Hi Monika, RFC 4447 says "Note that the PW label must always be at the bottom of the packet's label stack, and labels MUST be allocated from the per-platform label space." So I think we shouldn't use per-interface labels. The RFC says also " When PE2 receives a packet over a pseudowire, it must be able to determine that the packet was in fact received over a pseudowire, and it must be able to associate that packet with a particular pseudowire."
In my opinion it is better to use per-platform labels for example in networks implementing fast re-route. In that case the packets form PW can be received on different interfaces (depending on network state). Probably it is easier to build and maintain FEC. I think choosing per-interface or per-platform label space is not scaling problem too, because it has no influence on FEC size. As far as I'm concerned if multiple paths exists between PEs, the per-platform label space is closer to "per-peer". Regards Tomek Monika M pisze: > IETF-56 mpls minute's interop (ISOCORE) testing report claims following > point as one of the interop issue > > "Some vendors were advertising per-interface label space for targeted LDP > sessions when they should have been using per-platform labels due to an > ambiguity in the Martini signaling draft" > > Does that mean we can never use per-interface label space for PW label? > Can't we see "per-interface" as "per-peer" in L2VPN? (i.e) Use of same label > for two peers for scalability? May be we need extra intelligence in > hardware.!!! > > > Regards, > Monika. > _______________________________________________ > juniper-nsp mailing list juniper-nsp@puck.nether.net > https://puck.nether.net/mailman/listinfo/juniper-nsp > > -- Tomasz Szewczyk Poznan Supercomputing and Networking Center e-mail: [EMAIL PROTECTED] fax: +48 61 8525954 _______________________________________________ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
