[10742] in Commercialization & Privatization of the Internet
Re: "Fed **deal** may speed
daemon@ATHENA.MIT.EDU (Rick Adams)
Mon Mar 7 17:16:39 1994
From: rick@uunet.uu.net (Rick Adams)
To: dlynch@interop.com (Dan Lynch)
Date: Mon, 7 Mar 1994 10:25:50 -0500 (EST)
Cc: com-priv@psi.com
In-Reply-To: <9403061926.AA11877@polaris.interop.com> from "Dan Lynch" at Mar 6, 94 11:23:23 am
Yes. There is no technical reason (i.e. no increased connectivity).
Also, there is no financial reason (i.e. cost savings).
In reality, there is INCREASED cost. Not only do NAPs not save money
but they certainly cost more without delivering any benefit.
The "correct" technical solution is to have a "GIX" (for lack of a
better word) on each coast.
The east coast one is already operational and has lots of people
connected to it. NSF's NAP is DC is a particularly bad idea. It clearly
has no technical merit and clearly interferes with existing services.
A NAP is New York or Chicago serves no purpose other than perhaps to
subsidize regional providers in those areas (Remember how the original
list of NSF NAP sites so "accidentally" matched the location of ENSSs?)
Something in the Bay area is clearly needed. If Pac Bell had their act
together with reasonable SMDS pricing, a MAE-WEST would already exist.
All this will happen WITHOUT NSF funding, reporting or further meddling
if they would just get out of the way and stop distorting the fledgling
commercial market.
Remember MAE-EAST happened without NSF "help" in any way. MAE-WEST
would have already happened if the local telco providers in the bay
area offered a cost-effective solution, as they do in the DC area.
My rough guess is that it would cost me $100,000 per year to connect to
all four NAPS with no benefit.
Add to that NSF's proven track record of not enforcing any quality
standards on its awardees and you now have the problem of paying a
large amount of money for something that probably won't work.
No thanks. I'd rather bet my connectivity on someone that has a
financial incentive to make it work. (E.g. if MFS doesn't make MAE-EAST
work well enough, we'll switch to Bell Atlantic SMDS.)
NSF shoveled more and more money into ANS despite ANS never delivering
the contracted for (and widely falsely claimed) T3 speeds until over a year
AFTER the expiration of the original contract. Then NSF rewards this
non-performance with an extension of the extension.
Why should any NAP manager presume they have to make their NAPs work?
(For that matter, why should any NSF awardee of any kind presume they
need to deliver on their award)
That's the fundamental problem with the NSF shell game of using a
cooperative agreement instead of a contract. Use a contract and don't
deliver and GAO will be all over you. Use an co-operative agreement and
don't deliver and get a funding increase and an extension.
---rick
