes

500 seconds

.5 seconds

Computer Simulation

5 Gbytes

10,000 seconds

10 seconds

Table 7-4. Typical transmission times

The message here is the transmission time should be adequate for the application. A company that
produces animated computer images and simulations and requires remote connection to off-site
locations probably should not be connected to an ISDN at 64 Kbps. At 1 Mbps transmission rate
for a 20 Mbyte file, the flow of data seems like cold molasses clinging desperately to the bottom
of the jar.

ATM Competition
ATM is not without its competition. Leased lines, SONET over IP (X.25), Gigabit Ethernet
(X.25), High Speed Frame Relay, and Switched Multimegabit Data Service (SMDS) each offer
varying degrees of competitive pricing for ATM with similar port and PVC speeds. The next two
years will produce an amazing amount of technology hype as technologists, including yours truly,
harp on their favorite transmission technology for the 21st century.
Leased lines are legacy TDM networks that are slow and expensive. Just about everyone agrees
these cretaceous dinosaurs are pretty much a thing of the past. While they will be around and
serve useful purposes in a niche capacity, Frame Relay and ATM will gobble up this market as
leased line equipment ages and must be replaced.
Gigabit Ethernet at 100 Mbps to the desktop is already a done deal. But as a backbone
technology, well, 13 U.S. service providers have already installed ATM. Which do you think they
will market over the next two years? The theoretical maximum transmission lengths for Gigabit
Ethernet are 3 kilometers for single-mode fiber and 500 meters for multimode fiber. Since
Gigabit Ethernet cannot replace ADSL to the home over the ubiquitous copper pair, ATM in
partnership with ADSL is still the transmission technology of choice for delivery of video

products to the home. The consumer video market is considerably larger than the desktop market,
at least for now.
SMDS is already a dying niche technology with only one U.S. service provider. There are only
about 200 national customers who provide an average of $125,000 revenue per customer. Dont
bet on this horse!
1997 Features

ATM

Frame Relay

X.25

SMDS

Packet/Cell Size

53 octets

up to 4096 octets up to 4096 octets

Speed

up to 155Mbps 56 Mbps - 2 Gbps 9.6 - 64Kbps

1-34Mbps

Connectionless

Yes

No

No

Yes

PVCs

Yes

Yes

Yes

N/A

SVCs

Yes

Yes

Yes

N/A

Multicasting

Yes

Yes

No

Yes

53 octets

Table 7-5. Comparison of ATM and competing technologies

Why is ATM advantageous? Because competing technologies are soon going to be niche markets.
Remember, ATM is still an infant. Table 7-5 contrasts the features of ATM and the competing
technologies. While Frame Relay, with up to 2 gigabits transmission speed, has ATM beat in
1997, Frame Relays speed advantage will disappear by the year 2000. As more businesses
connect to the ATM network, the ATM port and PVC prices will fall. Additionally, ATM is the
only technology that can seamlessly integrate all types of data transmission into a single global
data pipe. This feature alone will make ATM attractive to global companies, who will drive the
rest of the world to adopt the technology.

Chapter 8
ATM and the Real World
Questions Answered in This Chapter
What are the real ATM benefits?
Who are the U.S. ATM players?
What does ATM cost?
Who are the ATM customers?
What hurdles must ATM leap to be the network technology of the world?
What is the future of ATM?

Introduction
This chapter provides a more in-depth look at ATM transmission features as found in the real
world. Every ATM service provider requires a user to sign a long-term contract, typically three
years. There are great differences in the features and services offered by the service providers.
Also, the ATM access charges vary greatly among the service providers. This chapter looks at the
data transmission market, where ATM is in that market, and where ATM is headed. Also, ATM
service features as offered by the various service providers are given an in-depth treatment. Dont
sign an access contract until you have read this chapter!

Explosions and the Growth of ATM

Figure 8-1. Networking explosion

Some brave souls are forecasting as many as 55 million people working remotely and over 300
million people with access to the Internet by the year 2000. Working remotely in this instance
means working at a location other than the employing companys premises. Remote office, remote
access, remote management, remote diagnostics, remote learning (distance learning), remote
switching system, and the list goes on. As we crowd around our desktops spending more and
more time with virtual space, we are becoming, no pun intended, a remote society.

Figure 8-2. Internet explosion

We are becoming virtualized at an ever increasing speed. Virtual in this context means almost,
but not quite. Virtual memory, virtual networks, virtual workplaces, virtual stores, virtual
routers, virtual classrooms, virtual disks, and the list goes on. We are becomingagain, no pun

intendeda virtual society. Virtually everything imaginable soon becomes virtualized if the
trend continues at the current rate. What waits for us at the moment we become virtual
humans? Wits, pundits, and soothsayers all have something to say about the end result. However,
it seems almost as if we are compelled to follow the path we are on to whatever final reward,
good or bad, awaits us. Remoting all the things that people once did in close association with
each other in something called work groups and virtualizing everything we can get our hands on
has given rise for the need to network like crazy. Information, and entertainment, must be easily
accessible to all the virtually remote inhabitants of the globe.
Every networked computer and every computer that connects to the Internet are candidates for
ATM service, now. No waiting, no kidding. Cant afford to buy DS-3 ATM access? No sweat. As
soon as telecommunications resellers figure out they can buy PVCs and resell the bandwidth to
the public and private sectors in smaller increments, we will be off on another wild
telecommunications technology ride.

Figure 8-3. ATM explosion

ATM service is still in its infancy. Currently, the approximately 500 worldwide ATM users are
corporate big-league hitters who are networking far-flung corporate sites into continental WANs
and worldwide GANs and large government agencies. In 1996 the Defense Information Services
Network (DISN) awarded an ATM contract to AT&T for ATM transmission and MCI for ATM
switching services. The contract specifies ATM services for 106 sites in the U.S. including Alaska
and Hawaii. The initial contract requires 120 DS-3 ports. As Big Business and Bigger
Government climb aboard the ATM train, ATM access prices will drop, fanning the flames of the
exponential growth fire.
As ATM matures, multimedia and video-based products offered by companies will either
augment or replace CATV companies mired in coaxial-based physical plants. The ATM market
focus will swing from corporate WANs and GANs to consumer products. The access speed
possible with ATM-based Internet service combined with ADSL delivery to the home and small
business anywhere in the world over the existing humble, copper wire physical plant, makes
ATM very attractive as the pump for transporting multimedia and video-based products into
every home in the world. Worldwide deregulation of telecommunications companies and markets
and the hands-off attitude assumed by many governments of Internet content opens up much of
the world as a market. What profit-loving, entrepreneurial company can resist such opportunity?
Two ATM factoids for the entrepreneurial readers:

In 1997 the U.S. was the largest region for ATM service, with Canada second and
Asia/Pacific third.
In 1998 the U.S. is the largest region for ATM service, with Asia/Pacific second and
Canada third.

Asia/Pacific is a potentially huge market for video-based products delivered over copper. Anyone
traveling to Hong Kong, Singapore, etc., cannot help observing all the satellite dishes hanging
from virtually every window of the skyscraper-sized apartment buildings.
Millions of households (apartment-holds?) are waiting for someone to show up with economical,
video-based services.

ATM Benefits and Network Evolution

Enterprises can use ATM as the local and campus backbone technology. ATM will deliver
desktop speeds of 25 Mbps and higher. Private enterprise ATM backbone networks in WANs
supporting legacy and emerging technologies are a reality. See Figure 8-4 for a typical ATM
WAN topology. ATM can painlessly take networks from:
1. regional to global
2. separated to integrated
3. bandwidth inefficient to bandwidth efficient
4. many platforms to fewer platforms
5. technology driven to service driven
6. complex to simple

Figure 8-4. Private enterprise WAN backbone network

ATM provides a worldwide standard offering scaleable, meaning any size bandwidth required by
the application, interoperability to all service providers and enterprises. The demand for WAN
switches to form private enterprise ATM backbone networks is growing, especially outside the
U.S. as Europeans and Asians are discovering the significant cost savings of ATM versus leased
lines.

ATM integrates LANs and WANs with little capital investment into bandwidth efficient
networks. Separate LANs and WANs are easily connected via ATM PVCs and inexpensive
customer premises equipment. And a single network monitoring and management center can
continuously watch over the combined network, providing peace of mind for little cost.
Since ATM is scaleable down to 16 Kbps (some service providers may allow access at 8 Kbps),
the user only pays for the bandwidth necessary to perform the intended network function. PVCs
are connected to the network continuously and the user pays for the access whether or not any
data is actually transmitted. If the application does not require continuous access, then SVCs may
be the answer since SVCs are built up for the duration of the transmission, then torn down. The
user only pays access charges for the time the connection was built up. Even for SVCs, the user
pays for the access while the circuit is connected whether or not any data was actually
transported. Besides allowing the user to utilize bandwidth efficiently, ATM also optimizes the
service provider backbone network bandwidth.
ATM consolidates voice, video, SNA, LAN, legacy data, and native ATM effortlessly into a
single transmission medium. No hassle, no worries, and no sweat. As native ATM networks
proliferate, there will be less need to aggregate non-native bit streams into a combined ATM bit
stream. In the interim, affordable concentrators and aggregators are available to combine any mix
of legacy bit streams into a single ATM bit stream.
Currently there are multiple transmission technologies and platforms available to choose from.
Network features and capabilities were confined within the limits of each particular technology.
Now, ATM goes far beyond the capabilities of current transmission technologies, allowing users
to focus on service issues and forget about technology limitations.
ATM enables transmission interworking and promotes pain-free communications evolution by
reducing interoperability problems with supporting multiple platforms and vendors. ATM
provides more services and features than competing transmission technologies. ATM provides a
global, seamless transmission technology, empowering GANs. And finally, ATM reduces user
long-term operating and capital costs since a significant reduction in trunking and support costs is
an ATM feature.

U.S. ATM Players

The ATM service providers offering access in national and large regional markets are MCI,
AT&T, Sprint, Worldcom/MFS Datanet, U.S. West, GTE, Nynex, Bell Atlantic, Bell South,
Pacific Bell, and Southwestern Bell. The ATM service features of import offered by each of these
companies is listed in Table 8-1. Listed underneath the service providers name is the region
covered by the service provider. Notice under GTE the region Shotgun. GTE does not offer a
regional or national ATM service. The company provides ATM access in several cities across the
country, coinciding with its presence in the local exchange markets.
For additional information regarding ATM suppliers, try the ATM Forum web page at
www.atmforum.com. The ATM Forum lists member companies and the particular ATM services
and equipment provided by them.

Company and
region

MCI
AT&T
Sprint
Worldcom/MFS Ameritech U.S. West
Nationwide Nationwide Nationwide Datacom
Northern West
Nationwide
Midwest

ATM UNI

NxDS-1, DS- T1/T3/OC- T1/T3/OC- DS-1/DS-3/OC- DS-1/DS- DS-1/DS3/OC-3, OC- 3/OC-12

3
3
3/OC3/OC-3
12 (ICB)
(ICB)
3/OC-12

Service Classes

VBR-nrt,
VBRCBR, VBR- VBR-nrt,voice
VBRnrt,VBR-rt nrt, VB-rt over VBR-nrt
rt,CBR, UBR CBR, ABR (4Q98),
CBR, ABR
ABR
(4Q98)

Customer
Network
Management

Plans to subcontract to
third-party
vendor

Yes

Backbone
Platform

GDC,
Lucent,
Newbridge, Cisco,
3Com Access Ascend
Builder
9600/9300

NEC,
Nortel

Frame Relay
Interworking

Yes

Limited

Managed CPE

ICB

Number of
Backbone
Switches

24 POPs

T1 UNI
300+
POPS,T3
UNI 240+
POPS

SVCs

4Q98

Yes

1Q98

Oversubscription 10xport

200%

IMA

Yes

Network Reports On-line

CBR,
VBR-nrt

VBR,
ABR,VBRnrt, VBR-rt

Cisco, GDC,
Lucent

Lucent

Newbridge

Yes

Yes

Yes

4Q98

4Q98

Yes

500%

200%

150%

None

No

Yes

No

No

1Q98

e-mail

e-mail

None

e-mail

none

Limited

Yes
T3 UNI 18 U.S.
POP. 4 Europe
POPs

Table 8-1. U.S. ATM service providers

Company and
region

GTE Shotgun

Bell Atlantic
Middle
Atlantic

ATM UNI

DS-1/DS-3/OC- DS-1/DS3/OC-12
3/OC-3

Bell South
South

Pacific Bell
California

DS-1/DS- DS-1/DS3/OC-3/OC- 3/OC-3

12

Southwestern
Bell South
Central
DS-1/DS-3/OC3

Service Classes

CBR,ABR,VBR- CBR, VBRRt,VBR-nrt,UBR rt,VBR-nrt

CBR, VBR- CBR, VBRrt,VBRnrt,UBR

nrt,UBR

CBR,VBR-nrt

Backbone
Platform

Lucent,
Newbridge

Ascend

Ascend

Cisco,
Newbridge

Fujitsu

Frame Relay
Interworking

Yes

No

Yes

No

Yes

Yes

Yes

4Q98

4Q98

3Q98

Oversubscription 200%

None

None

Unlimited

Negotiable

IMA

No

No

No

No

None

e-mail

Hard Copy

None

Customer Network
Management

Managed CPE
Number of
Backbone
Switches
SVCs

Yes

Network Reports e-mail

Table 8-1. U.S. ATM service providers (continued)

ATM Equipment Suppliers

See the ATM Forum list of equipment suppliers for a complete list of all ATM vendors. U.S. sales
of ATM equipment in 1997 were approximately $1 billion. This figure is expected to grow to $3
billion by 2000. Table 8-2 lists the top 12 ATM equipment vendors and their share of the 1997
ATM market. There is a great variety in the prices, reliability, and functionality among the ATM
equipment vendors products. Besides the usual evaluation of vendors before purchasing a piece
of equipment, ensure the vendor equipment is compliant to all ATM Forum specifications.
Vendor

Market %

Cisco

19%

Newbridge

16%

Nortel

14%

Fore

12%

Bay Networks

5%

Ascend

4%

Alcatel

3%

3Com

3%

Hughes

3%

IBM

2%

GDC

2%

Xylan

2%

Others

15%

Table 8-2. ATM equipment vendors

Figure 8-5. U.S. ATM revenues

Prior to 1998, many ATM specifications were still in the writing stage. ATM vendors either
attempted to second-guess the ATM Forum specification committees or just adopted a proprietary
solution and went forward with equipment development and deployment. To entice any reluctant
buyers, vendors promised to upgrade software and hardware to be specification compliant as
soon as the ink was dry on the specs. Now, some of them are finding it very costly and time
consuming to upgrade their products. Of course, anyone who purchased the proprietary
equipment will have to bear the cost of upgrading their equipment. Save yourself some
significant migraine headaches and purchase only specification-compliant equipment from day
one.
ATM equipment sales is the lions share of the ATM prize for the next two years as service
providers flesh out their network backbones and companies purchase ATM equipment for private
enterprise LANs and WANs. Of the estimated $3 billion U.S. ATM market for 2000, ATM
equipment will comprise $2.4 billion.

ATM Customers
The ATM customer base will grow dramatically between now and 2000. The following
discussion concerns the growth in service provider revenues and UNI connections. The dollars do
not include the ATM equipment market and the expansion of service provider networks. That
information is included elsewhere in total ATM revenue projections.
U.S. ATM service revenues have steadily climbed for the last five years. See Figure 8-5. The
ATM customer base is steadily increasing as users migrate from legacy systems. There are about
40 U.S. customers per service provider. ATM revenues are about $350,000 per customer per year.

Total worldwide ATM customer base at the end of 1997 was 1,000 or less customers. Between
1998 and 2000, U.S. ATM networks will add 1,200 new customers.
Outside the U.S., total ATM service revenues will increase from $65 million in 1997 to $650
million by 2000. Although Canada was second to the U.S. for ATM deployment in 1997, the
region of greatest growth from 1998 to 2000 will be Asia/Pacific, second only to the U.S.
Table 8-3 lists the top seven 1997 ATM revenue producing service providers. Not surprisingly,
due to its extensive national presence in communication markets, AT&T has the lions share of
the domestic ATM market. If everyone shopped for the best price though, AT&T would not be
leading the pack.
Service Provider

Market %

AT&T

28%

Sprint

22%

MCI

18%

Worldcom/MFS

13%

U.S. West

4%

Pacific Bell

3%

Ameritech

2%

Others

10%

Table 8-3. Who gets the ATM pie?

Table 8-4 lists the 1997 ATM service revenues by industry. Not too surprisingly, Internet Service
Providers (ISPs) are the users into ATM with the biggest bang. It is easy to imagine the huge
amount of data shipped around the country each day by the ISPs.
Industry

% Revenue

ISPs

41%

Government

32%

High Tech

16%

Manufacturing

4%

Retail

3%

Entertainment

2%

Utilities

1%

Insurance

1%

Table 8-4. Who bakes the ATM pie?

Perhaps not too surprising either are government and high tech providing the second and third,
respectively, sources of ATM revenue. While providing a paltry 2 percent of ATM revenue in
1997, expect entertainment to catch up and surpass all other categories by 2000.

ATM Ports By Industry

Table 8-5 lists the percentage of all ATM ports by industry. While ISPs outpace the government
in the revenue source category, as shown in Table 8-4, government has more ports than ISPs.
Guess it does pay to have prospective suppliers bid the job.
Industry

% of Ports

Government

37%

ISPs

28%

High Tech

21%

Manufacturing

6%

Retail

5%

Entertainment

1%

Utilities

1%

Durance

1%

Table 8-5. ATM users

From the information gleaned from service providers, it appears that if competitive bids are
sought, service providers will respond with better pricing. Let them know you are asking them to
competitively price their offering. They will sharpen their pencil a little more. Ask for their
preferred discount rate. Never, never, never pay the first asking price. Negotiate, negotiate,
negotiate.

ATM Port Sales By Type

Port Type

% Revenue

T1/DS-1

60%

NxDS-1

34%

DS-3

5%

OC-3

1%

Table 8-6. Where ATM revenue comes from

Table 8-6 lists the most popular ATM ports. One might conclude from Tables 8-4 and 8-5 that
ISPs and government prefer DS-1 ports. Someone certainly prefers them. Table 8-6 reflects the
current ATM service traffic profile. Most of the ATM traffic is data traffic. As entertainment

moves up the list of ATM revenue sources, expect the predominant port type to move toward DS3, as more bandwidth is required for video and multimedia than is required for ASCII character
and binary based data files.

ATM Services
Some ATM service providers are entering the managed network services market. Service
providers figure they are the network experts since they have recently installed two extensive
networks, Frame Relay and ATM, and have suffered the pain and agony of learning the ropes.
They are keen to share the knowledge gained, for a handsome profit, of course. However, their
solution to managed services is to contract the work out to multiple third parties and mark up the
service accordingly. Buyer beware! Often, the subcontractor is held accountable by the service
provider for installation, maintenance, service, training, network monitoring, and reporting only
by the casual business relationship between third-party vendors and the service provider.
However, contracting directly with a full service ATM managed services provider can guarantee
the users needs are met at a lower cost.

ATM Traffic
ATM traffic patterns are of concern to cost sensitive network managers. If the application is not
sensitive to the time of day, then knowing when the ATM backbones are most lightly loaded can
save some big bucks. With service providers offering liberal port oversubscription terms the
clever network manager can reduce the size of PVCs dramatically if data transmission occurs
during the right time of day and in the right direction.
Service provider networks are more heavily loaded with traffic going from the West Coast to the
East Coast. Possible reasons include the number of financial centers in Chicago and New York,
the number of corporate headquarters in New York, and the amount of government bureaucracies
in Washington, D.C. Notice the biggest single user of ATM ports identified in Table 8-5 is Uncle
Sam and Aunt Sarah. ISP, a close second to big government, is composed of many Internet
Service Providers with the majority of ISPs located in the eastern regions. Traffic from
Asia/Pacific enters western gateways and transits from west to east. Adding to the amount of
traffic bound for eastern ports (no pun intended) are the Canadian gateways located in eastern
regions. So, it seems reasonable that more traffic would flow from west to east.
ATM traffic peaks about 8 p.m. and begins to fall off rapidly after midnight until about 6 a.m.
when it begins to pick up again. Monday is the busiest day of the week for ATM networks which
seems reasonable as remote sites transmit weekend reports, data, etc., into corporate headquarters
on Monday morning. Friday afternoons are also a busy time as financial markets are abuzz with
closing market information. Avoid Monday morning and Friday afternoon, and transmit the bulk
of your data between midnight and 6 a.m. Eastern Standard Time, and you can cut your port sizes
in half. A significant cost reduction, indeed.

ATM Rate Structure

Service provider charges for ATM service include several elements. The amount charged by the
service provider for each element can vary widely from region to region and service provider to
service provider. The ATM pricing elements are port installation, port, port reconfiguration, PVC
fixed usage, usage per megabyte charge, and additional feature charges. The pricing factors are
number of PVCs, usage, distance, PCR, and SCR rates.
Usage-based billing is a feature of SVC service. The idea is simple enough: The user pays only
for the bandwidth used when it is used. However, SVCs are not yet offered by the service
providers. The issue with SVCs is not equipment availability to deal with SVC connections but
rather the service provider back office systems and cumbersome internal management processes.
Service providers must expend a herculean effort over several years just to plan, organize, and
implement the business systems to manage the provisioning, order entry, billing, and invoicing of
any service feature. Users can expect some service providers to offer SVCs by 1998 year-end.
Distance is a factor in ATM pricing for obvious reasons. The greater the distance to haul
anything, the greater the cost. A subtle cost factor not necessarily broken out separately in the
users contract is backhaul charges.
Additional features the user may be charged for include network monitoring and management,
CPE installation and maintenance, and network reports.
Non-Recurring Charges

A typical port installation charge averages $15 per port. The port installation charge is applicable
every time a new port is installed.
A typical port reconfiguration charge averages $15. The port reconfiguration charge is applicable
every time the user changes the parameters of the port such as SCR and/or PCR and VC/VP.
Recurring Charges

Service provider rate structures typically consist of five components. Each of these components
may or may not be broken out separately on the users contract and the users invoice. The five
rate components are access charge, access coordination, port charge, PVC charge, and backhaul
charge.
Access charges are the charges for connection to the local loop. Access charges are based on
standard tariff rates. Access charges are a feature of deregulation and are intended to compensate
the local operating telephone companies for the loss of subsidies once allegedly paid by AT&T.
This is the charge levied by the local service provider who is probably not the ATM service
provider this year. But, due to deregulation and the penetration of local markets by the long-haul
service providers, the local service provider very well could be the ATM service provider within
the next year or two.
The access charge is levied from the users premises to the ATM service providers Point of
Presence (POP). With the typical ATM backbone network having 25 POPs, the local access

charge can be a significant piece of the users monthly ATM cost. That is, if the users POP is not
colocated in the same community as the service providers POP, the user must pay to have the
cells hauled to and from the service providers POP. If the user is in Butte, Montana, and the
nearest ATM service provider POP is in Denver, then the user must pay the local access provider
its tariffed rate to haul the ATM cells from Butte to Denver. Even if the service providers POP
and the users POP are colocated, there is still a tariffed local access charge.

