Hi Erik, I'll pick the gauntlet, with great care :-) Erik Nordmark <[EMAIL PROTECTED]> writes: > > When other than Ethernet link layers are involved, probably the > > functionality of the u/g bits can be derived from the 8bit prefix? > > Maybe it's the 8bit prefix that should be tweaked to obtain the > > reservations proposed? > > Do you have a concrete proposal?
You seem to imply that the u/g bits are there not because of Ethernet, then I can observe that: universal bit =1 has similar meaning to Prefix =001 group bit =1 has similar meaning to Prefix =1111 1111 Many other prefixes are unassigned and as such can serve the purpose of about anything. For example, Prefix 010 could mean randomly generated Interface ID's. When doing this, I think it would be good to extract the commonalities of the usage scenarios of the current proposals, see how they fit the current Prefixes and what are the additional criteria needed to classify them. For example, how can one classify a randomly generated Interface ID and a Cryptographically Generated Address? Random IID serves privacy. CGA serves authentication of something. Both are security, BUT random IID also serves certain types of L2's, nothing to do with security. Both are global unicast and are not group. Since I find the classification difficult I think I'll refrain from proposing anything serious. What do you think? Alex -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------