|
The Digital Future
Of The Telephone Network
A Study of Evolving Technology
By Lee Goeller
Originally published by Probe Research Inc. 1979.
Reprinted by permission
Chapter 6
The Stored Program Controlled (SPC) Network
About the only thing
that can be said for certain about the system just described is that
it or anything even remotely resembling it will never come about. An
all-digital public network, through the Class 5 office or PBX and
perhaps even to the customer telephone, is simply not going to
happen. Why? There's a reason for it. Its just policy.
Non-PCM Developments
What leads me to this
conclusion? Several points, some of which have already been
mentioned. The most important is the existence of No. 1 ESS,
followed by No. 2 and No. 3. Although controlled by solid-state
computer-like processors, these switches have metallic reed-switch
matrices for connecting one line to another or to a trunk. Thus they
are in no position to handle digital signals in a digital mode.
No. 1 ESS, in
particular, poses a problem. The installation strategy for No. 1 has
been to put it where it will handle the maximum number of calls if
not the maximum number of lines;* thus it has been located primarily
in large metropolitan areas where business communication generates a
high calling volume.** Unfortunately, it is precisely this
environment where digital switching could make the most effective
use of end-to-end digital communications. Business users already
have a need for non-voice communication; residential users may
someday have such a need, but the cost of developing digital
techniques and equipment can be justified now only on the basis of
serving business. Then, later, just as lower rates after 5 let
residential customers ride in at discount rates on the toll network
sized for business traffic during the day, so lower rates for
pre-developed digital services can be offered to non-business
customers at modest cost. Anybody who thinks digital features and
services can be developed on the basis of residential needs simply
does not understand the telephone business.
[*Footnote: By 1985, it is expected that 60% of all customer
lines will be served by some form of ESS.]
[**Footnote: The 1A ESS, with a 1A Processor and Remreed matrix,
can switch 100,000 lines at a calling rate of 240,000 calls per
hour.]
Turning to the No. 2 and
No. 3 ESSs, for smaller communities, we find the same general design
philosophy: reed switch matrices with solid-state processors for
control. These machines, designed to go up to about 10,000 and 4500
lines, respectively, will effect some savings in floor space and
power, and bring Touch-Tone calling, call forwarding, call waiting
and three-way conferences to smaller towns and hamlets. Perhaps
their most important advantage lies in improved maintenance
operations, including remote testing and control. They are not,
however, digital in any sense of the word.
The existence of other
systems shows an anti-digital cast. Subscriber Loop Carrier systems*
recently announced, SLC-1, SLC-8 and SLC-40, are designed to reduce
the cost of connecting central offices to distant customer
telephones by substituting design sophistication for copper. SLC-1
and SLC-8 are analog systems, with SLC-1 adding one electronic pair
equivalent to a real copper pair and SLC-8 putting 8 analog voice
channels on a single pair. SLC-40 is something else again. It puts
40 voice terminals on two pairs of wires, using 24 voice channels
derived by digital techniques. However, there are various kinds of
digital systems. Although the one used here works over a line
designed for T-carrier, a completely different modulation technique
is used: delta modulation. Translation from delta-mod to PCM
requires considerable effort and would hardly preserve direct
digital data.
[*Footnote: SLC, Subscriber Loop Carrier, is a Bell System
trademark. After
years of effort, AT&T has apparently abandoned its attempts to
rename
subscribers more accurately as customers. "Subscriber" and
"exchange"
are two words that date from the late 19th century. An "exchange"
is a club or organization, and a "subscriber" is one who subscribes
to
the principles or purposes of the club. When local groups banded
together
to put in a telephone system, the club or exchange idea fitted
pretty well.
But let's face it: today we are all customers. Even when served by
SLC.]
Another non-T subscriber
system is the LSS, of Loop Switching System. In service in 1977, LSS
is a traditional concentrator. That is, a metallic matrix made up of
small relays gives 96 customers (subscribers?) access to 32 copper
pairs to the CO where a reverse unit expands back to 96 terminals on
a traditional switching system which will again provide
concentration. Controlled by a microprocessor via an additional two
pairs, two separate groups of 96 customers can be handled in an
overall system. These new systems, SLCs and LSSs, will doubtless be
quite useful with traditional switching systems such as SXS,
Crossbar, and ESS 1, 2 and 3, but they will certainly contribute
more than their share toward locking T-carrier PCM-type digital
communications out of the loop plant.
The final exhibit in the
anti-PCM display is the Dimension PBX. The Dimension 400, and the
concatenated 400s that make up the Dimension 2000, are two-wire
analog switches that just happen to use time division multiplexing.