Figure 8-6. Typical service provider PVC rate structure

Access coordination (AC) charge is one of those regulatory fees intended to subsidize for the loss
of monopolistic markets. Oh, well.
The PVC charge is also called the usage charge. The PVC charge is directly related to the number
of bits the user transmits over the ATM backbone. Typically, the more bits that are transmitted,
the higher the charge. The PVC charge varies according to the class of service contracted for and
the associated traffic parameters, such as SCR and PCR.
Port charge is the charge for the bandwidth chosen by the user to connect to the ATM backbone.
Various port bandwidths are available depending upon the service provider and the capabilities of
their ATM backbone. Typical bandwidths are DS-1, DS-3, OC-3, and NxDS-1. Notice the
bandwidth is specified as a port speed.
ATM UNI

Port Speed

Recurring Price per

Month

Price per Bit (cents)

DS-0

0.064 Mbps

$1,500

2.3

DS-1

1.544 Mbps

$2,000

0.13

2xDS-1

3.088 Mbps

$3,000

0.097

3xDS-1

4.50 Mbps

$3,325

0.074

4xDS-1

6.0 Mbps

$3,650

0.06

5xDS-1

7.50 Mbps

$3,975

0.053

6xDS-1

9.0 Mbps

$4,300

0.048

7xDS-1

10.50 Mbps

$4,625

0.044

8xDS-1

12.0 Mbps

$5,000

0.042

DS-2

6.312 Mbps

$3,500

0.055

DS-3

44.736 Mbps

$6,100

0.14

DS-4

274.176 Mbps

$24,000

0.0088

OC-3

155.520 Mbps

$10,000

0.0064

Table 8-7. Typical ATM port charges

The typical (meaning average) ATM port charges shown in Table 8-7 are list prices. Depending
upon the service provider and the users business relationship, pricing can be flexible. Negotiate.
Negotiate. Negotiate. Shop around and strike the best deal you can. The price of aDS-3 port can
vary from $400 to $12,000 per month. Do not share your pricing information gleaned from the
other service providers when dealing with any particular service providers account team. Keep
the account team in the dark about the competitions pricing. Believe it or not, not every service
providers ATM pricing is public knowledge and each service provider is attempting to determine
the competitions pricing and align their pricing with the competition.
Backhaul charges are applied to the customers whose traffic must be transported indirectly for
some portion of the transmission path to the intended destination. Due to the location of service
provider POPs, some ATM traffic may have to travel as much as twice the distance as the
physical distance between source and destination. Backhaul charges may be buried in the
recurring port or PVC charge and not broken out separately. Ask the service providers sales team
if backhaul charges are applicable for the ATM access locations in the users network. If backhaul
charges are applicable, determine how much. Remember, backhaul charges are recurring, that is,
the user pays the charges every month. The user may find that changing the location of ATM
access for one or more nodes may result in considerable monthly savings at the expense of
moving some equipment one time.
ATM PVC Rates

ATM PVC rates are variable from service provider to service provider. Rate structures are
typically not public information. Over half of the U.S. service providers have not disclosed their
pricing structure. So it is not surprising to find huge differences in prices for the same service. A
shopper can find as much as 25 times the difference in price for access, before discounts.
The rate structure shown in Figure 8-6 is for a PCR of 1.544 Mbps. Notice the dip in the PVC
cost at about 1024 Kbps. 1024 Kbps is cheaper than 895 Kbps and almost the same price as 830

Kbps. Why would any service provider sell 1024 Kbps cheaper than 895 Kbps? What logic
resides in the rate structure shown? Probably the inexperience of the person responsible for
analyzing network costs and compiling the rate structure. This particular rate structure is for
VBR-nrt service with PCR equal to 1.544 Mbps. There is a rate structure for each SCR/PCR and
for each class of service offered by a service provider. If the user can pry the rate structures from
the service providers (maybe an unrealistic expectation) and similarly graph the SCR/PCR rates,
other interesting economizing opportunities may become evident. I have analyzed other rate
structures with similar results.
Does it make any difference whether a user subscribes to CBR or VBR service? Well, Table 8-8
provides a resounding answer to the question. CBR class of service is much more expensive than
VBR. If the user wants the bandwidth reserved for exclusive use 24 hours per day, seven days per
week, the user must be willing to pay the hefty price tag. Ensure the application demands the
class of service before signing on the dotted line.
Bit Rate (Kbps)

CBR

VBR-rt
VBR-nrt
(PCR=1.544Mb (PCR=1.544Mbp
ps)
s)

1024

$1,650

$1,000

$450

1344

$1,950

$1,150

$600

VBR-rt
VBR-nrt
(PCR=6.18Mbp (PCR=6.18Mbps
s)
)
4608

$5,250

$4,450

$2,500

VBR-rt
VBR-nrt
(PCR=12.35Mb (PCR=12.35Mbp
ps)
s)
10800

$12,250

$9,200

$6,250

VBR-rt
VBR-nrt
(PCR=16.38Mb (PCR=16.38Mbp
ps)
s)
16000

$16,000

Table 8-8. CBR and VBR class of service cost comparison

$12,500

$8,150

One-stop Shopping

One-stop shopping is simply the inclusion of all telecommunications services and charges in one
customer bill, preferably with all customer services provided by the single service provider. For
many companies, one-stop shopping is a simple idea that has great market appeal. One-stop
shopping is appealing to service providers also. Besides being a market stratagem with broad
appeal, one-stop shopping integrates and streamlines service provider back office systems by
reducing the number of separate business systems. However, ATM is an infant transport
technology that requires at least two more years of growth before anyone attempts to potty train
it. Service providers order entry, provisioning, billing, invoicing, network management, and
managed customer premises equipment business systems remain stumbling blocks to a back
office suite of integrated ATM business management systems. Service providers are accepting
new ATM customers on an individual case basis (ICB), although they do not publicize the fact.
ICB customers require manual processing of all back office systems. ICBs are a labor-intensive
effort for service providers. To get around the internal back office integration hurdle, some
service providers are subcontracting as much of the work with third-party vendors as possible.
One-stop shopping will probably not be available until 1999 or 2000.

Figure 8-7. Economical ATM?

Managed Customer Premises Equipment

Service providers believe they can leverage their networking expertise and provide users
managed customer premises equipment services. The marketing approach utilized by the service
providers is to convince network managers their life is made more simple by allowing the service
provider to assume the responsibility for all CPE issues including site preparation, installation,
maintenance, training, and monitoring. CPE management provides the network manager a single
focal point for all network concerns and needs.

Managed CPE may not be installed, maintained, and managed by the service provider who signs
the ATM service contract with the user. As many as four or five other companies could be
involved in managing CPE. The relationship between these third-party vendors and the service
provider may depend largely upon the ambiguous business relationships between and among the
companies. Know before you go ...ahead and sign the contract. A safer approach is to contract
with a network management company to provide all facets of network management including
managed CPE.

Port Oversubscription
Port oversubscription is actually a service feature that each service provider may or may not
allow. Port oversubscription allows the user to burst above the contracted PCR for some amount
that is agreeable to the service provider. The practice is risky if the network is heavily subscribed.
Currently, some service providers allow users to oversubscribe, with no additional charge, up to
their access speed. The service provider does not guarantee oversubscribed traffic. After all, those
paying customers who are still transmitting at or below their PCR get first crack at the bandwidth.
Users can expect the freebie to disappear as the networks become more heavily loaded. However,
oversubscribing while the opportunity exists and transmitting in the wee hours of the morn can
save a financially ailing enterprise some big bucks.

ATM Service Contracts

A word to the wise. Read and understand all the fine print in ATM service contracts. Since the
duration of the typical service contract is two or three years with access charges as much as
$50,000 per month (or more), the smart network manager knows and understands the
implications of the finer points of the contract.
Typical ATM service contract clauses include the following service descriptors:

access interface and speed

service class

traffic parameters

quality of service parameters

Review the contract service descriptors for compliance to the user network requirements. Also,
read and understand the penalties for changing service. A three-year contract is a long time in the
network world where situations and prices change almost overnight.

Hurdles
Usage-based ATM billing is not yet a reality. Users are allowed to burst up to the access speed
without penalty. Service providers are still in the individual case basis (ICB) phase, and are just
rolling out service. Service providers still do not have equipment certified and order entry,
provisioning, billing, invoicing, and trouble management systems are not ready for automated
ATM service. One-stop shopping is not yet a reality. Only limited access speeds are available

such as DS-1, DS-3, and OC-3. Inverse multiplexing over ATM (IMA) is just coming on line as
this book is printed.
Certainly, we need Internet access faster than 9.6 Kbps if we wish to view graphical based web
sites without wasting a lot of time waiting for images to download. Figure 8-7 contrasts the
various methods of accessing the Internet and their associated costs. How can a user get Internet
access faster than 9.6 Kbps? The user can opt for 28.8 Kbps LAN access at $250 non-recurring
and $225 per month recurring. Or, a56 Kbps modem can be utilized for the price of the modem,
about $125. But, you really want ISDN which runs about $1,500 non-recurring and $250 per
month for a paltry 64 Kbps. Okay, so you want to fly at double ISDN rate, that is 128 Kbps, for
another $1,500 non- recurring and $475 per month. How about jetting around with Direct
Broadcast Satellite (DBS) which provides asynchronous access (downlink) at 400 Kbps and is
available for about $400 non-recurring cost forthe satellite dish and $150/month recurring cost?
All of these approaches to Internet access are much more economical than the $1,500 access
charge for 64 Kbps ATM. So why ATM?
Table 8-9 details the average price per bit for ATM service. The price per bit of ATM access at 64
Kbps is about 2.3 cents per bit. However, the price per bit at 1.544 Mbps is 0.13 cents per bit, a
significant reduction in cost. Still, at $2,000 per month access charge, it is pricey for the typical
Internet user. But, the $2,000 per month charge is for a 24 hour per day, 7 day per week constant
connection. Few consumers are that dedicated to Internet browsing. Lets say you are on the
Internet two hours per day, five days per week, in a four-week month (February?). Then for 1.544
Mbps access in a usage-based billing scheme, your cost would be $187. Still pricey. Okay, lets
go for 400 Kbps access on a 1.544 Mbps port. The price is then $48.50 per month. Now we are
beating the socks off DBS access at $150 per month for 400 Kbps access. This example is based
on a PVC of 1.544 Mbps.
ATM UNI

Port Speed

Recurring Price per

Price per Bit(cents)
Month

DS-0

0.064 Mbps

$1,500

2.3

DS-1

1.544 Mbps

$2,000

0.13

2xDS-1

3.088 Mbps

$3,000

0.097

3xDS-1

4.50 Mbps

$3,325

0.074

4xDS-1

6.0 Mbps

$3,650

0.06

5xDS-1

7.50 Mbps

$3,975

0.053

6xDS-1

9.0 Mbps

$4,300

0.048

7xDS-1

10.50 Mbps

$4,625

0.044

8xDS-1

12.0 Mbps

$5,000

0.042

DS-2

6.312 Mbps

$3,500

0.055

DS-3

44.736 Mbps

$6,100

0.14

DS-4

274.176 Mbps

$24,000

0.0088

OC-3

155.520 Mbps

$10,000

0.0064

Table 8-9. Typical ATM port charges

If a faster PVC is used for the analysis, even greater savings will result.A 7xDS-1 (10.5 Mbps)
connection costs 0.044 cents per bit. The hypothetical consumer usage would be $10.48 per
month for 400 Kbps access. Now were cooking! But we really want at least 1.544 Mbps access
so that we might overdose on all that wonderful multi-media stuff. Cost per month on an OC-3
PVC? A paltry $5.88 per month. Everyone who is satisfied with a 300 percent gross profit margin
(sell price to consumer is $40 per monthcompetitive with CATV) run to the nearest ATM
service provider and purchase your OC-3 PVC now and start shipping video around the nation.
Dont forget to acquire the usage-based billing software from someone. Those of you still reading
this book and not dashing off to the nearest ATM store, well, it just doesnt get any better, for
now.
So, the message is, usage-based billing (only pay for what you use) must become a reality for
consumer markets to find ATM transport economical and competitive. When resellers realize they
can purchase PVCs and resell the bandwidth using usage-based billing, the consumer market will
take off.

Network Evolution
Because of the presence of ATM in the global marketplace, networks will evolve under the
influence of ATM whether or not any particular group or company adopts the technology.
Network evolution follows the path from:
1. regional to global
2. separated to integrated
3. bandwidth inefficient to bandwidth efficient
4. many platforms to fewer platforms
5. technology driven to service driven
6. complex to simple

Crystal Balls
The total ATM market is expected to grow from a $1 billion market to a $7 billion market by
2000. While ATM equipment will comprise the bulk of the growth, network services will see
similar growth rates. Leased lines will see a dramatic decrease as ATM and Frame Relay begin to
gobble up the leased line market. Between 1998 and 2000 the ATM industry can expect 16 times
growth in the total worldwide ATM market.

Chapter 9
ADSL
Questions Answered in This Chapter
What is ADSL?
What is DSL?
What are the different flavors of DSL?
How is information coded into a DSL signal?
What are the DSL architectures?
What are POTS (and pans)?
How does ATM fit with ADSL?
What are the ADSL market hurdles?

What is ADSL?
What is ADSL? In terms of our OSI protocol model, ADSL is a physical layer transmission
protocol for unshielded twisted pair media and is recognized as such by the ATM Forum. A word
now about DSL to save, hopefully, some confusion later on. DSL is used to describe the general
transmission technology of transporting data over twisted pair copper lines and it is also used to
describe a specific flavor of DSL. So, there is a generic DSL that includes all flavors of DSL and
there is the specific flavored DSL. We (folks in the industry) pray that the reader will realize
which flavor is the topic by the usage in the literature. The usage of the term in this book is clear
from the context and by specifically referring to the particular usage.
Asymmetrical Digital Subscriber Line (ADSL) technologysounds kind of daunting. Well, if we
look at each of the terms separately, perhaps we will not be overwhelmed by the seemingly
incomprehensible terminology.
Symmetrical refers to the similarity of form or arrangement on either side of a dividing line or
plane. If one side of the symmetrical arrangement is discarded by some means, then the resultant
arrangement is called asymmetrical. So, asymmetrical describes some arrangement that had its
origin in a process that generated a symmetrical arrangement. Subsequent processes produced the
single asymmetrical arrangement. In the world of ADSL, asymmetrical refers to the fact that
communications is primarily one way. That is, the bulk of ADSL communications is from the
service provider, such as a video supplier or Internet service provider, to the user.

Digital refers to the practice of encoding information in digits. In this instance, we are encoding
the information in a binary manner. That is, every possible value of the individual information
elements, called bits, can assume only one of two values, either a zero or a one.
Subscriber is any person or company, or any other entity, that has telephone service provided by a
telephone company. In other words, a subscriber is a telephone company customer.
Line is the twisted pair copper lines that are connected from the subscribers telephone to the
local exchange. Now, for those few of you who already have fiber-to-the-curb, line refers to your
fiber connection to the local exchange. A subscriber line is the telephone line connecting the
telephone in your house to the switch at the local exchange.
Here is the confusing part. There really is no asymmetrical digital subscriber line. There is a
twisted copper pair of lines that is physically connected from the customer premises to the local
office. Only when an ADSL modem is connected to each end of the subscribers line is any such
thing as an asymmetrical digital subscriber line magically pulled from the hat. Without the
modems, it is just a humble, twisted pair of copper lines.
Okay, lets get it out in the open now. No more hiding behind technological foo-foo dust. ADSL
technology is modem technology. Whoopee! Big deal! Modems are a dime a dozen (figuratively
speaking). We are accustomed to seeing and using subscriber line modems now. Everybody with
dial-up Internet access must have a subscriber line modem. Even if you do not have Internet
access you still probably have a subscriber line modem in your computer (or sitting next to it).
So, whats the big deal about Another Danged Subscriber Line (ADSL) modem? Speed and
entertainment.

Figure 9-1. Modern analog modem communications

The current modems are voice frequency analog modems. Voice frequency means they operate
down in the frequency range of human voices, and at the low end of the range. Telephone
companies band-limited telephone lines long ago from 600 Hz to 4 KHz yielding a 3.4 KHz
bandpass on the PSTN. Therefore, current modem technology was developed to exploit the 3.4
KHz bandpass available in the voice frequency range. Now, only so much data can be transmitted
in the 600 Hz to 4 KHz spectrum. With the latest analog technology available, analog modem
manufacturers are pushing the envelope to 56.6 Kbps. Telephone companies treat voice frequency
analog modems just like they treat your telephone conversation with your broker, or your granny,
or anybody else you have a telephone conversation with. The analog modem signals are passed
through the bowels of the telephone network without alteration. With a low price tag (now but we
would not have made this claim several years ago), lots of phone lines to connect to, lots of slowspeed data to exchange, and the ability to connect them almost immediately anywhere in the
world, voice frequency analog modems were Rulers of the Subscriber Line (RSL)!

Figure 9-2. ADSL modem communications

But faster Internet access and video-on-demand are driving the desire to trash the slow analog
modems and replace them with something that will bring computer monitors and television sets
to life. ADSL modems will do just that. If you dont replace your RSL modem with an ADSL
modem PDQ, you will be SOL if you want VOD on your HDTV.

DSL Technology
Bellcore defined the basic DSL acronyms in 1989 when Bellcore engineers dreamed up ADSL.
Originally, these engineers envisioned a digital data stream transmitted over the humble twisted
pair lines that would be sufficient to transport video images. At the time, 1.5 Mbps upstream and
16 KHz downstream seemed sufficient to accommodate digital video traffic. Since its original
conception, ADSL data rates have increased significantly as the companies compete to provide
data rates sufficient to accommodate virtually every conceivable service subscribers may yearn
for.
The telephone companies who are competing with each other to bring ATM and ADSL
technology online are closemouthed about the status of development of the technologies. Other
than their life is a glorious parade and arent we proud to be a part of it marketing info, little can
be gleaned from their research and development departments. However, I can imagine how
ADSL was discovered. Someone, somewhere, connected a device to a telephone line, swept the
spectrum with a signal, and looked at the response. Obvious to the engineers who performed this
mundane task must have been the ability of that telephone line to transmit and receive analog
signals in a band of frequencies much higher than the 4 KHz upper limit imposed by the local
office equipment filters. Also, advances in digital signal processing in a noisy environment have
contributed significantly to the ability to utilize the telephone lines for digital transmission far
above the original 4 KHz. Due to the frequency response of telephone lines high above 4 KHz,
we can have ADSL, and the other flavors of DSL technology.
There exist several reasons why DSL technology is limited in the distance signals can be
transported. Attenuation, or signal loss, caused by the physical resistance of the medium (wire) to
electron current flow, cross-coupling of signals from line to line (cross-talk), and termination
impedance (like a rubber ball hitting a brick wall, the signal bounces back) influence just how far
any signal of a particular frequency can travel before its level becomes so low that it cannot be
distinguished from the ever present noise on the line. See Figure 9-3. Just like a car after it has
freely rolled down a hill and levels out, the signal level gradually "slows down" over distance
until the signal can no longer be distinguished from the ever present line noise. When the ratio of
the signal to the noise becomes less than 10dB, the signal begins to become "noisy" (listen to an
AM station almost out of range). As the signal continues to decrease in amplitude, it finally
becomes unusable, then indistinguishable from the noise.

Figure 9-3. Simplified signal-noise relationship

Twisted pair telephone lines coming out from the central office in the local loop may be bundled
into large cables of several thousands of twisted pair wires. As the wires are spread across the
distribution area (from the central office to your home), they are bundled into 50-pair cables. Of
course, along the way there are intermediate distribution equipment the wires are routed through,
resulting in any individual wire having to be spliced numerous times. It is estimated by one
source that the average U.S. subscriber line has 22 splices from the central office to the premises.
Each splice adds a little attenuation to the signal and the opportunity for cross-talk. Cross-talk is
the result of a signal coupling from its assigned or desired line into an adjacent line. Cross-talk is
unwanted interference in the adjacent wire and reduces the signal level in the desired line.
Unwanted signal coupling increases as a function of increasing frequency and increasing line
length. Ergo, at some intersection of frequency and line length the desired signal becomes
undistinguishable from the noise. Perhaps the intersection nearest your home? Anyway, for the
curious, the telephone folks found out long ago that twisting the two copper wires as they
traveled from the central office to your premises greatly reduced the amount of cross-talk
between pairs. So, they could bundle more wires together with less cross-talk. Now you know
why the world is wired with twisted pairs, and you just thought the world was twisted because it
was "wired."
Signal Category

Frequency (Hertz) Distance (feet)

DS-1/T1

1.544 Mbps

18,000

E1

2.048 Mbps

16,000

DS-2

6.312 Mbps

12,000

E2

9.448 Mbps

9,000

1/4 STS-1

12.960 Mbps

4,000

1/2 STS-1

25.920 Mbps

3,000

STS-1

51.840 Mbps

1,000

Table 9-1. Subscriber line practical frequency limitations

With splices, cable bundles, intervening distribution equipment, the length of the twisted pair, and
the resultant attenuation of the signal as it travels down the copper line, there is a maximum
practical distance limitation that the ADSL signal can be transmitted in one direction only. The

distance specified in Table 9-1 is the maximum distance an ADSL signal can be transmitted,
assuming 24 gauge twisted pair copper lines. Attenuation of the signal is caused by the physical
resistance of the wire to the moving electrons that comprise the electronic signal. The table is
valid only when there is a maximum of two subscriber telephones connected to the line at the
customer premises. If there are more telephones connected, the distance goes down. Also,
connecting other customer premises equipment, such as recording equipment, will reduce the
effective distance the ADSL signal can be transmitted. However, the ADSL modem will act as the
interface between the local office and the CPE, providing the isolation between the local office
and the CPE. The local office equipment will see only the ADSL modem as the line load.
Therefore, the ADSL modem makes it possible to connect televisions, computers, etc., to the line
via the modem connection.
Subscriber loop configurations vary all around the globe. In some places, 18,000 feet from the
local exchange to the subscriber premises may cover every subscriber in the loop. But in other
places, 18,000 feet may not. In the United States, it is estimated that 18,000 feet will cover about
80 percent of all subscribers. The 20 percent of subscribers outside the 18,000 feet zone use
loading coils that prevent the use of DSL technology on those lines. The coils will have to go.
Installing remote access nodes, thereby creating an ADSL distribution area, allows the removal of
the loading coils, virtually shrinking the distance from subscriber to local exchange to an average
of 6,000 feet. The practice of installing distribution areas is an ongoing event that telephone
companies implemented several years ago to increase the capacity of the existing central/local
office. Remote access nodes are connected to T1/E1 lines/fiber passing HDSL data. Distribution
areas cover an average of 1,500 suburban subscribers and an average of 3,000 urban subscribers.
The number of subscribers in a distribution area is greatly dependent upon the subscribed data
rates, as each distribution area has a maximum data rate ceiling in its service area that is, of
course, directly related to transport capacity.
Service

Required
Bandwidth

VOD (MPEG)

1.5 Mbps

Internet/multime 1.5 Mbps

dia
live digital TV

6.0 mbps

HDTV

20 Mbps

Table 9-2. Consumer services required bandwidth

Table 9-2 details the bandwidth requirements of several popular consumer services. VOD
requires a minimum 1.5 Mbps bandwidth, making 56.6 Kbps analog modems too lightweight to
supply VOD services. But ADSL modems, if you are within 18,000 feet of a distribution area,
have sufficient bandwidth muscle to provide VOD services.

An ADSL modem combines the total multiplexed data stream that includes the service data, such
as a request to view a particular VOD channel, and maintenance channels from the service
provider to the user into blocks of data; appends error correction codes, called cyclic redundancy
check (CRC), to each block of data; and ships the data block out to the user. The users modem
receives the data block, decodes the error correction codes, then repairs the data block, if
necessary, within limitations. The modems do have an option to allow the construction of
superblocks of data by interleaving data within portions of the blocks. The use of superblocks
allows the modem to correct specific combinations of bit errors within certain spans of bits which
helps the quality of data and video signals.

