On Monday, April 15, 2013 11:12:53 AM Nick Ryce wrote: > Hi Mark,
Hello Nick. > Re the control plane L2VPN interop issues. > > I believe this is meant to have been fixed in 15.3(2)S. > Currently about to start testing it in the lab and will > report back. > > http://www.cisco.com/en/US/docs/switches/metro/me3600x_38 > 00x/software/relea > se/15.3_2_S/configuration/guide/swmpls.html#wp128598 That's on the ME3600X/3800X. While IOS is not that different from IOS XE, I'm not sure whether they have feature parity (something Cisco have been working on, which is why the code is now numbered the same across many platforms now, i.e., 15.x). Would be good to find out whether they have this in IOS XE, as that is what the ASR1001 will run. I'm waiting for feedback from my SE and will report to the list if there is interest. Having said all that, the main issue wasn't that Cisco never supported VPLS signaling in a BGP control plane. The issue was that the last time I tested this, there was a difference in the BGP Update message size for this SAFI where Juniper are sending 21 bytes based on 'draft-kompella-ppvpn- l2vpn-03' while Cisco are expecting 17 bytes based on RFC 4761, causing session establishment for this SAFI to fail. I believe the motivation for Juniper here was that if you're running an all-Junos backbone, you can signal EoMPLS pw's as well as VPLS pw's using the same code. As Cisco do not support BGP-based signaling of EoMPLS pw's (they insisted on LDP, which is fine by me), one can see where the inter-op failed. Admittedly, I stopped being a VPLS fan many years ago, but I now see a use-case that might require inter-op for this to work, so trying to catch up on how far inter-op development has come re: this topic. Will be eager to hear how your testing goes, Nick. Thanks. Cheers, Mark.
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp