[e2e] Re: Bursty traffic and TCP flows (was: TOEs and related issues)

[RJ Atkinson]

On Mar 10, 2004, at 15:40, Marko Zec wrote:
>  However, it would be even more interesting
> to hear why people can't stop talking about "alternative designs" while
> avoiding to provide a pointer to a single example (vendor X router
> model Y) which _doesn't_ count buffer slots in # of packets?

None of mine do it the way you outline, in case that wasn't clear
earlier.  I know of several other implementations who count buffer
memory in bytes, not packets.

And as Noel alluded to earlier, our architecture uses custom hardware
for IP forwarding, not a general-purpose CPU.  Some other vendors
(e.g. Juniper, Procket) also have custom hardware support for IP
forwarding.  I can't speak to exactly how the internals of those
systems work, but I am reliably told that neither is a clone of the
Proteon achitecture you cited.[1]

Ran
rja at extremenetworks.com

[1] As Noel and some others already know, I was a very happy
Proteon user in the past and would not have changed away from
Proteon back then but for the small matter of DS-3 support.
So don't take my observation that modern routers are differently
architected as criticism of the Proteon design in its era.