DSL Types
A review of some basic terminology is appropriate at this point. Downstream means the
transmission path from the network, or local access, to the subscriber. Upstream means the
transmission path from the subscriber to the network. Duplex refers to the ability to carry on a
simultaneous two-way conversation over a single transmission path. Simplex refers to the ability
to carry on a one-way conversation over a transmission path. The DSL designations in the
following discussions are in accordance with ANSI and ETSI naming conventions.
Transmission
Technology

Maximum Data
Rate

Applications

analog modem

56.6 Kbps

low speed data communications

DSL

160 Kbps

ISDN voice and data communications

HDSL

1.544-2.048
Mbps

T1/E1, feeder plant, LAN/WAN access

SDSL

1.544-2.048
Mbps

T1/E1, feeder plant, LAN/WAN/premises access

ADSL

1.5-10 Mbps

Internet, VOD, Remote LAN, multimedia

VDSL

13-52 Mbps

Internet, VOD, Remote LAN, multimedia, HDTV

Table 9-3. Copper twisted pair transmission technologies

Compare any of the DSL modem technologies with the voice frequency analog modem. From
200 bps in the early days to 56.6 Kbps now, voice frequency analog modems have evolved over
the years in an almost herculean effort in the race to keep pace with the demand for faster data
communications. Voice frequency analog modems, while undoubtedly occupying a place in our
communications networks for some years to come, and a hallowed place in the Communications
Hall of Fame, located somewhere near Dallas, Texas (in my closet), have lost the race to DSL
modems.

The maximum bit rate is selectable by the user at the time service is ordered. This is a "you pay
for what you get" conceptthe more bandwidth you order, the more it will cost. Bit rates are
selectable in increments with 32 Kbps the most common increment. In actual practice, residential
users will probably not get an opportunity to select bit rates. Each type of service such as video or
Internet will come with some particular bit rate that is tailored to the needs of the service. Yet
those users who are interested in data transmission/reception service, typically business
customers such as remote LAN access to ATM, will probably be able to select the bit rate they
wish to transmit/receive.
DSL

DSL is, of course, the acronym for Digital Subscriber Line. This is the DSL that is a specific
transmission protocol and speed. This DSL really should be called SLMDSL for Slow Like (cold)
Molasses DSL. DSL transmission rate is 160 Kbps duplex. DSL is used for Basic Rate ISDN and
consists of two B channels of 64 Kbps each and one D channel of 16 Kbps. The two B channels
are used for voice and data traffic. The D channel is used for management (signaling and routing)
purposes.
Notice the transmission rate of 160 KHz duplex. Duplex mode is necessary since the B channels
are utilized for voice traffic. Imagine having a one-sided conversation (simplex mode) with the....
(fill in the blank). Standard ANSI T1.601 (ITU I.431) implementations employ echo cancellation
to separate the transmit and receive signals at each end of the transmission path. Without echo
cancellation, the users think they are in the bottom of the same well when talking over DSL.
Users should make sure DSL modems are ANSI compliant, else there will be problems of the
unwanted kind.
Since DSL bandwidth is 0 to 80 KHz (U.S.), analog POTS cannot be simultaneously provisioned
on the same circuit. But, a DSL modem can be connected to each end of a single POTS line to
effectively give two digital voice lines. The gain in voice channels is possible by using each B
channel as a separate voice line. The installation and use of two DSL modems to double the
number of voice channels is called an (au) pair gain application.

Figure 9-4. Maximum ADSL transmission distance

ADSL

Figure 9-5. ADSL streams

Various consumer services drive the demand for ADSL technology. We would love to have one
box in our homes, and businesses, that can connect to the Internet and video service providers
while receiving and sending faxes, providing videotelephony and video conferencing, and still
give us the same POTS we are accustomed to enjoying. Will the day arrive when all home
communications will be provided for out of a single box? Probably not. The ADSL modem will
be a termination box, similar to a CATV set-top box, with the various services transported from
the box through the home over appropriate cabling. However, some enterprising companies will
try to sell us a one-box-fits-all product with integrated TV, computer, and telephone functionality.
I can only envision an ugly fit.
ADSL is characterized by speeds of 1.5 MHz to 10 MHz downstream, 16 to 640 Kbps upstream,
and 1.5 kilometer maximum transmission distance. See Figure 9-4.
Since most of the data traffic is one way, from service provider to user, such as an Internet
connection, and there is much less traffic volume from the user to the service provider, the term
asymmetric was attached to DSL to distinguish the primarily one-way ADSL medium from the
full duplex, two-way DSL traffic medium. See Figure 9-5.
ADSL is used for Internet access, VOD, remote LAN access, and interactive multimedia. In the
future it will be used with circuit switched networks, packet switched networks, such as IP
routers, and ATM switched networks.
ADSL modems incorporate error correction algorithms, called CRCs, intended to reduce the
effect of impulse (white) noise on digitally compressed video signals. Impulse noise is noise
generated in a random fashion such as the noise heard on AM radio caused by lightning. The
error correction algorithm introduces approximately 20 microseconds (msec) of signal delay to
the bit stream. The artificially introduced 20 msec delay for digitally compressed video signals is
anathema to LAN- and Internet-based data communications applications. ADSL modems must
know whether to apply error correction or not to any particular signal.
ADSL modems incorporate adaptive line algorithms that enable the modem to deal with noisy
lines in an "intelligent" way. Let us meet the DSL family. One wifey, one hubby, no pets, and no
kids. Suppose the hubby is watching television, say the Super Bowl, one night when the wifey
decides to wash her hair. All is well until she plugs the 10,000 watt hair dryer with a 1.8 MHz AC
motor into the bathroom outlet. Now, on cable or broadcast TV the picture will get considerable
interference in it. The picture will be buggered up enough that the hubby will probably shout
some amount of unmentionable remarks. But our DSL family remains one happy bunch of loveydovey birds. The DSL modem actually detects the frequency of interference, which is 1.8 MHz,

and blocks out that frequency and so many above and below for use. The DSL hubby never
notices any picture degradation. Smart machine, eh? The preceding discussion is for educational
purposes only. I have not seen any 1.8 MHz, 10,000 watt hair dryers... yet. So, adaptive line
algorithm is a cryptic way of saying the machine does not use the portion of the bandwidth that
has a high noise level for the moment.
Connected to the ADSL modem on the customer premises side will be stereos, televisions,
computers, fax machines, telephones, and virtually any other audio, musical, video multimedia
gadget, and data contraption you can think of. Simultaneous reception of various services and
communications upstream and downstream are features of ADSL modem technology that analog
modems can only dream of. Want to overdose on technology? Hurry! Hurry! Hurry! Get your
ADSL modem connected to all your electronic goodies now!
Future use of ADSL technology is limited only by the demands of a communications-hungry
world. Circuit switched, packet switched, cell switched, operating in duplex mode, able to receive
and transmit simultaneously, monitoring and evaluating user input such as service specific
selections, ADSL modems must be much more intelligent than just plain vanilla flavor voice
frequency analog modems. ADSL modems must be smart modems, able to distinguish between
and among all the bewildering choices of services and service options available.
HDSL

High Data Rate Digital Subscriber Line (HDSL) is characterized by two transmission speeds,
1.544 Mbps and 2.048 Mbps, with a bandwidth dependent upon the particular information
encoding scheme used. A transmission speed of 1.544 Mbps (T1) duplex requires two twisted
pair lines from the source to the destination. A transmission speed of 2.048 Mbps (E1) duplex
requires three twisted pair lines from source to destination.
HDSL is used for interconnecting T1/E1 service, interoffice trunks, LAN/WAN access, server
access, cellular antenna stations, digital loop carrier systems, and interexchange POPS. HDSL
transmits 1.544/2.048 Mbps in bandwidths ranging from 80 KHz to 240 KHz, depending upon
the data encoding technique. The consumer will not really see HDSL as it will be used in local
exchange carrier (LEC) trunking.
VDSL

Very High Rate DSL (VDSL) is sometimes called BDSL, VADSL, or even ADSL. In June 1995,
the ADSL Forum T1E1.4 group chose VDSL as the official moniker. Confusing, eh? The key to
recognizing any particular flavor of DSL is the transmission distance and the rate of transmission.
In this case, VDSL is good for a maximum of 1,500 meters.
Service Downstream Data
Rate (Mbps)

Distanc
e
(meter
s)

1/4 STS- 12.96

1

1,500

1/2 STS- 25.92

1

1,000

STS-1

300

51.84

Table 9-4. VDSL modem characteristics

VDSL has an effective range of 1.5 kilometer radius from the local exchange. The transmission
rate for duplex operation is 13 to 52 Mbps downstream and 1.5 to 2.3 Mbps upstream. VDSL is
used for Internet access, VOD, remote LAN access, interactive multimedia, and HDTV. Future
VDSL lines may have equal upstream/downstream rates, albeit on shorter lines.
VDSL initially was known as VADSL with the A designating an asymmetrical flavor of VDSL.
The data rates are higher than ADSL but the modems are limited to shorter distances.
VDSL is designed for ATM networks, eliminating the channelization and packet handling
requirements ADSL must cope with in dealing with circuit switched and packet switched
networks. Since VDSL is limited to shorter lines than ADSL, fewer transmission constraints are
involved, resulting in simpler transceiver algorithms for VDSL modems. Since VDSL modems
can interface with passive line terminations, more than one VDSL modem can be connected to a
line. So, any user can connect three VDSL modems to one twisted pair and receive signals per the
frequencies listed in Table 9-4.
VDSL operates in conjunction with POTs and ISDN with the differing signals separated at the
opportune moment with passive filtering. Passive filtering is like a sand filter with one size of
holes in the filter material, perhaps a wire mesh. Everything smaller than the hole gets through.

Figure 9-6. DMT spectrum

Four data/information coding methods are proposed for VDSL: Carrierless Amplitude/Phase
Modulation (CAP), Discrete Multitone (DMT), Discrete Wavelet Multitone (DWMT), and
Simple Line Code (SLC).

SDSL

Single Line Digital Subscriber Line (SDSL) supports POTS and DS-1 simultaneously. The
transmission rate is either 1.544 Mbps or 2.048 Mbps duplex over a single copper line. SDSL is
used for DS-1 service, feeder plant, LAN/WAN access, server access, and premises access.
Symmetric, meaning as much data flows one way as the other, services such as remote LAN
users and servers are good candidates for SDSL.

Figure 9-7. ADSL FDM channelization bandwidth utilization

Discrete Multi-Tone
Discrete Multi-Tone (DMT) coding divides the available bandwidth into channels. Some of the
channels are used for upstream data transport and some of the channels are used for downstream
transport. One example reserves the frequency spectrum from 0 to 4 KHz for the POTS channel
and divides the spectrum from 26 KHz to 1.1 MHz into 246 distinct and separate channels. Each
channel carries its information as a discrete analog signal. DMT can dynamically adapt to the
changing line conditions to deliver the maximum throughput per channel with the least amount of
signal degradation. DMT was chosen over Carrierless Amplitude/Phase Modulation (CAP) and
Quadrature Amplitude/ Phase Modulation (QAM) because it is more robust over long distances.
Robust means it could go the distance and still deliver the goods.

Figure 9-8. ADSL echo cancellation bandwidth utilization

Two different techniques to divide the bandwidth are used. Frequency division multiplexing and
echo cancellation are used depending upon the modem manufacturer and the users needs. The
two techniques are discussed below.
FDM

FDM channelization consists of separate upstream and downstream paths. TDM techniques are
used to divide the upstream path and the downstream path into multiple channels. The
downstream path is multiplexed into both high-speed and low-speed channels while the upstream
path is just multiplexed into slow-speed channels. With FDM channelization, there is a guard
band between POTS and the upstream channels and another guard band between the upstream
channels and the downstream channels. The guard bands are necessary to ensure that the data
stream from each path does not interfere with the data in the other paths.

Figure 9-9. Signal adaptation

Echo Cancellation

Echo cancellation utilizes the available bandwidth between the upstream and downstream paths,
resulting in an overlap of paths. Echo cancellation can accomplish this feat of electronic dexterity
by the well-known technique of local echo cancellation. Local echo cancellation is used in the
V.32 and V.34 analog modem technology.

High-Level Online Adaptation

High-level online adaptation is also called bitswapping in some literature circles. Bitswapping is
the technique of determining which channels are usable due to line noise conditions and using
only those channels for communications.
The ADSL modem monitors the line and adapts its bit rate to the line conditions to provide the
most error free connection possible. Here is how signal adaptation occurs. The first diagram in
Figure 9-9 shows individual channels with the same amplitude level. The level of each channel is
sufficient for the equipment to recognize the signals as valid ADSL signals. The second diagram
in the figure shows a typical burst of noise on the line. As long as the desired signal is sufficiently
more (10 dB or more) than the noise signal, the equipment can recognize the desired signal.
When the noise signal rises higher and the ratio between the noise signal and desired signal is
sufficiently low (less than 10 dB), the desired signal becomes indistinguishable from the noise
signal. The third diagram shows the desired signal level after combining with the noise signal.
For those channels dipping down in the low spot, the signal-to-noise ratio is too low for
acceptable use. So, the equipment will recognize those channels as unusable and will not transmit
any data on the channels. The ADSL modem will monitor the line periodically and mark the
unusable channels as necessary.

Figure 9-10. ADSL network termination

The typical real-time adaptation rate is 32 Kbps. Just pray your modem doesnt decide to adapt in
the middle of the TV program you have waited all year to watch. Actually, to maintain a
minimum Quality of Service level acceptable to consumers, sufficient bandwidth margin is

engineered into the system to ensure that under average conditions, your viewing pleasure will
not be interrupted nor will the picture be degraded as the modem adapts to the usually changing
conditions of the line. If your viewing conditions fall outside the spectrum defined as average,
then you may not be a good candidate to receive ADSL service. By the way, just how often does
a UFO pass near your home?

Figure 9-11. ADSL modem

Application Architecture
The ADSL CPE set-top box is called the ADSL Network Termination (ANT). Figure 9-10 shows
a simplified diagram of an ANT. The ADSL Line Interface Module transmits signals to and
receives signals from the ADSL Access Adapter located at the CO. The ADSL Line Interface
Module conditions (adjusts the level and ensures proper timing intervals) the ADSL signals and
converts the ATM cell stream to/from the ADSL signal. Also, the ANT includes the POTS
splitter/combiner.
The ATM bus, operating at 10 Mbps, multiplexes/demultiplexes the cells received from or sent to
the individual service modules. The output of the service modules is the native (original)
electrical signal necessary for the service to work properly. So, the output of the LAN service
module is Ethernet packet data and it is sent to/from the business computer. The output of the
video service module to the television is the analog video signal and the output is ATM video
cells to the ATM bus.
The video, audio, Internet, and LAN boxes shown in Figure 9-10 comprise the ADSL modem
portion of the ANT. The functions of the ADSL modem are shown in Figure 9-11. The direction
of the arrows indicates this figure depicts a CPE modem. A CO modem is the reverse image of
this figure.
Lets follow a signal that is sent from the CO to the CPE. The downstream signal (from CO to
customer premises) comes into the receiver where the signal is converted from a high frequency
analog ADSL signal to a digital signal. The receiver applies the digital bit stream to the channel
separator. The channel separator separates the individual channel signals into their respective bit

streams. The native digital signals are sent out to their respective terminal equipment (computer).
The non-native digital signals (originally analog) are applied to the D/A converter which,
strangely enough, converts each digital bit stream into its native analog signal. The output of the
D/A converter is an analog reproduction of the original transmitted signal. The analog signals are
routed to the appropriate outputs with the analog telephone signal passed through the POTS
splitter.
The transmit path from CPE to CO is as follows. The analog telephone signal is sent to the A/D
converter through the POTS splitter. Also, video and audio signals are sent to the A/D converter.
The A/D converter, strangely enough, converts the native analog signals into digital bit streams.
The digital bit streams are combined with the native digital bit streams from the native digital
sources (computers) and are applied to the multiplexer. The multiplexed digital signal is now sent
to the transmitter. The transmitter converts the multiplexed digital bit stream into the analog
ADSL signal and sends it on its way to the CO.

Figure 9-12. ADSL access adapter

System Architecture
The central office includes ADSL access adapters for interfacing the ATM network to the ADSL
modems. Figure 9-12 shows a simplified diagram of a typical access adapter. The number of
ADSL line termination boards varies from CO to CO depending upon the traffic patterns of the
service area.
The ADSL access adapter shown in Figure 9-12 consists of a network interface, a 155 Mbps ATM
bus, four ADSL line termination boards, and POTS access. Data is received from and transmitted
to the ATM network through the network interface. The network interface performs signal
conditioning (signal level) and terminates the line, such as a fiber termination. The ATM bus
multiplexes the transmit data going out to the ATM network and demultiplexes data received

from the ATM network. The ATM bus multiplexes/demultiplexes the ATM VC/VPs, which can be
various bit rates. This feature allows the ADSL line termination boards to operate at different bit
rates, if desired. The ADSL line termination boards support a wide range of access speeds from
1.5 Mbps all the way up to the ATM bus speed of 155 Mbps.
The ADSL line termination boards terminate four ADSL lines. The ATM bus side of the boards
receives/transmits ATM cells from/to the ATM bus. The CPE side of the boards
transmits/receives ADSL traffic to/ from the CPE. The POTS splitter at the CO is physically
located on the ADSL line termination board. At this point the narrowband voice signal is
combined with (going to the CPE) or split from (coming from the CPE) the ADSL signal.
Local telephone companies install local access nodes remotely from the CO to reduce the length
of the transmission path from CO to customer premises. The area serviced by the remote nodes is
called a distribution area (DA) and is intended to service an average of 1,500 customers.
However, in urban areas the number of customer premises could double to 3,000. Some remote
access nodes are fed with DS-1 (T1) lines using HDSL. The particular system architecture
required for any given site is dependent upon the distance from the CO and the type of services
desired. All things must be considered before a site can be connected to DSL technology and
have an acceptable level of performance. Not all the capabilities of DSL technology are available
to everyone. However, the service of most interest is available to almost everyone, video-ondemand.

POTS Splitter
A CPE filter is necessary to split the narrowband voice channel (POTS channel) from the
broadband data channel. The filter can be either active or passive. An active filter uses integrated
circuits and the passive filter uses reactive components (inductors, capacitors) to perform the
filtering function. Both types of filters are designed so if they fail, the user will always receive
POTs and not broadband data (video/ Internet) service. The filter will most likely be physically
located in the set-top box.
A POTS splitter is really a combiner/splitter depending upon the direction the POTS signal is
traveling. A POTS signal that is added to an ADSL signal is really combined with the ADSL
signal. A POTS signal that is removed from an ADSL signal is really split from it. So, when the
literature refers to a "POTs splitter," we are really referring to a "combiner/splitter."

Transport Capacity
Frequency

Wire Size

Distance

1.544 Mbps

24 AWG

18,000 ft

2.048 Mbps

24 AWG

16,000 ft

2.048 Mbps

26 AWG

15,000 ft

6.312 Mbps

24 AWG

12,000 ft

6.312 Mbps

26 AWG

9,000 ft

8.448 Mbps

24 AWG

9,000 ft

12.960 Mbps

24 AWG

4,500 ft

25.920 Mbps

24 AWG

3,000 ft

51.84 Mbps

24 AWG

1,000 ft

Table 9-5. Data rates

The data rate is dependent upon the wire size and distance from the CO to the customer premises.
Table 9-5 compares multiple frequencies and the maximum distance the signal can be
transmitted, assuming a near perfect transmission line. Evident from the table is the relationship
between wire size and distance for any given frequency. As the wire size decreases, the maximum
distance decreases.

Channel

Frequency

1x1.536 Mbps 1.536 Mbps

2x1.536 Mbps 3.072 Mbps
3x1.536 Mbps 4.608 Mbps
4x1.536 Mbps 6.144 Mbps
1x2.046 Mbps 2.048 Mbps
2x2.046 Mbps 4.096 Mbps
Table 9-6. Bearer channels

Figure 9-14. ATM over ADSL

Various ADSL data rates, called bearer channels, are available for service providers to utilize for
service offerings. See Table 9-6. Perhaps the consumer can choose the quality of picture and only
pay for the bandwidth, and quality of picture, desired. Data rates depend upon several factors
including the length of the transmission path, the size of the wire, even the age of the wire, and
the number of taps and splices between the CO and customer premises. Table 9-5 lists some of

the data rate choices available to service providers and the maximum length of the transmission
path assuming near perfect conditions.

Framing
The ADSL channels are synchronized to the 4 KHz ADSL DMT frequency (also called symbol
rate). The data and service information passing over each channel has a specific format called
framing. ADSL first transmits service information, then a large number of individual data frames
before finally transmitting a sync frame. The modem combines all the data frames and service
information frames into one distinct data unit called a superframe. Figure 9-13 shows the
arrangement of the individual frames within the superframe.
Figure 9-13.ADSL superframe
ADSL framing consists of 68 data frames combined with an 8-bit CRC frame and a 24-bit
Operations Administration and Management (OAM) frame to produce the ADSL superframe.
Each data frame period is 250 microsec, yielding a 4 Kbaud rate for that particular data frame.
The sync frame is added on the end of the superframe to ensure the receiving circuitry is properly
timed to the signal.

ATM Over ADSL

ATM over ADSL is capable of delivering every conceivable video product to the home without
any changes to the current customer premises (my house) infrastructure (wiring, equipment, etc.)
other than connecting the incoming telephone line (the common twisted pair copper line) to an
ADSL modem. Just plug it in like it was another telephone, only plug the telephone into it.
Getting a video signal from a video service provider to your home over ATM is not very difficult.
Figure 9-14 shows a simplified diagram of the transmission path from the video service provider
to a home. The video service provider analog video signal is digitized, then cellutized, that is,
assembled into ATM cells. The ATM converter shown in Figure 9-14 is the digitizing and
cellutizing function. In the real world, the conversion of an analog video signal to ATM cells
could be performed by several pieces of equipment. The ATM converter sends the cells out to the
service providers ATM network source node for transport over the ATM network from the source
location to the destination location. This portion of the trip is represented by the "cloud." Ever
wonder why they use a cloud to represent the service providers network? Ever seen "inside" a
"cloud"?
At the destination location in the ATM network cloud, the ATM cell stream is converted to an
ADSL signal suitable for transmission over the humble pair to your home, represented by the
CPE. The ADSL modem converts the signal to first an ATM cell stream, then to the original video
signal. Thats all there is to it...more or less. The upstream channel can be used for service
requests, such as requesting access to a particular channel.

Figure 9-14 shows a single home receiving video from a single source. From an enterprise
perspective, the really beautiful aspect of ATM over ADSL is the ability to provide video
products to the whole country from one single location.
A video headend located in Chicago can send video products (most people may not want to see
what comes out of Chicago, though) all across the nation, to every household. All the homeowner
needs is an ATM-compatible ADSL modem. And they are getting cheaper by the month. Also, the
number of channels of service is not limited as the CATV industry is. The cable industry is band
(channel) limited due to the characteristics of the coaxial cable transport infrastructure. In the
ATM over ADSL approach, only one channel is transmitted but the user quickly and easily orders
a new channel over the upstream data channel and voila...just like changing channels on the old
set. But the old set remains on just one channel, say Channel 2. The video service provider
receives the change request from the viewer and immediately switches the video source from one
video signal to another. Guess what? Channels become limited by the number of channels that
can be represented by 16 (or 32) binary digits and the number of channels service providers can
supply. Does anyone have time to surf 65,536 channels?
From a consumer perspective, I can receive video products from any video products supplier
(remember the national/international ATM networks), regardless of their physical location. Okay,
so a one-hour show originating in Copenhagen and viewed in Dallas might cost the viewer $25,
but have you ever seen some of the shows originating in Copenhagen? Lets see, 65,536 possible
channels to view, one life to live. Where is a good genome project when you need one?

