Voice
Communication in Business Volume 1
Essays on telecommunications,
1969-1980
Chapter 19
Telecommunication
Technology of the Futures
A
magazine called Modern Asia asked me for an article on
communications that might be useful to their readers. At the time, I
was caught up in the wonders of fiber optics, a development that
seems to me more important than even the transistor. However, LSI
and satellites are also major factors; each device as its own
optimum scenario, and they are not quite the same. Thus the title of
the piece.
***
Telecommunications is a
field in wild ferment, with changes taking place at a rate that is
unprecedented. Older technologies such as microwave radio, solid
state circuitry and computer control have moved the field ahead in
the last 15 years. Microwave permitted many communication channels
to be installed inexpensively on heavy routes and small numbers to
be installed economically on thin routes, even over bad terrain. The
transistor and later solid state devices have revolutionized
transmission via cable as well as microwave by replacing most of the
vacuum tubes used in older systems but, equally important, they have
led to electronic switching in telephone central offices.
"Switching," or connecting one caller to another, is now going
electronic around the world.
The main impact of solid
state technology in switching has not been as great on the
electromechanical devices that actually make the speech path from
one caller's line to another as it has on the circuitry that
controls those devices. Computers made of solid state components are
smaller and far more reliable than their vacuum tube predecessors;
thus, their great flexibility and generality can now be used as the
"brain" of a telephone switch. Many new features have become
available as a result, and most standard services have also become
less expensive to implement and more effective to use.
These impacts, however,
can only be described as minor compared to what is in store when the
next wave of technology breaks over us. This wave of the future
includes communication satellites, large scale integration (LSI) and
fiber optics. Unfortunately, each of these developments controls a
future that is somewhat different from the futures of the others.
Thus, the shape of the one actual future that will emerge is by no
means clear at this point in history.
Satellites
The potential of
communication satellites has been recognized by many ever since 1943
when George O. Smith based a series of stories on Venus Equilateral,
a large, manned communication and research satellite at a LaGrange
point in Venus orbit. Smith's stories were science fiction*, of
course, but it hasn't taken technology too long to start to catch
up. We still don't have manned satellites that pay their own way,
but unmanned communication satellites have been making money for a
decade.
[* FOOTNOTE For the Arthur C Clarke
tans in the audience who are wondering lust who George O. Smith is,
let me refer them to Clarke's introduction to the paperback edition
of The Complete Venus Equilateral (Ballantine Books 1976)]
Satellites have been
particularly effective in facilitating communications between
scattered locations in remote areas such as northern Canada and
Alaska, and in areas such as Indonesia where other geographical
barriers make land-line circuits impractical. However, some kinds of
businesses also form a group of far-flung locations: an office here,
a factory there, a warehouse somewhere else. Thus, satellites are
ideally suited to providing private business communication systems
as well as extending public networks.
In either public or
private systems, a technique called "demand assignment" allows a
particular channel through a satellite to be used for a given
conversation and then, when the conversation terminates, to be made
available for other calls. The next call using that channel may very
well go between two completely different earth stations. The
satellite, then, becomes a switch as well as a transmission medium,
and calls can be routed directly, bypassing the complex hierarchies
of switching systems typical of land-line toll networks.
A major factor in
satellite communication is the impact on toll costs. Obviously, it
costs no more to move information through a satellite between San
Francisco and Tokyo than it does to connect San Francisco with Los
Angeles. This will tend to bring the cost of toll calls to a more or
less standard value, independent of distance, a trend already well
under way. That this will facilitate business goes without saying.
To get full utilization
from a satellite, certain modulation techniques are applied to
permit the largest possible number of voice channels to be made
available. These techniques are similar to those used on microwave
radio, and they are designated "analog" as will be explained
subsequently. Unfortunately, analog techniques do not work well with
LSI, or large scale integration, the second factor in the evolution
of the communications future. LSI is primarily a "digital" technique
and, although digital satellites and microwave are being developed,
there are many basic reasons to believe that they may never fully
replace analog techniques in long-distance communications.
Large scale integration
LSI allows very complex
circuits to be produced as a single device. Many hundreds of logic
"gates" can be built and interconnected as part of a piece of
silicon several millimeters square; such collections of gates can be
custom designed to provide specific functions or, when a given level
of flexibility and complexity is required, they can be organized
into microprocessors and memories that are as general in purpose as
a full scale computer system.
