[2602] in Commercialization & Privatization of the Internet
Re: Network World article Raises New Questions on T-3 Network
daemon@ATHENA.MIT.EDU (Susan Eldred)
Fri Mar 20 10:35:12 1992
Date: Fri, 20 Mar 92 10:24:17 EST
From: Susan Eldred <eldred@ans.net>
To: cook@tmn.com (Gordon Cook)
Cc: com-priv@psi.com
In-Reply-To: Your message of 18 Mar 92 21:11:59 EST (Wed)
>
> <<MESSAGE from>> Gordon Cook 18-MAR-92 21:11
> cook@tmn
> Internet migration to T-3 stalled by router problems, by Ellen Messmer
> Network World march 16th
>
> ".... ANS ... has been unable to obtain the Cisco Systems Inc T-3 routers
> it purchased last fall for shipping data at T-3 speeds. ANS said that the
> prototype IBM RISC System/6000 based T-3 routers it employs on the
> backbone have only been able to reach speeds of a bout 12 megabits per
> second at best. . . . . according to weis he main problem appears to
> center on the high volume of broadcast traffic required to communicate
> with the 5,000 nets attached to the NSF backbone. 'That's a difficult
> thing for routers to do' he said adding that the Cisco AGS+ routers are
> now being used at 56 KBS speeds on parts of the NSF Network while testing
> of routers at T-3 speeds continues on on the ANS T-3 test network. [GC:
> what is meant by the ANS T-3 *TEST* network?]
>
> However Cisco Officials were surprised to hear of the problems. 'It's
> news to me' said Christine hemrich, software product manager at Cisco.
> I'm surprised they are having difficulties.
>
> . . . .
>
> AnS officials have blamed the problems on the IBM prototype T-3 router
> interface cards. Last dceember ANS bought the cisco T-3 routers to
> supplement the IBM equipment on the T-3 backbone. the problems ANS is now
> experiencing with the Cisco T-3 router, which others have said can operate
> at only up tp 6.318 mega bits per second because of its v35 interface
> lmitations, raise the questions of whether technical limits to high spped
> routing exist in complex environments such as the internet. . .. "
>
> Phew! not the clearest thing i ever read. Is the Cisco AGS+ a t-3
> router? What's this about a v35 inteface limiting it to 6 plus megabits
> per sec.???> And what's the talk about limits to high speed routing in
> complex environments such as the internet? I thought we were in a march
> to gigabits. And how are the problems that ans has experienced now to be
> blamed on CISCO? Or are they? I am sure this is clear to memebers of the
> IETF. however it sure is *NOT* clear to me and I'd bet it isn't very
> clear to the majority of the list.
>
> Perhaps Peter Ford would be willing to comment. or Mark Kopper or any of
> the other dozens if not hundreds of very talented technical readers of
> this list?
>
>
Gordon,
The best thoughts I can give you regarding the Network World article is to
give up on understanding it -- it's a total mess. To give you just a sample,
-- We have had no delays in the delivery of any cisco routers.
-- We never told her that the blame for the earlier problems we had with
T3 was the fault of the IBM routers.
-- The Weis quote about routing problems "appear to center on the
high volume of broadcast traffic" -- I have no idea where that came
from.
We have written the editors of Network World expressing our
disappointment.
The story is just a total mess.
Susan Eldred
