Re: multi6-functional-dec and re-homing

[Jari Arkko <jari.arkko@piuha.net>]

john.loughney@nokia.com wrote:

> > So, the question we have now is whether the SHIM6 WG's RFCs
> > should specifically deal with the source address selection
> > problem when the other end is plain old IPv6 node. Or not.
>
> I might be confused here, but I think Brian was suggesting that
> to use shim6, both ends need to have shim6, otherwise both ends
> use 'normal' mechansims.  Do you see a case where one end would use
> shim6 and the other would not?

We need to separate the technical and the administrative
parts of the question. I _think_ we agree on what can be
done technically; the SHIM6 WG solution consists of a number
of 'pieces', and we could certainly improve the source
address selection rules without requiring the shim6 protocol
from either end. The question here is whether to bother
about it or not.

The administrative parts are (a) whether such rules
are within the shim6 protocol document or in a separate
RFC, and (b) whether such work would be within SHIM6 WG's
scope or a part of "normal" IPv6. I think Marcelo argued
well that the technical solution as such is necessary for
the SHIM6 solution to work, so unless we are going to rely
on some other WG to provide us some basic building blocks,
we're going to have to do it ourselves. So I think its
in scope. But maybe the remaining question is whether
we bother to document the case where the stuff is
used when shim6 protocol is not present.

--Jari