[171] in Commercialization & Privatization of the Internet
re: Was a "big Internet" needed to make TCP/IP useful?
daemon@ATHENA.MIT.EDU (steve@cise.nsf.gov)
Thu Nov 15 13:33:10 1990
To: nowicki@Legato.COM (Bill Nowicki)
Cc: com-priv@psi.com, cperry@gateway.mitre.org
Date: Thu, 15 Nov 90 13:07:43 EST
From: steve@cise.nsf.gov
> Collaboration is a great idea, so if NSF wants to fund research on
> building better collaboration tools, that would be wonderful.
We have a program in "Coordination Theory and Collaboration Technology";
it's funded by three divisions of the Computer and Information Sciences
and Engineering Directorate (CISE) at NSF and started two years ago.
> Also I have no objection to DS-3 speeds, if and when the network
> providers view it as cost-effective.
>From a pure bean counter perspective, whenever the load on a particular link
exceeds k*T1 (where k is distance sensitive) then T3 is cheaper. From the
science policy viewpoint T3 is cheaper if your estimate of the discounted
future value of the subset of the things discovered by those served which
could not be discovered at T1 exceeds the T3-T1 cost differential. And
that's where reasonable folk may differ.
> There are many ideas for making networks more efficient that could make not
> only NSFnet but other networks better, for many years to come.
Does "efficient" = "better" universally? Why aren't PSI and the other
network offerors using them? or are they? It *is* clear that many US
consumers are *not* interested in "efficiency."
-s
