Background
for Telephone Switching
2nd Edition (Revised and Expanded)
Webmaster's
note: We apologize for the poor quality of many of the diagrams
in this web edition. In some cases they were so poor that we left
them out entirely. If we are able to obtain better originals, we
post the omitted diagrams and clearer versions of those that are
here now.
The second
edition of Background for Telephone Switching was originally
published in 1997 by AVO Training Institute, who have kindly granted
permission to post it on LeeGoeller.com. A catalog of AVO's books
and training materials can be found online at
AVO Technical Resource Center.
This book is my
“magnum opus,” my nearest approach to a best seller. I started
writing it in 1975, almost as soon as I left RCA to go into
consulting and found myself free of constraints that tend to come
from employers. Joe Aiken, then editor of Telephony magazine, saw my
manuscript, and wanted it. He had just purchased the rights to Frank
E. Lee’s “ABC of the Telephone,” a four-volume set widely used for
technician training in the telephone industry, from Lee’s estate,
and wanted to modernize it, add to the series, and ultimately go
into the publishing business for himself.
My switching book
was just what he was looking for. The first edition, called “Design
Background for Telephone Switching,” came out in 1977 as volume 9 in
the ABC series, and ultimately sold over 17,000 copies, not bad
considering its esoteric nature and that it was sold only by direct
mail from ABC and Harry Newton’s Telecom Library. ABC ultimately
became “abc Teletraining, Inc.,” with a large catalog of books,
videos, workshops, etc.
For the second
edition, which I had pretty well finished by 1994, we dropped
“Design” from the title, and I expanded the book to almost twice it
s original size, adding much new material I had encountered in my
consulting practice, work with various manufacturers, business
telephone customers, and telephone companies. The new version was
directed to a wider range of telecom professionals than just the
design community, and was finally published in 1997.
In 2001, Joe
decided to retire and sold his abc Teletraining business to AVO
International, a subsidiary of Megger. AVO provides training
material for the electric power industry, and the expansion into
telecom was a logical move. They still have Background for Switching
available in book form, along with Voice Communication in Business
(Volume 1). I hope that some readers of this web site will want to
have these books, and others of the abc line, in actual paper form
which may be easier to read in an airplane or in an easy chair in
the living room.
But whether it is
read in electronic or paper form, I hope that some of those
designing telephony to run on the Internet will take advantage of my
efforts over the past 40 years or so. As the electronic generation
of 1975 found out the hard way, technology alone is not enough. You
have to know what it is expected to do if you want it to sell. I may
not be up on the latest trends in VoIP or TCP/IP, but I have
documented many of the things that telecom technology must do to
satisfy the customer. If the latest technological triumphs can’t do
them, the customer might as well buy cans and strings.
Good luck!
Lee Goeller, 2006
After a number of
years as an announcer-writer-technician ìn radio broadcasting and a
year in the Air Force, Lee Goeller obtained a bachelor's and
master's degree in Electrical Engineering from the University of
Virginia.
He spent the next
12 years at Bell Telephone Laboratories working on 1ESS and its
predecessor, the system used in the historic Morris, IL, field
trial. Moving to RCA, he spent some time in the design of military
switching systems, but as a member of the RCA Corporate Telecom
Group, he took advantage of a golden opportunity to learn about
telephony from the viewpoint of the large industrial customer.
Since 1974, he
has been an independent telecommunication consultant working with
business customers as well as designers of telecom equipment. During
this time, he has done seminars on PBXs, telecom management, and
traffic for Business Communications Review, and written many
articles for BCR, TeleConnect, TeleManagement and other industry
journals.
In addition to
the design and application of modern technology to solve human
problems, Lee has studied the historical development of
telecommunications, collecting a library of the works of early
telephone engineers in the process. Because technology, like
politics, is meaningless without its historical context, he has made
every effort to take advantage of the wisdom of the past as well as
what is currently fashionable in presenting technical developments
to both laymen and professionals.
Lee has spent the
last 20 years attempting to demonstrate two basic concepts: first,
that technical writing need not be dull, and second, it is possible
to convey the ideas behind modern technology without getting bogged
down in mathematics. This book is an example of his efforts.
Switching system
design is an art, not a science. And although it depends on various
branches of engineering, mathematics, cost accounting, industrial
psychology and data processing, it is something more. A switching
system is complex; it responds to stimuli from its environment in a
way that is not unlike a living creature, and it interacts with
others of its kind to form what may well be the most complex
creation of the human mind: the world communication network.
This book will
not teach you to design, operate or use switching systems. What it
is intended to do is provide background information--a context, if
you will--to facilitate understanding of the many far more technical
treatises available. Most of these books (for a selection, see the
bibliography) describe what has been done. What this book hopes to
do is elaborate on what is required, what is expected of a switching
system. It contains the information I wish I had had twenty years
ago when I first started in switch design. Not all the information,
of course. Just the information that I have been able to acquire so
far that applies to the overall problem.
