[2235] in Commercialization & Privatization of the Internet
Commercialization via 3 Acceptable Use Policies? Why not KISS?
daemon@ATHENA.MIT.EDU (Gordon Cook)
Sat Feb 1 22:44:48 1992
To: com-priv@psi.com
Date: 1 Feb 92 22:15:34 EST (Sat)
From: cook@tmn.com (Gordon Cook)
<<MESSAGE from>> Gordon Cook 01-FEB-92 22:15
cook@tmn
I had thought there were but two acceptable use policies. That of the
NSFnet for the NSFnet backbone and that of ANS for the ANSnet backbone.
It appears that I may be mistaken. There seem to be **THREE**. I have in
front of me copies of the acceptable use policy of ANS and a quite
different ANS CO+RE acceptable use policy. The third policy would be that
of the NSF.
I am confused as to why ANS has two policies. I thought there were only
two virtual backbones: NSFnet and ANSnet.... not three. But perhaps the
number of virtual backbones isn't the right question to focus on.
Apparently one can join ANS as either a CO+RE customer or as an R&E
customer. That would explain why Dialog was able to revert back to R&E
status. Apparently Abbot Labs and Union Carbide signed on during the
summer as R&E customers? Ah, it's getting clearer now. Purdue, VPI, and
CONCERT must be ANS R&E customers. (Someone stated that Dialog was the
first CO+RE customer.)
But if one joins ANS (which is a commercial provider), why on earth would
one join as an R&E customer when one could have the presumed benefit of
sending commercial traffic over the ANSnet backbone?
I suppose that we have seen the answer now. He who joins as CO+RE can
talk to only those relatively few mid-levels that are commercially
compliant with the ANS backbone rules. Why this is the case remains
unclear to me (and I guess to a few others) despite the reams of traffic
of a month ago. Since it is ANS' backbone, I guess ANS can make its own
rules for its own benefit. I do not see any force or law EXTERNAL to ANS
that requires it to make the rules that it does.
This lack of access exists because most mid-levels appear to find the ANS
agreements to be unacceptable. Nevertheless, with their own territories,
these mid-levels, for all practical purposes, allow self defined levels of
commercial use. They certainly allow use that would be unacceptable over
the NSFnet backbone. They allow within their networks (for the most part)
the commercialization that policy makers have been trying to get for the
net as a whole.
Now I thought that one advantage of giving ANS the right to use for its
own purposes a part of the here-to-fore government provided backbone was a
way commercialize the network. Customers of the mid-levels would now be
able to send accross the backbone the kind of traffic that they were free
to send within the midlevel. In the meantime ANS would add its own
commercial customers. Everyone could talkto everyone else and all would
live happily ever after.
Instead the ANS agreements, combined with the mid-levels reluctance to
sign them, appear to have recreated the same commercial use versus R&E use
dichotomy within ANSnet that exists within the NSFnet and that I thought
ANS had been created to avoid. How can it be said that these policies are
leading to the successful commercialization of the NSFnet when the Dialog
as the very first CO+RE customer throws up its hands in despair and bails
out?
If policy makers still desire the commercialization of the network and if
they cannot convince ANS to drop its agreements, then perhaps its time for
them to give the CIX a chance to perform?