The technique used is pulse amplitude modulation which could be
converted to PCM but isn't. In May 1978, the Bell System announced a
variety of new business services under the general heading of ETS,
or electronic tandem switching (Feature Package 8) based on the
Dimension 2000. These services are so well thought out and come so
close to meeting the needs of the larger business customer who
requires tandem tie lines for inter-location communication that they
will undoubtedly make another contribution to blocking the future of
digital communications. Gresham's Law rides again!
The Non-Digital Approach With No. 4 ESS
If the above illustrates
the kind of hardware that AT&T is using to build the future, what
will the overall system be like? Isn't No. 4 ESS going to save the
day? Maybe yes, maybe no. But to see the way the wind is blowing,
let's give No. 4 ESS a quick look.
No. 4 ESS is a
T-compatible PCM digital switching system designed to handle a total
of over 100.000 tandem, toll and tie trunks generating over half a
million calls per hour. If the trunks to be switched arrive via T
span-lines, they enter the switch directly without channel banks or
separate signaling equipment; indeed, one of the principal features
of No. 4 ESS is the way it reduces the prove-in distance of a
T-carrier trunk to zero miles. That is, the standard channel bank
that provides the A/D conversion can be located in the same room
with the time-division switch, a couple of floors away, in another
building, or 10, 20, 50 or more miles off in the distance. Analog
trunks can be connected to an A/D converter on a per-trunk basis,
but, since the A/D conversion is required no matter what, the saving
in cable suggests that short-haul analog carrier systems should be
replaced with T-carrier where the required A/D converter is placed
at the distant end rather than the No. 4 ESS end. This is expected
to drive most exchange and short-haul trunks to T in a very short
time, making available large quantities of SF signaling equipment
and analog carrier for non-digital areas.
The market for SF sets,
a $25 million a year business at Western Electric, the AT&T
manufacturing arm, illustrates some of the problems facing the
telephone industry when a revolutionary development comes about. One
of the great advantages of a regulated monopoly is the way nothing
is wasted. No. 4 ESS, replacing existing tandem and toll switching,
will make so many SF sets available for reuse between 1975 and 1979
that there will be no need to manufacture any at all until 1983.
Then the assembly line will have to be restarted, a feat that is
always much more difficult than outsiders would believe. Presumably
something similar will happen with short-haul analog carrier
systems. The stockyards claim to use everything but the squeal.
Here, even the SF squeal is reused. But one cannot help wondering,
in this instance, about the costs involved in reuse. Fantastic
computer-based economic studies have been carried out, however, so
proof of overall economy, whether real or not, is readily available.
It should be noted that,
where No. 4 ESSs terminate both ends of a trunk, no channel bank or
signaling equipment (or cross connect frames) at all is required.
This is an additional economy, but the size of No. 4 ESS makes small
the number of areas where two or more are presently within
short-haul carrier range of each other. As has been mentioned, to
conserve radio spectrum and to make use of existing coaxial cable
long-haul trunking systems, wide-band digital modulation will not
soon be used for trunks between more widely spaced No. 4 ESSs.
No. 4 ESS was purposely
made three times as large as the largest No. 4 Crossbar toll switch
so that one switch could serve an entire metropolitan area. If three
No. 4 Crossbar offices were used, each local switch would have to
have three groups of toll connecting trunks, one to each toll
switch, or else many terminals on each toll switch would have to be
tied up with trunks to the other nearby toll switches. The
alternative, having three long-haul toll groups to Chicago, say,
(one for each No. 4 Crossbar), from each and every other toll switch
in the country that can justify Chicago trunks, would be absurd. One
BIG No. 4 ESS eliminates the problem.
Because No. 4 ESS is
"essentially" non-blocking, traffic balancing is not required.
Whether or not connections can be patched through on a permanent or
semi-permanent basis is not clear. Because relatively few trunks
will enter the system on a per-trunk basis, the MDF problem, where
cross connections and traffic balancing might be carried out if
necessary, is changed to something different. The system contains
its own database with CRT data-set access, electronic updating, and
minimal paper. Between the database and the equipment layout, little
MDF effort is required.
One item conspicuous by
its absence in No. 4 ESS is an operator subsystem. Presumably TSPS
(Traffic Service Position System) will continue to be used in the
future, along with Nos. 1, 2 and 3 ESS. It is not generally
realized, but the reed switch access matrix for TSPS, connecting
operator positions to operator trunks, is approximately the same
size as the No. 4 Crossbar machine on which the trunks terminate.
