Margaret Wasserman wrote:
>
> Hi Alain,
>
> >There has been a very long discussion on the fate of Site Local addresses
> >in the wg. There are still two opposite views of what to do about them:
>
> Most of that discussion focused on whether or not to remove site-local
> addresses from the architecture, not on this draft. And, we have
> already determined that there was no consensus to remove them from
> the architecture. So, I don't know why it would make sense to remove
> them from this draft.
>
> >a) Do not require apps to support multi-sited nodes now, but
> > keep the address selection rules in place.
> > This means that in the future, if SL are available,
> > they will be used by ALL applications by default.
>
> An application will only use a site-local destination addresses if it
> is returned in the DNS and/or typed locally. So, it can't happen
> without the administrator or user doing something to make it happen.
>
> >b) Do not require apps to support multi-sited nodes now, and
> > remove the address selection rule that deal with scoped addresses.
> > This means that in the future, if SL are available,
> > they will not be used by default, but only by applications that request them.
>
> I don't understand how this is achieved by removing the default
> selection rules. If an application receives a site-local address
> from the DNS, how will removing scoped addressing rules from the
> default address selection draft prevent the app from using the site-local
> address? The assumption is that we are dealing with old apps that don't
> know anything about IPv6 scoped addressing, so we can't expect the app
> to recognize a site-local address and refuse to use it...
>
> > Keith has promised a draft to explain the trade-off between those two approaches,
> >I think we should not advance draft-ietf-ipv6-default-addr-select-08.txt
> >until there is a clear resolution on this question.
>
> Unfortunately, Keith did not submit a draft by the deadline, so there will
> not be a draft before Yokohama.
>
> The Default Address Selection document has been out for months, has
> been reviewed several times by the WG, has undergone a WG last call and
> has been reviewed by the IESG. It doesn't seem right to delay this
> draft indefinitely waiting for another draft, especially when we haven't
> been told about any specific problems that this draft will cause.
>
> If there are specific, technical objections to this draft, please state
> them clearly on the mailing list. It would also be helpful to suggest
> specific text changes to the draft. This will give others and opportunity
> to agree or disagree about whether the issues should block the publication
> of this draft.
I think that at this point we should advance the document as it is. While
I'm sympathetic to Keith's concerns, I think we have needed address selection
*defaults* for a long time, and the default behaviour recommended for SLs
is consistent and logical.
Defaults are made to be overridden, so the behaviour can always be adjusted.
Brian
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