Hurdles
Current applications for ADSL modems are limited to remote access to enterprise networks from
the few locations where telephone companies have installed ATM/ADSL switch gear. RBOCs
and others have ordered new CO equipment to enable ADSL technology. Be patient...okay, weve
waited long enough. Wheres the channel selector?
The current price of ADSL modems puts them out of reach of the average homeowner. However,
if you are considering installing ISDN, perhaps you should consider ADSL as an economical
alternative. Even with a $1,000 price tag, the ADSL modem is competitive with ISDN installation
(about $1,500) and the monthly charge for ADSL is the price of a copper wire telephone line.
Internet and remote LAN access via ADSL is possible today, if you have an ADSL modem.
No one is offering any video related services for ADSL transmission, yet. Since there are few, if
any, ADSL modems in the home, there is no market for ADSL-based video products, yet. This
market will bud in 1998 and blossom in 1999. By 2005, we will wonder what ever became of
cable TV.

Market Strategies
ATM over ADSL allows remote access to server-based networks from the home economically.
No, you dont have to pay for an ISDN line (fiber to the curb) to have Ethernet speed on the home
computer. Mobile workers, telecommuters, and branch offices have easy access to the enterprise

networks using dial-up servers. When laptop manufacturers begin installing ADSL boards in
laptops, workers will have dial-up access at Ethernet speed anywhere they can find a phone line
to connect to.
Audio, video, Internet, multimedia, LANs, video conferencing, video telephony, and the list goes
on and on and on and.... Now is the time for all good entrepreneurs to come to the aid of the
consumer. As citizens of this great and beautiful country we demand (nay, it is our birthrightit
must be in the Constitution somewhere, look harder) to be fattened on a feast of data
communication products that is only limited by the imagination of the creative minds, and the
larcenous souls, among us, and the depth of our own desire to be entertained by a visual and
audio extravaganza.

Crystal Balls
ADSL modems have been successfully tested in numerous trials by many telephone companies in
North America and Europe. Over 30 telephone companies are known to have conducted
successful technology trials. Various telephone companies are now engaged in market trials
involving Internet and video services to the customer premises. Over 10,000 ADSL modems are
now connected to the telephone networks around the world. In the spring of 1997, four RBOCs
announced the purchase of $300,000,000 worth of ADSL modems from Alcatel (France).
Semiconductor companies, including Motorola, offer DSL chipsets that consist of off-the-shelf
components, programmable digital signal processors, and application specific integrated circuits
(ASIC). There is a concerted effort by semiconductor manufacturers to compress the DSL
technology into a single chip. Industry analysts believe the cost of ADSL modems will approach
a $35 ceiling in large quantity buys, as the technology begins to pervade the networks around the
world. Current ADSL modem prices may range from $200 up, depending upon the quantity
ordered.
Service providers will keep POTS and N-ISDN services in place as much as possible and
introduce fiber as far as possible in the loop when the opportunity arises. Improving the existing
copper network may be necessary in those cases where there is feeder congestion, and
rehabilitation of poor quality local loops may be necessary. However, service providers will use
ATM and ADSL to enable broadband services.
Europe and Japan have mandated ATM as the preferred long-haul transmission technology.
Fortune 500 companies in this country will therefore utilize ATM to connect their far-flung
entities into seamless global networks, thereby paving the road for the rest of the pack. While
Frame Relay and Gigabit Ethernet will live on for a while, their time is limited. As the Fortune
500 connect to ATM, the price per bit will fall and smaller companies will then migrate to ATM.
Individual vendor ATM/ADSL products will soon offer an array of data speeds both upstream and
downstream. Once a subscriber has determined his or her data communications needs, a visit with
an equipment supplier will quickly narrow the plethora of choices available.

2000 is the year of the ADSL modem. CO switch gear is in place, ADSL modem prices are within
reach of the average homeowner, and service providers are offering interesting video products.
Say goodbye to CATV.

Appendix A
A Chronological Perspective of Communications Systems
prehistoric
1450
Late 18th
1839
1844
1845
1858
1860
1875
1876
1877
1881
1888
1891
1891
1896
1901
1907
1915
1918
1920s
1928
1929
1944
1946
1947
1949
1951
1951

Someone beats on a hollow tree trunk.

Gutenberg printing press invented.
Claude Chappe (France) organizes optical telegraph system.
Wheatstone (England) develops first electrical operating telegraph line.
Samuel Morse sends first U.S. public telegraph message.
Samuel Morse (United States) develops practical information encoding method
(Morse code) especially useful for telegraphic transmission.
First trans-Atlantic telegraph transmission by cable.
Pony Express rides the plains from St. Joseph, Missouri, to Sacramento, California.
Alexander Graham Bell invents the telephone.
Telephone patent issued to Bell.
First telephone in private home.
First long-distance line, from Boston to Providence, Rhode Island.
Heinrich Hertz demonstrates radio wave generation.
Almon B. Strowger patents step-by-step mechanical switch.
First undersea telephone cable, English Channel (France to England).
Marconi receives first radio patent.
First trans-Atlantic transmission and reception of radio waves
Lee De Forest invents three-element electron (vacuum) tube.
First United States transcontinental telephone transmission, by land line.
Bell System begins installing Strowger switches.
Radio amateurs demonstrate the usefulness of high frequency radio transmission for
long-distance communications.
Admiral Byrd takes an Art Collins (of Collins Radio fame) HF radio to Antarctica and
successfully communicates with the United States.
Coaxial cable invented.
First fully automatic calculator invented (MARK I - Harvard University).
First totally electronic digital computer invented (ENIAC - University of
Pennsylvania)
William Shockley, John Bardeen, and Walter Brattain discover the point contact
transistor (Bell Laboratories).
First electronic computer with an internal stored program (EDSAC - Cambridge,
England).
Direct, long-distance dialing in U.S. initiated.
First commercially delivered electronic digital computer delivered (UNIVAC -

1956
1959
1960
1963
1963
1965
1969
1970
1970
1971
1972
1973
1975
1975
1976
1976
1977
1977
1981
1984
1988
1988
1988
1990
1991

delivered to U.S. Bureau of the Census)

First trans-Atlantic telephone cable system is operational.
Patent for integrated circuit filed by J.S. Kilby
First electronic switch is tested.
First telephone conversation by geosynchronous orbit satellite transmission.
Touch-tone service is introduced.
First collect call service.
First light wave transmission.
Laser diode invented.
Packet switching introduced.
Intel introduces the first 4-bit microprocessor (Intel 4004).
Intel introduces the first 8-bit microprocessor (Intel 8008)
Intel introduces the 8080 chip.
MITS introduces the Altair personal computer kit in January issue of Popular
Electronics. Considered to be the first personal computer.
IBM introduces its first personal computer (Model 5100).
Apple Computer introduces the Apple I.
First digital electronic switch installed.
Apple Computer introduces the Apple II.
First fiber optic system installed.
On Wednesday, August 12, 1981, IBM debuted the IBM Personal Computer.
Divestiture of AT&T into AT&T (long distance, aka Ma Bell) and seven Regional Bell
Operating Companies (RBOCs or Baby Bells)
Frame Relay proposed by the ITU-T (Recommendation I.122).
First trans-Atlantic optical fiber cable.
ATM introduced.
2,000 km fiber optic link demonstration without using repeaters.
ADSL introduced.

Appendix B
Communications Standards Bodies
American National Standards Institute (ANSI)
ATM Forum
Canadian Standards Association (CSA)
Consulative Committee on International Telephony and
Telegraph (CCITT)
(now called the ITU-T)

European Telecommunications Standards Institute (ETSI)

Frame Relay Forum
International Telecommunication Union (ITU-T)
(formerly the CCITT)
Internet Engineering Task Force (IETF)

Appendix C
International System of Units Specifications
International System of Units (SI) specifies prefixes for certain multiples of 10. This practice is
followed in engineering and electronics as a shorthand way of representing large and small
numbers. The following chart provides the prefixes and their associated multiplier. As an
example, the number 1,200,000 could be represented as 1.2 Mega, or 1.2 M.
Prefix

Symbol

Multiplier

Giga

1,000,000,000

Mega

1,000,000

Kilo

1,000

milli

0.001

micro

0.000001

nano

0.000000001

pico

0.000000000001

Appendix D
ATM UNIs and Port Speeds
ATM UNIs and their associated port speeds are specified in the following chart.
ATM UNI

Port Speed

DS-0

0.064 Mbps

DS-1

1.544 Mbps

2xDS-1

3.088 Mbps

3xDS-1

4.50 Mbps

4xDS-1

6.0 Mbps

5xDS-1

7.50 Mbps

6xDS-1

9.0 Mbps

7xDS-1

10.50 Mbps

8xDS-1

12.0 Mbps

DS-2

6.312 Mbps

DS-3

44.736 Mbps

DS-4

274.176 Mbps

OC-3

155.520 Mbps

OC-12

622 Mbps

OC-48

2.4 Gbps

Glossary
This glossary is provided courtesy of the ATM Forum. Information about the ATM Forum is
available at http://www.atmforum.com.

16-CAP
Carrierless Amplitude/Phase Modulation with 16 constellation points: The modulation technique
used in the 51.84 Mb Mid-Range Physical Layer Specification for Category 3 Unshielded
Twisted-Pair (UTP-3).

64-CAP
Carrierless Amplitude/Phase Modulation with 64 constellation points.

AAL
ATM Adaptation Layer: The standards layer that allows multiple applications to have data
converted to and from the ATM cell. A protocol used that translates higher layer services into the
size and format of an ATM cell.

AAL Connection
Association established by the AAL between two or more next higher layer entities.

AAL-1
ATM Adaptation Layer Type 1: AAL functions in support of constant bit rate, time-dependent
traffic such as voice and video.

AAL-2

ATM Adaptation Layer Type 2: This AAL is still undefined by the International standards bodies.
It is a placeholder for variable bit rate video transmission.

AAL-3/4
ATM Adaptation Layer Type 3/4: AAL functions in support of variable bit rate, delay-tolerant
data traffic requiring some sequencing and/or error detection support. Originally two AAL types,
i.e., connection-oriented and connectionless, which have been combined.

AAL-5P
ATM Adaptation Layer Type 5: AAL functions in support of variable bit rate, delay-tolerant
connection-oriented data traffic requiring minimal sequencing or error detection support.

ABR
Available Bit Rate: ABR is an ATM layer service category for which the limiting ATM layer
transfer characteristics provided by the network may change subsequent to connection
establishment. A flow control mechanism is specified which supports several types of feedback to
control the source rate in response to changing ATM layer transfer characteristics. It is expected
that an end-system that adapts its traffic in accordance with the feedback will experience a low
cell loss ratio and obtain a fair share of the available bandwidth according to a network specific
allocation policy. Cell delay variation is not controlled in this service, although admitted cells are
not delayed unnecessarily.

ACM
Address Complete Message: A BISUP call control message from the receiving exchange to
sending exchange indicating the completion of address information.

ACR
Attenuation to Crosstalk Ratio: One of the factors that limits the distance a signal may be sent
through a given media. ACR is the ratio of the power of the received signal, attenuated by the
media, over the power of the NEXT crosstalk from the local transmitter, usually expressed in
decibels (db). To achieve a desired bit error rate, the received signal power must usually be
several times larger than the NEXT power or plus several db. Increasing a marginal ACR may
decrease the bit error rate.

ACR
Allowed Cell Rate: An ABR service parameter, ACR is the current rate in cells/sec at which a
source is allowed to send.

Address Prefix

A string of 0 or more bits up to a maximum of 152 bits that is the lead portion of one or more
ATM addresses.

Address Resolution
Address Resolution is the procedure by which a client associates a LAN destination with the
ATM address of another client or the BUS.

Adjacency
The relationship between two communicating neighboring peer nodes.

Administrative Domain
A collection of managed entities grouped for administrative reasons.

ADPCM
Adaptive Differential Pulse Code Modulation: A reduced bit rate variant of PCM audio encoding
(see also PCM). This algorithm encodes the difference between an actual audio sample amplitude
and a predicted amplitude and adapts the resolution based on recent differential values.

ADTF
ACR Decrease Time Factor: This is the time permitted between sending RM-cells before the rate
is decreased to ICR (Initial Cell Rate). The ADTF range is .01 to 10.23 sec. with granularity of 10
ms.

AFI
Authority and Format Identifier: This identifier is part of the network level address header.

Aggregation Token
A number assigned to an outside link by the border nodes at the ends of the outside link. The
same number is associated with all uplinks and induced uplinks associated with the outside link.
In the parent and all higher-level peer group, all uplinks with the same aggregation token are
aggregated.

AHFG
ATM-attached Host Functional Group: The group of functions performed by an ATM-attached
host that is participating in the MPOA service.

Ai

Signaling ID assigned by Exchange A.

AIM
ATM Inverse Multiplexer: A term discontinued because of conflict with an established product.
Refer to AIMUX.

AIMUX
ATM Inverse Multiplexing: A device that allows multiple T1 or E1 communications facilities to
be combined into a single broadband facility for the transmission of ATM cells.

AIR
Additive Increase Rate: An ABR service parameter, AIR controls the rate at which the cell
transmission rate increases. It is signaled as AIRF, where AIRF =AIR*Nrm/PCR.

AIRF
Additive Increase Rate Factor: Refer to AIR.

AIS
Alarm Indication Signal: An all ones signal sent down or up stream by a device when it detects an
error condition or receives an error condition or receives an error notification from another unit in
the transmission path.

Alternate Routing
A mechanism that supports the use of a new path after an attempt to set up a connection along a
previously selected path fails.

AMI
Alternate Mark Inversion: A line coding format used on T1 facilities that transmits ones by
alternate positive and negative pulses.

Ancestor Node
A logical group node that has a direct parent relationship to a given node (i.e., it is the parent of
that node, or the paren parent, ...).

ANI

Automatic Number Identification: A charge number parameter that is normally included in the
Initial Address Message to the succeeding carrier for billing purposes.

ANM
Answer Message: A BISUP call control message from the answer and that a through connection
should be completed in both directions.

ANSI
American National Standards Institute: A U.S. standards body.

API
Application Program Interface: API is a programmatic interface used for interprogram
communications or for interfacing between protocol layers.

API_connection
Native ATM Application Program Interface Connection: API_connection is a relationship
between an API_endpoint and other ATM devices that has the following characteristics:
1) Data communication may occur between the API_endpoint and the other ATM devices
comprising the API_connection.
2) Each API_connection may occur over a duration of time only once; the same set of
communicating ATM devices may form a new connection after a prior connection is released.
3) The API_connection may be presently active (able to transfer data), or merely anticipated for
the future.

APPN
Advanced Peer to Peer Network: IBM network architecture for building dynamic routing across
arbitrary network topologies. Intended as an eventual replacement for SNA, IBMs static routed,
hierarchical network architecture.

ARE
All Routes Explorer: A specific frame initiated by a source which is sent on all possible routes in
Source Route Bridging.

ARP
Address Resolution Protocol: The procedures and messages in a communications protocol which
determines which physical network address (MAC) corresponds to the IP address in the packet.

ASP
Abstract Service Primitive: An implementation-independent description of an interaction between
a service-user and a service-provider at a particular service boundary, as defined by Open
Systems Interconnection (OSI).

Assigned Cell
Cell that provides a service to an upper layer entity or ATM Layer Management entity (ATMMentity).

ATDM
Asynchronous Time Division Multiplexing: A multiplexing technique in which a transmission
capability is organized in a priori unassigned time slots. The time slots are assigned to cells upon
request of each application's instantaneous real need.

ATM
Asynchronous Transfer Mode: A transfer mode in which the information is organized into cells. It
is asynchronous in the sense that the recurrence of cells containing information from an
individual user is not necessarily periodic.

ATM Address
Defined in the UNI Specification as 3 formats, each having 20 bytes in length including country,
area and end-system identifiers.

ATM Layer Link

A section of an ATM Layer connection between two adjacent active ATM Layer entities (ATMentities).

ATM Link
A virtual path link (VPL) or a virtual channel link (VCL).

ATM Peer-to-Peer Connection

A virtual channel connection (VCC) or a virtual path connection (VPC).

ATM Traffic Descriptor

A generic list of traffic parameters that can be used to capture the intrinsic traffic characteristics
of a requested ATM connection.

ATM User-User Connection

An association established by the ATM Layer to support communication between two or more
ATM service users (i.e., between two or more next higher entities or between two or more ATMentities). The communications over an ATM Layer connection may be either bidirectional or
unidirectional. The same Virtual Channel Identifier (VCI) issued for both directions of a
connection at an interface.

ATS
Abstract Test Suite: A set of abstract test cases for testing a particular protocol. An "executable"
test suite may be derived from an abstract test suite.

Attenuation
The process of the reduction of the power of a signal as it passes through most media. Usually
proportional to distance, attenuation is sometimes the factor that limits the distance a signal may
be transmitted through a media before it can no longer be received.

B-ICI
B-ISDN Inter-Carrier Interface: An ATM Forum defined specification for the interface between
public ATM networks to support user services across multiple public carriers.

B-ICI SAAL
B-ICI Signaling ATM Adaptation Layer: A signaling layer that permits the transfer of connection
control signaling and ensures reliable delivery of the protocol message. The SAAL is divided into
a Service Specific part and a Common part (AAL5).

B-ISDN
Broadband ISDN: A high-speed network standard (above 1.544 Mbps) that evolved Narrowband
ISDN with existing and new services with voice, data and video in the same network.

B-LLI
Broadband Low Layer Information: This is a Q.2931 information element that identifies a layer 2
and a layer 3 protocol used by the application.

B-TE
Broadband Terminal Equipment: An equipment category for B-ISDN which includes terminal
adapters and terminals.

BBC
Broadband Bearer Capability: A bearer class field that is part of the initial address message.

BCD
Binary Coded Decimal: A form of coding of each octet within a cell where each bit has one of
two allowable states, 1 or 0.

BCOB
Broadband Connection Oriented Bearer: Information in the SETUP message that indicates the
type of service requested by the calling user.

BCOB-A
Bearer Class A: Indicated by ATM end user in SETUP message for connection-oriented, constant
bit rate service. The network may perform internetworking based on AAL information element
(IE).

BCOB-C
Bearer Class C: Indicated by ATM end user in SETUP message for connection-oriented, variable
bit rate service. The network may perform internetworking based on AAL information element
(IE).

BCOB-X
Bearer Class X: Indicated by ATM end user in SETUP message for ATM transport service where
AAL, traffic type and timing requirements are transparent to the network.

BECN
Backward Explicit Congestion Notification: A Resource Management (RM) cell type generated
by the network or the destination, indicating congestion or approaching congestion for traffic
flowing in the direction opposite that of the BECN cell.

BER
Bit Error Rate: A measure of transmission quality. It is generally shown as a negative exponent,
(e.g., 10-7 which means 1 out of 107 bits are in error or 1 out of 10,000,000 bits are in error).

BHLI

Broadband High Layer Information: This is a Q.2931 information element that identifies an
application (or session layer protocol of an application).

Bi
Signaling ID assigned by Exchange B.

BIP
Bit Interleaved Parity: A method used at the PHY layer to monitor the error performance of the
link. A check bit or word is sent in the link overhead covering the previous block or frame. Bit
errors in the payload will be detected and may be reported as maintenance information.

BIS
Border Intermediate System.

BISUP
Broadband ISDN Users Part: A SS7 protocol which defines the signaling messages to control
connections and services.

BN
Bridge Number: A locally administered bridge ID used in Source Route Bridging to uniquely
identify a route between two LANs.

BN
BECN Cell: A Resource Management (RM) cell type indicator. A Backwards Explicit Congestion
Notification (BECN) RM-cell may be generated by the network or the destination. To do so,
BN=1 is set, to indicate the cell is not source-generated, and DIR=1 to indicate the backward
flow. Source generated RM-cells are initialized with BN=0.

BOM
Beginning of Message: An indicator contained in the first cell of an ATM segmented packet.

Border Node
A logical node that is in a specified peer group, and has at least one link that crosses the peer
group boundary.

BPDU

Bridge Protocol Data Unit: A message type used by bridges to exchange management and control
information.

BPP
Bridge Port Pair (Source Routing Descriptor): Frame header information identifying a
bridge/LAN pair of a Source route segment.

Broadband
A service or system requiring transmission channels capable of supporting rates greater than the
Integrated Services Digital Network (ISDN) primary rate.

Broadband Access
An ISDN access capable of supporting one or more broadband services.

Broadcast
Data transmission to all addresses or functions.

BT
Burst Tolerance: BT applies to ATM connections supporting VBR services and is the limit
parameter of the GCRA.

Btag
Beginning Tag: A one octet field of the CPCS_PDU used in conjunction with the Etag octet to
form an association between the beginning of message and end of message.

BUS
Broadcast and Unknown Server: This server handles data sent by an LE Client to the broadcast
MAC address (`FFFFFFFFFFFF'), all multicast traffic, and initial unicast frames which are sent
by a LAN Emulation Client.

BW
Bandwidth: A numerical measurement of throughput of a system or network.

CAC
Connection Admission Control: Connection Admission Control is defined as the set of actions
taken by the network during the call set- up phase (or during call re-negotiation phase) in order to

determine whether a connection request can be accepted or should be rejected (or whether a
request for re-allocation can be accommodated).

Call
A call is an association between two or more users or between a user and a network entity that is
established by the use of network capabilities. This association may have zero or more
connections.

CAS
Channel Associated Signaling: A form of circuit state signaling in which the circuit state is
indicated by one or more bits of signaling status sent repetitively and associated with that specific
circuit.

CBDS
Connectionless Broadband Data Service: A connectionless service similar to Bellcore's SMDS
defined by European Telecommunications Standards Institute (ETSI).

CBR
Constant Bit Rate: An ATM service category which supports a constant or guaranteed rate to
transport services such as video or voice as well as circuit emulation which requires rigorous
timing control and performance parameters.

CCR
Current Cell Rate: The Current Cell Rate is an RM-cell field set by the source to its current ACR
when it generates a forward RM-cell. This field may be used to facilitate the calculation of ER,
and may not be changed by network elements. CCR is formatted as a rate.

CCS
Common Channel Signaling: A form signaling in which a group of circuits share a signaling
channel. Refer to SS7.

CD-ROM
Compact Disk-Read Only Memory: Used by a computer to store large amounts of data.
Commonly used for interactive video games.

CDF

Cutoff Decrease Factor: CDF controls the decrease in ACR (Allowed Cell Rate) associated with
CRM.

CDV
Cell Delay Variation: CDV is a component of cell transfer delay, induced by buffering and cell
scheduling. Peak-to-peak CDV is a QoS delay parameter associated with CBR and VBR services.
The peak-to- peak CDV is the ((1-a) quantile of the CTD) minus the fixed CTD that could be
experienced by any delivered cell on a connection during the entire connection holding time. The
parameter "a" is the probability of a cell arriving late. See CDVT.

CDVT
Cell Delay Variation Tolerance: ATM layer functions may alter the traffic characteristics of ATM
connections by introducing Cell Delay Variation. When cells from two or more ATM connections
are multiplexed, cells of a given ATM connection may be delayed while cells of another ATM
connection are being inserted at the output of the multiplexer. Similarly, some cells may be
delayed while physical layer overhead or OAM cells are inserted. Consequently, some
randomness may affect the inter-arrival time between consecutive cells of a connection as
monitored at the UNI. The upper bound on the "clumping" measure is the CDVT.

CE
Connection Endpoint: A terminator at one end of a layer connection within a SAP.

CEI
Connection Endpoint Identifier: Identifier of a CE that can be used to identify the connection at a
SAP.

Cell
A unit of transmission in ATM. A fixed-size frame consisting of a 5-octet header and a 48-octet
payload.

Cell Header
ATM Layer protocol control information.

Cells in Frames
Cells In Frames is a protocol established by the CIF Alliance which specifies how to transport
ATM protocol over Ethernet, Token Ring and other frame protocols. CIF uses software at the
workstation instead of a new hardware Network Interface Card to do QOS scheduling and ABR
flow control.

