Re: Issue on the SCTP draft

Jamal Hadi Salim <hadi@mojatatu.com> Mon, 24 November 2008 16:52 UTC

Message-Id: <MON.24.NOV.2008.115244.0500.>
Date: Mon, 24 Nov 2008 11:52:44 -0500
From: Jamal Hadi Salim <hadi@mojatatu.com>
Subject: Re: Issue on the SCTP draft
Comments: To: "Wang,Weiming" <wmwang@mail.zjgsu.edu.cn>
MIME-Version: 1.0
Content-Type: text/plain; charset="GB2312"
Content-Transfer-Encoding: base64
Content-Disposition: inline

2008/11/24 Wang,Weiming <wmwang@mail.zjgsu.edu.cn>:
> Patrick,
>
> ----- Original Message -----
> From: "Patrick Droz" <dro@zurich.ibm.com>
>> Weiming,
>>
>> it was a poll on the list without objections and therefore became a WG
>> document.
> Re:It was an incomplete process.
>
>> What kind of problems do you have in accepting that.
> Re: The problem is the process is unacceptable to me.

I will let Patrick respond - but i dont see why it is unacceptable.

> For more background, I think we just need more discussion on the TMLs. I do have
> had a discussion with Jamal during Czeck 68th IETF meeting Mar 2007 on TMLs,
> mentioning that my acceptace of things like SCTP as TML WG item is only the time
> when  there should be other TMLs that are easier to be deployed at the same time
> accepted, in order for more implementations able to go ahead easily.

I had the discussion with you because you proposed a TCP/UDP TML.
You remember i was the person who put that proposal/draft _ before_ you
did. I dropped it because the chairs told me that under no terms the IESG would
accept UDP.
I was hoping to convince you that we should pool our efforts together so we can
make progress in the WG. One of the most difficult challenges the WG has faced
is the slow progress in reaching milestones. The ADs had raised this
as an issue (and
upto recently we are still struggling to convince them otherwise).

> BTW, although not very strictly related to your this question, I have to mention that,
> since last year, my research team seemed become hard to contribute to the WG,
> even if we'd actually like to. Firstly, my team's implementation is greatly affected
> and blocked by the current TML decision.

I sympathize with you.

I can tell you though that this is a side-effect of standards
work when you implement something that is not totally ratified.
And if it makes you feel any better - you are not alone in this:
Znyx for example would have to redo many things as it stands now.
The TML needs redoing; the model used was very old; and the
protocol is out of sync with the many changes that went into the
document. If we had an interop today based on new documents, it will
fail miserably to inter-operate.

>  Secondly,the TML Service Primitive draft work was temporarily halted by
> Jamal's thought.

As a co-author of the doc, I explained to you why - many times, Weiming.
There have been a lot of complaints (whether right or wrong) that
we/ForCEs are trying
to do APIs. There have been claims that ForCES work should not be done at IETF
at all because of this.
Its a "branding" problem. The IETF is about protocols. It has been
known to stand
a few APIs - but in our case that stood out very visibly.
My suggestion to you was we drop this until we get our core work done. It could
then be pushed as an informational RFC. It was just too distracting.

> Re: Patrick, I just think the WG Charter is different from homeworks that should be strictly > scheduled.

The ADs would strongly disagree with you on this point.

In any case, please go ahead and publish a draft on TML -
<chair hat off>my earlier opinions on UDP stand.

cheers,
jamal