Re: [rtcweb] -transports- discussion of DSCP

Harald Alvestrand <harald@alvestrand.no> Thu, 14 August 2014 08:19 UTC

Return-Path: <harald@alvestrand.no>
X-Original-To: rtcweb@ietfa.amsl.com
Delivered-To: rtcweb@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 0FD1D1A0921 for <rtcweb@ietfa.amsl.com>; Thu, 14 Aug 2014 01:19:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.568
X-Spam-Level:
X-Spam-Status: No, score=-2.568 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RP_MATCHES_RCVD=-0.668] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id dfhOoWXxRgtG for <rtcweb@ietfa.amsl.com>; Thu, 14 Aug 2014 01:18:53 -0700 (PDT)
Received: from mork.alvestrand.no (mork.alvestrand.no [IPv6:2001:700:1:2::117]) by ietfa.amsl.com (Postfix) with ESMTP id B5EA31A095D for <rtcweb@ietf.org>; Thu, 14 Aug 2014 01:18:47 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by mork.alvestrand.no (Postfix) with ESMTP id 095747C3D6F for <rtcweb@ietf.org>; Thu, 14 Aug 2014 10:18:47 +0200 (CEST)
X-Virus-Scanned: Debian amavisd-new at alvestrand.no
Received: from mork.alvestrand.no ([127.0.0.1]) by localhost (mork.alvestrand.no [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yYWT1PICw5kc for <rtcweb@ietf.org>; Thu, 14 Aug 2014 10:18:46 +0200 (CEST)
Received: from hta-hippo.lul.corp.google.com (unknown [IPv6:2620:0:1043:1:480e:b098:4917:7a16]) by mork.alvestrand.no (Postfix) with ESMTPSA id 31E847C3C47 for <rtcweb@ietf.org>; Thu, 14 Aug 2014 10:18:46 +0200 (CEST)
Message-ID: <53EC70E5.7070105@alvestrand.no>
Date: Thu, 14 Aug 2014 10:18:45 +0200
From: Harald Alvestrand <harald@alvestrand.no>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.0
MIME-Version: 1.0
To: rtcweb@ietf.org
References: <CA+9kkMBBfPyZdq0uzo96sxYaWK75bAHu8madAd5P4OTXLtb=Wg@mail.gmail.com> <CABkgnnVBtyb2wNZfo7Fto_X0Rhr_e2wAHWUxqcgBksj6YMLgFA@mail.gmail.com>
In-Reply-To: <CABkgnnVBtyb2wNZfo7Fto_X0Rhr_e2wAHWUxqcgBksj6YMLgFA@mail.gmail.com>
Content-Type: text/plain; charset="windows-1252"; format="flowed"
Content-Transfer-Encoding: 7bit
Archived-At: http://mailarchive.ietf.org/arch/msg/rtcweb/JBoeIfeMyilJT6RBRZHfmf8_7fc
Subject: Re: [rtcweb] -transports- discussion of DSCP
X-BeenThere: rtcweb@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Real-Time Communication in WEB-browsers working group list <rtcweb.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/rtcweb>, <mailto:rtcweb-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/rtcweb/>
List-Post: <mailto:rtcweb@ietf.org>
List-Help: <mailto:rtcweb-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/rtcweb>, <mailto:rtcweb-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 14 Aug 2014 08:19:01 -0000

On 08/13/2014 09:07 PM, Martin Thomson wrote:
> On 13 August 2014 09:10, Ted Hardie <ted.ietf@gmail.com> wrote:
>> My concern here is that access to the ability to set different DSCP markings
>> on a per packet basis for a flow may not be available generally on all
>> platforms.  Even if we want to make a forward looking push toward that
>> ability, naming what we think is the appropriate alternative may make sense.
> Yeah, requiring per-packet DSCP seems overly enthusiastic.

I know it's possible to do this on Linux (for UDP, of course). I think 
it still requires an extra sysctl before sending each packet, which is 
reasonably high overhead if your app is sysctl-limited. But I don't know 
of any case where you would be allowed to set per-flow DSCP markings and 
would not be allowed to set per-packet DSCP markings; all the reasons to 
not do marking seem to be reasons to not do marking.

I'd like to know if the APIs exist on other platforms (whether they're 
overridden by default policy or not).

The two alternatives if we don't get per-packet setting are:

- Deaggregate, so that per-port-flow classifiers can do their thing
- Don't ask the network to differentiate, keep precedence of flow a 
sender-stack thing only

I don't know which is the more (or less) distasteful / useful of those 
two paths.


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