CER
Cell Error Ratio: The ratio of errored cells in a transmission in relation to the total cells sent in a
transmission. The measurement is taken over a time interval and is desirable to be measured on
an in-service circuit.

CES
Circuit Emulation Service: The ATM Forum circuit emulation service interoperability
specification specifies interoperability agreements for supporting Constant Bit Rate (CBR) traffic
over ATM networks that comply with the other ATM Forum interoperability agreements.
Specifically, this specification supports emulation of existing TDM circuits over ATM networks.

Child Node
A node at the next lower level of the hierarchy which is contained in the peer group represented
by the logical group node currently referenced. This could be a logical group node, or a physical
node.

Child Peer Group

A child peer group of a peer group is any one containing a child node of a logical group node in
that peer group. A child peer group of a logical group node is the one containing the child node of
that logical group node.

CI
Congestion Indicator: This is a field in a RM-cell, and is used to cause the source to decrease its
ACR. The source sets CI=0 when it sends an RM-cell. Setting CI=1 is typically how destinations
indicate that EFCI has been received on a previous data cell.

CIP
Carrier Identification Parameter: A 3 or 4 digit code in the initial address message identifying the
carrier to be used for the connection.

CIR
Committed Information Rate: CIR is the information transfer rate which a network offering
Frame Relay Services (FRS) is committed to transfer under normal conditions. The rate is
averaged over a minimum increment of time.

CL

Connectionless Service: A service which allows the transfer of information among service
subscribers without the need for end-to-end establishment procedures.

CLP
Cell Loss Priority: This bit in the ATM cell header indicates two levels of priority for ATM cells.
CLP=0 cells are higher priority than CLP=1 cells. CLP=1 cells may be discarded during periods
of congestion to preserve the CLR of CLP=0 cells.

CLR
Cell Loss Ratio: CLR is a negotiated QoS parameter and acceptable values are network specific.
The objective is to minimize CLR provided the end-system adapts the traffic to the changing
ATM layer transfer characteristics. The Cell Loss Ratio is defined for a connection as: Lost
Cells/Total Transmitted Cells. The CLR parameter is the value of CLR that the network agrees to
offer as an objective over the lifetime of the connection. It is expressed as an order of magnitude,
having a range of 10-1 to 10-15 and unspecified.

CMIP
Common Management Interface Protocol: An ITU-TSS standard for the message formats and
procedures used to exchange management information in order to operate, administer maintain
and provision a network.

CMR
Cell Misinsertion Rate: The ratio of cells received at an endpoint that were not originally
transmitted by the source end in relation to the total number of cells properly transmitted.

CNR
Complex Node Representation: A collection of nodal state parameters that provide detailed state
information associated with a logical node.

COD
Connection Oriented Data: Data requiring sequential delivery of its component PDUs to assure
correct functioning of its supported application, (e.g., voice or video).

COM
Continuation of Message: An indicator used by the ATM Adaptation Layer to indicate that a
particular ATM cell is a continuation of a higher layer information packet which has been
segmented.

Common Peer Group

The lowest level peer group in which a set of nodes is represented. A node is represented in a peer
group either directly or through one of its ancestors.

Communication Endpoint
An object associated with a set of attributes which are specified at the communication creation
time.

Configuration
The phase in which the LE Client discovers the LE Service.

Connection
An ATM connection consists of concatenation of ATM Layer links in order to provide an end-toend information transfer capability to access points.

Connection
In switched virtual connection (SVC) environments the LAN Emulation Management entities set
up connections between each other using UNI signaling.

Connectionless
Refers to ability of existing LANs to send data without previously establishing connections.

Control Connections
A Control VCC links the LEC to the LECS. Control VCCs also link the LEC to the LES and
carry LE_ARP traffic and control frames. The control VCCs never carry data frames.

Corresponding Entities
Peer entities with a lower layer connection among them.

CPCS
Common Part Convergence Sublayer: The portion of the convergence sublayer of an AAL that
remains the same regardless of the traffic type.

CPCS-SDU

Common Part Convergence Sublayer-Service Data Unit: Protocol data unit to be delivered to the
receiving AAL layer by the destination CP convergence sublayer.

CPE
Customer Premises Equipment: End user equipment that resides on the customer premise which
may not be owned by the local exchange carrier.

CPN
Calling Party Number: A parameter of the initial address message that identifies the calling
number and is sent to the destination carrier.

Crankback
A mechanism for partially releasing a connection setup in progress which has encountered a
failure. This mechanism allows PNNI to perform alternate routing.

CRC
Cyclic Redundancy Check: A mathematical algorithm that computes a numerical value based on
the bits in a block of data. This number is transmitted with the data and the receiver uses this
information and the same algorithm to insure the accurate delivery of data by comparing the
results of algorithm and the number received. If a mismatch occurs, an error in transmission is
presumed.

CRF
Cell Relay Function: This is the basic function that an ATM network performs in order to provide
a cell relay service to ATM end-stations.

CRF
Connection Related Function: A term used by Traffic Management to reference a point in a
network or a network element where per connection functions are occurring. This is the point
where policing at the VCC or VPC level may occur.

CRM
Missing RM-Cell Count: CRM limits the number of forward RM-cells which may be sent in the
absence of received backward RM cells.

CRM

Cell Rate Margin: This is a measure of the difference between the effective bandwidth allocation
and the allocation for sustainable rate in cells per second.

CRS
Cell Relay Service: A carrier service which supports the receipt and transmission of ATM cells
between end users in compliance with ATM standards and implementation specifications.

CS
Convergence Sublayer: The general procedures and functions that convert between ATM and
non-ATM formats. This describes the functions of the upper half of the AAL layer. This is also
used to describe the conversion functions between non-ATM protocols such as frame relay or
SMDS and ATM protocols above the AAL layer.

CSU
Channel Service Unit: An interface for digital leased lines which performs loopback testing and
line conditioning.

CT
Conformance Test: Testing to determine whether an implementation complies with the
specifications of a standard and exhibits the behaviors mandated by that standard.

CTD
Cell Transfer Delay: This is defined as the elapsed time between a cell exit event at the
measurement point 1 (e.g., at the source UNI) and the corresponding cell entry event at
measurement point 2 (e.g., the destination UNI) for a particular connection. The cell transfer
delay between two measurement points is the sum of the total inter-ATM node transmission delay
and the total ATM node processing delay.

DA
Destination Address: Information sent in the forward direction indicating the address of the called
station or customer.

DA
Destination MAC Address: A six octet value uniquely identifying an endpoint and which is sent
in IEEE LAN frame headers to indicate frame destination.

Data Connections

Data VCCs connect the LECs to each other and to the Broadcast and Unknown Server. These
carry Ethernet/IEEE 802.3 or IEEE 802.5 data frames as well as flush messages.

DCC
Data Country Code: This specifies the country in which an address is registered. The codes are
given in ISO 3166. The length of this field is two octets. The digits of the data country code are
encoded in Binary Coded Decimal (BCD) syntax. The codes will be left justified and padded on
the right with the hexadecimal value "F" to fill the two octets.

DCE
Data Communication Equipment: A generic definition of computing equipment that attaches to a
network via a DTE.

Default Node Representation

A single value for each nodal state parameter giving the presumed value between any entry or
exit to the logical node and the nucleus.

Demultiplexing
A function performed by a layer entity that identifies and separates SDUs from a single
connection to more than one connection.

DES
Destination End Station: An ATM termination point which is the destination for ATM messages
of a connection and is used as a reference point for ABR services. See SES.

Dijkstra's Algorithm
An algorithm that is sometimes used to calculate routes given a link and nodal state topology
database.

DIR
This is a field in an RM-cell which indicates the direction of the RM- cell with respect to the data
flow with which it is associated. The source sets DIR=0 and the destination sets DIR=1.

Direct Set
A set of host interfaces which can establish direct layer two communications for unicast (not
needed in MPOA).

DLPI
UNIX International, Data Link Provider Interface (DLPI) Specification: Revision 2.0.0, OSI
Work Group, August 1991.

Domain
Refer to Administrative Domain.

DS
Distributed Single Layer Test Method: An abstract test method in which the upper tester is
located within the system under test and the point of control and observation (PCO) is located at
the upper service boundary of the Implementation Under Test (IUT) - for testing one protocol
layer. Test events are specified in terms of the abstract service primitives (ASP) at the upper tester
above the IUT and ASPs and/or protocol data units (PDU) at the lower tester PCO.

DS-0
Digital Signal, Level 0: The 64 kbps rate that is the basic building block for both the North
American and European digital hierarchies.

DS-1
Digital Signal, Level 1: The North American Digital Hierarchy signaling standard for
transmission at 1.544 Mbps. This standard supports 24 simultaneous DS-0 signals. The term is
often used interchangeably with T1 carrier although DS-1 signals may be exchanged over other
transmission systems.

DS-2
Digital Signal, Level 2: The North American Digital Hierarchy signaling standard for
transmission of 6.312 Mbps that is used by T2 carrier which supports 96 calls.

DS-3
Digital Signal, Level 3: The North American Digital Hierarchy signaling standard for
transmission at 44.736 Mbps that is used by T3 carrier. DS-3 supports 28 DS-1s plus overhead.

DS3 PLCP
Physical Layer Convergence Protocol: An alternate method used by older T carrier equipment to
locate ATM cell boundaries. This method has recently been moved to an informative appendix of
the ATM DS3 specification and has been replaced by the HEC method.

DSE
Distributed Single-Layer Embedded (Test Method): An abstract test method in which the upper
tester is located within the system under test and there is a point of control and observation at the
upper service boundary of the Implementation Under Test (IUT) for testing a protocol layer, or
sublayer, which is part of a multi-protocol IUT.

DSS1
Digital Subscriber Signalling System #1: N-ISDN UNI Signalling

DSS2 Setup
DSS2 Digital Subscriber Signalling System #2: B-ISDN UNI Signalling

DSU
Data Service Unit: Equipment used to attach users computing equipment to a public network.

DTE
Data Terminal Equipment: A generic definition of external networking interface equipment such
as a modem.

DTL
Designated Transit List: A list of nodes and optional link IDs that completely specify a path
across a single PNNI peer group.

DTL Originator
The first switching system within the entire PNNI routing domain to build the initial DTL stack
for a given connection.

DTL Terminator
The last switching system within the entire PNNI routing domain to process the connection and
thus the connection's DTL.

DXI
Data Exchange Interface: A variable length frame-based ATM interface between a DTE and a
special ATM CSU/DSU. The ATM CSU/DSU converts between the variable-length DXI frames
and the fixed-length ATM cells.

E.164
A public network addressing standard utilizing up to a maximum of 15 digits. ATM uses E.164
addressing for public network addressing.

E1
Also known as CEPT1, the 2.048 Mbps rate used by European CEPT carrier to transmit 30 64
kbps digital channels for voice or data calls, plus a 64 kbps signaling channel and a 64 kbps
channel for framing and maintenance.

E3
Also known as CEPT3, the 34.368 Mbps rate used by European CEPT carrier to transmit 16
CEPT1s plus overhead.

Edge Device
A physical device which is capable of forwarding packets between legacy interworking interfaces
(e.g., Ethernet, Token Ring, etc.) and ATM interfaces based on data-link and network layer
information but which does not participate in the running of any network layer routing protocol.
An Edge Device obtains forwarding descriptions using the route distribution protocol.

EFCI
Explicit Forward Congestion Indication: EFCI is an indication in the ATM cell header. A network
element in an impending-congested state or a congested state may set EFCI so that this indication
may be examined by the destination end-system. For example, the end-system may use this
indication to implement a protocol that adaptively lowers the cell rate of the connection during
congestion or impending congestion. A network element that is not in a congestion state or an
impending congestion state will not modify the value of this indication. Impending congestion is
the state when a network equipment is operating around its engineered capacity level.

EFS
Error Free Seconds: A unit used to specify the error performance of T carrier systems, usually
expressed as EFS per hour, day, or week. This method gives a better indication of the distribution
of bit errors than a simple bit error rate (BER). Also refer to SES.

ELAN
Emulated Local Area Network: A logical network initiated by using the mechanisms defined by
LAN Emulation. This could include ATM and legacy attached end stations.

EMI

Electromagnetic Interference: Equipment used in high speed data systems, including ATM, that
generate and transmit many signals in the radio frequency portion of the electromagnetic
spectrum. Interference to other equipment or radio services may result if sufficient power from
these signals escape the equipment enclosures or transmission media. National and international
regulatory agencies (FCC, CISPR, etc.) set limits for these emissions. Class A is for industrial use
and Class B is for residential use.

EML
Element Management Layer: An abstraction of the functions provided by systems that manage
each network element on an individual basis.

EMS
Element Management System: A management system that provides functions at the element
Management Layer.

End Station
These devices (e.g., hosts or PCs) enable the communication between ATM end stations and end
stations on "legacy" LAN or among ATM end stations.

Entry Border Node

The node which receives a call over an outside link. This is the first node within a peer group to
see this call.

EOM
End of Message: An indicator used in the AAL that identifies the last ATM cell containing
information from a data packet that has been segmented.

ER
Explicit Rate: The Explicit Rate is an RM-cell field used to limit the source ACR to a specific
value. It is initially set by the source to a requested rate (such as PCR). It may be subsequently
reduced by any network element in the path to a value that the element can sustain. ER is
formatted as a rate.

ES
End System: A system where an ATM connection is terminated or initiated. An originating end
system initiates the ATM connection, and terminating end system terminates the ATM
connection. OAM cells may be generated and received.

ESF
Extended Superframe: A DS1 framing format in which 24 DS0 times lots plus a coded framing
bit are organized into a frame which is repeated 24 times to form a superframe.

ESI
End System Identifier: This identifier distinguishes multiple nodes at the same level in case the
lower level peer group is partitioned.

ETSI
European Telecommunications Standards Institute: The primary telecommunications standards
organization.

Exception
A connectivity advertisement in a PNNI complex node representation that represents something
other than the default node representation.

Exit Border Node

The node that will progress a call over an outside link. This is the last node within a peer group to
see this call.

Exterior
Denotes that an item (e.g., link, node, or reachable address) is outside of a PNNI routing domain.

Exterior Link
A link which crosses the boundary of the PNNI routing domain. The PNNI protocol does not run
over an exterior link.

Exterior Reachable Address

An address that can be reached through a PNNI routing domain, but which is not located in that
PNNI routing domain.

Exterior Route
A route which traverses an exterior link.

Fairness

As related to Generic Flow Control (GFC), fairness is defined as meeting all the agreed quality of
service (QOS) requirements, by controlling the order of service for all active connections.

FC
Feedback Control: Feedback controls are defined as the set of actions taken by the network and
by the end-systems to regulate the traffic submitted on ATM connections according to the state of
network elements.

FCS
Frame Check Sequence: Any mathematical formula which derives a numeric value based on the
bit pattern of a transmitted block of information and uses that value at the receiving end to
determine the existence of any transmission errors.

FDDI
Fiber Distributed Data Interface: A 100 Mbps Local Area Network standard that was developed
by ANSI that is designed to work on fiber-optic cables, using techniques similar to token-ring.

FEBE
Far End Block Error: A maintenance signal transmitted in the PHY overhead that a bit error(s)
has been detected at the PHY layer at the far end of the link. This is used to monitor bit error
performance of the link.

FEC
Forward Error Correction: A technique for detection and correction of errors in a digital data
stream.

FG
Functional Group: A collection of functions related in such a way that they will be provided by a
single logical component. Examples include the Route Server Functional Group (RSFG), the
IASG (Internetwork Address Sub-Group), Coordination Functional Group (ICFG), the Edge
Device Functional Group (EDFG) and the ATM attached host Behavior Functional Group
(AHFG).

Flush Protocol
The flush protocol is provided to ensure the correct order of delivery of unicast data frames.

Foreign Address

An address that does not match any of a given nodes summary addresses.

Forwarding Description
The resolved mapping of an MPOA Target to a set of parameters used to set up an ATM
connection on which to forward packets.

FRS
Frame-Relay Service: A connection oriented service that is capable of carrying up to 4096 bytes
per frame.

FRTT
Fixed Round-Trip Time: This is the sum of the fixed and propagation delays from the source to
the furthest destination and back.

G.703
ITU-T Recommendation G.703, "Physical/Electrical Characteristics of Hierarchical Digital
Interfaces".

G.704
ITU-T Recommendation G.704, "Synchronous Frame Structures Used at Primary and Secondary
Hierarchy Levels".

G.804
ITU-T Recommendation G.804, "ATM Cell Mapping into Plesiochronous Digital Hierarchy
(PDH)".

GCAC
Generic Connection Admission Control: This is a process to determine if a link has potentially
enough resources to support a connection.

GCRA
Generic Cell Rate Algorithm: The GCRA is used to define conformance with respect to the traffic
contract of the connection. For each cell arrival the GCRA determines whether the cell conforms
to the traffic contract. The UPC function may implement the GCRA, or one or more equivalent
algorithms to enforce conformance. The GCRA is defined with two parameters: the Increment (I)
and the Limit (L).

GFC
Generic Flow Control: GFC is a field in the ATM header which can be used to provide local
functions (e.g., flow control). It has local significance only and the value encoded in the field is
not carried end-to-end.

H-Channel
H-Channels are ISDN bearer services that have pre-defined speeds, starting and stopping
locations on a PRI and are contiguously transported from one PRI site through networks to
another PRI site.

H0 Channel
A 384 kbps channel that consists of six contiguous DS0s (64 kbps) of a T1 line.

H10 Channel
The North American 1472 kbps channel from a T1 or primary rate carrier. This is equivalent to
twenty-three (23) 64 kbps channels.

H11 Channel
The North American primary rate used as a single 1536 kbps channel. This channel uses 24
contiguous DS0s or the entire T1 line except for the 8 kbps framing pattern.

H12
The European primary rate used as a single 1920 kbps channel (30 64 kbps channels or the entire
E1 line except for the 64 kbps framing and maintenance channel.

HBFG
Host Behavior Functional Group: The group of functions performed by an ATM-attached host
that is participating in the MPOA service.

HDLC
High Level Data Link Control: An ITU-TSS link layer protocol standard for point-to-point and
multi-point communications.

Header
Protocol control information located at the beginning of a protocol data unit.

HEC
Header Error Control: Using the fifth octet in the ATM cell header, ATM equipment may check
for an error and corrects the contents of the header. The check character is calculated using a
CRC algorithm allowing a single bit error in the header to be corrected or multiple errors to be
detected.

Hello Packet
A type of PNNI Routing packet that is exchanged between neighboring logical nodes.

Hierarchically Complete Source Route

A stack of DTLs representing a route across a PNNI routing domain such that a DTL is included
for each hierarchical level between and including the current level and the lowest visible level in
which the source and destination are reachable.

Hop-by-Hop Route
A route that is created by having each switch along the path use its own routing knowledge to
determine the next hop of the route, with the expectation that all switches will choose consistent
hops such that the call will reach the desired destination. PNNI does not use hop-by-hop routing.

Horizontal Link
A link between two logical nodes that belong to the same peer group.

Host Apparent Address

A set of internetwork layer addresses which a host will directly resolve to lower layer addresses.

I.356
ITU-T Specifications for Traffic Measurement.

I.361
B-ISDN ATM Layer Specification.

I.362
B-ISDN ATM Layer (AAL) Functional Description.

I.363

B-ISDN ATM Layer (AAL) Specification.

I.432
ITU-T Recommendation for B-ISDN User-network Interface.

IASG
Internetwork Address Sub-Group: A range of internetwork layer addresses summarized in an
internetwork layer routing protocol.

ICD
International Code Designator: This identifies an international organization. The registration
authority for the International Code Designator is maintained by the British Standards Institute.
The length of this field is two octets.

ICR
Initial Cell Rate: An ABR service parameter, in cells/sec, that is the rate at which a source should
send initially and after an idle period.

IDU
Interface Data Unit: The unit of information transferred to/from the upper layer in a single
interaction across the SAP. Each IDU contains interface control information and may also contain
the whole or part of the SDU.

IEC
Inter-exchange Carrier: A long distance telephone company.

IEEE
Institute of Electrical and Electronics Engineers: A worldwide engineering publishing and
standards-making body for the electronics industry.

IEEE 802.3
A Local Area Network protocol suite commonly known as Ethernet. Ethernet has either a 10
Mbps or 100 Mbps throughput and uses Carrier Sense Multiple Access bus with Collision
Detection CSMA/CD. This method allows users to share the network cable. However, only one
station can use the cable at a time. A variety of physical medium dependent protocols are
supported.

IEEE 802.5
A Local Area Network protocol suite commonly known as Token Ring. A standard originated by
IBM for a token passing ring network that can be configured in a star topology. Versions
supported are 4 Mbps and 16 Mbps.

IETF
Internet Engineering Task Force: The organization that provides the coordination of standards
and specification development for TCP/IP networking.

ILMI
Integrated Local Management Interface: An ATM Forum defined interim specification for
network management functions between an end user and a public or private network and between
a public network and a private network. This is based on a limited subset of SNMP capabilities.

Induced Uplink
An uplink "A" that is created due to the existence of an uplink "B" in the child peer group
represented by the node that created uplink "A". Both "A" and "B" share the same upnode, which
is higher in the PNNI hierarchy than the peer group in which uplink "A" is seen.

Inside Link
Synonymous with horizontal link.

Instance ID
A subset of an objects attributes which serve to uniquely identify a MIB instance.

Interior
Denotes that an item (e.g., link, node, or reachable address) is inside of a PNNI routing domain.

Internal Reachable Address

An address of a destination that is directly attached to the logical node advertising the address.

IOP
Interoperability: The ability of equipment from different manufacturers (or different
implementations) to operate together.

IP

Internet Protocol: Originally developed by the Department of Defense to support interworking of

dissimilar computers across a network. This protocol works in conjunction with TCP and is
usually identified as TCP/IP. A connectionless protocol that operates at the network layer (layer
3) of the OSI model.

IPX
Novell Internetwork Packet Exchange: A built-in networking protocol for Novell Netware. It was
derived from the Xerox Network System protocol and operates at the network layer of the OSI
protocol model.

IS
Intermediate System: A system that provides forwarding functions or relaying functions or both
for a specific ATM connection. OAM cells may be generated and received.

ISO
International Organization for Standardization: An international organization for standardization,
based in Geneva, Switzerland, that establishes voluntary standards and promotes global trade of
90 member countries.

ITU-T
International Telecommunications Union Telecommunications: ITU-T is an international body of
member countries whose task is to define recommendations and standards relating to the
international telecommunications industry. The fundamental standards for ATM have been
defined and published by the ITU-T. (Previously CCITT.)

ITU H.222
An ITU-T Study Group 15 standard that addresses the multiplexing of multimedia data on an
ATM network.

ITU Q.2100
B-ISDN Signaling ATM Adaptation Layer Overview.

ITU Q.2110
B-ISDN Adaptation Layer Service Specific Connection Oriented Protocol.

ITU Q.2130

B-ISDN Adaptation Layer Service Specific Connection Oriented Function for Support of
Signaling at the UNI.

ITU Q.2931
The signaling standard for ATM to support Switched Virtual Connections. This is based on the
signaling standard for ISDN.

ITU Q.931
The signaling standard for ISDN to support SVCs. The basis for the signaling standard developed
for Frame Relay and ATM.

ITU Q.933
The signaling standard for Frame Relay to support SVCs. This is based on the signaling standard
for ISDN.

IUT
Implementation Under Test: The particular portion of equipment which is to be studied for
testing. The implementation may include one or more protocols.

IWF
Interworking Function

Joining
The phase in which the LE Client establishes its control connections to the LE Server.

JPEG
Joint Photographic Experts Group: An ISO Standards group that defines how to compress still
pictures.

