On 2010-10-24 12:01, Fred Baker wrote: > On Oct 23, 2010, at 3:23 PM, Brian E Carpenter wrote: > >> Lower, at the moment, since we have people very interested in the flow label >> for ECMP/LAG. > > well, yes, but 6man does everything in its power to prevent it from being > useful for that. You said that after 15-20 years of history with IPv6 things > were pretty well set in stone. Well, I can show you 15 years use of the > source and destination addresses, the next header field, the traffic class > aka DSCP/ECN, hop count, and so on. Can you show me 15 years actual use of > the flow label? > > If not, the rules for taking a protocol to draft standard would suggest the > use be deprecated, no?
Well, if you're talking about 2460, it basically says: set the flow label to zero, anything else is experimental. That's fine for DS, imho. Indeed, the Proposed Standard 3697 is on the table. My opinion is that it needs to be fixed, but deprecation would be an alternative. Brian -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