Microprocessors can be
programmed to do almost anything; they are the basis of a variety of
items from TV games (the first computer in the home) to telephone
dialers (that let callers select a desired party by depressing a
single button) to word processing systems (that may expand the
telephone system to incorporate typewriters and electronic filing
cabinets). Because microprocessor systems are highly reliable and
relatively inexpensive, they are a principal factor in applying
computer control to small switching systems such as PBXs or private
switchboards used in business. Because businesses need many features
and services that are not required by residential telephones,
microprocessor controlled PBXs are often appreciably more
sophisticated than the electronic switches used in telephone company
central offices.
But in all such
instances, LSI is best applied in "digital" circuits; that is, it
can be designed, fabricated and applied when the signals it must
process can be treated as ones and zeros, pulses present or pulses
absent, voltage high or voltage low. Such signals are the natural
form for computer communications, but they do not match the
communications of human beings.
Humans function in an
"analog" world. When we speak, we create variations in the pressure
of the air around us. These variations are changing continuously and
smoothly, and blend from one pattern to another as we talk. Our
communication systems follow these smoothly varying patterns,
creating electrical analogs that can be sent over wires and radio.
If we were to look at an electrical analog of a voice signal on a
test instrument called an oscilloscope, we couldn't tell the
difference between the patterns produced by the analog and the
actual measurement of air pressure variations.
It is possible to
convert analog speech signals to digital signals similar to those
used by computers. One technique, called Pulse Code Modulation or
PCM, was first used commercially in "T Carrier," a system for
putting 24 voice channels on two pairs of wires in local telephone
cable. There is a very great deal of "exchange cable" in place;
further, adding to this cable is expensive because streets must be
dug up at great cost and inconvenience. Relatively simple and
inexpensive electronics that can be located at cable ends and can
multiply the number of circuits by a factor of 12 or more has a
great potential for economies.
The analog to PCM
digital conversion is made by sampling an analog signal very
rapidly, measuring its size or amplitude, and coding that amplitude
in a binary number which can be represented by the presence or
absence of 8 pulses in a "byte" or word. A pulse that can be present
or absent is called a "bit," and 8 bits can define up to 256
different values. To convert back to analog, each 8 bit word is used
to reconstruct the signal and smooth it into the analog form we are
used to hearing.
The reconstruction is
not exact. In the coding process, the analog signal can be anywhere
in the region between two levels that is defined by an 8 bit word.
Reconstruction is to the point halfway between the two levels. Thus,
the reconstructed signal may be off by as much as half a level.
Fortunately, the levels are very close together and this difference
is small. Thus the "graininess" of digital transmission (called
"quantizing noise") cannot usually be detected. T Carrier is
relatively free from the noise and distortion that plagues most
other forms of communication, however, so there is an appreciable
overall improvement in transmission in most instances.
Although analog carrier
systems have existed for years and are used on cable, microwave and
satellite, T Carrier PCM is much less expensive in areas where many
relatively short cables already exist. Indeed, in the United States
there are more telephone trunks on T Carrier than any other form of
inter-switch transmission, and the United States is not unique. But
for long haul transmission, the situation is quite different.
Coaxial cable, microwave and satellite circuits have strong economic
reasons for remaining analog. And there are more circuit-MILES of
analog carrier than there are of digital carrier. So we have a
problem. Where "bandwidth," or the spectrum needed to carry signals,
is expensive and/or scarce, analog techniques are more economical
than digital.
To see just how this
comes about, consider some numbers. One voice channel in T Carrier
is made up of 8000 samples per second coded into 8 bits, for a total
of 64,000 bits per second. In the United States and Japan, 24 voice
channels are combined for transmission over one pair of wires in
each direction and the total pulse rate on the facility, allowing
for synchronizing information, etc., is 1.544 million bits per
second. CCITT international standards have 30 voice channels and two
signaling channels as a basic building block and run at a pulse rate
that is proportionally higher.
