[ih] Dave Farber's DCS
Brian E Carpenter
brian.e.carpenter at gmail.com
Thu Jul 23 22:30:36 PDT 2020
What was the date, approx?
The Cambridge Ring (1974 onwards) avoided that particular defect by having
a packet size of 4 bytes (2 byte address, 2 byte payload) so really it was
more of a memory bus than a network. (Not that I ever worked on it.)
Of course IBM didn't care about you people, they only cared about the Söderblom
On 23-Jul-20 13:44, Dave Crocker via Internet-history wrote:
> On 7/22/2020 4:48 PM, Karl Auerbach via Internet-history wrote:
>> (IBM had to concede to DCS that IBM did not invent the token ring
>> network concept.)
> IBM's token product came during my time at Univ. of Delaware, almost 10
> years after the DCS work, and Farber had an IBM contract. They sent a
> senior tech out to present their token ring design to him and his
> students, all of whom were EEs (other than me). This was clearly an
> intent to get the august professor to bless their work.
> In spite of not being all that well versed in the fine-grained technical
> aspects of local area networking design choices, I could tell that their
> design had a basic problem: The Irvine ring had a cost of one bit-time
> per node the packet went through. The IBM insertion ring design imposed
> an entire packet-time per node. So much for scaling.
> After the presentation, it was interesting watching the discomfort among
> Dave and the other students, non-verbally negotiating who was going to
> deliver the assessment to the IBM person.
More information about the Internet-history