Re: [rtcweb] Protesting the QoS document decision

Harald Alvestrand <> Thu, 14 November 2013 02:16 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 50BF421E80D8 for <>; Wed, 13 Nov 2013 18:16:57 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -110.556
X-Spam-Status: No, score=-110.556 tagged_above=-999 required=5 tests=[AWL=0.043, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id kUgyZLrqQclG for <>; Wed, 13 Nov 2013 18:16:45 -0800 (PST)
Received: from ( []) by (Postfix) with ESMTP id DEF7D21E80D9 for <>; Wed, 13 Nov 2013 18:16:43 -0800 (PST)
Received: from localhost (localhost []) by (Postfix) with ESMTP id F011339E216; Thu, 14 Nov 2013 03:16:42 +0100 (CET)
X-Virus-Scanned: Debian amavisd-new at
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id cX3kJfiwiphq; Thu, 14 Nov 2013 03:16:41 +0100 (CET)
Received: from [] (unknown []) by (Postfix) with ESMTPSA id 81CCB39E209; Thu, 14 Nov 2013 03:16:41 +0100 (CET)
Message-ID: <>
Date: Thu, 14 Nov 2013 03:16:40 +0100
From: Harald Alvestrand <>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.1.0
MIME-Version: 1.0
To: James Polk <>
References: <> <>
In-Reply-To: <>
X-Enigmail-Version: 1.6
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Cc:, "" <>,
Subject: Re: [rtcweb] Protesting the QoS document decision
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: Real-Time Communication in WEB-browsers working group list <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 14 Nov 2013 02:16:58 -0000

On 11/13/2013 11:11 PM, James Polk wrote:
> I, speaking at the mic, was merely pointing out that this draft had
> moved to TSVWG back in the Atlanta timeframe (though, thinking back, I
> think I left the timing piece out of my comment), so it wasn't a
> recent decision. I believe this came directly from talks between the
> above mentioned ADs.

This only reinforces my procedural comment. The fact that the decision,
and the reasons for it, has been hidden from the group for 8 (?) months
only makes it worse, not better.

When ADs make decisions by talking among themselves, they have a special
duty to make sure those decisions are public and open to the community.
That duty has not been executed in this case.
I don't know who dropped the ball - the ADs or the chairs. But the ball
was definitely dropped.

> As for RTCweb work being more appropriately inside the RTBweb WG
> instead of another group, I'd like to point out that much of the SDP
> work related to this WG is in fact being done in MMUSIC. Much of the
> RTP work being done by this WG is being done in AVTCORE (or AVTEXT?).
> All of the base SCTP work being done by this WG is being done in TSVWG.

I have issues with those too. Especially the speed at which they're
progressing. But those decisions were made and announced in a timely

The argument that has been used is that the pieces we need defined
(BUNDLE, MSID, circuit-breakers, RMCAT) have general applicability, and
should therefore be processed by groups that are chartered to define
extensions with general applicability.

> (this is not an attack, but...)
> ...why are you arguing for something as far away down the stack as
> DSCP assignments to be done in RTCweb, and not where all other
> DiffServ work is being done in the IETF (in TSVWG)? Do you not trust
> that TSVWG will do an appropriate just with a DiffServ ID but you will
> trust TSVWG with SCTP IDs?

My understanding of the QoS draft has always been that it should define
no new codepoints. It should point out the language we need to translate
RTCWEB requirements ("audio should go faster than video, except when the
app says that it should be the other way around") into already-defined
DSCP code points.

I have not seen anyone arguing for new DSCP codepoints, so I really
wonder what work there is for TSVWG to do.

Again, the arguments may be there, but since they were never exposed to
the mailing list, I cannot evaluate them.

> Color me confused...
> James
> BTW - Full disclosure, I'm an listed author of this draft, but had no
> voice in it getting moved between the WGs.

My biggest practical issue with this draft is that it's functionally
dead. I'm getting put in a place where I have to either delay shipping
this feature in product or ship without even the guidance of a live draft.

I want this group to be done. As long as I can't even point to the
document that describes how we do QoS, I have no chance of getting it to
that stage.

> At 02:21 PM 11/13/2013, Harald Alvestrand wrote:
>> This mail concerns both administrative and technical issues, which is
>> why it is explicitly copied to the ADs of RAI and TSV. I hope I have
>> managed to keep them separate in the message.
>> Magnus said in an email yesterday, concerning draft-ietf-rtcweb-qos:
>> > Okay, we might not have been public enough on this. It was
>> requested by
>> > the Transport ADs quite some time ago that doing the QoS document
>> in our
>> > WG was not appropriate and requested us to direct the document to
>> > Which was done, and where it hasn't made progress.
>> >
>> > You might have noted that James Polk did comment in the milestone part
>> > in the monday session of IETF88 about our QoS milestone should be
>> killed.
>> I want to protest this - both practically and formally.
>> To get the formal stuff out of the way first:
>> Changing the deliverables of the working group *without telling the
>> working group* is totally inappropriate in an open, consensus-driven
>> process.
>> Changing the deliverables of the working group *without telling the
>> working group why* and *without allowing those reasons to be debated* is
>> totally inappropriate in an open, consensus-driven process.
>> I protest against this action.
>> ACTION REQUEST 1: I request that this decision be declared null and
>> void, and that the relevant ADs send out a message to RTCWEB (and TSVWG
>> if appropriate) *PROPOSING* such an action, and giving a reasonable
>> timeline for when they will make a decision based on mailing list
>> feedback.
>> In practice:
>> The draft as it existed before its untimely demise consisted of two
>> things:
>> - A description of how QoS mechanisms could be useful in the RTCWEB
>> use case
>> - A description of existing mechanisms that could be appropriate for the
>> RTCWEB use case
>> The first one is clearly something that needs RTCWEB consensus. It seems
>> to have no need for, nor likelyhood of gathering interest enough for, a
>> TSVWG consensus.
>> There could be some argument that the second part needs TSVWG consensus,
>> especially if it was redefining any mechanisms, or it was making choices
>> between mechanisms where TSVWG had strong opinions about which
>> mechanisms should be chosen, but had not chosen to document that in any
>> document on which IETF consensus had been declared (that is to say,
>> existing RFCs).
>> My archive shows 36 messages where the title refers to this draft. It
>> shows no messages declaring that feedback has been incorporated and
>> calling for new review - something that is usually necessary to get a WG
>> consensus on any document. The debate hasn't been conclusive, but then,
>> it hasn't been pushed hard either.
>> The people who know how RTCWEB works are in this working group. Some of
>> them may be in TSV, but I think the locus of knowledge for saying what
>> QoS mechanisms are appropriate for RTCWEB are here, not in TSV.
>> This results in my second request.
>> ACTION REQUEST 2: I request that the chairs declare that based on the
>> debate about the QoS functionality so far, the document should be
>> resurrected, and will continue to be  worked on in the RTCWEB working
>> group, bringing in advice from TSVWG as required to make sure the
>> descriptions of underlying mechanisms, and the choice of such
>> mechanisms, is correct and appropriate.
>> -- 
>> Surveillance is pervasive. Go Dark.
>> _______________________________________________
>> rtcweb mailing list

Surveillance is pervasive. Go Dark.