On Wednesday, August 6, 2003, at 01:30 AM, Aidan Williams wrote:
Alain Durand wrote:
IMHO, what need to happen is the following:
-1. Make an in-depth study of the consequences of introducing addresses with different ranges.
How would this different from the material that has been presented already in the SL debate? The whole anti-site-local argument is based around the consequences of a scoped addressing architecture that people don't like.
Many arguments in the SL discussion were about their ambiguity. This is now well understood, at least I dare to hope ;-)
Here the issue is range/scope. Some of the SL characteristics remain, not all, and this need to be understood better,
in particular, in the case where "local" and "global" addresses are intertwined:
- impact on node/application and address selection rules
- impact on routing protocols (because those addresses are not ambiguous, it might be simpler than SL,
but this remain to be proven)
- impact on DNS, where you now have a strong coupling between this "local" routing and your views of the DNS.
-2. Realize that if the issue at stake here has more to do with getting addresses
than with their actual scope/range, something probably can be
done working with the registries.
I don't think that is true. The local-ness of these addresses avoids the issue of having to scalably route the PI space. I can't see significant differences in process between globally unique local address allocation and a globally unique PI address allocation.
See Geoff's answer.
- Alain.
-------------------------------------------------------------------- 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] --------------------------------------------------------------------
