Of all the new 1960s wonders of telephone technology -
satellites, ultra modern Traffic Service Positions (TSPS) for
operators, the picturephone, and so on - the one that gave Bell
Labs the most trouble, and unexpectedly became the greatest
development effort in Bell System's history, was the perfection
of an electronic switching system, or ESS.
It may be recalled that such a system was the specific end in
view when the project that had culminated in the invention of the
transistor had been launched back in the 1930s. After successful
accomplishment of that planned miracle in 1947-48, further delays
were brought about by financial stringency and the need for
further development of the transistor itself. In the early 1950s,
a Labs team began serious work on electronic switching. As early
as 1955, Western Electric became involved when five engineers
from the Hawthorne works were assigned to collaborate with the
Labs on the project. The president of AT&T in 1956, wrote
confidently, "At Bell Labs, development of the new electronic
switching system is going full speed ahead. We are sure this will
lead to many improvements in service and also to greater
efficiency. The first service trial will start in Morris, Ill.,
in 1959." Shortly thereafter, Kappel said that the cost of the
whole project would probably be $45 million.
But it gradually became apparent that the developement of a
commercially usable electronic switching system - in effect, a
computerized telephone exchange - presented vastly greater
technical problems than had been anticipated, and that,
accordingly, Bell Labs had vastly underestimated both the time
and the investment needed to do the job. The year 1959 passed
without the promised first trial at Morris, Illinois; it was
finally made in November 1960, and quickly showed how much more
work remained to be done. As time dragged on and costs mounted,
there was a concern at AT&T and something approaching panic at
Bell Labs. But the project had to go forward; by this time the
investment was too great to be sacrificed, and in any case,
forward projections of increased demand for telephone service
indicated that within a phew years a time would come when,
without the quantum leap in speed and flexibility that electronic
switching would provide, the national network would be unable to
meet the demand. In November 1963, an all-electronic switching
system went into use at the Brown Engineering Company at Cocoa
Beach, Florida. But this was a small installation, essentially
another test installation, serving only a single company.
Kappel's tone on the subject in the 1964 annual report was, for
him, an almost apologetic: "Electronic switching equipment must
be manufactured in volume to unprecedented standards of
reliability.... To turn out the equipment economically and with
good speed, mass production methods must be developed; but, at
the same time, there can be no loss of precision..." Another year
and millions of dollars later, on May 30, 1965, the first
commercial electric centeral office was put into service at
Succasunna, New Jersey.
Even at Succasunna, only 200 of the town's 4,300 subscribers
initially had the benefit of electronic switching's added speed
and additional services, such as provision for three party
conversations and automatic transfer of incoming calls. But after
that, ESS was on its way. In January 1966, the second commercial
installation, this one serving 2,900 telephones, went into
service in Chase, Maryland. By the end of 1967 there were
additional ESS offices in California, Connecticut, Minnesota,
Georgia, New York, Florida, and Pennsylvania; by the end of 1970
there were 120 offices serving 1.8 million customers; and by 1974
there were 475 offices serving 5.6 million customers.
The difference between conventional switching and electronic
switching is the difference between "hardware" and "software"; in
the former case, maintenence is done on the spot, with
screwdriver and pliers, while in the case of electronic
switching, it can be done remotely, by computer, from a centeral
point, making it possible to have only one or two technicians on
duty at a time at each switching center. The development program,
when the final figures were added up, was found to have required
a staggering four thousand man-years of work at Bell Labs and to
have cost not $45 million but $500 million!
No comments:
Post a Comment