[10446] in North American Network Operators' Group
Re: Another UUNET Explanation
daemon@ATHENA.MIT.EDU (Dorian R. Kim)
Wed Jul 2 03:32:49 1997
Date: Wed, 2 Jul 1997 03:27:33 -0400 (EDT)
From: "Dorian R. Kim" <dorian@blackrose.org>
To: Robert Bowman <rob@elite.exodus.net>
cc: Alex Rubenstein <alex@nac.net>, joe@via.net, nanog@merit.edu
In-Reply-To: <199707020255.TAA15763@elite.exodus.net>
On Tue, 1 Jul 1997, Robert Bowman wrote:
> On the inverse, routers--if designed properly, can be 10 times better
> because of lower overhead, cost effectiveness, etc. I don't think this
> is going to be a debate that one will win in this forum. There are
> many router only based backbones, such as ours, Digex, etc. Then there
> are the many on the fr/router side.. just as "effective" one might
> say..
There is nothing inherently good or bad about switches either. The problem
comes in, as Sean stated far more eloquently than I can, when layer 2 and
layer 3 each have their own intelligence and do not work together. This is
probably the single most problematic part of the layer 2 + layer 3 designs
I've seen implemented.
While this has in large part caused by the lack of correct tools to build such
a beast, when you are trying to tie a necktie with two hands and your left and
right hands don't coordinate, you sometime end up hanging yourself instead.
-dorian
