"Evan Koblentz" <evan@...
>>>> technologies that came out of d|i|g|i|t|a|l. One of those technologies is Ethernet
>Please clarify DEC's role vs. PARC's.
From the "DIX" specification:
A Local Area Network
Data Link Layer
d|i|g|i|t|a|l intel XEROX
(note who was responsible for all documentation)
This document contains the specification of the Ehternet, a local area
network developed joinly by Digital Equipment Corporation, Intel Corp-
oration and Xerox Corporation. The Ethernet specification is the result
of an extensive collaborative effort of the three corpporations, and
several years of work at Xerox on an earlier prototype Ethernet.
The early ethernet from Xerox was a 3 mbit/sec. prototype. It was based
on slotted-Aloha. The Ethernet of V2 was CSMA-CD and 10mbits. Intel did
and supplied the transceiver chips. The controllers for the transceivers
(in the early VAX 11-780 days, the DEUNA -- DEC Etherner Unibus Network
Adapter -- with one DMA driver developed by yours truly) were designed and
built by DEC. [**]
The data link layer was very much influenced by DEC -- developers of the
UART -- who'd understood issues of serial communications protocols. The
Ethernet's FCS (Frame Check Seq.), as an example, uses a CRC AutoDIN-II
polynomial. For those not so VAX instruction set familiar, the VAX has a
single CRC instruction to efficiently perform this calculation as part of
its instruction set. The VAX predated the Ethernet. ;)
Xerox maintained the preamble address assignments and was responsible for
the seed development. Metcalfe left Xerox, if my memory serves, before it
was involved in DIX and the "Ethernet"
I'm sure there's actual history on all of this published somewhere. I'm
only familiar with what I known from working with/on it at the time. I'm
a programmer/developer, not a historian. ;)
[**] I don't have a DEUNA card but I do have a DEBNT (DEC Ethernet BI-BUS
Network Transceiver) pinned to my office wall.