[1698] in Commercialization & Privatization of the Internet
nsfnet as large-scale testbed
daemon@ATHENA.MIT.EDU (Stephen Gould)
Wed Dec 11 12:02:05 1991
Date: Wed, 11 Dec 91 12:02:35 -0500
From: sgou@seq1.loc.gov (Stephen Gould)
To: com-priv@psi.com
In monitoring this discussion over the past few months, I have seen very
little acknowledgement of what I have always understood to be a key
characteristic of the nsfnet backbone--that, among other things, it is
intended by NSF to be a large-scale testbed for high-capacity computer
network technology.
For better or worse, NSF (in cooperation with DARPA) has a mandate from
some segments of its research community constituency, the Bush administration
and now Congress to steadily increase the information carrying capacity of
computer networks available to the research and education communities into
the gigabits per second range. While I have never read the cooperative
agreement NSF has with Merit, I assume that steady progress in increasing
the carrying capacity of the backbone--starting with 1.5 megabits but not
ending there--has been part of the deal.
The purpose of the subsidies offered by NSF to potential nsfnet backbone users
is at least twofold in this regard: 1) to attract a sufficient volume of usage
to adequately test each new generation of carrying capacity on a large enough
scale to prove their suitability for use in the real world (and get the bugs
worked out until they are), and 2) to compensate users for the inconveniences
associated with using not-yet-ready-for-primetime technology. Users who need
a higher degree of reliability and who do not need bleeding-edge transmission
speeds should go elsewhere. As the saying goes, there is no such thing as a
free lunch. So if you need reliability, or do not qualify for getting the
subsidy because you are not sufficiently R&E oriented, take your business to
one of the growing number of non-nsfnet providers of internet services.
If my understanding is wrong, I would appreciate an official rebuke from Steve
Wolff since this is how I explain what nsfnet is and is not to my Congressional
staff clients (I am an analyst with the Congressional Research Service).
Thanks.
