[11385] in Commercialization & Privatization of the Internet
vBNS commercial use - something doesn't compute
daemon@ATHENA.MIT.EDU (Gordon Cook)
Thu Mar 31 01:11:27 1994
From: cook@path.net (Gordon Cook)
Date: Thu, 31 Mar 1994 05:08:32 GMT
To: com-priv@psi.com
Cc: cook@pandora.sf.ca.us
Question for Steve Wolff or anyone else who thinks they can see answers.
I have been reading the set of questions asked about NSF 93-52 last May.
Interesting to see that the most egregious thing that NSF has done with the
vBNS is to allow MCI commercial resale. . .blatently expressed by Steve to
Ellen Messmer in her article in Network World over a month ago.
Last May, in writing, I asked the NSF to clarify what its commercial use and
cost sharing policies for the vbns would be. So did Ameritech, Advantis,
Bellcore, and NCAR.
NSF choose to ignore these critical questions. It provided no answers. Why?
NSF knew this would be controversial. Why then was its intent to grant MCI
commercial use so critical to its game plan? Why would MCI once more seem to
want to be rewarded for what it bid with commercial use? This is supposed a
small market....why the insistance on the part of all parties on commercial
use? And Steve's heavy emphasis on it with Ellen Messmer?
Someone from a mid-level recently raised the following point in a conversation
with me: what would NSF do to ensure that two commercial customers of MCI's
vBNS did not use their transit of the vBNS to talk to each other directly? In
which case what would prevent many commercial entities from taking out
contracts with the NSF funded super-computer centers for the *sole*purpose of
talking to each other rather than the supers?? I have requested by FOIA
details of NSF audits of MERIT and ANS - audits which would provide public
assurance that Merit and ANS are complyng with the requirements of Sept 10
1990 and May 24 1991. The NSF has responded that it has NO SUCH Documents.
If NSF has done nothing to verify compliance with the last sweatheart deal,
why should we expect it to do so now. I also asked for evidence of what cost
sharing MCI had promised in return for this new round of benefits. The FOIA
response last week was "no documents responsive to your request have been
found."
If Sprint wins its protest and this may be a BIG *if*, I believe it will be on
the grounds of NSF's unwarranted grant of commercial use to MCI......for whom
$50 million over five years apparently wasn't enough compensation. If MCI is
really serious about the research involved, why the insistence on the extra
commercial use sweetener?
Or as a major corporation with several thousand employees put it to NSF last
may:
"in many places in the solicitation, the VBNS is described as supporting
'meritorious high bandwidth network applications.' The implication from these
statements could be that the vBNS will only be used in research applications
requiring experimental high bandwidth networking capabilities. However, in
the background section the solicitation states that 'any traffic which is in
support of research and education will be permitted on the vbns.' This quote
could imply that it would be acceptable for all the research and education
traffic on the current NSFnet to flow over the vbns even though the
preponderence of that traffic may not be high bandwidth applications requiring
high bandwidth network capabilities. What if any restrictions on access
and/or useage will be imposed on the vbns? What mechanism will be used to
enforce this policy? Will such mechanisms be a part of the final criteria for
vbns selection?"
NSF answered neither this nor any of several closely related questions from
five different respondents. In view of this and steve's actions in the
messmer interview, how can NSF claim to have run a fair and competitive
solicitation? Now I'll bet Steve will say he has authorization from Congress
via the Boucher ammendment for his approval of commercial use. He made that
assertion here on Dec 29, 1993. I asked by FOIA whether there was any
document anywhere in the foundation supporting Steve's interpretation. The
response was no.
Something doesn't compute.
