[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Draft IPv6 Minutes from Atlanta IETF

>Of course, a "Site Locals Considered Harmful" would also be useful (Keith, 
>know it makes sense, and less effort than 100 emails to the list :)

I agree that such a document would be useful.  I doubt, though, that
it will prevent 100's of e-mails to the list -- I've just become
resigned to spending ~45 minutes a day reading e-mail on the IPv6

As part of deprecating site-local addressing, we agreed in the
meeting that, in addition to deprecating site-local addressing in
the addressing architecture and removing it from other places (scoped
addressing architecture, address selection rules, etc.), a document
would be written that would do two things:

         - Explain why site-local addressing was deprecated
         - Outline alternative means to address some of the
                 problems that could have been solved by
                 site-local addressing.

A "site-locals considered harmful" document could be a part of
this document or be referenced from it.

One of the problems of writing such a document is that "site-locals"
are a moving target.  We have (at least) five different proposals on
the table for site-local address "usage" (only four of which are
documented), and most of those proposals can be modified by one or
more of the proposals to create "unique" site-local addresses.

I attempted to capture the issues associated with site-local
addressing in my site-local impact draft:


This document specifically addresses the benefits and issues
associated with the "full" site-local model currently documented
in the scoped addressing architecture WG I-D.  I did not consider
other models, because at the time this document was written, the
"full" model was the only documented model...  A subset of these
benefits and issues exists for each of the less complete site-local
usage models.

Bob and I attempted to capture, in our IPv6 WG presentation, the
benefits and issues associated with each of the three usage models
we presented -- "limited", "exclusive" and "moderate".  We focused
on the differences between those three approaches -- mainly
whether or not they require split DNS and whether or not they
require address selection logic in applications that pass addresses
to other nodes.

I don't believe that our presentation is on the minutes site yet,
but it can be found at:


It might serve as useful input for a "considered harmful" document,
or for an attempt to explain the decision to deprecate site locals.


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 majordomo@sunroof.eng.sun.com