Re: [DNA] [Issue 15] [Issue 16] DAD and MLD Interaction

[James Kempf <kempf@docomolabs-usa.com>]

> OptiDAD got a "discuss" from IESG, so there's a last round
> of revision happening some time Real Soon Now.  If there's something
> you reckon needs clarification, now's the time to discuss it!
>

I'll take another look at the draft and see if I can come up with some
suggested changes. Where should I send them?

> Yeah, I think the safest thing is to Deprecate any addresses at Link
> Down, and Tentative/Optimistic them at Link Up.  That should allow
> source address selection to behave correctly in the case of multiple
> interface devices as well.
>

The problem with Deprecated is that any apps with such addresses active
still continue to send on the address, though new socket connections will
not be opened, so new connections will favor interfaces with Confirmed
addresses. I suppose one could take the position that this is OK, and it is
up to some additional mechanism (seamless handover support?, buffering in
the IP stack while the link transition is occuring?) to handle this problem;
that is, it is outside the scope of DNA to handle it.

            jak