IPv6 folk,
I tried to summarize below the GUSL / GUPI deal.
We are pursuing two tracks here:
1. Uniqueness/limitation of site-locals: Two currents:
1.1 Limit site-locals to disconnected sites.
Margaret is preparing a draft.
1.2 make site-locals unique (remove ambiguity).
Also called GUSL. Two drafts to come:
Bob and Michel+Charlie.
Non-routability of site-locals could be guaranteed
By the combination of default discards and of course
by the local scope of such addresses.
In *both* cases, SLs or GUSLs can *not* be routed
outside the site, which means that for inter-site
communications we need GUPIs.
2. Not-making-a-mess of GUPIs.
GUPI = Globally unique, not globally routable.
The first sad truth is that there is no consensual
enforcement mechanism can guarantee that GUPIs will
not end up being a global PI mess, IPv4-style.
The second sad truth is that there is no consensus
for un-aggregated PI today.
The third sad truth is that there is no consensus
either in giving away un-aggregated PI now and try
to clean it later.
The combination of the second and third sad truths
is why we don't have IPv6 PI today.
Therefore, GUPI candidates are indeed scalable
(which very likely means aggregatable) global PI
solutions. As a side note, this topic has been the
quest for the Holy Grail of IPv6 multihomers since
1995 for what I have references of.
Michel.
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
