Re: [netlmm] Last Call: draft-ietf-netlmm-mip-interactions (Interactions betweenPMIPv6 and MIPv6: scenarios and related issues) to Informational RFC

Charlie,
Thank you for your review and comments. Please note that the WG has spent a lot of time on this topic of same vs. separate BCEs. We have had two consensus calls on it after discussion at a meeting.  As you have seen from the thread, the chairs did see rough consensus to move on (specifically, please see http://www.ietf.org/mail-archive/web/netlmm/current/msg05551.html).  Explicit text on this topic was also provided to our AD along with the shepherding write-up: 

"There have been some disagreements on the approach used to solve the collocated LMA and HA case in the document.  The present approach uses logically separate binding cache entries for the LMA and HA, which is based on consensus.  Some members of the WG strongly preferred the approach of having the same binding cache entry for the LMA and HA - however, the approach needed additional solutions to address race conditions that arose from it and it did not gain consensus in the WG. It should be noted that this was not raised by anyone in the WG during WGLC of the document and the document as a whole had strong consensus to be published."

As you see, the disagreement has been noted and we've moved on. We don't plan on re-opening this discussion at this time. Let's resolve any remaining issues with the document. 

Thanks,
Vidya

-----Original Message-----
From: netlmm-bounces@ietf.org [mailto:netlmm-bounces@ietf.org] On Behalf Of Charles E. Perkins
Sent: Monday, May 17, 2010 9:49 AM
To: Giaretta, Gerardo
Cc: netlmm@ietf.org; ietf@ietf.org
Subject: Re: [netlmm] Last Call: draft-ietf-netlmm-mip-interactions (Interactions betweenPMIPv6 and MIPv6: scenarios and related issues) to Informational RFC

Hello Gerardo,

Comments below...

On 5/17/2010 8:17 AM, Giaretta, Gerardo wrote:

> You have one comment on the recommendation in the draft to have
> separate binding cache entries. This was extensively discussed
> in the NETLMM WG and also at the IETF Dublin meeting. There was
> a mailing list discussion after that in September/October 2008
> which led to the conclusion in the current version of the draft.
>
> You can find more information in the archives at:
> http://www.ietf.org/mail-archive/web/netlmm/current/msg05533.html

Thanks for that link.  It was most enlightening,
especially in the context of the ensuing discussion.

Having reviewed the latter, it seems to me quite likely
that the consensus call was (at least) premature.

For instance:

> I object to this. There was absolutely no consensus on this for
> you guys to decide. There were clarifying questions that people
> had on what exactly you meant by multi-homing. You didn't respond
> to any of those emails.

and

> I am sorry, but I thought the discussion was either incomplete or did
> not steer towards one particular way or the other. For instance, I
> didn't get a clear answer for my question on why there would be a single
> BCE when two different interfaces are in use. Could you please clarify?

I could go on.  And, without naming names, I want to emphasize
that the abovementioned objections were made by some real experts.

Do you have any links to discussion that _supports_
the consensus call?

Furthermore, I still suggest (constructively) that
_at the minimum_ a system architect ought to be allowed
to have the design freedom to identify the two mobile
node identities (and thus the relevant BCEs).
What is the downside of enabling new systems to
offer such obvious improvements?

Or, would it be better to start writing the ...bis
document already (just kidding...)?

Regards,
Charlie P.

_______________________________________________
netlmm mailing list
netlmm@ietf.org
https://www.ietf.org/mailman/listinfo/netlmm