From tsv-area-bounces@ietf.org Fri Apr 06 04:02:42 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HZjOd-0004H6-Sd; Fri, 06 Apr 2007 04:02:11 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HZjOc-0004Gt-CT; Fri, 06 Apr 2007 04:02:10 -0400
Received: from smtp.nokia.com ([131.228.20.170] helo=mgw-ext11.nokia.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1HZjOZ-0000SL-Pm; Fri, 06 Apr 2007 04:02:10 -0400
Received: from esebh105.NOE.Nokia.com (esebh105.ntc.nokia.com [172.21.138.211])
	by mgw-ext11.nokia.com (Switch-3.2.5/Switch-3.2.5) with ESMTP id
	l3681VbQ022383; Fri, 6 Apr 2007 11:02:00 +0300
Received: from esebh104.NOE.Nokia.com ([172.21.143.34]) by
	esebh105.NOE.Nokia.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 6 Apr 2007 11:01:43 +0300
Received: from esebh102.NOE.Nokia.com ([172.21.138.183]) by
	esebh104.NOE.Nokia.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 6 Apr 2007 11:01:43 +0300
Received: from mgw-int01.ntc.nokia.com ([172.21.143.96]) by
	esebh102.NOE.Nokia.com over TLS secured channel with Microsoft
	SMTPSVC(6.0.3790.1830); Fri, 6 Apr 2007 11:01:43 +0300
Received: from [192.168.255.2] (essapo-nirac253239.europe.nokia.com
	[10.162.253.239])
	by mgw-int01.ntc.nokia.com (Switch-3.2.5/Switch-3.2.5) with ESMTP id
	l3681dNm020333; Fri, 6 Apr 2007 11:01:40 +0300
In-Reply-To: <26B285A9-3D4C-43B1-9C32-37EBC4AF39BE@nokia.com>
References: <26B285A9-3D4C-43B1-9C32-37EBC4AF39BE@nokia.com>
Mime-Version: 1.0 (Apple Message framework v752.3)
Content-Type: multipart/signed; micalg=sha1; boundary=Apple-Mail-56--29203158;
	protocol="application/pkcs7-signature"
Message-Id: <0260FF1D-E29B-4D91-991D-612E77F53970@nokia.com>
From: Lars Eggert <lars.eggert@nokia.com>
Date: Fri, 6 Apr 2007 11:01:32 +0300
To: TSV Area <tsv-area@ietf.org>
X-Mailer: Apple Mail (2.752.3)
X-OriginalArrivalTime: 06 Apr 2007 08:01:43.0060 (UTC)
	FILETIME=[D07F3540:01C77821]
X-eXpurgate-Category: 1/0
X-eXpurgate-ID: 149371::070406110200-50BF6BB0-1224CEE6/0-0/0-1
X-Nokia-AV: Clean
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 10d3e4e3c32e363f129e380e644649be
Cc: iccrg@cs.ucl.ac.uk, tsvwg WG <tsvwg@ietf.org>
Subject: [tsv-area] consensus call: bringing new congestion control to the
	IETF
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: TSV Area <tsv-area@ietf.org>
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org


--Apple-Mail-56--29203158
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=US-ASCII;
	delsp=yes;
	format=flowed

Hi,

On 2007-3-12, at 10:06, ext Lars Eggert wrote:
> the attentive reader of IETF agendas will have picked up on the  
> following presentation during the TSVAREA meeting:
>
> 20 min	Bringing Experimental High-Speed Congestion Control to the IETF
> 	Lars Eggert

after the presentation at TSVAREA, there was strong consensus in the  
room for the presented process to bring new congestion control to the  
IETF. I'd like to confirm that the community as a whole has consensus  
on this.

Please take a look at the following draft ION, which describes the  
process from the presentation: http://www.ietf.org/IESG/content/ions/ 
drafts/ion-tsv-alt-cc.txt

The consensus call will run until April 18, 2007. Please review the  
draft ION and send comments by the deadline.

Assuming the consensus from Prague is confirmed, I will ask the IESG  
to approve the ION after April 18.

Lars

PS: Please discuss on tsv-area@ietf.org. Reply-to set accordingly.

