> Just run MPLS across the ring. That's what these boxes are > built for :-).
The primary reason it's L2 was because we had a slow migration from ME3400s/MRV OS912s to ME3600s. There were connection points in the ring that required L2 because of intra-ring L2 services we were providing to customers. Secondly, the 50ms-failover times of G.8032 ERPS was a big decision point. That said, I've recently become aware of LFA and plan to test that in the lab. Finally, the L2 topology *seemed* to work fine until just yesterday when discovering it's really the SVI/EVC combination that seems to be causing the problem. I am amazed that no one at TAC, not a single one of the half dozen people I've been dealing with, even mentioned that the SVI might be the issue because its only a fairly "recent" feature addition. I do read documentation - half my day is often spent reading documentation - but even the most studious readers can miss a single line or feature addition once in a while. > Also, since you own the ring (I presume), just run a Jumbo > frames. Could be a selling point for your customers (and > gives you better goodput). The other thing that I just remembered is that we still have a lot of ME3400s in the network that don't support >1998B IP MTUs. While I can certainly boost the Junipers and ME3600s to 1998, I don't see how it really gains much. Thoughts? -evt _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
