The Routing Over Large Clouds Mailing List Archive by date

Cell Relay Retreat>List Archive>month:1996-May> msg00159



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

My personal take on cell switching routers

  • From: Sam Wilson <ercm20@tattoo.ed.ac.uk>
  • Date: 30 May 96 18:15:20 BST
  • cc: ion@nexen.com

> ...
> Given the existence of CSRs, does the need for NHRP go away?  My own
> opinion is that the need for NHRP probably goes away when ALL routers
> on an NBMA network are integrated into the NBMA-layer switching fabric
> (so, for example, every ATM switch is also a router), ...

Surely it's when every router is also an ATM switch rather than the
other way round?

> ... and router hops
> are no more expensive than non-router hops (because there AREN'T any
> non-router hops).

... and router hops are no more expensive than non-router hops because
each router hop is (or can be) a non-router hop.

> However, I also think the reality is that neither non-cell-switching
> routers nor non-router ATM switches will be going away for quite some
> time, if ever, thus the need for NHRP.

NHRP arranges for router hops to be avoided by the edge devices
bypassing the routers; CSRs arrange for router hops to be avoided by the
routers redirecting the traffic.  In fact you could see it as just a
development of router fast-pathing where the router's fast path just
happens to be the same technology as a link level switch (please indulge
me - I DON'T want to get into the 'is ATM a link level' argument). 


Sam Wilson
Network Services Division
Computing Services, The University of Edinburgh
Edinburgh, Scotland, UK