>
> What exactly will be reasonable assurance? If your "site ID registry"
> allocates you a unique site ID, why would the site local address
> created with that site ID not be globally routable?
>
It would not be globally routable "by definition"... That is, the prefix
would be recognized by routers as non-globally routable and they would
refuse to route them. But as you are suggesting since they are in fact
globally unique, ISPs could decide some day to route on them. But they'd
pretty much all have to agree to do it. By the same token they could agree
to route on larger globally aggregatable prefixes if they so desired, so the
potentiallity of huge routing tables is there whether or not a site-id
registry is created.
>
> If a "site ID registry" is giving out the ID's, I would think it would
> be in their best interest to allocate unique numbers. If two site
> local addresses have the same ID in them, why would they not be in
> the same site, given the above assumption?
Because they might be from a previous single site that split, for instance.
Conversely, if two different sites merged you would have a situation where
two machines with different "site ids" were in fact in the same site. So
you'd need other mechanisms to determine when in fact two nodes were in the
same site.
In this regard, "site id" is the wrong term (and not the term I originally
started with). A better term is "globally unique locally routable" address,
or GULR address.
PF
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
