On 2009-11-20 21:06, Fred Baker wrote: > > On Nov 20, 2009, at 11:16 AM, Brian E Carpenter wrote: > >> If you are polite, and use some sort of exponential backoff like >> REAP does, the time to discover a working pair can be quite long, >> unless your first or second guess is correct. > > That's kind of the point of the caching angle...
Of course. (REAP is used as part of SHIM6, which continues using the successful pair as long as possible.) But when a major failure occurs, you have to assume that thousands of hosts may all start reachability probes simultaneously, whatever they had previously cached. Brian -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------