One voice channel on T
Carrier requires 64,000 bits per second. However, in most telephone
networks, 12 voice channels in an analog carrier "group" are
required to handle high speed data at 50,000 bits per second. Thus
one voice channel in a T Carrier system can handle as much computer
data as 12 voice channels in an analog system. On the other hand, 24
voice channels in an analog carrier require only 100,000 cycles per
second of bandwidth. Although cycles per second and bits per second
do not equate directly, it is evident that 1.54 million bits per
second in T Carrier uses a lot more spectrum than the analog
counterpart. Where we have lots of wires in place, this spectrum is
available. Where we already have our spectrum fully occupied, as in
long haul radio systems, we're out of luck. There is one more factor
to consider, however. And that is digital switching.
Digital switching
Because so many of the
short haul trunks in the United States (and elsewhere) are digital,
it would be highly desirable to connect one circuit to another
without converting back to analog to accommodate the switch. This
would minimize the quantizing noise that could result from repeated
analog to digital and digital to analog conversions and, more
important, would allow digital techniques to be used to actually
switch signals as well as to control the switching mechanisms. If
signals are already in pulse form, it should be possible to switch
them by moving them from one register to another, just as data is
manipulated in a digital computer.
Digital switching, in a
form compatible with T Carrier, is being studied and applied around
the world. In the United States, the Bell System has already
installed many No. 4 Electronic Switching Systems (ESS) in its toll
network. These switches can handle 100,000 high usage trunks;
because only a few hundred will be needed for the entire toll
network, digital toll switching will be nearly universal in the
United States well before the turn of the century. Short haul trunks
will not need any special equipment at the No. 4 ESS end; indeed,
one of the principal motivations for No. 4 ESS is the way the analog
to digital conversion can take place many miles away in regular T
Carrier terminal equipment, and the T lines, with many channels
multiplexed together, can enter the switch directly. When No. 4 ESS
machines are close enough together, no terminal equipment will be
needed on either end of the T lines, increasing savings even
further.
The long haul trunks
will remain analog, and will need analog to digital conversion at
each No. 4 ESS. Many thousands of local telephone switches (which
connect customers to each other and to the toll network), will also
remain analog, and will access the toll network through analog to
digital converters just as they do now. But there will be digital
"islands" where digital switches are interconnected by T Carrier
trunks, and digital signals will go a long way from one point to
another without losing their digital integrity.
As far as the average
telephone user is concerned, he cannot tell the difference between
digital and analog. If we were just concerned with voice signals,
analog transmission would continue to serve us well and digital
techniques would simply remain a good way to reduce costs for short
haul trunks and very large toll switches. However, data
communications, ever increasing as a result of the applications of
LSI to computers and related technology, needs desperately to use
the ubiquity of the public, analog telephone network. Because so
much of this network actually operates in a digital manner in both T
Carrier systems and digital switches such as No. 4 ESS, there ought
to be a way for digital and analog signals to have the best of both
worlds at the same time. Fiber optics just may make this possible.
Fiber optics
The best of both worlds
requires bandwidth to be so inexpensive that voice and data can
operate in a digital mode that is suitable to both. The eight bit
voice coding of T Carrier just happens to fit the 8 bit bytes used
to describe letters and other symbols in a number of standard codes
(ASCII with parity, for instance). Thus T Carrier is a good format
to start with. T Carrier on fiber optics may be the answer, because
glass fibers have such a wide bandwidth and so many other desirable
properties that they appear to approach utopia.
Glass fibers as thin as
a human hair may not seem like the stuff from which to build the
future, but they are already off to a good start. Fiber systems have
been tested in Atlanta, Chicago and Las Vegas, and a very big fiber
system is going in between White Plains, the regional center and
international switching point just north of New York, and downtown
Manhattan. And Japan is doing even more.
Because many fibers make
up one cable and, even so, the cable has a very small cross section,
little space will be required in existing cable ducts. Further, the
installation will have the effect of greatly expanding the capacity
of existing ducts as ancient copper cables are removed and replaced
with additional glass. Glass is not bothered by lightning or crosses
to power mains, and it is only slightly affected by moisture; it
should last even longer than the cables it will replace. And, as has
been mentioned, NOT having to place new ducts under the streets in a
major metropolitan area is a factor of considerable importance. It
is quite likely that much metropolitan trunking, all over the world,
will be T Carrier running in glass within a very few years.
