On Wed, 27 Nov 2002, Keith Moore wrote: > > > > Globally unique PI addresses are a step in entirely wrong direction. > > > > I should have said "globally unique PI site-local addresses" but that does > > not probably change your statement. > > > > > I entirely disagree. > > > > Care to elaborate a bit? > > we've been talking about this for weeks. the worst problems with > site-locals are their ambiguity; we need global uniqueness to get > rid of that ambiguity and also to allow them to be reliably routed > between sites. > > however we may just be using the words in slightly different ways - > globally unique vs. probably unique, > site-local (FEC0::/10) vs. some site-specific address space TBD, > etc.
I think your "may" may have hit the nail on the head. I'm _only_ _specifically_ thinking of "how do we fix site-locals to make them usable". I do _not_ want this w.g. to try to bang its head against the wood, trying to get workable globally scoped unique PI, possibly routable, addresses to work (in all layers from L1 to L10). Seems like a total waste of energy to me, as most efforts are _not_ in the layers we can really control. You seem to be thinking "how do we provide alternative globally unique, possibly routable, PI addresses so that site-locals would not be needed". Looking at this from your perspective: - I think GUPI, possibly routable, global-scope addresses might be ok, I have lots and lots of reservations about them, but they're not for this working group to try to cope with. - If we just stick to modifying site-locals somehow (and leave the PI global scopes out of the problem statement), I would definitely push for "nearly unique" site-locals Does this seem like a fair characterization? > I currently think > > - we need a new address space for globally-unique PI addresses (GUPIs) > > - some mechanism for ensuring global uniqueness is probably better > than having those addresses be probably unique > > - we need an understanding that GUPIs are not globally routable > unless/until we know how to make that routing scale - but NOT > an architectural limitation against it > > - use of site-locals (FEC0:://10) should be discouraged in general > with a possible exception for isolated/test networks > > -------------------------------------------------------------------- > 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] > -------------------------------------------------------------------- > -- 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] --------------------------------------------------------------------