The first draft
was written with young design engineers in mind, but, with the help
of Joseph J. Aiken, abc publisher, and Frank Reese of North
Pittsburgh Telephone Co., I have tried to broaden the scope so that
both operating telephone company personnel and business
communication managers will also find, if not the answers to their
questions, at least the background they need to lead them to the
answers elsewhere. Knowing what question to ask is often half the
battle.
However, a word
of warning is needed. This is a very personal book. It represents,
to a large extent, my own opinions about how things are and how they
ought to be. As a person with many years of fairly varied
experience, not only as a circuit and system designer, but also as a
customer and user of switching equipment, my opinion has something
to recommend it. But it is still open to challenge.
In the vastness
of the telephone business, there are almost infinite combinations
and permutations of "standard" methods of operation. My experience
is limited, and I have written about what I know, and what many
friends and associates have been willing to share with me. Doubtless
many readers will take issue with some point or another, and many
will have experience that is directly counter to mine. It is likely
that, in such cases, we are both right--and both wrong. There are
many ways to skin a cat, and a great many cats in the telephone
business.
But if you
disagree with some point, or feel that I have omitted something of
importance, I would like very much to hear from you. Progress comes
from applied experience; it is not deduced by mathematics or logic
from abstract principles. And although we learn from our own
experience, it is much faster and cheaper to learn from the
experiences of others. There is a need for dialogue so that those of
us presently in the business can capture the basic points and
preserve them for those who will come along in future years. If we
communicate, it won't be hard to minimize the reinvention of the
wheel--or at least to limit praise for square corners. Let me hear
from you.
Lee Goeller
1977
When I wrote the
first edition of Background for Telephone Switching, my intent was
to stress the functions which must be performed by switching systems
rather than to catalog all known ways these functions have been
implemented in the past. I hoped to make the book more or less
hardware independent; after all, if you know what you want to do,
you can do it with any devices, but if you don't know what you want
to do, neither hardware nor software can save you.
From a different
point of view, it seemed to me that a hardware-independent book, in
addition to stressing fundamentals rather than ephemerals, would
have a longer useful life. That I succeeded to a certain extent is
demonstrated by sales of over 1000 copies a year for many years
since 1977.
Since the first
edition came out, however, the telephone industry had been turned
upside down by political and technological changes the like of which
it has never before experienced. These changes have not altered the
functions a telephone switch must perform, but they have provided
both constraints on performance as well as new ways to expand well
known functions into something that appears quite different. Thus
“Background for Telephone Switching” needs an update.
The political
changes brought about by the FCC's deregulation of the telephone
industry and the DOJ's divestiture of the Bell Operating Companies
from AT&T, all intended by lawyers and economists to make a "level
playing field" for the benefit of new competitors, have had a
powerful impact of the services actually available to the customer.
The changes that resulted have not always been beneficial and, in
some cases, have almost certainly delayed advances that technology
might have brought us.
The technology,
on the other hand, has been a joy to behold. My grandfather, an
engineer who specialized in the construction of steel frames for
skyscrapers, used to subscribe to a magazine called "Iron Age." We
have now left the age of iron and steel and entered the Glass Age,
the age of silicon, with its transistors, integrated circuits,
microprocessors, ROM, RAM, and all the other wonders we now take for
granted.
Although the
transistor was invented at Bell Labs in December, 1947, used
commercially in telephony in 1953 (in a translator for the 4XBAR
toll switch), and made possible "shirt pocket radios" in the early
1960s, it is only since first edition of this book came out that
evolution into LSI has made the personal computer a standard office
(and household) item. Control systems for telephone switches have
used this same technology, taking advantage of really big programs
running on processor chips more powerful than mainframe computers
used to be, and using more memory in a tiny PBX than was found in
the original version of AT&T's 1ESS Central Office Switch.
But silicon
devices for switching and control, important as they are, have to
take second place to optical fiber used for transmission. Optical
fiber is glass, also made from silicon, and it has driven
satellites, the delight of futurists, along with coaxial cable and
microwave, into niche markets, taking over almost the entire field
of interoffice trunking. Indeed, Sprint, one of the new long
distance carriers, blew up one of its nearly new microwave towers in
a TV commercial, just to illustrate how much better optical fiber
really is.
Had the
device-makers not gotten optical fiber to work (they made two or
three unsuccessful attempts in the 1960s and 70s), it is very likely
that the digital revolution in telecommunications would have been
severely limited. The reason is simple: long distance networks based
on microwave can develop many more channels in the limited radio
spectrum available by using analog rather than digital modulation.
