[2699] in Commercialization & Privatization of the Internet
Frustrations re the T-3 Backbone, Cisco Routers, and Leadership
daemon@ATHENA.MIT.EDU (Gordon Cook)
Thu Apr 23 22:51:04 1992
To: com-priv@psi.com
Date: 23 Apr 92 22:20:29 EDT (Thu)
From: cook@tmn.com (Gordon Cook)
<<MESSAGE from>> Gordon Cook 23-APR-92 22:20
cook@tmn
The T-3 backbone continues to confound most observers. Consider what I
learned from Cisco at the Telestrategies Conference April 21 and 22nd. In
November of 1991 Cisco began shipping a working clear channel T-3 router.
I mentioned that I had heard people from the NSF say in March of 1992 that
there were no commercially released routers that would do what had to be
done on the T-3 backbone. Did Cisco's product have some deficiency?
Absolutely not. It would do everything necessary on the NSFnet ANSnet T-3
backbone.
Well, it was being used? ANS had said that it was using Ciscos colocated
at MCI POPs on the T-3? Did Cisco agree?
Well.... sort of. The T-1 safety net backbone that connects to the T-3 as
a backup for T-3 failures is linked to the T-3 at 56kbs. Cisco routers
are used on *that* linkage.
But what about the T-3 routers?
Some were used in the ANS test network was the answer but none in the
"production" network.
One wonders why.
But perhaps we will hear that finally all 16 nodes are peering with each
other? And therefore the net is complete and that IBM's betatest T-3s are
no longer betatest? But please someone refresh my memory....when I last
brought this discussion up a month ago someone was saying that for
technical reasons that were /are unclear to me the maximum throughput on
the t-3 backbone was only a bit more than 5 megabits per second. Is this
still correct? And if it is still correct then aren't we facing a
situation where ANS - presumably on behalf of IBM - is using the NSFnet
backbone as a TESTBED for continued T-3 router development rather than
supplying the higher bandwidth *NOW* that supporters of the NREN vision
continue to assert is necessary?
The person to whom I was talking said that my interpretation that ANS was
using the backbone primarily as a testbed was perhaps unfair to ANS
because in Dec 90 when they started to develop their T-3 routers, Cisco
did not have its product out and ANS had made an awfully large investment
in its prototypes. He could understand why it would hurt to dump them now
and replace them with Ciscos. The desire to complete the testbed would
obviously be strong.
But consider that the NSF is paying ANS an extra $7 million a year for a
backbone that some believe is being used primarily to benefit *product*
development on behalf of IBM. Some might think it strange that the
government, having announced that it needs much higher speeds on the net,
is paying much more money for slightly higher speeds to an awardee chosen
almost five years ago. I have heard some wonder given the benefits to IBM
that IBM through ANS enjoys, ANS (IBM) is not PAYING the National Science
Foundation for the privileges it enjoys.
Now one of the staffer's from the House Science committee acknowledged
that the concern that led to the hearings in March was ignited by the
appearance that the NSF was not in control. Yet here we are at a major
meeting where the NSF was the ONLY major player that did not even SHOW UP.
And here we are with some suggestions that the draft solicitation that
was due by the NSF's own schedule in February is not likely to make it out
in May and may be significantly further delayed.
One wonders whether the NSF will be asking the science board to extend the
Merit agreement this fall by 24 to 30 months instead of 12 to 18? I have
the feeling that if the NSF were a private corporation we would have seen
some major shakeups by now.
How did we get in such a log jam that representatives of the congress are
saying that they can do nothing until they see what the NSF is going to do
and the NSF responds by doing nothing?
I submitt that the network is too important to be "run" this way. We KNOW
what the problems are although we all don't agree that they are the same.
Why can't we find a LEADER to start SOLVING them? Surely I am not the
only person who is deeply frustrated?
Any one have some ideas on how to solve the problem?