LAN
Local Area Network: A network designed to move data between stations within a campus.

LANE
LAN Emulation: The set of services, functional groups and protocols which provide for the
emulation of LANS utilizing ATM as a backbone to allow connectivity among LAN and ATM
attached end stations.

LAPD
Link Access Procedure D: A layer 2 protocol defined by CCITT (original name of ITU-T). This
protocol reliably transfers blocks of information across a single Layer 1 link and supports
multiplexing of different connections at Layer 2.

Layer Entity
An active element within a layer.

Layer Function
A part of the activity of the layer entities.

Layer Service
A capability of a layer and the layers beneath it that is provided to the upper layer entities at the
boundary between that layer and the next higher layer.

Layer User Data

Data transferred between corresponding entities on behalf of the upper layer or layer management
entities for which they are providing services.

LB
Leaky Bucket: Leaky Bucket is the term used as an analogous description of the algorithm used
for conformance checking of cell flows from a user or network. See GCRA, UPC and NPC. The
"leaking hole in the bucket" applies to the sustained rate at which cells can be accommodated,
while the "bucket depth" applies to the tolerance to cell bursting over a given time period.

LE
LAN Emulation. Refer to LANE.

LE_ARP
LAN Emulation Address Resolution Protocol: A message issued by a LE client to solicit the ATM
address of another function.

Leadership Priority
The priority with which a logical node wishes to be elected peer group leader of its peer group.
Generally, of all nodes in a peer group, the one with the highest leadership priority will be elected
as peer group leader.

Leaky Bucket
An informal term for the Generic Cell Rate Algorithm.

LEC
Local Exchange Carrier: A telephone company affiliate of a Regional Bell Operating Company or
an Independent Telephone Company.

LEC
LAN Emulation Client: The entity in end systems which performs data forwarding, address
resolution, and other control functions.

LECID
LAN Emulation Client Identifier: This identifier, contained in the LAN Emulation header,
indicates the ID of the ATM host or ATM-LAN bridge. It is unique for every ATM Client.

LECS
LAN Emulation Configuration Server: This implements the policy controlled assignment of
individual LE clients to different emulated LANs by providing the LES ATM addresses.

LES
LAN Emulation Server: This implements the control coordination function for the Emulated
LAN, examples are enabling a LEC to join an ELAN, resolving MAC to ATM addresses.

LGN
Logical Group Node: LGN is a single node that represents the lowest level peer groups in the
respective higher level peer group.

LIJP
Leaf Initiated Joint Parameter: Root screening options and Information Element (IE) instructions
carried in SETUP message.

Link
An entity that defines a topological relationship (including available transport capacity) between
two nodes in different subnetworks. Multiple links may exist between a pair of subnetworks.
Synonymous with logical link.

Link Aggregation Token

Refer to Aggregation Token.

Link Attribute
A link state parameter that is considered individually to determine whether a given link is
acceptable and/or desirable for carrying a given connection.

Link Connection
A link connection (e.g., at the VP-level) is a connection capable of transferring information
transparently across a link without adding any overhead, such as cells for purposes for
monitoring. It is delineated by connection points at the boundary of the subnetwork.

Link Constraint
A restriction on the use of links for path selection for a specific connection.

Link Metric
A yenk parameter that requires the values of the parameter for all links along a given path to be
combined to determine whether the path is acceptable and/or desirable for carrying a given
connection.

Link State Parameter

Information that captures an aspect or property of a link.

LNNI
LANE NNI: The standardized interface between two LAN servers (LES-LES, BUS-BUS, LECSLECS and LECS-LES).

LOC
Loss of Cell Delineation: A condition at the receiver or a maintenance signal transmitted in the
PHY overhead indicating that the receiving equipment has lost cell delineation. Used to monitor
the performance of the PHY layer.

LOF
Loss of Frame: A condition at the receiver or a maintenance signal transmitted in the PHY
overhead indicating that the receiving equipment has lost frame delineation. This is used to
monitor the performance of the PHY layer.

Logical Group Node

A logical node that represents a lower level peer group as a single point for purposes of operating
at one level of the PNNI routing hierarchy.

Logical Link
An abstract representation of the connectivity between two logical nodes. This includes
individual physical links, individual virtual path connections, and parallel physical links and/or
virtual path connections.

Logical Node
An abstract representation of a peer group or a switching system as a single point.

Logical Node ID
A string of bits that unambiguously identifies a logical node within a routing domain.

LOP
Loss of Pointer: A condition at the receiver or a maintenance signal transmitted in the PHY
overhead indicating that the receiving equipment has lost the pointer to the start of cell in the
payload. This is used to monitor the performance of the PHY layer.

LOS
Loss of Signal: A condition at the receiver or a maintenance signal transmitted in the PHY
overhead indicating that the receiving equipment has lost the received signal. This is used to
monitor the performance of the PHY layer.

LPF
Low Pass Filter: In an MPEG-2 clock recovery circuit, it is a technique for smoothing or
averaging changes to the system clock.

LSAP
Link Service Access Point: Logical address of boundary between layer 3 and LLC sublayer 2.

LSB
Least Significant Bit: The lowest order bit in the binary representation of a numerical value.

LSR

Leaf Setup Request: A setup message type used when a leaf node requests connection to existing
point-to-multipoint connection or requests creation of a new multipoint connection.

LT
Lower Tester: The representation in ISO/IEC 9646 of the means of providing, during test
execution, indirect control and observation of the lower service boundary of the IUT using the
underlying service provider.

LTE
SONET Lite Terminating Equipment: ATM equipment terminating a communications facility
using a SONET Lite Transmission Convergence (TC) layer. This is usually reserved for end user
or LAN equipment. The SONET Lite TC does not implement some of the maintenance functions
used in long haul networks such as termination of path, line and section overhead.

LUNI
LANE UNI: The standardized interface between a LE client and a LE Server (LES, LECS and
BUS).

M1
Management Interface 1: The management of ATM end devices.

M2
Management Interface 2: The management of Private ATM networks or switches.

M3
Management Interface 3: The management of links between public and private networks.

M4
Management Interface 4: The management of public ATM networks.

M5
Management Interface 5: The management of links between two public networks.

MAC
Media Access Control: IEEE specifications for the lower half of the data link layer (layer 2) that
defines topology dependent access control protocols for IEEE LAN specifications.

MAN
Metropolitan Area Network: A network designed to carry data over an area larger than a campus
such as an entire city and its outlying area.

Managed System
An entity that is managed by one or more management systems, which can be either Element
Management Systems, Subnetwork or Network Management Systems, or any other management
systems.

Management Domain
An entity used here to define the scope of naming.

Management System
An entity that manages a set of managed systems, which can be either NEs, subnetworks or other
management systems.

MaxCR
Maximum Cell Rate: This is the maximum capacity usable by connections belonging to the
specified service category.

MBS
Maximum Burst Size: In the signaling message, the Burst Tolerance (BT) is conveyed through
the MBS which is coded as a number of cells. The BT together with the SCR and the GCRA
determine the MBS that may be transmitted at the peak rate and still be in conformance with the
GCRA.

MCDV
Maximum Cell Delay Variance: This is the maximum two-point CDV objective across a link or
node for the specified service category.

MCLR
Maximum Cell Loss Ratio: This is the maximum ratio of the number of cells that do not make it
across the link or node to the total number of cells arriving at the link or node.

MCR

Minimum Cell Rate: An ABR service traffic descriptor, in cells/sec, that is the rate at which the
source is always allowed to send.

MCTD
Maximum Cell Transfer Delay: This is the sum of the fixed delay component across the link or
node and MCDV.

Metasignaling
ATM Layer Management (LM) process that manages different types of signaling and possibly
semipermanent virtual channels (VCs), including the assignment, removal and checking of VCs.

Metasignaling VCs
The standardized VCs that convey metasignaling information across a User-Network Interface
(UNI).

MIB
Management Information Base: A definition of management items for some network component
that can be accessed by a network manager. A MIB includes the names of objects it contains and
the type of information retained.

MIB Attribute
A single piece of configuration, management, or statistical information which pertains to a
specific part of the PNNI protocol operation.

MIB Instance
An incarnation of a MIB object that applies to a specific part, piece, or aspect of the PNNI
protocols operation.

MIB Object
A collection of attributes that can be used to configure, manage, or analyze an aspect of the PNNI
protocols operation.

MID
Message Identifier: The message identifier is used to associate ATM cells that carry segments
from the same higher layer packet.

MIR

Maximum Information Rate: Refer to PCR.

MMF
Multimode Fiberoptic Cable: Fiberoptic cable in which the signal or light propagates in multiple
modes or paths. Since these paths may have varying lengths, a transmitted pulse of light may be
received at different times and smeared to the point that pulses may interfere with surrounding
pulses. This may cause the signal to be difficult or impossible to receive. This pulse dispersion
sometimes limits the distance over which a MMF link can operate.

MPEG
Motion Picture Experts Group: An ISO Standards group dealing with video and audio
compression techniques and mechanisms for multiplexing and synchronizing various media
streams.

MPOA
Multiprotocol over ATM: An effort taking place in the ATM Forum to standardize protocols for
the purpose of running multiple network layer protocols over ATM.

MPOA Client
A device which implements the client side of one or more of the MPOA protocols, (i.e., is a SCP
client and/or an RDP client. An MPOA Client is either an Edge Device Functional Group (EDFG)
or a Host Behavior Functional Group (HBFG).

MPOA Server
An MPOA Server is any one of an ICFG or RSFG.

MPOA Service Area

The collection of server functions and their clients. A collection of physical devices consisting of
an MPOA server plus the set of clients served by that server.

MPOA Target
A set of protocol address, path attributes, (e.g., internetwork layer QoS, other information
derivable from received packet) describing the intended destination and its path attributes that
MPOA devices may use as lookup keys.

Mrm

An ABR service parameter that controls allocation of bandwidth between forward RM-cells,
backward RM-cells, and data cells.

MSB
Most Significant Bit: The highest order bit in the binary representation of a numerical value.

MT
Message Type: Message type is the field containing the bit flags of a RM-cell. These flags are as
follows: DIR = 0 for forward RM-cells = 1 for backward; RM-cells BN = 1 for Non-Source
Generated (BECN), RM-cells = 0 for Source Generated RM-cells CI = 1 to indicate congestion =
0 otherwise NI = 1 to indicate no additive increase allowed = 0 otherwise RA Not used for ATM
Forum ABR.

MTP
Message Transfer Part: Level 1 through 3 protocols of the SS7 protocol stack. MTP 3 (Level 3) is
used to support BISUP.

Multicasting
The transmit operation of a single PDU by a source interface where the PDU reaches a group of
one or more destinations.

Multiplexing
A function within a layer that interleaves the information from multiple connections into one
connection.

Multipoint Access
User access in which more than one terminal equipment (TE) is supported by a single network
termination.

Multipoint-to-Multipoint Connection
A Multipoint-to-Multipoint Connection is a collection of associated ATM VC or VP links, and
their associated nodes, with the following properties:
1) All Nodes in the connection, called endpoints, serve as a Root Node in a Point-to-Multipoint
connection to all of the (N-1) remaining endpoints.
2) Each of the endpoints on the connection can send information directly to any other endpoint,
but the receiving endpoint cannot distinguish which of the endpoints is sending information
without additional (e.g., higher layer) information.

Multipoint-to-Point Connection
A Point-to-Multipoint Connection may have zero bandwidth from the Root node to the Leaf
Nodes, and non-zero return bandwidth from the Leaf Nodes to the Root Node. Such a connection
is also known as a Multipoint-to-Point Connection. Note that UNI 4.0 does not support this
connection type.

N-ISDN
Narrowband Integrated Services Digital Network: Services include Basic rate interface (2B+D or
BRI) and primary rate interface (30B+D - Europe and 23B+D - North America or PRI). Supports
narrowband - speeds at/or below 1.5 Mbps.

Native Address
An address that matches one of a given nodes summary addresses.

NDIS
Network Driver Interface Specification: Refer to 3COM/Microsoft, LAN Manager: Network
Driver Interface Specification, October 8, 1990.

NE
Network Element: A system that supports at least NEFs and may also support Operation System
Functions/Mediation Functions. An ATM NE may be realized as either a standalone device or a
geographically distributed system. It cannot be further decomposed into managed elements in the
context of a given management function.

NEF
Network Element Function: A function within an ATM entity that supports the ATM based
network transport services, (e.g., multiplexing, cross-connection).

Neighbor Node
A node that is directly connected to a particular node via a logical link.

NEL
Network Element Layer: An abstraction of functions related specifically to the technology,
vendor, and the network resources or network elements that provide basic communications
services.

NEXT

Near End Crosstalk: Equipment that must concurrently receive on one wire pair and transmit on
another wire pair in the same cable bundle must accommodate NEXT interference. NEXT is the
portion of the transmitted signal that leaks into the receive pair. Since at this point on the link the
transmitted signal is at maximum and the receive signal has been attenuated, it may be difficult to
maintain an acceptable ACR with the received signal if the cable media allows large amounts of
crosstalk leakage to occur. Foiled or shielded cables generally have less crosstalk than unshielded
varieties.

NM
Network Management Entity: The body of software in a switching system that provides the
ability to manage the PNNI protocol. NM interacts with the PNNI protocol through the MIB.

NML
Network Management Layer: An abstraction of the functions provided by systems which manage
network elements on a collective basis, so as to monitor and control the network end-to-end.

NMS
Network Management System: An entity that implements functions at the Network Management
Layer. It may also include Element Management Layer functions. A Network Management
System may manage one or more other Network Management Systems.

NMS Environment
A set of NMS which cooperate to manage one or more subnetworks.

NNI
Network Node Interface: An interface between ATM switches defined as the interface between
two network nodes.

Nodal Attribute
A nodal state parameter that is considered individually to determine whether a given node is
acceptable and/or desirable for carrying a given connection.

Nodal Constraint
A restriction on the use of nodes for path selection for a specific connection.

Nodal Metric

A nodal parameter that requires the values of the parameter for all nodes along a given path to be
combined to determine whether the path is acceptable and/or desirable for carrying a given
connection.

Nodal State Parameter

Information that captures an aspect or property of a node.

Node
Synonymous with logical node.

NPC
Network Parameter Control: Network Parameter Control is defined as the set of actions taken by
the network to monitor and control traffic from the NNI. Its main purpose is to protect network
resources from malicious as well as unintentional misbehavior which can affect the QoS of other
already established connections by detecting violations of negotiated parameters and taking
appropriate actions. Refer to UPC.

Nrm
An ABR service parameter, Nrm is the maximum number of cells a source may send for each
forward RM-cell.

NSAP
Network Service Access Point: OSI generic standard for a network address consisting of 20
octets. ATM has specified E.164 for public network addressing and the NSAP address structure
for private network addresses.

NSR
Non-Source Routed: Frame forwarding through a mechanism other than Source Route Bridging.

NT
Network Termination: Network Termination represents the termination point of a Virtual
Channel, Virtual Path, or Virtual Path/Virtual Channel at the UNI.

NTSC
National Television System Committee: An industry group that defines how television signals are
encoded and transmitted in the U.S.

Nucleus
The interior reference point of a logical node in the PNNI complex node representation.

nx64K
This refers to a circuit bandwidth or speed provided by the aggregation of nx64 kbps channels
(where n=integer &gt; 1). The 64K or DS0 channel is the basic rate provided by the T Carrier
systems.

OAM
Operations Administration and Maintenance: A group of network management functions that
provide network fault indication, performance information, and data and diagnosis functions.

Octet
A term for eight (8) bits that is sometimes used interchangeably with "byte" to mean the same
thing.

ODI
Open Data-Link Interface: This refers to Novell Incorporated, Open Data-Link Interface
Developer Guide, March 20, 1992.

One Hop Set

A set of hosts which are one hop apart in terms of internetwork protocols TTLs (TTL=0 -on the
wire+).

OOF
Out of Frame. Refer to LOF.

OSI
Open Systems Interconnection: A seven (7) layer architecture model for communications systems
developed by the ISO for the interconnection of data communications systems. Each layer uses
and builds on the services provided by those below it.

OSPF
Open Shortest Path First: A link-state routing algorithm that is used to calculate routes based on
the number of routers, transmission speed, delays and route cost.

OUI
Organizationally Unique Identifier: The OUI is a three-octet field in the IEEE 802.1a defined
SubNetwork Attachment Point (SNAP) header, identifying an organization which administers the
meaning of the following two octet Protocol Identifier (PID) field in the SNAP header. Together
they identify a distinct routed or bridged protocol.

Outlier
A node whose exclusion from its containing peer group would significantly improve the accuracy
and simplicity of the aggregation of the remainder of the peer group topology.

Outside Link
A link to an outside node.

Outside Node
A node which is participating in PNNI routing, but which is not a member of a particular peer
group.

PAD
Packet Assembler and Disassembler: A PAD assembles packets of asynchronous data and emits
these buffers in a burst to a packet switch network. The PAD also disassembles packets from the
network and emits the data to the non-packet device.

Parent Node
The logical group node that represents the containing peer group of a specific node at the next
higher level of the hierarchy.

Parent Peer Group

The parent peer group of a peer group is the one containing the logical group node representing
that peer group. The parent peer group of a node is the one containing the parent node of that
node.

Path Constraint
A bound on the combined value of a topology metric along a path for a specific connection.

PBX

Private Branch eXchange: PBX is the term given to a device which provides private local voice
switching and voice-related services within the private network. A PBX could have an ATM API
to utilize ATM services, for example Circuit Emulation Service.

PC
Protocol Control: Protocol Control is a mechanism which a given application protocol may
employ to determine or control the performance and health of the application. Example, protocol
liveness may require that protocol control information be sent at some minimum rate; some
applications may become intolerable to users if they are unable to send at least at some minimum
rate. For such applications, the concept of MCR is defined. Refer to MCR.

PCM
Pulse Code Modulation: An audio encoding algorithm which encodes the amplitude of a
repetitive series of audio samples. This encoding algorithm converts analog voice samples into a
digital bit stream.

PCO
Point of Control and Observation: A place (point) within a testing environment where the
occurrence of test events is to be controlled and observed as defined by the particular abstract test
method used.

PCR
Program Clock Reference: A timestamp that is inserted by the MPEG-2 encoder into the
Transport Stream to aid the decoder in the recovering and tracking the encoder clock.

PCR
Peak Cell Rate: The Peak Cell Rate, in cells/sec, is the cell rate which the source may never
exceed.

PDH
Plesiochronous Digital Hierarchy: PDH (plesiochronous means nearly synchronous), was
developed to carry digitized voice over twisted pair cabling more efficiently. This evolved into
the North American, European, and Japanese Digital Hierarchies where only a discrete set of
fixed rates is available, namely, nxDS0 (DS0 is a 64 kbps rate) and then the next levels in the
respective multiplex hierarchies.

PDU

Protocol Data Unit: A PDU is a message of a given protocol comprising payload and protocolspecific control information, typically contained in a header. PDUs pass over the protocol
interfaces which exist between the layers of protocols (per OSI model).

Peer Entities
Entities within the same layer.

Peer Group
A set of logical nodes which are grouped for purposes of creating a routing hierarchy. PTSEs are
exchanged among all members of the group.

Peer Group Identifier

A string of bits that is used to unambiguously identify a peer group.

Peer Group Leader

A node which has been elected to perform some of the functions associated with a logical group
node.

Peer Group Level

The number of significant bits in the peer group identifier of a particular peer group.

Peer Node
A node that is a member of the same peer group as a given node.

PES
Packetized Elementary Stream: In MPEG-2, after the media stream has been digitized and
compressed, it is formatted into packets before it is multiplexed into either a Program Stream or
Transport Stream.

PG
Peer Group: A set of logical nodes which are grouped for purposes of creating a routing
hierarchy. PTSEs are exchanged among all members of the group.

PGL
Peer Group Leader: A single real physical system which has been elected to perform some of the
functions associated with a logical group node.

PHY
OSI Physical Layer: The physical layer provides for transmission of cells over a physical medium
connecting two ATM devices. This physical layer is comprised of two sublayers: the PMD
Physical Medium Dependent sublayer, and the TC Transmission Convergence sublayer. Refer to
PMD and TC.

Physical Layer (PHY) Connection

An association established by the PHY between two or more ATM entities. A PHY connection
consists of the concatenation of PHY links in order to provide an end-to-end transfer capability to
PHY SAPs.

Physical Link
A real link which attaches two switching systems.

PICS
Protocol Implementation Conformance Statement: A statement made by the supplier of an
implementation or system stating which capabilities have been implemented for a given protocol.

PID
Protocol Identification. Refer to OUI.

PIXIT
Protocol Implementation eXtra Information for Testing: A statement made by a supplier or
implementor of an IUT which contains information about the IUT and its testing environment
which will enable a test laboratory to run an appropriate test suite against the IUT.

Plastic Fiber Optics

An optical fiber where the core transmission media is plastic in contrast to glass or silica cores.
Proposed plastic fibers generally have larger attenuation and dispersion than glass fiber but may
have applications where the distance is limited. Plastic systems may also offer lower cost
connectors that may be installed with simple tools and a limited amount of training.

PLCP
Physical Layer Convergence Protocol: The PLCP is defined by the IEEE 802.6. It is used for DS3
transmission of ATM. ATM cells are encapsulated in a 125 microsecond frame defined by the
PLCP which is defined inside the DS3 M-frame.

PLL
Phase Lock Loop: Phase Lock Loop is a mechanism whereby timing information is transferred
within a data stream and the receiver derives the signal element timing by locking its local clock
source to the received timing information.

PM
Physical Medium: Physical Medium refers to the actual physical interfaces. Several interfaces are
defined including STS-1, STS-3c, STS-12c, STM-1, STM-4, DS1, E1, DS2, E3, DS3, E4, FDDIbased, Fiber Channel-based, and STP. These range in speeds from 1.544Mbps through 622.08
Mbps.

PMD
Physical Media Dependent: This sublayer defines the parameters at the lowest level, such as
speed of the bits on the media.

PNI
Permit Next Increase: An ABR service parameter, PNI is a flag controlling the increase of ACR
upon reception of the next backward RM-cell. PNI=0 inhibits increase. The range is 0 or 1.

PNNI
Private Network-Network Interface: A routing information protocol that enables extremely
scalable, full function, dynamic multi-vendor ATM switches to be integrated in the same network.

PNNI Protocol Entity

The body of software in a switching system that executes the PNNI protocol and provides the
routing service.

PNNI Routing Control Channel

VCCs used for the exchange of PNNI routing protocol messages.

PNNI Routing Domain

A group of topologically contiguous systems which are running one instance of PNNI routing.

PNNI Routing Hierarchy

The hierarchy of peer groups used for PNNI routing.

PNNI Topology State Element

A collection of PNNI information that is flooded among all logical nodes within a peer group.

PNNI Topology State Packet

A type of PNNI Routing packet that is used for flooding PTSEs among logical nodes within a
peer group.

POH
Path Overhead: A maintenance channel transmitted in the SONET overhead following the path
from the beginning multiplexer to the ending demultiplexer. This is not implemented in SONET
Lite.

Point-to-Multipoint Connection
A Point-to-Multipoint Connection is a collection of associated ATM VC or VP links, with
associated endpoint nodes, with the following properties:
1) One ATM link, called the Root Link, serves as the root in a simple tree topology. When the
Root Node sends information, all of the remaining nodes on the connection, called Leaf Nodes,
receive copies of the information.
2) Each of the Leaf Nodes on the connection can send information directly to the Root Node. The
Root Node cannot distinguish which Leaf is sending information without additional (higher
layer) information. (See note below for UNI 4.0 support.)
3) The Leaf Nodes cannot communicate directly to each other with this connection type.
Note: UNI 4.0 does not support traffic sent from a Leaf to the Root.

