Keith, I haven't done an analysis, but it seems to me that deprecating
site local is a significant simplification of the scoping problem,
and thereby greatly increases the number of cases where the default
algorithm will work.

   Brian

Keith Moore wrote:
> 
> > Multiple addresses per host are a real-world feature of IPv4 and have
> > always been a design goal of IPv6.
> >
> > We've already defined the default address selection algorithm (RFC
> > 3484).
> > Applications are welcome to use a different algorithm, but they
> > don't need to, which is why the default was defined.
> 
> it's one thing to allow multiple addresses per host in the
> architecture, quite another to expect hosts and apps to be able to
> tolerate different addresses with different scopes and force them to
> choose between these addresses in order to interoperate.
> 
> the default address selection algorithm will only work in a fairly
> narrow set of cases.
> it's simply not the case that applications "don't need to" use a
> different algorithm.
--------------------------------------------------------------------
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]
--------------------------------------------------------------------

Reply via email to