Zefram wrote:
> I oppose your idea of having allocated prefixes be numerically related, > because it defeats the aim of having no discernable structure in the > prefix IDs. Whoever generates the prefixes must do so randomly and > independently.
I don't see why this is a problem. The prefix generation could be done sequentially, it seems to me. Randomness makes it a _lot_ harder to guarantee uniqueness, for no good reason that I can see. Who cares if there is structure or not, when the prefixes aren't going to be routed?
The observation is that you cannot guarantee "aren't going to be routed",
If you want to make it "very hard to route globally" then the Hinden draft
makes the assumption that if you assure that there is no guessable
structure of addresses this space. No guessable structure (i.e.
random selection) is intended to ensure that
a) its impossible to undertake any form of aggregation of
adjacent prefixes and
b) its impossible to create a reverse DNS structure that exploits
structureOf course you may not care that these prefixes leak into the global routing table, in which case it becomes a different scenario altogether.
Geoff
-------------------------------------------------------------------- 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] --------------------------------------------------------------------
