Jacking in from the Protest Port: by "Brock N. Meeks" <brock () well sf ca us>

[David Farber <farber () central cis upenn edu>]

CyberWire Dispatch // Copyright (c) 1994 //


Jacking in from the Protest Port:


Washington, DC --  The General Accounting Office (GAO) recently filled a
temporary pothole on the road to what will eventually become the nation's
foremost research and development tool: A high speed computer network that
will link six supercomputer centers located throughout the U.S.


On May 9, the GAO issued a six-page finding in which it dismissed Sprint's
protest of a $50 million award given to MCI by the National Science
Foundation (NSF) to build its very high speed backbone network service
(vBNS).  The NSF awarded MCI the job of building what, for all intents and
purposes, will become a hands-on driver training course for the information
superhighway.


The job of building this network out of what the NSF (and MCI) likes to
call "bleeding edge technology" will be done under a specialized grant
called a "cooperative agreement."  Such agreements don't carry any of the
performance penalties of a hard and fast contract.  If MCI (or its
subcontractors) screw up -- and both the NSF and MCI have more or less said
screw ups will occur, given the state of "experimental" technology that
will go into this flagship network -- MCI still gets its check in the mail.


No shirt, no shoes, no data?  So what?!  Cut those guys a check anyway.
Because, well, hell, at least they're trying.  That, in essence, is the
soft white underbelly of the so-called cooperative agreement.


It was the choice of a cooperative agreement as the funding vehicle that
pissed off Sprint, who was a losing bidder, along with AT&T, among others.
Sprint also alleged that federal funds were essentially going to underwrite
the development of MCI's commercial Asynchronous Transfer Mode (ATM)
technology.  ATM is a technology some say is crucial to high speed
networks.  Sprint and the rest of the world thought MCI didn't have any ATM
technology, until May 4 when MCI pulled the proverbial ATM rabbit out of
its hat by announcing a six-node precommerical ATM network backed by
Northern Telecom.


Sprint sweetened its protest by claiming that there also was a conflict of
interest involving an MCI subcontractor and a member of the National
Science Board, the group that rides herd over the NSF checkbook.


But Sprint's cries fell on deaf ears.  The GAO punted.  Instead of showing
some political will, it chose to simply dismiss the protest out of hand
claiming it has no jurisdiction to review a protest filed when a
cooperative agreement is involved.  (Didn't we tell you this funding
creature had a soft white underbelly?)


In other words:  GAO didn't have the balls to consider the protest.


Apparently the GAO comes up short in technology wonk department, too.  When
discussing Sprint's protest, the GAO writes that Sprint alleged "that
federal funds are improperly being used to underwrite MCI's commercial
development of an AUTOMATED TELLER MACHINE service."  (Emphasis not in the
original.)


Now that is a scoop.  Or it would be if the GAO lawyers had a brain. What
Sprint meant -- and what the GAO obviously has no clue about -- is ATM, as
in the high speed data transfer technology which has been unceremoniously
hyped as the Christ of High Speed Computing.


GAO said under the Competition in Contracting Act (CICA) of 1984 and its
own Bid Protest Regulations, it's only allowed to review protests of real
contracts.  Those ironclad, fuck-up-and-you-pay-the-piper procurement
agreements that aren't suited for our nation's most prestigous, most
expensive computer resources.


Sprint tried to make the case in its protest that the High Performance
Computing Act, in which Congress ponied up billions in funding for various
high speed computing projects, is actually a procurement statute.  GAO said
"we disagree."  End of discussion.


GAO said that because the High Performance Computing Act doesn't
specifically tell the NSF to use a contract, well, then it can pretty much
damn well chose whatever funding mechanism it wants. And in this case, it's
a cooperative agreement.


And because GAO doesn't have jurisdiction over cooperative agreements, it
also chose not to review the allegations of conflict of interest.  But that
didn't stop them from commenting on the allegations just the same.  You
see, the president of the National Science Board (remember, they hold the
NSF purse strings) is James Duderstadt.  He's also the president of the
University of Michigan (an institution that also happens to be among the
top 5 in total taxpayer dollars received every year as a result of NSF
grants).  The university of Michigan is also home to Merit, which is a
consortium of Michigan universities and is an MCI subcontractor to help
build the vBNS with $50 million of your money.


As the GAO decision points out, Duderstadt is entitled to a seat on Merit
board of directors.  A seat he has apparently chosen not to fill, GAO says.
Still, that brings up the pesky question of if Duderstadt is president of
the National Science Board and the National Science Board approved $50
million to go to MCI, some of which will flow to Merit, does that portend a
conflict of interest?


You make the call the GAO went mute on the subject.  GAO said "we do not
review protests of cooperative agreement awards based on allegations of
conflict of interest." Oh, and by the way, GAO is kind enough to inform us
that Duderstadt (who isn't formally identified in the ruling expect by
title) "did not participate in the selection of MCI for the award."


With the Sprint protest out of the way, MCI and NSF can get down to
negotiating the final details of the vBNS contract.  The technology will
choke and sputter and maybe even ignite a spark or two of technological
advancement.  Let's hope that, unlike the current flagship network, the
operators  of the NSFNet (which is Merit, for those keeping score), MCI
keeps both hands on the table and is open and honest about its own
network's shortcomings.  When the technology stumbles, let's hope they
admit it and provide details the entire computing network can learn from.


It's the least they can do with real world test lab that would cost
hundreds of millions to build and maintain if they had to do it themselves.
Instead, MCI gets a token $50 million for driving lessons on the so-called
information superhighway.  Let's hope they remember the correct hand
signals when the turn indicators burn out.


Meeks out...