[TUHS] About the SPIDER network in V6

Paul Ruizendaal pnr at planet.nl
Wed Feb 15 20:16:01 AEST 2017

I have found a video by Sandy Fraser from 1994 which discusses the Spider network (but not the related Unix software). The first 30 min or so are about Spider and the ideas behind it, then it moves on to Datakit and ATM:

Although the thinking behind them is very different, the "switch" on the Spider network seems to have been somewhat similar to an Arpanet IMP.



On page 3 of the Research Unix reader (http://www.cs.dartmouth.edu/~doug/reader.pdf).

"Sandy (A. G.) Fraser devised the Spider local-area ring (v6) and the Datakit switch (v7) that have served in the lab for over a decade. Special services on Spider included a central network file store, nfs, and a communication package, ufs."

I do not recall ever seeing any SPIDER related code in the public V6 source tree. Was it ever released outside Bell Labs?

From a bit of Googling I understand that SPIDER was a ATDM ring network with a supervisor allocating virtual circuits. Apparently there was only ever one SPIDER loop with 11 hosts connected, although Fraser reportedly intended to create multiple connected loops as part of his research.

The papers that Fraser wrote are hard to find: lots of citations, but no copies, not even behind pay walls. The base report seems to be:
A. G. FRASER, " SPIDER-a data communication experiment", Tech Report 23 , Bell Lab, 1974.

Is that tech report available online somewhere?



More information about the TUHS mailing list