Re: [clouds] CloudApps BoF (IETF-80) proposal for your review and comments

[Peter Saint-Andre <[e-mail]>]

On 1/29/11 2:43 PM, Bhumip Khasnabish wrote:
> I plan to submit it by the deadline (Monday-31-January-2011).
>  
> http://trac.tools.ietf.org/area/app/trac/attachment/wiki/Clouds/Khasnabish-et-al-CloudApps-BoF-Proposal-v7-30Jan2011.doc 

It would have been good to post actual text to the list. Attachments
that require support for proprietary formats are not welcome.

> So, pls review and send me your comments and suggestions ASAP. Many thanks.

Let me say that you are making some progress. At least you now have
separate proposals that are focused on more particular problem domains
(e.g., applications vs. operations).

However, when I read this proposal I see several big warning flags.

First, as various folks explained in Beijing, IETF participants prefer
to develop protocols, because that's what we do best. Your proposal is
to work on:

Phase 1 - Use cases, reference framework, survey

Phase 2 - Requirements

Phase 3 - New protocol, profile of existing protocols, and/or APIs

That approach might work well in other SDOs, but I don't think it is a
recipe for success at the IETF.

Second, everything in this proposal still reeks of the old "clouds are
special" attitude. As folks explained in Beijing, we care about solving
specific engineering problems, not building any kind of cloud-related
application that some people might find interesting.

Several of the referenced drafts talk about the problem of building
virtual desktop infrastructure (VDI):

http://tools.ietf.org/id/draft-wang-clouds-vdi-problem-statement-00.txt

http://www.ietf.org/id/draft-ma-clouds-vdi-survey-00.txt

http://www.ietf.org/id/draft-yokota-cloud-service-mobility-00.txt

http://tools.ietf.org/id/draft-cloud-log-00.txt

Those documents seem to provide a good start toward defining a more
focused engineering problem that IETF participants could work on (my
apologies, but I have yet to read them in depth). I would suggest
throwing out Phases 1 and 2 to some extent (using them only as
background to the engineering problem) and trying to get to Phase 3 as
quickly as possible.

Holding a Bof about forming a VDI WG might be appropriate at IETF 80 in
Prague. At the least it would help IETF folks understand whether there
really is a problem here that the IETF needs to solve, how many people
are willing to work on building solutions, etc. See RFC 5434 for some
*very* helpful information about the thinking (and work!) necessary to
hold a successful BoF.

Next steps:

a. Read RFC 5434 and restructure your proposal a bit to answer the
questions posed in that document.

b. Complete the BoF request template here:

http://trac.tools.ietf.org/bof/trac/

Because I will be the continuing area director at IETF 80, I will
volunteer to be your responsible AD.

Also, do you plan to request additional BoF slots at IETF 80 (e.g., for
CloudCenter/DataCenter Operations)? It takes a lot of work to hold a
successful BoF. It even takes a lot of work to hold an unsuccessful BoF!
Therefore I counsel you to plan accordingly, and to hold only one BoF
per IETF meeting if you want to form multiple working groups in
different IETF areas.

Peter

-- 
Peter Saint-Andre
https://stpeter.im/