Re: draft-lee-mpls-path-request
<leecy@sympatico.ca> Thu, 31 January 2002 18:49 UTC
Envelope-to: ccamp-data@psg.com
Delivery-date: Thu, 31 Jan 2002 10:51:07 -0800
From: leecy@sympatico.ca
To: jeff.pickering@caspiannetworks.com
CC: ccamp@ops.ietf.org
Subject: Re: draft-lee-mpls-path-request
Date: Thu, 31 Jan 2002 13:49:01 -0500
MIME-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit
Message-Id: <20020131184901.JTFI23714.tomts9-srv.bellnexxia.net@[209.226.175.20]>
Jeff, Thanks for your feedback. Comments inline. regards cheng-yin Jeff Pickering wrote: > cheng-yin, > > I have a couple questions/comments on your draft: > > 1) The reference to [OSPF-TE-EXT] used in the resource class definition > is not in the reference list. > 2) It would seem reasonable to deploy PCE capability inside ABRs and > let the IGP instance in the LER determine how to find/recurse PCEs > to > make a complete ER. I think this is what the document means, but its > not > clear. Yes, this is what it means. The LERs and LSRs would know where the ABRs are. > On the other hand, the draft-ash-multi-area-te-reqmts seems > to > state that your idea requires out-of-band signaling to locate PCEs. > I'm > confused, can you clarify? You may be referring to draft-lee-mpls-te-exchange which allows non ABRs to be PCE as well. > 3) It would be nice to explicitly state how a "no ER available" is > returned by > a path reply. Will do. > 4) I presume the disjoint route tlv is used to retry after LSP setup > failure. It would > help clarity if something to this effect was put in the text. The disjoint route tlv is used to specify that the path to be computed should be disjoint from the specified disjoint route tlv (not necessarily after an LSP setup failure). We'll add more clarification as you suggest. > > Cheers, > Jeff