This would seem to suggest that operator trunks will have to remain
analog so that the analog TSPS switching matrix can be used to
bridge on the operator. Perhaps the operator trunk market will
consume all those SF sets after 1983.
There is, of course, an
alternative to TSPS that requires almost no switching matrix at all.
As has been mentioned, Northern Telecom, in Canada, has developed
TOPS (Traffic Operator Position System, shown in Figure 2), an
operator system that uses the ports on the toll switch to connect in
the operators. Designed for use with Northern Telecom's
computer-controlled crossbar SP systems, TOPS will soon be adapted
for service with digital (PCM) switching matrices. This approach
allows local offices to access the toll network by one group of
(digital) trunks for both operator-handled and non-operator traffic,
and replaces an expensive, bulky switching system with an
incremental addition of relatively small size on the toll switch
itself.
With the exception of
this lack of an operator subsystem, No. 4 ESS appears to be an
excellent switch, ideally suited to the development of the digital
future. However, because voice communications in the public switched
network is still appreciably greater than non-voice traffic, all
cost justification is based on serving analog traffic. Since No. 4
ESS can connect directly to long-haul analog trunks and, by putting
its A/D converters in nearby Class 5 central offices, it can connect
quite economically to analog local switches, a great deal of digital
equipment can be installed AS LONG AS IT WORKS ENTIRELY IN AN ANALOG
MODE. That is, digital equipment, including the No. 4 ESS, is being
used ONLY to be cost effective, and it is NOT being used to be
DIGITAL for the sake of a digital future.
Common Channel Interoffice Signaling (CCIS)
If No. 4 ESS must be
given good marks, CCIS has to come in for a somewhat lower level of
praise. CCIS has been floating around for 20 years or more, waiting
for the right technology to make it viable. Now, its time has come.
AT&T is installing CCIS, using 2400 b/s data channels (!) between
No. 4 ESSs and No. 4 Crossbar systems as rapidly as possible, and
will add CCIS connections to TSPS, No. 1 ESS (particularly when used
for toll switching) and the Dimension PBXs in the future. Although
CCIS will improve telephone service somewhat, a realistic appraisal
indicates the main benefits will be to the telephone company's
internal network management and, as a result, will impact customer
service only indirectly.
The general idea behind
CCIS is that the digital computers that control switching systems
should, be able to "talk" to each other in their own language and at
a rate approaching their own natural operating speeds. They should
not, as systems have in the past, pretend that they are manual or
electromechanical systems just to maintain the stately interfaces
via SF, DP and MF signaling on copper and analog carrier trunks.
The simplest form of
CCIS is simply a data link between two common-control switches,
paralleling the trunk group between the same two switches. The extra
expense of the data link* is offset somewhat by the opportunity to
remove SF signaling units and MF senders and registers; SF sets must
be supplied on a per-trunk basis, but MF (and DP) senders and
registers, although supplied on a traffic basis and serving a large
number of trunks, are large, expensive, and require a lot of
manipulation by the switching system, particularly in No. 4
Crossbar. In any event, signaling and supervision can go directly
from one common control to the other and, because the signaling
channel is always available (registers and senders are only attached
at the beginning of each call), additional information can be
exchanged. Further, because the formats and contents of the messages
are a function of system programs, changes can be made and new
features can be added without affecting the hardware. Note that, in
particular, if both ends of the same trunk are seized
simultaneously, the common controls can talk about it and come to an
amicable solution.
[*Footnote: On T-carrier systems, a separate data link can be
eliminated.
Framing bits already available and no longer needed to identify
signaling
bits associated with each voice channel become a CCIS data stream.]
Another point of no
small interest is the fact that speech-type signals are NOT used on
the trunk in question. This immediately pulls the teeth of the Blue
Boxers, since it will take more than a premium whistle from a
Captain Crunch cereal box to gain access to a CCIS link. And even if
a way can be found to release a toll trunk while holding the
toll-connecting trunk, simulating MF tones will get the cheat
exactly nowhere.
But there is another
side to this. In the old days (like now), it used to be possible to
assume that if you could signal over a trunk, you could talk over
it. This was not always true, considering the automatic gain control
available in both SF and MF signaling systems,* but voice frequency
signaling did have a lot in common with actual voice frequency
communications. It became necessary, therefore, to add back some
sort of a test to be sure the trunk could be used. The procedure is
to attach a transceiver to the outgoing trunk when it is seized.