--Apple-Mail-56--29203158
Content-Transfer-Encoding: base64
Content-Type: application/pkcs7-signature;
	name=smime.p7s
Content-Disposition: attachment;
	filename=smime.p7s
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==

--Apple-Mail-56--29203158--




From tsv-area-bounces@ietf.org Fri Apr 06 11:00:09 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HZpuc-0007mO-Cc; Fri, 06 Apr 2007 10:59:38 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1HZpub-0007gL-GZ
	for tsv-area@ietf.org; Fri, 06 Apr 2007 10:59:37 -0400
Received: from smtpout.mac.com ([17.250.248.185])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1HZpuP-0004L3-EE
	for tsv-area@ietf.org; Fri, 06 Apr 2007 10:59:37 -0400
Received: from mac.com (smtpin05-en2 [10.13.10.150])
	by smtpout.mac.com (Xserve/smtpout15/MantshX 4.0) with ESMTP id
	l36ExIb1029763
	for <tsv-area@ietf.org>; Fri, 6 Apr 2007 07:59:22 -0700 (PDT)
Received: from [192.168.1.64] (adsl-70-231-237-202.dsl.snfc21.sbcglobal.net
	[70.231.237.202]) (authenticated bits=0)
	by mac.com (Xserve/smtpin05/MantshX 4.0) with ESMTP id l36ExGTk017654
	for <tsv-area@ietf.org>; Fri, 6 Apr 2007 07:59:17 -0700 (PDT)
Mime-Version: 1.0 (Apple Message framework v624)
In-Reply-To: <0260FF1D-E29B-4D91-991D-612E77F53970@nokia.com>
References: <26B285A9-3D4C-43B1-9C32-37EBC4AF39BE@nokia.com>
	<0260FF1D-E29B-4D91-991D-612E77F53970@nokia.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <03b5c35065a64acb1934ab9ac81637d4@mac.com>
Content-Transfer-Encoding: 7bit
From: Sally Floyd <sallyfloyd@mac.com>
Subject: Re: [tsv-area] consensus call: bringing new congestion control to
	the	IETF
Date: Fri, 6 Apr 2007 07:59:16 -0700
To: TSV Area <tsv-area@ietf.org>
X-Mailer: Apple Mail (2.624)
X-Brightmail-Tracker: AAAAAA==
X-Brightmail-scanned: yes
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 8ac499381112328dd60aea5b1ff596ea
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

I agree with the consensus on "Bringing Experimental High-Speed 
Congestion Control to the IETF".

- Sally
http://www.icir.org/floyd/





From tsv-area-bounces@ietf.org Fri Apr 20 05:58:20 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hepsh-0006pI-8z; Fri, 20 Apr 2007 05:58:19 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hepsf-0006p1-Pt; Fri, 20 Apr 2007 05:58:17 -0400
Received: from smtp.nokia.com ([131.228.20.170] helo=mgw-ext11.nokia.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1Hepse-0003Nb-05; Fri, 20 Apr 2007 05:58:17 -0400
Received: from esebh105.NOE.Nokia.com (esebh105.ntc.nokia.com [172.21.138.211])
	by mgw-ext11.nokia.com (Switch-3.2.5/Switch-3.2.5) with ESMTP id
	l3K9vrwI024175; Fri, 20 Apr 2007 12:58:09 +0300
Received: from esebh104.NOE.Nokia.com ([172.21.143.34]) by
	esebh105.NOE.Nokia.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 20 Apr 2007 12:57:48 +0300
Received: from esebh101.NOE.Nokia.com ([172.21.138.177]) by
	esebh104.NOE.Nokia.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 20 Apr 2007 12:57:48 +0300
Received: from mgw-int02.ntc.nokia.com ([172.21.143.97]) by
	esebh101.NOE.Nokia.com over TLS secured channel with Microsoft
	SMTPSVC(6.0.3790.1830); Fri, 20 Apr 2007 12:57:47 +0300
Received: from [10.183.156.223] (essapo-nirac253236.europe.nokia.com
	[10.162.253.236])
	by mgw-int02.ntc.nokia.com (Switch-3.2.5/Switch-3.2.5) with ESMTP id
	l3K9vjvY013040; Fri, 20 Apr 2007 12:57:46 +0300
In-Reply-To: <0260FF1D-E29B-4D91-991D-612E77F53970@nokia.com>
References: <26B285A9-3D4C-43B1-9C32-37EBC4AF39BE@nokia.com>
	<0260FF1D-E29B-4D91-991D-612E77F53970@nokia.com>
Mime-Version: 1.0 (Apple Message framework v752.3)
Content-Type: multipart/signed; micalg=sha1; boundary=Apple-Mail-14--960130484;
	protocol="application/pkcs7-signature"
Message-Id: <36B3B805-BCD5-4490-9F5F-87BC56FC5453@nokia.com>
From: Lars Eggert <lars.eggert@nokia.com>
Subject: Re: [tsv-area] consensus call: bringing new congestion control to the
	IETF
Date: Fri, 20 Apr 2007 12:57:28 +0300
To: TSV Area <tsv-area@ietf.org>
X-Mailer: Apple Mail (2.752.3)
X-OriginalArrivalTime: 20 Apr 2007 09:57:47.0794 (UTC)
	FILETIME=[5995C720:01C78332]
X-Nokia-AV: Clean
X-Spam-Score: 0.0 (/)
X-Scan-Signature: b280b4db656c3ca28dd62e5e0b03daa8
Cc: iccrg@cs.ucl.ac.uk, tsvwg WG <tsvwg@ietf.org>
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: TSV Area <tsv-area@ietf.org>
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org


