At 06:37 PM 8/6/2003 +0930, Mark Smith wrote:
I'd think it is pretty unlikely that two people, at separate sites, with the same birthday, were executing this algorithm at the same millisecond.
And even if they were, what do you think the chance is that their companies will ever acquire each other, requiring them to merge their private address space? :-)
Actually, this risk is exactly why I would like to make the selection algorithm for bits in the FD/8 space as bulletproof as possible. In a previous life, I have spent too much time in the house of pain merging and renumbering networks that were behind IPv4 NATs when companies acquired each other. Little companies do that quite a lot. I also have little faith in user supplied input, and not a lot of faith in the implementation quality of an arbitrary random number generator. Hence I would like to see local address disambiguation make use of any "good quality unique bits" we already have lying around, like EUI-48s. I don't think that is a hard thing to do, so why don't we?
- aidan
-------------------------------------------------------------------- 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] --------------------------------------------------------------------
