Re: [pim] PIM recharter

[Lucy yong <[e-mail]>]

Alia,

Please see inline below.

From: Alia Atlas [mailto:[e-mail]]
Sent: Thursday, February 26, 2015 8:18 PM
To: Lucy yong
Cc: [e-mail]; joel jaeggli; Brian Haberman; Deborah Brungard ([e-mail])
Subject: Re: [pim] PIM recharter

Lucy,

Just to clarify and finish up this conversation.  In the proposed charter, I have:

"The WG should discuss and develop new ideas related to multicast
protocols and distributing of multicast-related information.  The WG must
be explicitly rechartered to adopt any."

and the multicast work that I was specifically thinking of is the IGP-based multicast work that you've
been working on.  I can certainly see and understand its potential usefulness in data-center scenarios
as well as the trade-offs implied.

I am sorry if the previously floated charter raised expectations.  I was busy at NANOG and didn't
have time to rapidly revise the preliminary draft version of the charter.

This really isn't about trying to have you jump through hoops.  It's trying to make sure that
there's enough interest in the work that others will collaborate, that others will review, and that it will
become a useful standard.

I would appreciate hearing from others on the mailing list about interest in pursuing this work.

At this point, I am not expecting to have a final rechartering until after Dallas and there should be time
in the meeting to discuss thoughts on the charter too.
Lucy: OK, we will submit the arch draft, solicit comments/feedbacks, and present/discuss it in PIM in Dallas. Hope we reach a consistent consensus then.

Which WG should the draft be entitled for now?

Thanks,
Lucy

Regards,
Alia

On Wed, Feb 25, 2015 at 10:45 AM, Lucy yong <[e-mail]<mailto:[e-mail]>> wrote:
[resend without too BIG message body]

Alia,
[snip]
I am disappointed this charter text. It is very different from the proposed and discussed charter text on the mailing list. The following charter text only allows working on PIM/IGMP/MLD protocol. Any other item is in “may” category and requires WG rechartering to work on it.

 Yes, I do not think that having a broadly open charter to come work on anything is healthy for the WG.  We are working on what should be clearly in charter and what areas the WG may explore but cannot work on unless the recharter occurs.

You know that IGP multicast proposal and homenet were discussed in PIM in Honolulu.  There were interests to work on these technologies and you suggested that PIM WG is the right place to work on IGP multicast (was discussed in IS-IS WG several times).  But the new charter text does not accept it as a work item and make it as “may or may not”.  This does not reflect the outcome from Honolulu meeting. Why do you want make this gate for IGP multicast and homenet?

Because there has been NO DISCUSSION or WORK done on the IGP multicast proposal since Honolulu.  The WG said that it was interesting - but has not yet demonstrated active consensus and willingness to work on it.  I see a message from Nov 10 on the mailing list ( http://www.ietf.org/mail-archive/web/pim/current/msg03055.html ) that was never responded to.  This does not indicate to me that there is currently active interest in developing the work in the WG.
Lucy:  I did not response that email because that was in IETF meeting week and before the PIM meeting. Some questions and comments are addressed in our presentation and Ice were there. I did not know that you are waiting on some discussion on mailing list while We are waiting on the next charter text. In fact, we are working it offline.

[Alia] You know that conversations and decisions happen on the list.  One of the critical aspects of IETF is that we have an open consensus process.  That means open discussion on the mailing list.
[Lucy] This is my thought too. The charter text was openly discussed on the mailing list based on the WG consensus in Honolulu. ADs do not express any concern but change it now.
 There is no reason to "wait on the next charter text".  If you do the work and have the active consensus and enthusiasm, that is key to the WG getting the work done.  Of course, off-line and individual conversations happen and are also useful - but doing them solely doesn't forward the standardization process.  This particular email raised a concern about looping that I haven't seen addressed.
[Lucy] Yes, we need do our job and also need AD help. If you could ping us early on this concern, we would do that to avoid it.  I have replied Ice’s comments about looping concern.
 Similarly, the homenet work has not been discussed on the mailing list.  It has been raised by one other in this thread as something of potential interest.  Sure - but I need more information on what that work should entail and whom would actually do it.

That's why both of these items are not explicitly in the charter.   There is the possibility to add them later - but I must see active consensus to work on them and some actual work.  The IGP multicast proposal is interesting - but it is also clearly another way of doing multicast.
Lucy: The proposal is supported by many vendors as of draft co-authors. Huawei has some implementation. I am not sure what kind of active consensus you look for at this point. We already clearly pointed out the use case for this in DC and why PIM can’t fit for the use case.

[Alia] I'm sure that you don't need me to explain that draft authors should be writing the draft.  Support can be expressed in many ways - including discussions.  I am happy to hear that Huawei has an implementation; that's a good sign of serious interest and that the basic ideas can work.  I just read through the draft again; it could be improved significantly to be clearer.

[Alia] As I've said, the reception at the last IETF in the room was good but nothing has confirmed that on the list.  I look forward to hearing more discussion and opinions on the list and in Dallas.
[Lucy] This is my misunderstand. The charter text discussed on the mailing list was the confirmation in my mind.  Hope you /chairs can give some warning ahead. We will submit the architecture draft and present it in Dallas meeting.

Regards,
Lucy