The MPLS WG Archive[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] Question of RRO behavior
David, Okay, thanks for the refresh on pinning. :-) It might be a bad idea for the ingress to infer from an RRO that an LSP did not follow the path given in the ERO. This is not being 'liberal in what you receive' since the specification doesn't say that the address used in the RRO MUST be the same one that was provided in the ERO for the same Path message (it does give guidelines as to what address SHOULD be used, and it RECOMMENDS consistent use from Path message to Path message but, it if it also says that the address MUST be the same as the one contained in the ERO, I didn't see it). I think the least that an LSR should do, if it is actually checking to see if it is expecting an RRO when it does not get one, is record this event at least once so that the operator might be able to discover that an inconsistency exists. I suspect that most LSRs are not doing this check, so the absence of an RRO in a Resv message will mean that it will not include an RRO in a Path message sent upstream - automagically. Eric W. Gray Systems Architect Celox Networks, Inc. egray@celoxnetworks.com 508 305 7214 -----Original Message----- From: David Charlap [mailto:david.charlap@marconi.com] Sent: Friday, July 26, 2002 3:38 PM To: 'mpls@UU.NET' Subject: Re: Question of RRO behavior Gray, Eric wrote: > > Can you give an example of an appropriate action > that might be driven by RRO, even supposing it has been > correctly constructed? My impression is that the RRO > is intended to be used for information only and is not > intended to incur some activity in response... It may be used to verify that the LSP is actually following the route specified by an ERO. It may also be used as a part of route pinning - generate a Path without ERO, or with a loose ERO, and then compose a strict ERO from the RRO in order to pin the route in place. -- David |
|