[8315] in Commercialization & Privatization of the Internet
A reply to HWB's Comment re IBM & Cost Sharing
daemon@ATHENA.MIT.EDU (Gordon Cook)
Sat Nov 13 13:39:42 1993
From: cook@path.net (Gordon Cook)
Date: Sat, 13 Nov 1993 18:39:03 GMT
To: com-priv@psi.com
Hans Werner,
Your post has merit and raises some issues worthy of serious consideration.
Yes I must agree that a pre and post ANS demarcation should be made. I have
also been told by some others that IBM received no funds prior to the T-3
upgrade.
What I object to is the ex post facto justification of the privatization of
the backbone and the bestowal of ANS's commercial monopoly on the grounds that
the parties to the cooperative agreement have spent most of the money and
Uncle Sam only a pittance. Figures from Steve to me were 30 to 60 million in
July 1991. By the time Brock Meeks got to doing his article in the spring of
1992, the figures being offered had escalated to 100 million. And then when I
asked by FOIA to see the data on which they were based, NSF had no data and
suggested I interview IBM and MCI. I did and reported the results in the COOK
Report. (IBM and MCI declined to quantify their contributions.)
Certainly the shell set up by IBM and MCI and Merit received not only very
substantial money beginning January 1991, but also a government sanctioned
advantage in the marketplace. Its and Merits Form 990s from the second half
of 1990 through September and July of last year respectively show that the
money expended on providing backbone services matched within less than 5% that
provided by the NSF. ANS and MERIT's cost sharing in these documents has been
miniscule.
IBM has certainly continued to be involved in router development, but, for
presumably legitimate proprietary corporate reasons, declines to quantify its
cost sharing. We have had a discussion before that the NSF and, I believe,
IBM view the development costs of the routers as cost sharing. And as I have
said before I cannot accept that line of argument. IBM has major commercial
strategic reasons that are forcing it to develop this technology. It shares
the technology with the government in return for which it is now getting
several million dollars a year. It also gets quite an outstanding testbed
that it can use to work out the technology kinks. This I submit should have a
dollar value in return to IBM.
So prior to September 10, 1990 was there a situation on the NSFnet backbone
for which IBM deserves positive credit? Sure. That seems to be the major
point of your post and I think I can accept that. Did this situation still
exist at the time of Mike Nelson's remark to me on Feb 28, 1992? I think NOT.
When the marchng orders came down to grant ANS exclusive commercial resale of
the backbone at the end of 1990, the US gov't warped the playing field. Folk
have been dealing with the headaches of the ever since.
At long last we have a chance to recover from these mistakes. Will we? The
solicitation now being evaluated seems to be ready to repeat the same errors.
Why? Will the award go again to IBM and an IXC? Or to ANS where public money
will further the development of an IBM/Northern Telecom "magellan" based
strategic alliance? I have a very well palced Telco source who told me last
week that the telco's find Fujitsu and Newbridge and to a lesser extent Fore
ATM switches acceptable. Magellan apparently doesn't make the cut. Does
someone in washington feel it is in the national interest to use federal money
to keep IBM and a Canadian company positioned in the NREN and NII gold rush
and thereby help the value of IBM stock recover. I wonder? So far there
appears to be a good deal of circumstantial evidence in this direction. I
believe that this is NOT a proper role for the US gov't to be taking. I hope
it doesn't happen and that my concerns will be shown to be groundless. Then
my critics can have their last laugh about Gordon's "ravings". Will they get
the opportunity? I wonder?
PS. you mention a survey of which I am unaware. Please tell me how I can
obtain a copy.
HWB Wrote: IBM also did an industry survey for T3 nodes back before the RS6k
based solution was put in place, with a stong consideration for rather using
outside products, had they been available and performed in a rather complex
environment (including advanced routing functionality at the packet level, not
just T3 switching). IBM, even within the project partnership, did not make the
RS6k decision lightly.