What will happen with
long haul trunks? Where toll switches are close together, as in the
Boston-Washington Corridor, fiber optics will take over and "digital
islands" will appear. But longer circuits pose a problem. For the
immediate future, they will have to stay on microwave, existing
analog cable systems, and satellites. In some places, where radio in
one form or another is the only possible mode of transmission, wide
enough spectrum may never be available. In other areas (Canada is a
good case in point), existing microwave routes are being expanded
and "overbuilt" with digital capability. And sophisticated coding
techniques are being investigated with all possible speed to make
digital satellites economically competitive. Eventually, however,
end to end digital paths will become available if not universal. And
that will be the major factor in business communications.
Business communications
In many industrialized
countries, residential and small business telephones greatly
outnumber the telephones of large businesses. In the United States,
the ratio of individual telephones to those served by PBX and
Centrex systems is somewhat greater than 4 to 1. Thus telephone
designers the world over have tended to ignore business
communications. They have even attempted to use legislation and
other approaches to limit the application of facsimile and data
transmission. And, when they have designed new PBXs, they have
seldom provided features and services, particularly in non-voice
areas, that businesses require. In many cases, they have provided
features that are very good for residential service but of limited
utility in an office.
There are many PBXs on
the market today. Most of them allow the individual telephone user
to transfer calls, add on third parties for conferences, and pick up
telephones ringing unanswered by dialing a "magic number" on his own
phone. They also permit one to "forward" calls to another telephone,
either immediately or after so many rings. "Hunting," available in
PBXs for decades, has been improved in various ways so that calls
encountering busy lines will still be completed.
These features are
sometimes useful and are all easily provided now that most PBXs are
controlled by microprocessors or other computer-like devices. But
they make no particular concession to non-voice traffic.
Some PBXs are "digital."
Unfortunately, digital means different things to different people.
Some feel that a PBX controlled by a digital computer is a digital
PBX. Some feel that any transmission approach that uses pulses is
digital. Others use technically correct definitions of digital as it
refers to the signal being connected through the PBX, but fail to
warn the potential customer that there are several different types
of true digital systems and not all of them are compatible with the
PCM transmission facilities that are widely used around the world.
For digital PBX
technology to be useful to the telecommunication customer, it must
reduce the cost of handling voice communications, and it should also
facilitate the handling of non-voice communications. If digital
technology is used just to reduce analog costs, it may fail. Savings
in business systems at present just aren't all that great.
It is the possibility of
handling non-voice signals that offers the overall chance to come
out ahead. If voice and non-voice communications can be coded the
same way, and if non-voice signals can be transmitted through the
system without expensive external modems to make the analog to
digital conversion, the digital nature of the system will, indeed,
be useful. If, further, the PBX is compatible with T Carrier span
lines as is the No. 4 ESS, further savings are possible
(ultimately), and (also ultimately), direct digital communication at
high speeds can be carried on through the public telephone network.
Here is where the real future lies.
Unfortunately, T Carrier
PCM uses one set of standards in America and Japan and another in
the rest of the world. Further, although both of these will work
well and give fiber optics extensive motivation, neither appears to
be as satisfactory on satellite systems as another form of digital
coding called "delta modulation." Delta modulation uses only half as
many bits per second as PCM (and sometimes uses less). Thus, a
digital satellite can handle twice as many voice channels in the
same bandwidth, but these satellite circuits will not be compatible
with land-line PCM.
Some PBXs on the market
today are compatible with T Carrier and, when it is permitted, will
be able to connect directly to T span lines in the public network.
Some use delta modulation in various forms and mayor may not be
compatible with delta modulation actually used in satellite and
radio systems. And some digital PBXs use coding schemes that are not
compatible with either T Carrier or delta modulation.
The several futures
This leads us right back
to where we started. Satellites, digital components and fiber optics
are the waves of the future, but they do not look like the waves of
the same future. How it will all work out is anybody's guess.
Different countries and different parts of the world have different
needs. Indonesia, for instance, with its thousands of islands, will
find satellites more useful for internal communications than will
Japan, while China, a large land mass like North America and Europe,
may find a considerable need for both.
The dominance of Europe
and America in establishing requirements may or may not produce
results that are the best possible for other parts of the world. But
businessmen everywhere must communicate with each other quickly,
easily and economically. Uniform standards are desperately needed,
even in widely differing areas. Thus, businessmen must understand
their stake in the digital future, and make the effort required to
keep up with developments. Even though the future is not as clear as
we all would like, it will be upon us before we know it. There is
still time to shape it to meet our needs.
[
Top ] [ Next ] [
Table of Contents
] |
