[ih] Early History of the Internet

[Karl Auerbach]

(This is a second repost, this time via another of my email providers, 
the original and first repost both seem to have vanished, silently, into 
the luminiferous ether. Between odd problems such as this and the slow 
accumulation of various kinds of anti-spam/anti-something-else blocks I 
begin to wonder if the net isn't perhaps beginning to get a case of 
arterial blockage.  But so much for my whining ... now onto the topic at 
hand:)

Let's not forget the work of Dave Farber and his students on the 
Distributed Computer Network (DCS) during that late 1960s at UC Irvine.

I first became aware of that work in the rather early 1970s when I was 
at SDC.  Frank Heinrich (one of Farber's students) had worked on a 
distributed file system for DCS.  Frank joined our network/operating 
security research group (Dave Kaufman, Jerry Cole, and myself) and 
introduced us to the notion of an operating system and applications in 
which the various parts would exist on separate computers of disjoint 
architectures and that work would be allocated via a bid-quote-contract 
exchange, That idea was very much like what has developed today with 
today's web based computing where pieces of applications go forth on the 
net to locate service APIs (often reified today via things like REST).

(We can also look inwards at how similar ideas, but on a different and 
smaller scale are used inside modern processors as they schedule work 
among collections of processing units.)

Sun's slogan "The network is the computer" can just as easily turned 
around to be "The computer is the network."

To me this notion of the network as a collection of functions forming a 
distributed process (filled with all of the kinds of feedback loops and 
emergent properties of many distributed processes) was a major 
revelation and a big break from my prior view of a network as a 
collection of interconnected computers. It had a huge impact on my 
approach to network management - from dealing with individual computers 
(and routers and such) to dealing with the overall distributed process.  
(That kind of alternate view has, I believe, been of use in some very 
useful recent ideas, such as the work of Dave Taht and others on 
bufferbloat along network paths.  I have my own own hopes to introduce 
kinds of homeostatic self-diagnosing, self-adjusting system to the net 
that try to deal with the net much the same way that medicine often 
measures indirect affects and applies indirect influences to adjust and 
correct the health of individual people.)

I nudged and cajoled Dave F. for several years to create something to 
help us remember the oft overlooked DCS project. Here's the resulting 
online session:

https://youtu.be/Noqf36Fx20s

         --karl--