--Apple-Mail-14--960130484
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=US-ASCII;
	delsp=yes;
	format=flowed

On 2007-4-6, at 11:01, ext Lars Eggert wrote:
> after the presentation at TSVAREA, there was strong consensus in  
> the room for the presented process to bring new congestion control  
> to the IETF. I'd like to confirm that the community as a whole has  
> consensus on this.
>
> Please take a look at the following draft ION, which describes the  
> process from the presentation: http://www.ietf.org/IESG/content/ 
> ions/drafts/ion-tsv-alt-cc.txt
>
> The consensus call will run until April 18, 2007. Please review the  
> draft ION and send comments by the deadline.

I have seen very few comments during this consensus call, but all  
where positive. Together with the strong positive feedback from  
Prague, my take is that we have community consensus on this  
procedure. If you think I'm misjudging the consensus, please yell.

Lars



--Apple-Mail-14--960130484
Content-Transfer-Encoding: base64
Content-Type: application/pkcs7-signature;
	name=smime.p7s
Content-Disposition: attachment;
	filename=smime.p7s
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==

--Apple-Mail-14--960130484--




From tsv-area-bounces@ietf.org Thu Apr 26 11:54:10 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh6IJ-0001uo-Bn; Thu, 26 Apr 2007 11:54:07 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh6IH-0001uB-Ah
	for tsv-area@ietf.org; Thu, 26 Apr 2007 11:54:05 -0400
Received: from sj-iport-6.cisco.com ([171.71.176.117])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh6IF-0004Dw-RA
	for tsv-area@ietf.org; Thu, 26 Apr 2007 11:54:05 -0400
Received: from sj-dkim-3.cisco.com ([171.71.179.195])
	by sj-iport-6.cisco.com with ESMTP; 26 Apr 2007 08:54:03 -0700
X-IronPort-AV: i="4.14,455,1170662400"; 
	d="scan'208"; a="140938139:sNHT47796246"
Received: from sj-core-2.cisco.com (sj-core-2.cisco.com [171.71.177.254])
	by sj-dkim-3.cisco.com (8.12.11/8.12.11) with ESMTP id l3QFs31N032513; 
	Thu, 26 Apr 2007 08:54:03 -0700
Received: from dwingwxp ([10.32.240.196])
	by sj-core-2.cisco.com (8.12.10/8.12.6) with ESMTP id l3QFs2ZT014177;
	Thu, 26 Apr 2007 15:54:02 GMT
From: "Dan Wing" <dwing@cisco.com>
To: <tsv-area@ietf.org>
Date: Thu, 26 Apr 2007 08:54:02 -0700
Message-ID: <07f401c7881b$1cd9aa70$05716b80@amer.cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3028
Thread-Index: AceIGxw695NYzQ5TSEebTJP6EGlSXw==
DKIM-Signature: v=0.5; a=rsa-sha256; q=dns/txt; l=4175; t=1177602843;
	x=1178466843; c=relaxed/simple; s=sjdkim3002;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=dwing@cisco.com;
	z=From:=20=22Dan=20Wing=22=20<dwing@cisco.com>
	|Subject:=20ICE=20pacing=20connectivity=20checks |Sender:=20;
	bh=TFly5WSuB8fpL6YRR6YseEPjaOLHOrwCdhl1/55/WfA=;
	b=RCAP01vdaa9zB/gK3mIV3EOqvcVaFl335WcyA1LjCI7Ve1OIiZBc032cBoAzBO6JmdHxvavz
	l8S6l6UoFV+6OlyfGry/e1/ee3ToC5ZkeGp19QYRCOMSr7rHCidu6D7j;
Authentication-Results: sj-dkim-3; header.From=dwing@cisco.com; dkim=pass (s
	ig from cisco.com/sjdkim3002 verified; ); 
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 10d3e4e3c32e363f129e380e644649be
Cc: "'Jonathan Rosenberg \(jdrosen\)'" <jdrosen@cisco.com>
Subject: [tsv-area] ICE pacing connectivity checks
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Lars and Magnus raised an issue in MMUSIC that ICE's timers may not be TCP
friendly
<http://www1.ietf.org/mail-archive/web/mmusic/current/msg05571.html>.  We
are moving this discussion to TSVWG where there is better interest in this
specific matter.