When the called end receives the word via CCIS, it connects the
incoming side of the trunk to its outgoing side (loop around) and
the calling end detects the returned signal, measuring its level to
be sure the loss in the trunk is within acceptable limits. In the
very rare cases where 2-wire trunks are used, more elaborate
procedures are required. This test, however, when coupled with
several messages over the CCIS line, insures trunk integrity.
[*Footnote: In Europe, a system called compelled MF can signal
with even higher reliability over a channel much too noisy for
conversations.]
It has doubtless
occurred to the reader by now that switching systems, particularly
those used in the toll network, have very large numbers of trunk
groups to many other switches, and some of these groups may be quite
small (like 1 trunk, when used to skim the cream while the rest of
the traffic is alternate-routed via a higher-level hub). Obviously,
all possible pairs of toll switches cannot be connected by data
links; only those with large enough direct trunk groups could
justify such an "associated" signaling channel. To deal with all the
rest, heroic measures are required.
It will be recalled that
the United States is divided into 10 regions, each served by a Class
1 switching system. It was decided to add to each region two Signal
Transfer Points, or STPs, separated for security but essentially
duplicating each other for reliability. Each CCIS switch would home
on BOTH of the STPs in its area; STPs would all be interconnected.
Links between the two STPs in one area would be primarily for
keeping each other updated so that, if one had to take over the
whole load due to failure of the other, its information would be
correct. Between each pair of regions, four CCIS channels exist:
each of the two switches in each region connects to each switch in
the other region. This scheme of separation with duplication
provides a very high degree of reliability and allows a lot of
information to be moved from one switching system to another through
no more than two STPs.
To process a call
through a built-up connection in the toll network, the first CCIS
switch will look at the dialed number, check in its routing tables,
seize a trunk to the next indicated switch, send the required
information to the second switch via one or both STPs, and connect a
continuity-check transceiver. The second switch will acknowledge,
put the trunk into a loop-around condition, and check its routing
tables for a path to the next switch where the process will be
repeated. When each switch gets an indication from its transceiver
that the trunk it seized is OK, the transceiver is dismissed and a
check-signal is sent forward on the CCIS channel.
Since CCIS will
initially run only among toll switches, there will come a time when
traditional signaling will have to be used. Traditional senders will
be available to pump the called number out via MF or dial pulsing.
The last digit will not be sent, however, until a continuity check
signal is passed forward for all of the preceding trunks. It would
not do to have the customer answer before the path through several
trunks is completed.
When everything is all
set, all the switches in the talking path complete the connection
from incoming trunk to outgoing trunk, the final CCIS switch sends
forward the last digit, and the called line is rung. Upon answer,
the terminating CCIS switch picks up the signal and routes it via
the data links to the point where the AMA (Automatic Message
Accounting) records are kept. At hang up, another message is sent.
The individual trunks are actually released from the originating end
toward the terminating end with yet another CCIS message.
There are several
interesting aspects to this whole sequence. First, the switches
start to establish the path before all the dialed digits from the
user have been received. Additional digits, as received by the first
CCIS switch, are sent along as they come in. This technique, called
"overlap operation," has been used with dial pulsing for many years;
it reduces the post-dialing delay — the time between the dialing of
the last digit and the start of ringing. If the customer abandons
(as when he knows he has dialed an incorrect digit in an SXS office
homing on the toll CCIS switch), a lot of labor is wasted.
Second, continuity is
checked on a trunk-by-trunk basis; it is NOT checked end to end. In
electromechanical offices, where continuity and false cross and
ground (FCG) tests are common, both the path through the trunk and
the path through the switch are checked not only for the required
continuity, but also for lack of continuity to other calls in
progress. In electronic switches, particularly time-division,
continuity checks are difficult and FCG tests are almost impossible.
It would seem that the existence of No. 4 ESS would require, at the
very least, an end-to-end continuity check, just to maintain some
sort of parity in the reliability department with present methods of
connection.
On the plus side, CCIS
makes all 8 bits in T-carrier available all the time, eliminating
digit robbing. Further, the built-up connection can apparently be
established quite rapidly and, if congestion is found after several
links have been set up, the overall system can back off and retry
via a second or third choice route. This ability to retry, coupled
with set-up speed, will greatly reduce the nonproductive use of
trunks, allowing the same number of circuits to carry appreciably
more traffic during a given month, although hang up time may be
slowed somewhat. Further, more elaborate routing schemes can be
built into the switches; the old hierarchal arrangement can give way
to more flexible approaches since the CCIS messages can carry along
a list of the switches through which the call has already
progressed.