When there is no way to pay more and get more, you have to do the
best you can with what you have. Thus digital transmission, although
it dominated short haul trunks on copper from its introduction in
1962, could hardly have been economically competitive on long-haul
microwave, either satellite or terrestrial, except in very special
circumstances.
Optical fiber,
however, has enormous bandwidth in each hair-thin glass strand, and
there is no limit to the number of strands which can be installed on
a given route (actually, labor costs being what they are, the
incremental cost of adding another strand has to be zilch, and "dark
fiber" abounds, waiting for future activation). This lets
bandwidth-consuming digital techniques be used for long distance as
well as local trunks, and T-carrier on glass has taken over.
Starting in 1976,
five years before long-haul digital trunks were available, AT&T
began putting in 4ESS digital toll switches. Northern Telecom and
other companies followed almost immediately; although there are many
ways to design digital switches, manufacturers made a point of using
the same digital format as T-carrier so that multiplexed bit streams
could be taken directly into the switch without being de-multiplexed
into individual trunks or decoded back to analog. When analog
(microwave) trunks were encountered, their channel banks had to be
connected back-to-back with T-carrier channel banks for the
appropriate conversion. Even with this added expense, digital
switching proved in; when optical fiber replaced microwave, AT&T had
digital trunk switching already in place, and the back-to-back
channel banks were retired.
The result was
the possibility of end-to-end full duplex digital connections at 64
thousand bits per second, something of potential interest for data
and image transmission as well as voice. Because some PBXs, starting
about 1975, had also been designed to be compatible with T-carrier,
it began to look as though the millennium was at hand. However,
there were a few hitches.
First, by 1977,
most local customer telephone lines in the U.S. were served by
relatively new 1ESS switches and the smaller 2ESS and 3ESS; older
crossbar and SXS switches still existed in quantity, and all made
connections via two-wire metallic matrices. Thus an analog barrier
stood between digital PBXs and the increasingly digital toll network
of AT&T.
Second, competing
long distance carriers such as MCI and Sprint were based, as was
AT&T, on microwave (analog) long-haul trunks. However, most of the
microwave of the new long distance carriers (Unlike AT&T's
microwave, much of which had been in service for 20 years or more)
had just been installed, and its replacement, like that of the new
ESS switches in the Bell System, posed an economic problem. Sprint
and some of the regional carriers bit the bullet and went quickly to
all-fiber networks, but at a considerable cost.
Third, T-carrier,
although digital, had problems of its own. Designed to provide
low-cost voice transmission, it went through several stages of
improvement and cost reduction, each of which made the handling of
non-voice signals more difficult. In the United States, some of the
bits used for signaling and supervision were taken to improve voice
coding, expanding five out of six voice samples from 7 to 8 bits.
The European version of T-carrier used a separate signaling channel
so that all 8 bits per sample were always available, but both the
European and American systems had a further problem: the inability
to operate when continuous strings of digital zeros, or absences of
pulses, were sent. Repeaters along the line depended on pulses for
synchronization and without them would quickly get out of step,
causing the line to go down. Use of common channel signaling in the
United States is slowly making all 8 bits available as in Europe,
but the solution to the "ones density" problem will take a while
longer.
To further
increase the number of voice channels on T-carrier, two additional
techniques are available: ADPCM and TASI. Adaptive Differential
Pulse Code Modulation can take several samples of digitized speech,
extrapolate what the next one will be, and then send a 4 bit code
representing the difference between the extrapolation and the actual
measurement. This doubles the number of channels from 24 to 48. Time
Assignment Speech Interpolation goes a little further. Based on the
concept that half the customers are listening while the other half
are talking, it uses one conversation's listen path for another's
talk path, keeping all channels busy in both directions and doubling
to 96 the number of simultaneous conversations on a T-span set up
for ADPCM.
Unfortunately,
neither ADPCM nor TASI work very well with anything but speech;
non-speech signals from modems for data and facsimile are
effectively excluded.
But the potential
is there. Digital switches have replaced many SXS and Crossbar local
central offices, and are even replacing early 1ESS installations.
Further, the excess of T-carrier capacity has led business customers
to contract directly for T-spans so that they can build their own
digital networks for voice, data, and combined voice/data.
To add to the
problem, manufacturers of customer premises equipment such as
modems, fax machines, voice mail, etc., are tending to make these
devices look, to the CO or PBX switch on which they home, like
ANALOG telephones. Even digital PBXs and electronic key systems
still depend to a large extent on analog CO trunks, although digital
CO trunks are becoming more common. Because customers are required
to own their own telephones and other equipment, there is no way the
telephone company can speed up replacement of something belonging to
others to take advantage of end-to-end digital transmission, even
when local CO switches are digital. One result of all this is to
lock the customer into ultra-modern devices built to 1950s analog
standards.