Background:
-----------

ICE is http://tools.ietf.org/html/draft-ietf-mmusic-ice-15, and ICE's
primary purpose is for NAT traversal of RTP media flows.  To do this, ICE
gathers a list of candidate addresses from both media endpoints (private
IPv4 address, public IPv4 address, IPv6 addresses, VPN'd addresses, and so
on), exchange those addresses with each other, and then performs
connectivity checks.  A more detailed explanation of the steps is the
Overview section of the ICE specification.

Details:
--------

The primary concern is that ICE's connectivity checks are worse to the
network than an RTP flow with RTCP Feedback (RFC4585).  RFC4585 allows an
RTP receiver to issue an RTCP message after 500ms, and that message can tell
the RTP sender to 'slow down'; this message might mean switch to a more
bandwidth-efficient codec, for example.  For the first 500ms, however, the
sender will send RTP packets towards the receiver.

The current ICE specification (ice-15) uses two timers to control its
pacing; one is "Ta" which controls when the first connectivity check is sent
for each candidate address and is 20ms.  The other is STUN's retransmit
interval which is defined in draft-ietf-behave-rfc3489bis-06 to start at
100ms and double thereafter (100ms, 300ms, 700ms, 1500ms, 3100ms, 6300ms).

I was able to model ICE with two a=candidate lines (4 candidate pairs) and
STUN's currently-specified timers are modeled against an RTP payload size of
40 bytes and RTP "ptime" (packetization interval) of 20ms, ICE uses more
bandwidth for the first 440ms, at which time RTP consumes more bandwidth.
If we assume both endpoints use RTCP Feedback (RFC4585), in approximately
another 60ms the sender's RTP will be informed of packet loss and slow down.

Of course, if any of ICE's connectivity checks complete (that is, we receive
a STUN Response), ICE processing stops and soon thereafter RTP would start.
This, however, is difficult to model and I did not attempt to model this.


With more conservative RTP payload sizes or RTP packetization intervals, or
with a lot more candidate pairs, ICE becomes quite a lot worse than RTP.  


Proposal:
---------

I propose the following formulas which prevent ICE from exceeding the
bandwidth RTP would have used:

  Initial STUN transaction delay ("Ta") = 
     stun_packet_size / rtp_packet_size * rtp_ptime

  STUN retransmit interval ("RTO") = 
    Ta * candidate_count

These formulas cause us to first send a STUN Binding Request to each
a=candidate, at a rate that consumes the same bandwidth as the RTP flow we
will eventually be establishing with the remote endpoint.  If we don't get
any responses we then start retransmitting with a 2x backoff.


I have modeled these formulas in the Excel spreadsheet at
http://www.employees.org/~dwing/ice-rtp-backoff.xls, where you can also
input Ta=20ms and RTO=100ms and various numbers of candidate pairs to see
how ICE compares with RTP.  The yellow fields are where you can make
adjustments without breaking the spreadsheet.  You can see the bandwidth
used by RTP (including IP, UDP, and PPPoE overhead) versus ICE in the middle
of the spreadsheet, as time increases from 0ms to 1000ms on the left.  The
second worksheet in that workbook has a graph of the same data.  


The advantage of this proposal is that ICE's bandwidth consumption is the
same as RTP's bandwidth consumption.  The disadvantage is that packet loss
of the ICE connectivity checks will likely cause us to learn more slowly
that we do, in fact, have connectivity with a more-preferred candidate.  Of
course, injecting packets into the network faster than the network expected
would, itself, be likely to contribute to packet loss causing the same
situation.  Being kind to the network seems the best approach.

Please let me know if this approach seems reasonable.