Another major advantage
of CCIS is the way it eliminates massive seizures of trunks in
switches when carrier systems fail. It should be remembered that SF
signaling has the tone ON when the trunk is idle; tone off means the
far end has seized the circuit. In the event of a carrier failure,
the channels therein go dead and the SF tone is no longer received.
The SF units on each end know only that the tone has gone away and
they demand that their associated switches connect a digit receiver.
Carrier Group Alarms can sometimes minimize this problem, but not
always. A separate signaling system is far superior.
Going one step beyond,
failure of the continuity check on two or three trunks in the same
trunk group can, via CCIS messages, alert both switches to the
likelihood of carrier failure, and suitable alarms can be sent to
appropriate emergency forces. If the failure turns out to be
mometary, and attempts on the same group, a few minutes later, meet
with success, the alarm condition can be removed and both switches
can use CCIS to update each other.
But perhaps the most
important use of CCIS will be to send additional signaling messages
between switches. In particular, routing tables in each switch can
be updated automatically, and can be changed when trouble
conditions, congestion, etc., require. Further, "traveling class
marks" indicating the presence of an echo suppressor, for instance,
or a path via a satellite, can be added to the signaling message.
This will prevent two echo suppressors or two satellite paths from
being used back to back; in either case, one is good, but two are a
disaster. Note, also, that, since an echo suppressor must be used in
a satellite circuit, it is important to prevent an echo-suppressor
land circuit from connecting via a satellite.
Ultimately, when CCIS is
extended into the Class 5 ESS offices, it will be possible to look
ahead to see if the called line is busy. If so, all the trunks in
the built-up connection can be dismissed immediately and busy tone
can be returned from the CCIS office nearest the calling party. If
the originating office also has CCIS, the busy tone can be returned
directly to its calling customer; in any event, use of trunks to
return busy tone (or reorder, or other signals when a call cannot be
completed) will be greatly reduced with considerable savings
accumulating.
At that time, it will
also be possible to see if the called party has the call forwarding
feature in effect; presumably the most economical routing from the
calling party to the called number as defined by forwarding could
then be used. Other features sometimes mentioned include carrying
the calling number to the called office to match against a list of
numbers stored by the called party. If the calling number matches
one on such a list, distinctive ringing (or call waiting or,
perhaps, barge-in) could be applied. If no answer is obtained, the
calling number could be stored and the called party could eventually
obtain it via a voice-announce system by dialing a code. Few of
these features would be useful via PBX trunks, although some
versions of Centrex might incorporate them. Since the majority of
calls are PBX or Centrex calls, the overall impact of such features
may not be great.
The screaming need, on
the part of the customer, is for automating credit card, third party
and collect calls (that is, calls from some phone other than his
own). If such calls could be made without incurring the high
"operator handled" charge, businessmen in general, and salesmen in
particular, would find the telephone an instrument of much greater
utility. Similarly, in-WATS could certainly use improved information
handling to facilitate telephone company operations. Apparently all
these improvements are in the works; someday they will be
implemented. Where CCIS-equipped ESS Class 5 offices are not
available, TSPS will take over the job.
Although CCIS offers
considerable opportunity for saving money and even offers some
useful service features, the main pitch being used in its behalf is
the time it is going to save the customer. Saving five or six
seconds a call (sometimes) for people who make less than 10 calls a
day is an "advantage" that approaches the absurd. It is almost as
ridiculous as the vast amounts of leisure that were supposed to
accrue through the use of Touch-Tone telephones. Meaningful savings
can only come where they add up from hundreds or thousands of calls.
A switchboard or console operator at a PBX or toll operator position
can, indeed, save appreciable time and increase productivity if a
few seconds are shaved off each call; similarly, the productivity of
trunks can be increased, as has been mentioned, if nonproductive
signaling time is reduced. But few telephone customers will notice
any difference at all when CCIS goes in.
Sadly, even the
telephone company may not achieve time-reduction savings for many
years to come. Consider a call going all the way across the country
via CCIS, and ending on a standard, non-CCIS switch. The switch gets
the call and connects ringing...and ringing may not start for up to
FOUR SECONDS. This may be longer than the entire time to set up the
connection! (Ringing is on for two seconds and off for four
seconds—a pattern designed to produce maximum effort on the part of
the called party to get to the phone and answer it. ESS offices
start ringing immediately upon connection to the called line (my
patent) but most other offices do not.)
