The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2000-May> msg00390



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

MPLS Performance analysis.....

  • From: smd@ebone.net (Sean Doran)
  • Date: Wed, 31 May 2000 22:17:45 +0200 (CEST)
  • Cc: mpls@UU.NET

Brijesh - 

| ... using CR-LDP for certain traffic (or whatever else you have),
| or by going to each router in the network and creating suitable entries
| (manually or auto-magically under the control of a TE process) in the
| routers so that a particular class of traffic goes over a statically
| configured path.

One thing over which I am sometimes accused of being provocative
is the question of whether one can take the control plane (i.e.,
the manual or automagic process that behaves in a CR-LDP or RSVP
fashion) and use it with a "label" that is in effect an IP datagram.

I think it can, and that it addresses the possibility that this:

| the core assumption is, lots of TE is required.

is only partly true (i.e., "lots" but not nearly as much as normal
best-efforts Internet traffic _in some situations_.  (Such as a
well-provisioned backbone's core)).

	Sean.