The MPLS WG Archive[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] [mpls] One LDP Implementation specific question of receivelabelmapping for prefix FECs
Hi Pranjal, all, That's a good question; thanks for asking ;-) Currently, as per RFC 3036 section 3.5.7.1, Ru should perform a longest match. But we believe that in some situation (eg use of IGP areas), we could benefit from doing a longest match as this would relax the need for leaking in the IGP all the /32 prefixes across area boundaries. To initiate discussion on this point, we've written draft-decraene-mpls-ldp-interarea-01.txt and presented it in Vancouver ( http://www3.ietf.org/proceedings/05nov/slides/mpls-4.pdf ) Right after this presentation, it was suggested to gather feed-back on the usefulness of this ID before moving forward. Pranjal, all, could you please tell us whether you think that it is worth being pursed? Best regards, Bruno ________________________________________ From: Dutta, Pranjal [mailto:pdutta@riverstonenet.com] Sent: Saturday, April 08, 2006 10:11 PM To: mpls@ietf.org Subject: [mpls] One LDP Implementation specific question of receive labelmapping for prefix FECs Hi, I have a basic confusion on the "Label Mapping Receive Procedure" explained in the appendix of RFC 3036. One step says "Find the FEC next hop". When Ru receives a label mapping for a FEC of type IPv4/v6 prefix from Rd, the next hop for the FEC can be determined after route lookup. Now here, for looking up the FEC at Ru do we need to do exact match or Longest prefix match? It is quite possible that the prefix as FEC mapped by Rd may be within a supernetted prefix at Ru. The reason why I am asking is I guess the implications could be different based on whether LSR is merge capable or non-merge capable. Thanks in advance, Pranjal _______________________________________________ mpls mailing list mpls@lists.ietf.org https://www1.ietf.org/mailman/listinfo/mpls
|
|