[Pana] RE: Sam's IESG comments

["Alper Yegin" <alper.yegin at yegin.org>]

> > -> Would you suggest we re-introduce what we had but only with a one-bit
> > info that says "PaC shall configure a new IP address" (without
> enumerating
> > any specific address config mechanisms)?
> >


> Sam, the notification of an IP address change in the PANA protocol
> itself was among the items that we removed in order to reduce PANA's
> complexity. If the IP address changes for some reason, whether it is
> configured via DHCP, SA, IKE, IPCP, or manually, PANA would need to
> react like any other protocol that would be affected (e.g., TCP) by
> this. I believe we discussed that ending the current session and
> restarting PANA would be a perfectly reasonable thing to do, but not
> certain that this made it into the text.

Somehow I must have missed this. I don't remember discussing termination of
current session and start of a new one when the IP address changes. Section
5.6 describes how the PaC's IP address (as one of the PANA session
attributes) change is handled.


> You suggest that PANA try and indicate when address configuration is
> required. Trying to enumerate this is due to the number of environments
> PANA wishes to applicable in and the number of ways an address may be
> configured sends the author and the reader out into the weeds rather
> quickly. Rather, I believe in the base framework and protocol spec it is
> sufficient to say that the IP address may change, and if it does PANA
> needs to know this and may need to restart. Specific "PANA with Foo"
> specifications may make this recommendation more precise if need be.

I agree enumeration of IP address configuration mechanisms is not a good
idea. But how about that one-bit indication of "PaC shall configure a new IP
address"? Sam agrees that his comment can be handled that way. If you are
fine with it, we can do that.

Alper



_______________________________________________
Pana mailing list
Pana at ietf.org
https://www1.ietf.org/mailman/listinfo/pana