In the meantime,
the telephone industry is trying to develop standards for the
Integrated Services Digital Network, or ISDN. The word "Services"
was inserted into the original concept of an IDN, apparently to
satisfy those who wanted residential customers to buy new,
high-profit services using digital switches and trunks already
cost-justified, bought, and installed; as a result, even an IDN has
been delayed from year to year while participants try to achieve
standardization for their company's proprietary approach. In the
meantime, business customers who, unlike residential customers,
actually need digital transmission, are expanding their private
digital networks on rented T-spans.
Now that optical
fiber and digital central office switches have made possible
end-to-end dial-up digital transmission, the telephone industry's
delays of ISDN, encouraging the defection of business customers to
private digital networks, coupled with the computer and software
industry's desire to sell advanced services itself via the
Internet*, may leave the rest of us with a digital telephone system
that carries only analog signals.
[*Footnote.
The Internet is a large packet network for data that grew out of
ARPAnet, developed by the Department of Defense in the 1970s for
use by military contractors and universities doing DOD
"research." Many are lured to access the Internet with tales of
countless files which can be downloaded free from the multitude
of data bases already on the net, but those who hope to make
money by selling data files seem to look upon the Internet as
something here and now as opposed to ISDN which may exist at
some time in the future.]
At present, we
sit on the cusp not only between analog and digital communications,
but also between a voice telephone network which has learned to
carry data and could, in the near future, do this job even better,
and two or more networks, one for voice but the others for various
types of data, the latter hoping to learn to carry voice (where the
money is) as well. But even with digital networks, the huge amount
of new equipment owned by customers which looks to the telephone
network like 40 year old analog telephones, will be with us for some
time.
Perhaps ISDN or
some other new standard will encourage the replacement of the 2500
type telephone set and its look-alikes, as well as answering
machines and the modems for fax and PCs. Or perhaps one of the new
public or private digital networks will take over and develop its
own interface standards. But something is clearly needed to make
world-wide digital end-to-end connectivity happen; the potential of
digital telephone technology to serve the customer is so great that
a considerable effort is demanded.
I hope this
second edition of Design Background will serve as both a prod and
guide to the right direction.
Lee Goeller
1995
About 30 years
ago, Lee Goeller and I first worked together at RCA in the corporate
telecom group. Our job was to modernize the far-flung RCA internal
telephone network which consisted of step-by-step PBXs
interconnected with enormous numbers of tie-trunks obtained under
the Telpak tariff. Although we soon discovered that we could replace
the maintenance-intensive electromechanical equipment with
electronics, the new electronic PBXs of the day turned out to be
utterly unable to provide the features and functions that RCA relied
upon for efficient communications. We quickly came to realize that
no matter how "modern" an electronic PBX might be, it had, at the
very lest, to be able to perform the fundamental functions of its
predecessors.
A glance through
this book might give the casual browser the impression that it is
concerned with older technology that has no place in the current
world. Nothing could be further from the truth. It is concerned with
the fundamental functions and principles on which switching systems
are based, sometimes illustrated in historical context. Knowledge of
these fundamentals is a prerequisite for successful
telecommunications system design, selection and use.
When Frank Reese
wrote the foreword to the first edition, he stated "This book is
more in the nature of a conversation with an experienced telephone
man than it is a text...It affords...an appreciation of what has
gone before, and what may be just around the corner...Any future
plant craftsman, telephone engineer or designer should find (it) a
very helpful preliminary to (the) study of his specialty..(in) our
challenging industry."
Deregulation of
telecommunications and the divestiture of AT&T from the local
telephone operating companies has greatly increased opportunities
for designers and business people outside the telephone industry.
Further, there is a need for deeper understanding on the part of the
customer who must now take a more active role in system selection
and management. Therefore, it seems likely that the second edition
of this book will find an even broader field of application than
Frank Reese originally suggested.
Richard Frank,
Siemens Business Communications Systems
1997
Although it would
take a whole book to thank all the people who have done their best
to explain switching system design to me, I want to take this
opportunity to express my particular thanks to Richard Frank who
read the entire manuscript of the second edition, saved me from a
number of embarrassing errors, and offered logical resistance to
many of my more extreme opinions. Dick and I worked together at RCA,
and when I went into consulting in 1974, he went to Siemens where he
did system engineering for the SD192 and Saturn PBX product lines.
He is now in Technical Planning for ROLM, a Siemens Company, and
chairman of the Telecommunications Industry Association TR41.1
Committee on PBX and KTS Standards, a vantage point from which he
gets a closer look at in-progress R&D than I do. Errors and
eccentricities that remain in the book after all his efforts are, of
course, my responsibility and should not attributed to him.
Lee Goeller
1997
This book is
dedicated to the memory of the pioneers who left a trail of paper
through the wilds of telephone switching:
[ Top ] [
Next Chapter ] [
Table of
Contents ] |