Point-to-Point Connection
A connection with only two endpoints.

Port Identifier
The identifier assigned by a logical node to represent the point of attachment of a link to that
node.

PRI
Primary Rate Interface: An ISDN standard for provisioning of 1.544 Mbit/s (DS1 - North
America, Japan, et al) or 2.048 Mbit/s (E1 - Europe) ISDN services. DS1 is 23 "B" channels of

64 kbit/s each and one signalling "D" channel of 64 kbit/s/ E1 is 30 "B" channels of 64 kbit/s
each and one signalling "D" channel of 64 kbit/s.

PRS
Primary Reference Source

Primitive
An abstract, implementation independent, interaction between a layer service user and a layer
service provider.

Private ATM Address

A twenty-byte address used to identify an ATM connection termination point.

Protocol
A set of rules and formats (semantic and syntactic) that determines the communication behavior
of layer entities in the performance of the layer functions.

Protocol Control Information

Information exchanged between corresponding entities, using a lower layer connection, to
coordinate their joint operation.

PT
Payload Type: Payload Type is a 3-bit field in the ATM cell header that discriminates between a
cell carrying management information or one which is carrying user information.

PTI
Payload Type Indicator: Payload Type Indicator is the Payload Type field value distinguishing the
various management cells and user cells. Example: Resource Management cell has PTI=110, endto-end OAM F5 Flow cell has PTI=101.

PTMPT
Point-To-Multipoint: A main source to many destination connections.

PTS

Presentation Time Stamp: A timestamp that is inserted by the MPEG-2 encoder into the
packetized elementary stream to allow the decoder to synchronize different elementary streams
(i.e., lip sync).

PTSE
PNNI Topology State Element: A collection of PNNI information that is flooded among all
logical nodes within a peer group.

PTSP
PNNI Topology State Packet: A type of PNNI Routing packet that is used for flooding PTSEs
among logical nodes within a peer group.

PVC
Permanent Virtual Circuit: This is a link with static route defined in advance, usually by manual
setup.

PVCC
Permanent Virtual Channel Connection: A Virtual Channel Connection (VCC) is an ATM
connection where switching is performed on the VPI/VCI fields of each cell. A Permanent VCC
is one which is provisioned through some network management function and left up indefinitely.

PVPC
Permanent Virtual Path Connection: A Virtual Path Connection (VPC) is an ATM connection
where switching is performed on the VPI field only of each cell. A Permanent VPC is one which
is provisioned through some network management function and left up indefinitely.

QD
Queuing Delay: Queuing Delay refers to the delay imposed on a cell by its having to be buffered
because of unavailability of resources to pass the cell onto the next network function or element.
This buffering could be a result of oversubscription of a physical link, or due to a connection of
higher priority or tighter service constraints getting the resource of the physical link.

QoS
Quality of Service: Quality of Service is defined on an end-to-end basis in terms of the following
attributes of the end-to-end ATM connection:
1) Cell Loss Ratio
2) Cell Transfer Delay

3) Cell Delay Variation

Q.SIG
A symmetrical adaptatation of N-ISDN signalling (DSS1) for inter-PBX signalling.

RBOC
Regional Bell Operating Company: Seven companies formed to manage the local exchanges
originally owned by AT&T. These companies were created as a result of an agreement between
AT&T and the United States Department of Justice.

RD
Routing Domain: A group of topologically contiguous systems which are running one instance of
routing.

RDF
Rate Decrease Factor: An ABR service parameter, RDF controls the decrease in the cell
transmission rate. RDF is a power of 2 from 1/32,768 to 1.

RO
Read-Only: Attributes which are read-only cannot be written by Network Management. Only the
PNNI Protocol entity may change the value of a read-only attribute. Network Management
entities are restricted to only reading such read-only attributes. Read-only attributes are typically
for statistical information, including reporting result of actions taken by auto-configuration.

RW
Read-Write: Attributes which are read-write cannot be written by the PNNI protocol entity. Only
the Network Management Entity may change the value of a read-write attribute. The PNNI
Protocol Entity is restricted to only reading such read-write attributes. Read-write attributes are
typically used to provide the ability for Network Management to configure, control, and manage
a PNNI Protocol Entity's behavior.

Registration
The address registration function is the mechanism by which Clients provide address information
to the LAN Emulation Server.

Relaying

A function of a layer by means of which a layer entity receives data from a corresponding entity
and transmits it to another corresponding entity.

RFC
Request For Comment: The development of TCP/IP standards, procedures and specifications is
done via this mechanism. RFCs are documents that progress through several development stages,
under the control of IETF, until they are finalized or discarded.

RFC1695
Definitions of Managed Objects for ATM Management or AToM MIB.

RFI
Radio Frequency Interface: Refer to EMI.

RIF
Rate Increase Factor: This controls the amount by which the cell transmission rate may increase
upon receipt of an RM-cell. The additive increase rate AIR=PCR*RIF. RIF is a power of 2,
ranging from 1/32768 to 1.

RISC
Reduced Instruction Set Computing: A computer processing technology in which a
microprocessor understands a few simple instructions thereby providing fast, predictable
instruction flow.

RM
Resource Management: Resource Management is the management of critical resources in an
ATM network. Two critical resources are buffer space and trunk bandwidth. Provisioning may be
used to allocate network resources in order to separate traffic flows according to service
characteristics. VPCs play a key role in resource management. By reserving capacity on VPCs,
the processing required to establish individual VCCs is reduced. Refer to RM-cell.

RM-Cell
Resource Management Cell: Information about the state of the net work like bandwidth
availability, state of congestion, and impending congestion, is conveyed to the source through
special control cells called Resource Management Cells (RM-cells).

Route Server

A physical device that runs one or more network layer routing protocols, and which uses a route
query protocol in order to provide network layer routing forwarding descriptions to clients.

Router
A physical device that is capable of forwarding packets based on network layer information and
that also participates in running one or more network layer routing protocols.

Routing Computation
The process of applying a mathematical algorithm to a topology database to compute routes.
There are many types of routing computations that may be used. The Djikstra algorithm is one
particular example of a possible routing computation.

Routing Constraint
A generic term that refers to either a topology constraint or a path constraint.

Routing Protocol
A general term indicating a protocol run between routers and/or route servers in order to
exchange information used to allow computation of routes. The result of the routing computation
will be one or more forwarding descriptions.

RS
Remote single-layer (Test Method): An abstract test method in which the upper tester is within
the system under test and there is a point of control and observation at the upper service boundary
of the Implementation Under Test (IUT) for testing one protocol layer. Test events are specified
in terms of the abstract service primitives (ASP) and/or protocol data units at the lower tester
PCO.

RSE
Remote Single-layer Embedded (Test Method): An abstract test method in which the upper tester
is within the system under test and there is a point of control and observation at the upper service
boundary of the Implementation Under Test (IUT) for testing a protocol layer or sublayer which
is part of a multi-protocol IUT.

RSFG
Route Server Functional Group: The group of functions performed to provide internetworking
level functions in an MPOA System. This includes running conventional interworking Routing
Protocols and providing inter-IASG destination resolution.

SA
Source Address: The address from which the message or data originated.

SA
Source MAC Address: A six octet value uniquely identifying an end point and which is sent in an
IEEE LAN frame header to indicate source of frame.

SAAL
Signaling ATM Adaptation Layer: This resides between the ATM layer and the Q.2931 function.
The SAAL provides reliable transport of Q.2931 messages between Q.2931 entities (e.g., ATM
switch and host) over the ATM layer; two sublayers: common part and service specific part.

SAP
Service Access Point: A SAP is used for the following purposes:
1) When the application initiates an outgoing call to a remote ATM device, a destination_SAP
specifies the ATM address of the remote device, plus further addressing that identifies the target
software entity within the remote device.
2) When the application prepares to respond to incoming calls from remote ATM devices, a
local_SAP specifies the ATM address of the device housing the application, plus further
addressing that identifies the application within the local device.
There are several groups of SAPs that are specified as valid for Native ATM Services.

SAR
Segmentation and Reassembly: Method of breaking up arbitrarily sized packets.

Scope
A scope defines the level of advertisement for an address. The level is a level of a peer group in
the PNNI routing hierarchy.

SCCP
Signaling Connection and Control Part: A SS7 protocol that provides additional functions to the
Message Transfer Part (MTP). It typically supports Transaction Capabilities Application Part
(TCAP).

SCP

Service Control Point: A computer and database system which executes service logic programs to
provide customer services through a switching system. Messages are exchanged with the SSP
through the SS7 network.

SCR
Sustainable Cell Rate: The SCR is an upper bound on the conforming average rate of an ATM
connection over time scales which are long relative to those for which the PCR is defined.
Enforcement of this bound by the UPC could allow the network to allocate sufficient resources,
but less than those based on the PCR, and still ensure that the performance objectives (e.g., for
Cell Loss Ratio) can be achieved.

SDH
Synchronous Digital Hierarchy: The ITU-TSS International standard for transmitting information
over optical fiber.

SDT
Structured Data Transfer: An AAL1 data transfer mode in which data is structured into blocks
which are then segmented into cells for transfer.

SDU
Service Data Unit: A unit of interface information whose identity is preserved from one end of a
layer connection to the other.

SE
Switching Element: Switching Element refers to the device or network node which performs
ATM switching functions based on the VPI or VPI/VCI pair.

SEAL
Simple and Efficient Adaptation Layer: An earlier name for AAL5.

Segment
A single ATM link or group of interconnected ATM links of an ATM connection.

SEL
Selector: A subfield carried in SETUP message part of ATM endpoint address Domain specific
Part (DSP) defined by ISO 10589, not used for ATM network routing, used by ATM end systems
only.

Semipermanent Connection
A connection established via a service order or via network management.

SES
Severely Errored Seconds: A unit used to specify the error performance of T carrier systems. This
indicates a second containing ten or more errors, usually expressed as SES per hour, day, or
week. This method gives a better indication of the distribution of bit errors than a simple Bit
Error Rate (BER). Refer also to EFS.

SES
Source End Station: An ATM termination point, which is the source of ATM messages of a
connection, and is used as a reference point for ABR services. Refer to DES.

SF
SuperFrame: A DS1 framing format in which 24 DS0 timeslots plus a coded framing bit are
organized into a frame which is repeated 12 times to form the superframe.

Shaping Descriptor
N ordered pairs of GCRA parameters (I,L) used to define the negotiated traffic shape of a
connection.

SIPP
SMDS Interface Protocol: Protocol where layer 2 is based on ATM, AAL and DQDB. Layer 1 is
DS1 and DS3.

SMDS
Switched Multi-Megabit Data Services: A connectionless service used to connect LANs, MANs
and WANs to exchange data.

SMF
Single Mode Fiber: Fiber optic cable in which the signal or light propagates in a single mode or
path. Since all light follows the same path or travels the same distance, a transmitted pulse is not
dispersed and does not interfere with adjacent pulses. SMF fibers can support longer distances
and are limited mainly by the amount of attenuation. Refer to MMF.

SN

Sequence Number: SN is a 4 octet field in a Resource Management cell defined by the ITU-T in
recommendation I.371 to sequence such cells. It is not used for ATM Forum ABR. An ATM
switch will either preserve this field or set it in accordance with I.371.

SN cell
Sequence Number Cell: A cell sent periodically on each link of an AIMUX to indicate how many
cells have been transmitted since the previous SN cell. These cells are used to verify the sequence
of payload cells reassembled at the receiver.

SNA
Systems Network Architecture: IBMs seven layer, vendor specific architecture for data
communications.

SNC
Subnetwork Connection: In the context of ATM, an entity that passes ATM cells transparently,
(i.e., without adding any overhead). A SNC may be either a stand-alone SNC, or a concatenation
of SNCs and link connections.

SNMP
Simple Network Management Protocol: Originally designed for the Department of Defense
network to support TCP/IP network management. It has been widely implemented to support the
management of a broad range of network products and functions. SNMP is the IETF standard
management protocol for TCP/IP networks.

SONET
Synchronous Optical Network: An ANSI standard for transmitting information over optical fiber.
This standard is used or accepted in the United States and Canada and is a variation of the SDH
International standard.

Source Route
As used in this document, a hierarchically complete source route.

Source Traffic
A set of traffic parameters belonging to the ATM Traffic Descriptor Descriptor used during the
connection set-up to capture the intrinsic traffic characteristics of the connection requested by the
source.

SPE

SONET Synchronous Payload Envelope.

Split System
A switching system which implements the functions of more than one logical node.

SPTS
Single Program Transport Stream: An MPEG-2 Transport Stream that consists of only one
program.

SR
Source Routing: A bridged method whereby the source at a data exchange determines the route
that subsequent frames will use.

SRF
Specifically Routed Frame: A Source Routing Bridging Frame which uses a specific route
between the source and destination.

SRT
Source Routing Transparent: An IETF Bridging Standard combining Transparent Bridging and
Source Route Bridging.

SRTS
Synchronous residual Time Stamp: A clock recovery technique in which difference signals
between source timing and a network reference timing signal are transmitted to allow
reconstruction of the source timing at the destination.

SSCF
Service Specific Coordination Function: SSCF is a function defined in Q.2130, B-ISDN
Signaling ATM Adaptation Layer-Service Specific Coordination Function for Support of
Signaling at the User-to- Network Interface.

SSCOP
Service Specific Connection Oriented Protocol: An adaptation layer protocol defined in ITU-T
Specification: Q.2110.

SSCS

Service Specific Convergence Sublayer: The portion of the convergence sublayer that is
dependent upon the type of traffic that is being converted.

SS7
Signal System Number 7: A family of signaling protocols originating from narrowband
telephony. They are used to set-up, manage and tear down connections as well as to exchange
non-connection associated information. Refer to BISUP, MTP, SCCP and TCAP.

STC
System Time Clock: The master clock in an MPEG-2 encoder or decoder system.

STE
Spanning Tree Explorer: A Source Route Bridging frame which uses the Spanning Tree algorithm
in determining a route.

STE
SONET Section Terminating Equipment: SONET equipment that terminates a section of a link
between a transmitter and repeater, repeater and repeater, or repeater and receiver. This is usually
implemented in wide area facilities and not implemented by SONET Lite.

STM
Synchronous Transfer Module: STM is a basic building block used for a synchronous
multiplexing hierarchy defined by the CCITT/ITU-T. STM-1 operates at a rate of 155.52 Mbps
(same as STS-3).

STM-1
Synchronous Transport Module 1: SDH standard for transmission over OC-3 optical fiber at
155.52 Mbps.

STM-n
Synchronous Transport Module "n": (where n is an integer) SDH standards for transmission over
optical fiber (OC-'n x 3) by multiplexing "n" STM-1 frames, (e.g., STM-4 at 622.08 Mbps and
STM-16 at 2.488 Gbps).

STM-nc
Synchronous Transport Module "n" concatenated: (where n is an integer) SDH standards for
transmission over optical fiber (OC-'n x 3) by multiplexing "n" STM-1 frames, (e.g., STM-4 at

622.08 Mbps and STM-16 at 2.488 Gbps, but treating the information fields as a single
concatenated payload).

STP
Signaling Transfer Point: A high speed, reliable, special purpose packet switch for signaling
messages in the SS7 network.

STP
Shielded Twisted Pair: A cable containing one or more twisted pair wires with each pair having a
shield of foil wrap.

STS-1
Synchronous Transport Signal 1: SONET standard for transmission over OC-1 optical fiber at
51.84 Mbps.

STS-n
Synchronous Transport Signal "n": (where n is an integer) SONET standards for transmission
over OC-n optical fiber by multiplexing "n" STS-1 frames, (e.g., STS-3 at 155.52 Mbps STS-12
at 622.08 Mbps and STS-48 at 2.488 Gbps).

STS-nc
Synchronous Transport Signal "n" concatenated: (where n is an integer) SONET standards for
transmission over OC-n optical fiber by multiplexing "n" STS-1 frames, (e.g., STS-3 at 155.52
Mbps STS-12 at 622.08 Mbps and STS-48 at 2.488 Gbps but treating the information fields as a
single concatenated payload).

Sublayer
A logical sub-division of a layer.

Subnet
The use of the term subnet to mean a LAN technology is a historical use and is not specific
enough in the MPOA work. Refer to Internetwork Address Sub-Group, Direct Set, Host Apparent
Address Sub-Group and One Hop Set for more specific definitions.

Subnetwork
A collection of managed entities grouped together from a connectivity perspective, according to
their ability to transport ATM cells.

subNMS
Subnetwork Management System: A Network Management System that is managing one or more
subnetworks and that is managed by one or more Network Management Systems.

Summary Address
An address prefix that tells a node how to summarize reachability information.

SUT
System Under Test: The real open system in which the Implementation Under Test (IUT) resides.

SVC
Switched Virtual Circuit: A connection established via signaling. The user defines the endpoints
when the call is initiated.

SVCC
Switched Virtual Channel Connection: A Switched VCC is one which is established and taken
down dynamically through control signaling. A Virtual Channel Connection (VCC) is an ATM
connection where switching is performed on the VPI/VCI fields of each cell.

SVE
SAP Vector Element: The SAP address may be expressed as a vector, (ATM_addr, ATM_selector,
BLLI_id2, BLLI_id3, BHLI_id), where:
1) ATM_addr corresponds to the 19 most significant octets of a devices 20-octet ATM address
(private ATM address structure) or the entire E.164 address (E.164 address structure)
2) ATM_selector corresponds to the least significant octet of a devices 20-octet ATM address
(private ATM address structure only)
3) BLLI_id2 corresponds to an octet in the Q.2931 BLLI information element that identifies a
layer 2 protocol
4) BLLI_id3 corresponds to a set of octets in the Q.2931 BLLI information element that identify
a layer 3 protocol
5) BHLI_id corresponds to a set of octets in the Q.2931 BHLI information element that identify
an application (or session layer protocol of an application)
Each element of the SAP vector is called a SAP Vector Element, or SVE. Each SVE consists of a
tag, length, and value field.

SVPC
Switched Virtual Path Connection: A Switched Virtual Path Connection is one which is
established and taken down dynamically through control signaling. A Virtual Path Connection
(VPC) is an ATM connection where switching is performed on the VPI field only of each cell.

Switched Connection
A connection established via signaling.

Switching System
A set of one or more systems that act together and appear as a single switch for the purposes of
PNNI routing.

Symmetric Connection
A connection with the same bandwidth value specified for both directions.

T1E1
An ANSI standards sub-committee dealing with Network Interfaces.

T1M1
An ANSI standards sub-committee dealing with Inter-Network Operations, Administration and
Maintenance.

T1Q1
An ANSI standards sub-committee dealing with performance.

T1S1
An ANSI standards sub-committee dealing with services, architecture and signaling.

T1X1
An ANSI standards sub-committee dealing with digital hierarchy and synchronization.

TB
Transparent Bridging: An IETF bridging standard where bridge behavior is transparent to the data
traffic. To avoid ambiguous routes or loops, a Spanning Tree algorithm is utilized.

TBE
Transient Buffer Exposure: This is a negotiated number of cells that the network would like to
limit the source to sending during startup periods, before the first RM-cell returns.

TC
Transaction Capabilities: TCAP (see below) plus supporting Presentation, Session and Transport
protocol layers.

TC
Transmission Convergence: The TC sublayer transforms the flow of cells into a steady flow of
bits and bytes for transmission over the physical medium. On transmit, the TC sublayer maps the
cells to the frame format, generates the Header Error Check (HEC), sends idle cells when the
ATM layer has none to send. On reception, the TC sublayer delineates individual cells in the
received bit stream, and uses the HEC to detect and correct received errors.

TCAP
Transaction Capabilities Applications Part: A connectionless SS7 protocol for the exchange of
information outside the context of a call or connection. It typically runs over SCCP and MTP 3.

TCP
Transmission Control Protocol: Originally developed by the Department of Defense to support
interworking of dissimilar computers across a network. A protocol which provides end-to-end,
connection-oriented, reliable transport layer (layer 4) functions over IP controlled networks. TCP
performs the following functions: flow control between two systems, acknowledgements of
packets received and end-to-end sequencing of packets.

TCP
Test Coordination Procedure: A set of rules to coordinate the test process between the lower tester
and the upper tester. The purpose is to enable the lower tester to control the operation of the upper
tester. These procedures may, or may not, be specified in an abstract test suite.

TCR
Tagged Cell Rate: An ABR service parameter, TCR limits the rate at which a source may send
out-of-rate forward RM-cells. TCR is a constant fixed at 10 cells/second.

TCS

Transmission Convergence Sublayer: This is part of the ATM physical layer that defines how
cells will be transmitted by the actual physical layer.

TDF
An ABR service parameter, TDF controls the decrease in ACR associated with TOF. TDF is
signaled as TDFF, where TDF = TDFF/RDF times the smallest power of 2 greater or equal to
PCR. TDF is in units of 1/seconds.

TDFF
Refer to TDF. TDFF is either zero or a power of two in the range 1/64 to 1 in units of 1/cells.

TDM
Time Division Multiplexing: A method in which a transmission facility is multiplexed among a
number of channels by allocating the facility to the channels on the basis of time slots.

TE
Terminal Equipment: Terminal equipment represents the endpoint of ATM connection(s) and
termination of the various protocols within the connection(s).

TLV
Type / Length / Value: A coding methodology which provides a flexible and extensible means of
coding parameters within a frame. Type indicates parameter type. Length indicates parameters
value length. Value indicates the actual parameter value.

TM
Traffic Management: Traffic Management is the aspect of the traffic control and congestion
control procedures for ATM. ATM layer traffic control refers to the set of actions taken by the
network to avoid congestion conditions. ATM layer congestion control refers to the set of actions
taken by the network to minimize the intensity, spread and duration of congestion. The following
functions form a framework for managing and controlling traffic and congestion in ATM
networks and may be used in appropriate combinations:
1) Connection Admission Control
2) Feedback Control
3) Usage Parameter Control
4) Priority Control

5) Traffic Shaping
6) Network Resource Management
7) Frame Discard
8) ABR Flow Control

TMP
Test Management Protocol: A protocol which is used in the test coordination procedures for a
particular test suite.

TNS
Transit Network Selection: A signaling element that identifies a public carrier to which a
connection setup should be routed.

TOF
Time Out Factor: An ABR service parameter, TOF controls the maximum time permitted between
sending forward RM-cells before a rate decrease is required. It is signaled as TOFF where
TOF=TOFF+1. TOFF is a power of 2 in the range: 1/8 to 4,096.

TOFF
Time Out Factor: Refer to TOF.

Topology Aggregation
The process of summarizing and compressing topology information at a hierarchical level to be
advertised at the level above.

Topology Attribute
A generic term that refers to either a link attribute or a nodal attribute.

Topology Constraint
A topology constraint is a generic term that refers to either a link constraint or a nodal constraint.

Topology Database
The database that describes the topology of the entire PNNI routing domain as seen by a node.

Topology Metric
A generic term that refers to either a link metric or a nodal metric.

Topology State Parameter

A generic term that refers to either a link parameter or a nodal parameter.

TP-MIC
Twisted-Pair Media Interface Connector: This refers to the connector jack at the end user or
network equipment that receives the twisted pair plug.

TPCC
Third Party Call Control: A connection setup and management function that is executed from a
third party that is not involved in the data flow.

Trail
An entity that transfers information provided by a client layer network between access points in a
server layer network. The transported information is monitored at the termination points.

Trailer
Protocol control information located at the end of a PDU.

Transit Delay
The time difference between the instant at which the first bit of a PDU crosses one designated
boundary and the instant at which the last bit of the same PDU crosses a second designated
boundary.

Trm
An ABR service parameter that provides an upper bound on the time between forward RM-cells
for an active source. It is 100 times a power of two with a range of 100*2-7 to 100*20.

