On Mon, 25 Nov 2002, Keith Moore wrote: > actually I'm fairly convinced that "nearly unique" would be good enough > as long as we could convince people to actually assign random site > prefixes instead of just choosing one (like all zeros or whatever).
Yes, this is the problem.. but the way it's written down or helper apps included in the implementations to do this for you would help, I suggest. > and they seem to have the "right" property regarding routability - > the chance of collisions is so low that you can hook them up with > a large number of other networks via private arrangement without > fear of collision, but high enough that there's a clear incentive > for ISPs to filter these things from routing advertisements. Exactly. I don't think we should be providing totally unique PI addresses. There's just no need for them. Or actually there is, _but that's not in the site-local context_. So I fear if we provide totally unique PI addresses, people would just abuse them, sooner or later. -- Pekka Savola "Tell me of difficulties surmounted, Netcore Oy not those you stumble over and fall" Systems. Networks. Security. -- Robert Jordan: A Crown of Swords -------------------------------------------------------------------- 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] --------------------------------------------------------------------
