[1637] in Commercialization & Privatization of the Internet
NSF/ANS Chronology
daemon@ATHENA.MIT.EDU (William Schrader)
Sat Dec 7 23:18:11 1991
Date: Sat, 7 Dec 91 23:15:29 -0500
From: wls@psi.com (William Schrader)
To: com-priv@psi.com
To: The Internet Community
>From: Bill Schrader
In early November, Dave Hughes and Joel Maloff debated on this list
the contents of my presentation to the Westnet Technical Group in
Boulder, CO. In that talk, I presented a chronologically ordered
review of all actions concerning the NSFNet Backbone (economics/
techology/politics) taken by the team of NSF-MERIT-IBM-MCI-ANS.
ANS's quiet listening in Boulder, but strong response in the email
debate confused me, but could be attributed to the NREN Buy American
provision which was also discussed.
However, since ANS asserts that all the actions of this team were
totally legitimate, it seems to run counter to my analysis of
history. Given the extension of this lucrative contract for an
additional $15M, and recent postings which support ANS's purportedly
exclusive position in the market, I am posting the chronology section
of my presentation here for your reivew and consideration in making
your own determination. I presume someone from ANS or NSF will
attempt to set my chronology right, since I saw it as an outsider.
For those with patience to review these transactions carefully,
consider whether they were fair and legitimate from the position of
the law, the taxpayer, other federal agencies, international networks,
the network users, institutions connected through regionals, regional
networks, other national or international providers of service.
[And, yes, there are some loaded words in the remarks, which were
meant to entice listeners and the ANS panelist to debate.]
Bill Schrader
--------------------------------------------------------------------
[from WLS presentation to Westnet, Nov 6, 1991.]
Current Status of NREN activity
Through its actions, the NREN lead agency, the National Science
Foundation, has positioned its current contractor (Merit) together
with its joint study partners (IBM and MCI), and subcontractor (ANS)
to dominate the NREN and, because of it, potentially monopolize the
emerging market.
Lets us review the chronology of events, both public and private,
which have resulted in the above conclusion.
'86 NSF issued its Solicitation for a T1 network, to be called
the NSFNet Backbone, and received several competitive bids.
'87 NSF awards the contract (called a Cooperative Agreement) to
Merit, with IBM and MCI as its Joint Study Partners, to build
the T1 NSFNet Backbone network. Their bid is considered
by some to include excessively costly and unnecessary
equipment, which is donated by the Joint Study Partners.
'87-89 Merit uses proprietary IBM hardware and software
including RT/PCs to build the T1 NSFNet Backbone. It takes
months to stabilize using this gear instead of the commercially
available hardware and software specifically designed for the
task from vendors such as Proteon and Cisco.
'90 NSF negotiates with Merit/IBM/MCI a T3 upgrade to the
NSFNet Backbone, which was not in the original solicitation,
and injects $6,000,000 of NSF funds into the Merit/IBM/MCI
team in FY 91 for T3 and $10,000,000 more in FY 92. This
constitutes more than a 300% increase over earlier payments.
'90 NSF/Merit/IBM/MCI begin their public relations campaign
to justify the T3 upgrade, without any significant traffic data
to back it up. They purported to "lead the state of the art"
by using equipment and software not designed for this
purpose (IBM RS6000s) at a time when T3 was not required.
'90 Merit/IBM/MCI announces ANS, with NSF in attendance and
Senator Gore in full public support. Clear statements are
made by all that ANS is a "not-for-profit" and has no
intention of competing with for-profit enterprise.
'90 NSF changes the NSFNet T1 Backbone from an NSF "resource"
to an ANS owned T3 facility to which NSF had purchased "access".
This appears to have been a privatization of the Backbone without
public comment, without open competitive bidding, and without
(presumably) the concurrence of the National Science Board.
'90 NSF gives Merit/IBM/MCI/ANS the exclusive right to carry
commercial traffic on the NSF funded T1 or sponsored T3
connections to the Backbones. This agreement also appears
to have been reached without public comment, competitive
bidding, or the concurrence of the NSB.
'91 ANS/Merit/IBM/MCI use proprietary hardware/software to
build ANS's T3 Backbone which the NSF has access to. After
nine months of "production operation" it remains unstable.
Even after changing the primary routers on the T3 to Cisco
hardware and software, network remains unstable due to
other proprietary hardware and software.
'91 ANS announces a for-profit subsidiary called ANS CO&RE to
carry commercial traffic, with exclusive connectivity to NSF's
network. ANS CO&RE is espoused to market to the commercial
segment in response to continuing demand. ANS
is pushing Fortune 1000 firms to connect "directly to the
NSFNet backbone without any risk of violating the NSFNet
Acceptable Use Policy" since it alone can make that promise.
'91 ANS offers Regional Networks an opportunity to sign a
contract with ANS which allows commercial traffic to flow, and
charges the Regionals up to twice as much as academic traffic.
'91 ANS develops a pricing strategy where commercials who
sign directly with ANS pay less than academics, due to
competitive pressure. Despite rhetoric in T3 and gigabits,
ANS primary sales are in 56 kbps and T1, which for the time
being appears to be the highest speed anyone in the market
requires and is willing to pay for.
'91 The Gore Bill passes, NREN money is possible, and will be
directed toward gigabits and T3, all the things which ANS
claims to represent.
In addition to this chronology, the NSF/Merit/IBM/MCI/ANS team
has led the marketplace into believing that the following untruths
are facts:
1. ANS runs the NSFNet Backbone. (In truth, Merit does.)
2. ANS is pushing the "state of the art in T3 networking". (In
truth, it is delaying the state of the art, using taxpayer money,
installing and operating inefficient proprietary T3 equipment
without any reasonable justification.)
3. NSFNet Backbone equals the NSFNet. (In truth, the "NSFNet" is
made up of Regional Networks and this Backbone. The Backbone is
one of the smallest networks in the country in terms of customers
and connected sites. They are the largest network in the country
when measured by unused bandwidth.)
4. NSFNet equals the US Internet. (In truth, DOE, NSI are much
larger, and commercial nets make up nearly 40% of the US Internet.)
5. ANS's network equals the US Internet's "IP core". (In truth,
Merit is.)
6. "GIGABIT speeds are required" is the largest single incorrect
statement. (In truth, no one needs gigabits per second except as an
aggregation of other traffic.)
The Merit/IBM/MCI/ANS/NSF lobbying effort continues as a "unit"
supporting their own collective benefit at the House, the Senate, at
the White House Office of Science and Technology Policy, at the Office
of Technology Assessment (OTA), and in some state legislatures.
Their goal appears to be to keep the money directed toward the
backbone contractor, rather than the Regionals or Institutions
needing connectivity.
===============================================================