TS
Transport Stream: One of two types of streams produced by the MPEG-2 Systems layer. The
Transport Stream consists of 188 byte packets and can contain multiple programs.

TS

Traffic Shaping: Traffic Shaping is a mechanism that alters the traffic characteristics of a stream
of cells on a connection to achieve better network efficiency, while meeting the QoS objectives,
or to ensure conformance at a subsequent interface. Traffic shaping must maintain cell sequence
integrity on a connection. Shaping modifies traffic characteristics of a cell flow with the
consequence of increasing the mean Cell Transfer Delay.

TS
Time Stamp: Time Stamping is used on OAM cells to compare time of entry of cell to time of
exit of cell to be used to determine the cell transfer delay of the connection.

TTCN
Tree and Tabular Combined Notation: The internationally standardized test script notation for
specifying abstract test suites. TTCN provides a notation which is independent of test methods,
layers and protocol.

UBR
Unspecified Bit Rate: UBR is an ATM service category which does not specify traffic related
service guarantees. Specifically, UBR does not include the notion of a per-connection negotiated
bandwidth. No numerical commitments are made with respect to the cell loss ratio experienced
by a UBR connection, or as to the cell transfer delay experienced by cells on the connection.

UDP
User Datagram Protocol: This protocol is part of the TCP/IP protocol suite and provides a means
for applications to access the connectionless features of IP. UDP operates at layer 4 of the OSI
reference model and provides for the exchange of datagrams without acknowledgements or
guaranteed delivery.

UME
UNI Management Entity: The software residing in the ATM devices at each end of the UNI
circuit that implements the management interface to the ATM network.

Unassigned Cells
A cell identified by a standardized virtual path identifier (VPI) and virtual channel identifier
(VCI) value, which has been generated and does not carry information from an application using
the ATM Layer service.

UNI
User-Network Interface: An interface point between ATM end users and a private ATM switch, or
between a private ATM switch and the public carrier ATM network; defined by physical and

protocol specifications per ATM Forum UNI documents. The standard adopted by the ATM
Forum to define connections between users or end stations and a local switch.

Unicasting
The transmit operation of a single PDU by a source interface where the PDU reaches a single
destination.

UPC
Usage Parameter Control: Usage Parameter Control is defined as the set of actions taken by the
network to monitor and control traffic, in terms of traffic offered and validity of the ATM
connection, at the end-system access. Its main purpose is to protect network resources from
malicious as well as unintentional misbehavior, which can affect the QoS of other already
established connections, by detecting violations of negotiated parameters and taking appropriate
actions.

Uplink
Represents the connectivity from a border node to an upnode.

Upnode
The node that represents a border node outside neighbor in the common peer group. The upnode
must be a neighboring peer of one of the border nodes ancestors.

UT
Upper Tester: The representation in ISO/IEC 9646 of the means of providing, during test
execution, control and observation of the upper service boundary of the IUT, as defined by the
chosen Abstract Test Method.

UTOPIA
Universal Test & Operations Interface for ATM: Refers to an electrical interface between the TC
and PMD sublayers of the PHY layer.

UTP
Unshielded Twisted Pair: A cable having one or more twisted pairs, but with no shield per pair.

VBR
Variable Bit Rate: An ATM Forum defined service category which supports variable bit rate data
traffic with average and peak traffic parameters.

VC
A communications channel that provides for the sequential unidirectional transport of ATM cells.

VCC
Virtual Channel Connection: A concatenation of VCLs that extends between the points where the
ATM service users access the ATM layer. The points at which the ATM cell payload is passed to,
or received from, the users of the ATM Layer (i.e., a higher layer or ATM-entity) for processing
signify the endpoints of a VCC. VCCs are unidirectional.

VCI
Virtual Channel Identifier: A unique numerical tag as defined by a 16 bit field in the ATM cell
header that identifies a virtual channel, over which the cell is to travel.

VCL
Virtual Channel Link: A means of unidirectional transport of ATM cells between the point where
a VCI value is assigned and the point where that value is translated or removed.

VCO
Voltage Controlled Oscillator: An oscillator whose clock frequency is determined by the
magnitude of the voltage presented at its input. The frequency changes when the voltage changes.

VD
Virtual Destination. Refer to VS/VD.

VF
Variance Factor: VF is a relative measure of cell rate margin normalized by the variance of the
aggregate cell rate on the link

Virtual Channel Switch

A network element that connects VCLs. It terminates VPCs and translates VCI values. It is
directed by Control Plane functions and relays the cells of a VC.

Virtual Path Switch

A network element that connects VPLs. It translates VPI (not VCI) values and is directed by
Control Plane functions. It relays the cell of the VP.

VLAN
Virtual Local Area Network: Work stations connected to an intelligent device which provides the
capabilities to define LAN membership.

VP
Virtual Path: A unidirectional logical association or bundle of VCs.

VPC
Virtual Path Connection: A concatenation of VPLs between Virtual Path Terminators (VPTs).
VPCs are unidirectional.

VPI
Virtual Path Identifier: An eight bit field in the ATM cell header which indicates the virtual path
over which the cell should be routed.

VPL
Virtual Path Link: A means of unidirectional transport of ATM cells between the point where a
VPI value is assigned and the point where that value is translated or removed.

VPT
Virtual Path Terminator: A system that unbundles the VCs of a VP for independent processing of
each VC.

VS
Virtual Scheduling: Virtual Scheduling is a method to determine the conformance of an arriving
cell. The virtual scheduling algorithm updates a Theoretical Arrival Time (TAT), which is the
"nominal" arrival time of the cell assuming that the active source sends equally spaced cells. If
the actual arrival time of a cell is not "too" early relative to the TAT, then the cell is conforming.
Otherwise the cell is non-conforming.

VS
Virtual Source. Refer to VS/VD.

VS/VD
Virtual Source/Virtual Destination: An ABR connection may be divided into two or more
separately controlled ABR segments. Each ABR control segment, except the first, is sourced by a

virtual source. A virtual source implements the behavior of an ABR source endpoint. Backwards
RM-cells received by a virtual source are removed from the connection. Each ABR control
segment, except the last, is terminated by a virtual destination. A virtual destination assumes the
behavior of an ABR destination endpoint. Forward RM-cells received by a virtual destination are
turned around and not forwarded to the next segment of the connection.

VTOA
Voice and Telephony Over ATM: The ATM Forum voice and telephony over ATM service
interoperability specifications address three applications for carrying voice over ATM networks;
desktop (or LAN services), trunking (or WAN services), and mobile services.

WAN
Wide Area Network: This is a network which spans a large geographic area relative to office and
campus environment of LAN (Local Area Network). WAN is characterized by having much
greater transfer delays due to laws of physics.

XDF
Xrm Decrease Factor: An ABR service parameter, XDF controls the decrease in ACR associated
with Xrm. It is a power of two in range:
[0, 1].

Xrm
An ABR service parameter, Xrm limits the number of forward RM-cells which may be sent in the
absence of received backward RM-cells. The range is 0-255.

Bibliography
ATM Forum, The. The Anchorage Accord. New York: McGraw-Hill, Inc., 1997.
Bigelow, Stephen J. Understanding Telephone Electronics. Carmel, Indiana: SAMS, 1992.
Black, Uyless. Data Link Protocols. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1993.
Black, Uyless. TCP/IP and Related Protocols. New York: McGraw-Hill, Inc., 1992.
Dungan, Frank R. Electronic Communications Systems. Albany, N.Y.: Delmar Publishers, Inc.,
1993.
Handel, Rainer, and Manfred Huber. Integrated Broadband Networks: An Introduction to ATMBased Networks. Reading, Mass.:Addison-Wesley, 1991.

Newton, Harry. Newtons Telecom Dictionary. New York: Telecom Library, 1989.
Pecar, Joseph A., Roger J. O'Conner, and David A. Garbin. Telecommunications Factbook. New
York: McGraw-Hill, Inc., 1993.
Smouts, M. Packet Switching Evolution From Narrowband to Broadband ISDN. Boston: Artech
House, 1992.
Stallings, William. Networking Standards: A Guide to OSI, ISDN, LAN, and MAN Standards.
Reading, Mass.: Addison-Wesley, 1993.
Stallings, William. Handbook of Computer-Communications Standards. vol. 1. New York:
Macmillan, 1987.
Summers, Charles K. ISDN Implementors Guide. New York: McGraw-Hill, Inc., 1995.
Taylor, D. Edgar. The McGraw-Hill Internetworking Handbook. New York: McGraw-Hill, Inc.,
1995.
Wakerly, John F. Digital Design Principles and Practice. Englewood Cliffs, N.J.: Prentice Hall,
1990.

Index
A
AAL1, 115-116, 131
AAL2, 115-116
AAL3/4, 115-116
AAL5, 115-116, 131
AAL Layer, 113, 115-116
ABR, 126, 130
AC, 24, 25, 182
access coordination, 182
Access Tandem, 101
ADC, 30
A/D converter, 98
ADSL, 9-14, 16, 21-23, 54, 100, 102, 155, 191, 200
framing, 213
modem, 209-210
system architecture, 210-211
transport capacity, 212
ADSL Network Termination, 207-210
aggregators, 157
alternating current, 24
analog modem, 10, 15-16, 99

analog signal, 25-26, 30, 32, 39, 55

analog-to-digital converter, 30, 98
ANT, 207-208
API, 74, 87
Application Layer, 73, 87-88
application programming interface, 74, 87
Asymmetric Digital Subscriber Line, 9 See also ADSL
asynchronous transmission, 47, 49-50
Asynchronous Transmission Mode, 7 See also ATM
ATM, 7, 10-14, 16-18, 21-23, 41, 54, 89, 95, 99, 102-103, 155-159, 167
ATM
AAL1, 115-116, 131
AAL2, 115-116
AAL3/4, 115-116
AAL5, 115-116, 131
AAL Layer, 113, 115-116
Adaptation Layer, 116
backbone network, 152
benefits, 11-12, 170-172
cell, 101, 118-124
cell header, 119, 125
cell types, 119
"cloud," 151-152
customers, 176
equipment suppliers, 174-175
fiber based, 20
growth, 168-170
header, 113, 120
interfaces, 119-121, 135
interworking, 156-159
layer, 114
network, 21
network evolution, 169
network interfaces, 119, 135-136
non-recurring charges, 181
over ADSL, 214-215
physical layer, 114
planes, 116-118
players, 172-174
ports, 178-179
protocol layers, 113-116
PVC, 12
rate structure, 180-185
recurring charges, 181-183
router, 101
service revenues, 177
services, 14, 179

switch, 125, 148-150

traffic, 179-180
traffic management and control, 132-135
ATM "cloud," 151-152
ATM planes, 116-118
Control Plane, 118
Layer Management subplane, 79, 113, 118
Management Plane, 118
Plane Management subplane, 118
User Plane, 117
audio, 15, 16, 142-143
available bit rate, 126
average, 25

B
B-ISDN, 18, 40, 57
backbone network, 11, 151-152
backhaul charges, 183
bandwidth, 16-18, 33, 38-40, 49, 141, 145
big endian, 89
binary, 27
bit, 28
bitswapping, 206
bit timing, 115
bridge, 97, 103
Broadband ISDN, 18, 40, 57-58
broadband services, 18
broadcast, 7
buffers, 79
burstiness, 35-36
bursty, 33, 67
business drivers, 12
bus topology, 107-108
byte, 28

C
carrier, 19, 26
CATV, 6, 7, 16
CBR, 66, 69, 130-131
CCITT, 73
cell
address, 114
delay, 131
delineation, 115

demultiplex, 115
header, 114, 115
loss, 131
multiplex, 115
rate decoupling, 115
relay, 69
VPI/VCI translation, 115
cell delay variation, 127
cell loss priority, 114, 120, 121, 126
cell loss probability, 127
central office, 9
switch, 62, 101
channel designations, 37
circuit emulation, 14, 162
circuit switching, 66
class of service, 15
CLP, 114, 121, 126
coaxial cable, 4, 7, 16
common carrier, 7
common control, 63
communication
networks, 7, 33, 91
protocols, 109-110
services, 6, 33
systems, 4, 5, 6, 19, 27, 97
technology, 23
communications, 2, 3, 5, 19
Community Antenna Television, 6
competition, 163
compression, 40
computer, 18
computer technology, 19
concentrators, 157
congestion control, 113
connection oriented connection, 115-116, 124, 129, 133
connection switching, 63
connectionless oriented connection, 115-116, 124, 129, 133
connectionless switching, 66, 67
constant bit rate, 69, 129
Control Plane, 118
copper wire, 11, 16, 18, 58
CPE, 49, 128, 153
managed, 185
CRC, 92
crossbar switching, 63
CSU, 136
customer premises equipment, 49

convergence, 115
cyclic redundancy check, 92

D
DA, 210
DAC, 30
D/A converter, 99
data, 16, 19
access, 15
communication, 5, 6, 19, 99
integrity, 77
radio network, 99
rate, 17, 18, 37, 38
transmission, 33
database access sharing, 13
Data Exchange Interface, 136
datagram, 67
Data Link Layer, 73, 77-80
DBS, 6, 7, 60, 188
DC, 23
DE, 114, 139
defense, 12
Defense Information Services Network, 169
Delta-2 Switching Matrix, 65
demodulation, 26
demultiplexing, 53-54
destination, 8, 19
digital
bits, 40
channel bank, 46
data, 4, 5, 6
devices, 29
modulation, 99
pulses, 31
radio, 99
signal, 27-30, 32, 39, 55
signal processing, 9
terminal interface, 98
Digital Subscriber Line, 9, 191 See also DSL
digital-to-analog converter, 30
Direct Broadcast Satellite, 6, 7, 189 See also DBS
direct current, 24
Discard Enable, 114, 134
Discrete Multi-Tone coding, 204
DISN, 169

DMT, 204
DOS, 70, 86
downstream, 16
DS-1, 35
DSL, 9, 198-199
technology, 194-198
types, 198-204
DSU, 136
DTI, 98, 99
duplex, 49, 98
DXI, 136

E
echo cancellation, 205-206
effective transparency, 66
electrical
current, 24
voltage, 24
electromagnetic spectrum, 38
e-mail, 90
encoding information, 9
endian
big, 89
little, 89
end-to-end, 7
error control, 69
error recovery, 68
Ethernet, 16
II framing, 94
Gigabit, 163, 164
SNAP framing, 94
equipment manufacturers, 10

F
FDM, 37, 44, 52-54, 205
fiber transmission, 7
flow control, 82
frame, 89-91
adaptation, 115
IEEE 802.3, 93
Ethernet II, 94
Ethernet SNAP, 94
generation, 115
lengths, 118

recovery, 115
structures, 93-95
Token Ring, 95
frame relay, 7, 68, 118, 163-165
high speed, 163
framing, 79, 213
frequency, 26
frequency division multiplexing, 37, 44, 52-53, 205

G
GAN, 7, 11, 17, 102
gates
AND, 29
NAND, 29
OR, 29
NOR, 29
gateway, 97, 99, 103
generic flow control, 119, 121
GFC, 119, 120, 121
global area network, 7 See also GAN
global
backbone network, 11
community, 11
information transmission system, 11
language, 11
unification, 22

H
half duplex, 44
HDLC, 80
HDSL, 202
HDTV, 5
High Data Rate Digital Subscriber Line, 202
high level online adaptation, 206
header, 74
header error check, 124
header error control, 120-121
HEC, 121
hierarchical topology, 105-106
high logic level, 27
high-speed data services, 18, 141-148
hub, 97, 99, 103

idle cells, 119

IDN, 56
IEC, 100
IEEE 488, 72
IEEE 802, 93, 94
IEEE 803, 72, 79
IMA, 150
information, 19
encoding, 4
services, 7
superhighway, 11, 12, 21
transport, 6, 7
Integrated Services Digital Network, 6, 55, 56 See also ISDN
Interexchange Carriers, 100, 102
International Standardization Organization, 111
Internet, 11, 15-16, 91, 168
Internet services, 16
interworking,
ATM, 156-159
platform, 14
inverse multiplexing over ATM, 150
IP, 158
IPX, 90
IPX/SPX, 90
ISDN, 6, 15, 18, 55-58
Broadband, 40, 57-58
Narrowband, 40, 56-57
ISO, 111
ISP, 15
ITU-T, 18, 73

L
LAN, 6, 11, 16-17, 97, 102
applications, 16
emulation, 130
LANE, 130
LATA, 100
latency, 131
Layer Management Plane, 79, 113, 118
leading edge, 29
leased lines, 164
LEC, 100, 102
legacy networks, 11
light wave, 4, 61-62
little endian, 89

LLC, 79
LLD, 78
Local Access and Transport Area, 100
local area network, 6, 7, 16 See also LAN
Local Exchange Carriers, 100
local loop, 11
local subscribers, 101
Logical Link Control Sublayer, 79
long-haul carrier, 19
low level driver, 75, 78
low logic level, 27
low-speed data, 141

M
MAC, 79
MAN, 7, 11, 17, 97, 102
managed customer premises equipment, 185-186
Management Information Base, 81, 141
Management Plane, 118
maximum burst size, 126
maximum cell delay, 126-127
MBS, 126, 128
MCR, 126
mean time to repair, 152-153
Media Access Control, 79
mesh topology, 107
metropolitan area network, 7 See also MAN
MIB, 81, 141
microwave transmission, 4, 60
minimum cell rate, 126
modulating frequency, 26
modulation, 26, 100
MRCS, 68
MSP, 156
MTTR, 152-153
multicasting, 134-135
multimedia, 5, 16, 146-147
applications, 16
services, 10
multiple services platform, 156
multiplexing, 33-37, 45-46, 50-53, 68, 114
multirate circuit switching, 68

N-ISDN, 18, 40, 77

Narrowband ISDN, 40, 56-57
network interface See NNI and UNI
Network Layer, 73, 80-82
network management, 155-156
Network Node Interface, 135, 151
networks, 16, 19
network topologies, 103-109
nibble, 28
NNI, 119, 135, 148, 151
noise generation, 46
numbering system, 27
Nyquist rate, 31

O
OAM, 74
octet, 28
one-stop shopping, 185
open-shortest-path-first, 80
Open Systems Interconnection model, 73, 111 See also OSI model
Operations Administration and Management, 74
optical transmission, 7, 47
OSI model, 73-88, 111-113
oversubscription, 186

P
packet, 89, 91
assembler/disassembler, 87
switching, 7, 67
paging systems, 60
PAM, 32
passband, 39-40
payload, 118
type indicator, 121
PCM, 45, 47
PCR, 125, 128, 130, 139-140
PDH, 47, 66
PDM, 33
PDU, 90-92
peak cell rate, 125, 139-140
peak-to-peak, 25
peer-to-peer communications, 112
period, 28
permanent virtual circuit, 14, 91 See also PVC

personal communications, 5, 60
Physical Layer, 36, 73, 75-77
physical link control, 77
physical medium, 115
Physical Service Access Points, 75
Plain Old Telephone Service, 5, 34
Plane Management Subplane, 118
platform interworking, 14
Plesciosynchronous Digital Hierarchy, 47
Point of Presence, 181
point-to-point topology, 104
POP, 181
port
charges, 182-183, 189
oversubscription, 186
speed, 132
POTS, 5, 16, 34, 35
splitter, 211-212
Presentation Layer, 73, 86-87
primary rate, 18
primitive, 78
Protocol Data Unit, 90
protocols, 71-72, 89-90, 109-110
provisioning, 124
PSAP, 75
PSTN, 5, 35
user characteristics, 34
PTM, 32
Public Switched Telephone Network, 5, 19, 34, 97, 100
pulse
amplitude modulation, 31
code modulation, 45, 47
demodulation, 31
duration modulation, 33
modulation, 31
time modulation, 32
train, 31
width modulation, 33
PVC, 91, 129, 133
charge, 182
rate, 183
PWM, 33

Q
QoS, 33-34, 67, 129-130, 159-160

quality of service See QoS

quantizing error, 46

R
radio wave, 4
radio wave propagation, 19, 60
RBOC, 100
reassembly, 115
redundancy enhancement, 68
reed relay, 63
Regional Bell Operating Company, 100
registers, 78
reliability levels
simple, 82
multiplexing, 82
basic error recovery, 82
error recovery and multiplexing, 82
error detection and recovery class, 82
remote
file transfers, 90
logon services, 90
request
for circuit activation, 78
for circuit deactivation, 78
to transfer data, 78
RF communications, 19
RF propagation, 59
ring topology, 108-109
RMS, 35
router, 97, 99, 103

S
sampling period, 32
SAP, 93
satellite transmission, 4
scaleable, 12, 171
SCR, 125, 128, 130, 140
SDH, 47, 61
SDSL, 204
seamless, 55, 56
segmentation, 90, 115
serial communication, 49
Service Advertising Protocol, 93
service classes, 14, 128-131

service provider, 14, 19, 21

Session Layer, 73, 84-85
set-top box, 10
shared access, 162
shielded twisted pair, 59, 77
signal
conditioning, 9
conversion, 30
multiplexing, 33
signaling, 124-125
Single Line Digital Subscriber Line, 204
sinusoidal waveform, 25
SMDS, 163, 164
SONET, 47, 61
source, 8, 19
space division switching, 63
spectrum, 38
splitter, POTS, 211
star topology, 104-105
step-by-step switching, 63
stereophonic audio, 15
STM, 47
STP, 59, 77
Strowger switch, 62-63
sustainable cell rate, 125, 140-141
SVC, 91, 133, 160-161
SVCC, 160
SVPC, 160, 162
switch, central office, 101
switched
channel, 36
multimegabit data service, 163
networks, 34
virtual circuit, 14, 91, 160-161
virtual circuit connection, 160
virtual path connection, 160, 162
switches, 19
switching, 19, 23, 42, 62-68, 114
switching technology, 7
synchronous transmission, 47-49
Synchronous Digital Hierarchy, 47 See also SDH
Synchronous Optical Network, 47 See also SONET
synchronous time division multiplexer, 47
system capacity, 36
systems, 19

T
TCP/IP, 89-90
TDM, 37, 46-47, 52-54, 66, 68
carrier standards, 37
telecommunications, 19
industry, 21
telegraph, 3, 22
telephone, 3-4, 22
telephone company, 7
time division multiplexing, 37 See also TDM
token bus, 16
token ring, 16
framing, 95
topology, 103-109
bus, 107-108
hierarchical, 105-106
mesh, 107
point-to-point, 104
ring, 108-109
star, 104-105
traffic shaping, 127-128
trailer, 74
trailing edge, 29
transistors, 45
transmission systems, 2-4, 7, 43-50, 56
transmission technology, 5, 7
transparency, 8
Transport Layer, 73, 82-84
trunk, 101

U
UART, 99
UBR, 130-131
unassigned cells, 119
UNI, 119, 135, 153, 160
Universal Asynchronous Receiver/Transmitter, 99
UNIX, 70, 86
unshielded twisted pair, 59, 76
unspecified bit rate, 130
UPC, 134
upper cutoff frequency, 40
usage based billing, 187-189
usage parameter control, 134
User Network Interface, 135, 148, 153-154

User Plane, 117

UTP, 59, 76

V
variable bit rate, 68, 129 See also VBR
VBR, 68, 125, 130-131
VBR-nrt, 115, 130
VBR-rt, 115, 130
VCI, 120-124
VDSL, 203-204
Very High Rate DSL, 203-204
video, 5, 12, 15-16, 143-146
applications, 16
virtual
channel, 14, 121-123
channel identifier, 120-124
circuit, 67
path, 121-123
path identifier, 14, 120-124
terminal, 88
voice, 16
Voice Over Frame Relay protocol, 119
VPI, 120-124
VP OAM traffic cells, 119
VP traffic cells, 119

W
WAN, 7, 11, 17, 97, 102
wide area network, 7 See also WAN
wireless systems, 45, 46, 60
word, 28