As if this weren't
enough, consider the case of DID to PBXs. Direct Inward Dialing
requires the Class 5 office (or, sometimes a tandem office which, in
the future, might be a No. 4 ESS) to send two, three or four digits
into the PBX to identify the called party. As things are set up
today, dial pulsing, at an average of one second per digit, is the
only way that DID is implemented. Considering the frequency of
calling to PBXs, noted often above, such a delay will help to reduce
the impact of CCIS appreciably. CCIS will someday be extended to
some Bell PBXs, but the priority is doubtless low.
Implementation of CCIS
is relatively interesting. The general idea is to maximize the
impact of CCIS as quickly as possible, and the way to do that is to
attack the toll network first and only later reach out to the much
larger number of local central offices. No. 4 ESS is being installed
with CCIS capability, but for the next few years, No. 4 Crossbar
will dominate the toll-switching field. The trick, then, is to
retrofit No. 4 Crossbar in such a way that a maximum number of
trunks will be served by CCIS; that is, that "connectivity" will be
maximized.
No. 4 Crossbar exists in
two forms: the old card-translator version, and the newer version
using ETS, the Electronic Translator System. ETS is actually an
electronic computer-like arrangement that can easily be expanded to
handle CCIS signaling; of the 183 Bell System No. 4 Crossbar
switches in service at the end of 1976, all but 19 use ETS. The last
new Bell System No. 4A/ETS was installed in Madison, Wisconsin, in
May 1976. Future new installations will all presumably be No. 4 ESS.
In any event, an old No.
4 Crossbar, near the limits of its capacity, will already be set up
for SF, MF and DP. More important, there will be little room for
future additions where new trunks, taking advantage of CCIS, can be
installed. Thus small No. 4 Crossbars, with high potential for
growth, will provide the greatest return. However, don't count out
the big old systems. These are the ones that are scheduled for
replacement by No. 4 ESS, since they cannot grow any larger
themselves. If they are equipped for ETS, it appears that they can
be modified for CCIS so that, in selected cases, they can become
Signal Transfer Points or STPs. Then, when No. 4 ESS takes over, the
crossbar switching matrix can be removed and the ETS is an STP and
nothing else. Nothing is wasted.
With regard to CCIS, it
might have been hoped that end-to-end connection checks would have
been used, that digital signaling channels would have been employed,
and that a "common control" rather than a "progressive control"
analogy for the toll network would have worked out. It would
maximize paid use of trunks if the called line could be tested for
busy first and the path across the country selected and operated
only if the called line were free. Then an end-to-end integrity
check would guarantee operation of the entire path including both
switches and trunks prior to start of ringing. Perhaps some of this
will be possible someday.
CCIS is supposed to save
money to some extent by making available for reuse SF, DP and MF
signaling equipment on old trunk groups, and to a larger extent by
simplifying office wiring and trunk circuits and never even
installing SF, MF and DP equipment on new trunk groups. But in the
future, CCIS will really make money by providing new administrative
services within the telco. Faster connections and better network
management will be of great value to the telephone industry. On the
other hand, features designed for extracting new revenues from
customers seem both trivial and frivolous. Hopefully, nobody is
counting on them to justify CCIS.
As can be seen, CCIS has
nothing to do with an all-digital network. It is a signaling system
that interconnects computer-like control units, and is independent
of the type of transmission path used by customers. One of the main
cost advantages of T-carrier in the past has been inexpensive
built-in signaling. Since signaling built into trunks will no longer
be needed, whether the trunks are analog or digital, CCIS will tend
to slow, slightly, the stampede to T-carrier.
Summary
Common Control stored
program switching systems interconnected by CCIS make up the SPC
Network, the public communication system of the future as visualized
by AT&T. The fact that the Stored Program Controlled Network will
include the digital No. 4 ESS and digital T-carrier trunks is purely
coincidental. All of this hardware is justified by providing voice
service, and all of the facilities are used in an analog form.
Apparently there is little place for digital equipment used in a
digital mode anywhere in the future of the public switched network.
But, we are told, data
is growing by leaps and bounds. Data communications is the wave of
the future. Most large businesses already have elaborate data
networks, electronic word processing is upon us, small
micro-processor-based systems are coming on the market so that even
the smallest business can handle its typing and accounting
electronically, and idiot-type electronic games are found
everywhere, illustrating just how easily each home can be equipped
with a data terminal when suitable motivation is provided. Sooner or
later, the dam will burst and the digital flood will become a tidal
wave. What will carry it away?
[
Top ] [ Table of Contents ] [
Next Chapter ] |
