telecom@eecs.nwu.edu (TELECOM Moderator) (05/26/90)
Here was a topic of major interest in the fall of 1981, as reported in
TELECOM Digest, during September that year.
------------------------------
Date: 5 Sep 1981 14:06:36-PDT
From: IngVAX.geoff at Berkeley
Subject: Judge Approves AT&T Expansion
Copyright (C) 1981, Washington Post Company, Saturday, September 5, 1981
A federal judge yesterday gave American Telephone and Telegraph
Co. permission to compete for the first time in the computer and
data-processing business -- a ruling that could have an explosive
impact on the future of the telecommunications industry.
Handing the Communications giant a major victory in its bid to
enter lucrative new business areas, Judge Vincent P. Biunno of the
U.S. District Court in New Jersey ruled that AT&T could offer
data-processing because it is a communications service.
Under a consent decree AT&T signed with the government 25 years
ago, the company has been barred from offering any noncommunications
service. Until yesterday's ruling, that meant AT&T could not offer
computer services.
"It seems to the court beyond dispute that AT&T ... will be
engaging in the business of furnishing communications services and
facilities" by providing data-processing services, Biunno wrote. The
ruling was handed down from the New Jersey court because that was here
the original 1956 consent decree was filed.
The decision means that beginning as early as March 1, AT&T will
be able to become a direct competitor with International Business
Machines Corp. and other major computer companies.
What's more, the Bell System will be able to offer a wide variety
of telephone equipment and services, ranging from the black
rotary-dial phone to highly sophisticated computer services, without
any of the government pricing restraints it now must follow.
The judge's decision, however, by no means puts to rest the
current congressional debate over AT&T's future structure and its role
in the telecommunications industry. Congress is considering
legislation to allow AT&T to offer data communications services,
although it would bar the communications giant from offering
electronic newspapers and up-to-the minute advertising. Debate on the
issue is expected to continue shortly after Congress returns from its
August recess.
------------------------------
Date: Friday, 11 Sep 1981 16:04-PDT
To: unix-wizards at SRI-UNIX
Cc: mike at RAND-UNIX
Subject: New computer company?
From: mike at RAND-UNIX
The following rumor was news to me. I have no way of knowing if there
is any truth to it but my source is well connected with the computer
industry.
He claims that the Bell System computer subsidiary, whatever it will
be called, will announce a computer system within six months. The
computer looks like "a cross between a vax and a MAC 32". (The MAC 32
was an array processor, I understand). Speed is said to be about 70%
faster than a vax. It runs Unix. It has 512 32-bit registers.
No information about what it will cost, when it will be announced,
what sort of bus, etc. Supposedly it will be used in the ESS in place
of the PDP 11.
Does anyone care to refute or elaborate upon this rumor?
Michael Wahrman
------------------------------
Date: 14 September 1981 09:22-EDT
From: Andrew Tannenbaum <TRB at MIT-MC>
Subject: New Bell System Computer
cc: "mike@rand-unix, unix-wizards" at SRI-UNIX
Dear, dear. The Bell System is very touchy about its employees
divulging proprietary information about its products. You can't get
fired from Bell Labs for being incompetent, just for screwing your
secretary or smoking dope or drinking beer in your office, or
divulging proprietary information.
Sooner or later, the Bell System will be able to release its computers
for sale. Sooner will be in more than six months though, you can rest
assured. The wheels of justice turn S-L-O-W-L-Y.
If some Bell System drone reports on proprietary products he does so
at personal risk (possibly great). I can say that the MAC32 is NOT an
array processor, and that Bell Labs is working on processors to use in
their switching systems, even ones that run UNIX (our
telecommunications support operating system). Have fun generating
rumors, just don't attach your names to them. And do try to make them
somewhat accurate. Should the Bell System just start offering its
computers for sale (without clearing up small legal details), it would
be sued several times. When the Bell System decides let you know
about its computers (as soon as the government lets us sell them to
you), we'll let you know.
Andy Tannenbaum
Bell Labs Whippany, NJ
------------------------------
Date: 12 Sep 1981 02:51:28-PDT
From: purdue!cak at Berkeley
Subject: the Bell Computer
I suspect that the computer that Bell will announce as their first
product is the 3B<simplex>. As far as I have heard, the 3B is a
machine that was patterned somewhat after the VAX, but with high
reliability for ESS applications in mind. It can run either in a
simplex/single processor mode, which is probably what will be sold, or
in a duplex/dual processor mode, each processor watching the other
(like Tandem NONSTOP systems), for ESS applications. They hope for
something like 1 day in 40 years downtime. It does run unix, I talk to
people in the Labs who use it every day.
Chris Kent (purdue!cak)
------------------------------
Date: 14 Sep 1981 13:36:46-PDT
From: menlo70!hao!cires!harkins at Berkeley
re: Mike Wahr's query about the "bell machine"
I have not seen anything announced, BUT... about 3 months ago there
was an ad in Computerworld I think that was asking for marketing types
for a "new line of mini and micro computers" ergo, the rumor nearly
has to be true.
------------------------------
Date: 12 Sep 1981 03:30:47-PDT
From: CSVAX.dmr at Berkeley
Subject: Bell Computer Company
I can't confirm or deny the rumors reported by Mike@Rand-Unix relating
to the purported Bell System computer company, but his informant
doesn't have good gen on the potential hardware.
There are two processors. One (currently called the 3B-20) comes in
two forms: simplex and duplex. Both are built of commercial MSI. The
simplex is a conventional midicomputer in packaging and the like. The
duplex runs on 48 volts, looks like an ESS machine, and has two
mutually-checking processors. It will be used as the processor for #5
ESS. It runs DMERT, a real-time kernel (successor to MERT, see the
Unix BSTJ issue). A version of Unix is one of the supervisors that
can run under DMERT.
The simplex, on the other hand, is being pushed reasonably hard
internally (inside BTL) as an alternative to the Vax both for
conventional computing and especially for OSSs ("operational support
systems", machines for trouble reporting, record-keeping, and the
like). It runs a version of Unix that is a straightforward port of the
internally supported system.
Then, there is the 3B-5. This is based on an internally developed LSI
processor chip lately called the BELLMAC(tm)-32, previously called the
MAC-32. It does not exist as a "system" yet, but there are working
chips. Although its existence has been announced, I suspect most of
the details about design rules and the like are still proprietary--
fortunately I don't remember them. It is, however, quite large in
area. It gave rise to the joke that whereas the early, non-working
LSI chips from most projects are made into souvenir tie tacks, the
MAC-32 was being turned into belt buckles.
The two processors are "assembly-language compatible" in that there is
an assembler for their common machine language, which is called IS-25.
(IS- instruction set; 25- the old number of the organization that
developed it. A year ago all organizations in BTL were renumbered.)
The compatibility extends to the instructions and address modes,
though there can be differences in bit encoding, and some of the odder
instructions aren't in the MAC-32.
IS-25 is very strongly influenced by the Vax. The instructions and
address modes are, in fact, nearly identical, though some of the
especially recondite Vaxisms were dropped. I don't know how many
registers there are internally, but the programmer sees 16 (32 bits
each). Neither machine has anything to do with array processing.
Dennis Ritchie
----------------------------
I thought Digest readers today would enjoy seeing what the Digest
readers in the fall of 1981 were discussing in this forum each day.
And of course, soon thereafter the "Bell System Computers" did come
around, in a big way.
Happy holiday!
Patrick Townson