-d




From tsv-area-bounces@ietf.org Thu Apr 26 12:48:56 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh79M-0006s4-Hi; Thu, 26 Apr 2007 12:48:56 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh79L-0006rz-Ub
	for tsv-area@ietf.org; Thu, 26 Apr 2007 12:48:55 -0400
Received: from mailgate.siemenscomms.co.uk ([195.171.110.225]
	helo=bemg01.siemenscomms.co.uk)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh79L-0005qB-J5
	for tsv-area@ietf.org; Thu, 26 Apr 2007 12:48:55 -0400
Received: from GBNTHT12009MSX.gb002.siemens.net ([137.223.219.235])
	by siemenscomms.co.uk (PMDF V6.0-24 #40642)
	with ESMTP id <0JH4001266PBG2@siemenscomms.co.uk> for tsv-area@ietf.org;
	Thu, 26 Apr 2007 17:48:48 +0100 (BST)
Date: Thu, 26 Apr 2007 17:48:55 +0100
From: "Elwell, John" <john.elwell@siemens.com>
In-reply-to: <080c01c7881e$6755a150$05716b80@amer.cisco.com>
To: Dan Wing <dwing@cisco.com>, tsv-area@ietf.org
Message-id: <0D5F89FAC29E2C41B98A6A762007F5D00AFA03@GBNTHT12009MSX.gb002.siemens.net>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft Exchange V6.5
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7BIT
Thread-Topic: ICE pacing connectivity checks	(wasRE:[MMUSIC]
	WGLC:draft-ietf-mmusic-ice-15.txt)
Thread-Index: AceCmOYTyAWMEsXOTcySp9+m8ZV0YgAAdgbgAWAcIdAAAbFS4A==
Content-class: urn:content-classes:message
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
References: <00e201c7829d$99cf1430$c4f0200a@amer.cisco.com>
	<080c01c7881e$6755a150$05716b80@amer.cisco.com>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: bdc523f9a54890b8a30dd6fd53d5d024
Cc: 
Subject: [tsv-area] RE: ICE pacing connectivity checks	(wasRE:[MMUSIC]
	WGLC:draft-ietf-mmusic-ice-15.txt)
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Dan,

All these calculations are a bit academic if they don't take into
account SIP forking. Suppose a SIP INVITE request containing an ICE
offer gets forked to 10 UASs, and each agrees to use ICE. There is then
10 times the amount of connectivity check traffic. The UAC might be on a
low bandwidth link, yet it has no control over the degree of forking
that might be encountered downstream.

John 

> -----Original Message-----
> From: Dan Wing [mailto:dwing@cisco.com] 
> Sent: 26 April 2007 17:18
> To: mmusic@ietf.org
> Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> WGLC:draft-ietf-mmusic-ice-15.txt)
> 
> (Jonathan had also listed this as ICE issue #11.)
> 
> I have moved discussion of ICE pacing to the tsv-area mailing list
> <https://www1.ietf.org/mailman/listinfo/tsv-area>, as the 
> tsv-area list is
> more appropriate for discussing TCP friendliness.  
> 
> My post to tsv-area
> <http://www1.ietf.org/mail-archive/web/tsv-area/current/msg00077.html>
> proposed new calculations for ICE's Ta and STUN's RTO.  These changes
> prevent ICE from exceeding the bandwidth RTP would have consumed.
> 
> -d
> 
> 
> > -----Original Message-----
> > From: Dan Wing [mailto:dwing@cisco.com] 
> > Sent: Thursday, April 19, 2007 9:13 AM
> > To: 'Lars Eggert'; 'Magnus Westerlund'
> > Cc: mmusic@ietf.org
> > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > WGLC: draft-ietf-mmusic-ice-15.txt)
> > 
> > I have created an Excel spreadsheet analyzing ICE's behavior 
> > compared to
> > RTP's behavior.  I am tweaking this spreadsheet and expect it 
> > to be useful
> > in helping us tune ICE's algorithms.  I want to make this a 
> > more useful
> > spreadsheet before sharing it; it doesn't quite do everything 
> > I like yet.  I
> > expect it will remain awkward to use, as I'm not an excel wizard.
> > 
> > In my initial analysis using this spreadsheet, I do agree 
> > that ICE's default
> > 20ms pacing and STUN's default 100ms retransmit timers are 
> > too aggressive if
> > we assume severely bandwidth-constrained links.  I expect 
> > that adjusting
> > these values based on the ptime, RTP packet size, STUN 
> > Request packet size,
> > and number of candidates will result in ICE behaving as well 
> > as RTP (with
> > RTCP feedback) in all cases, and better than RTP (with RTCP 
> > feedback) in
> > some cases.  The drawback of such tuning, however, is that 
> > ICE will take
> > longer to complete if there are many candidates.
> > 
> > -d
> > 
> > _______________________________________________
> > mmusic mailing list
> > mmusic@ietf.org
> > https://www1.ietf.org/mailman/listinfo/mmusic
> > 
> 
> _______________________________________________
> mmusic mailing list
> mmusic@ietf.org
> https://www1.ietf.org/mailman/listinfo/mmusic
> 




From tsv-area-bounces@ietf.org Thu Apr 26 12:56:24 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh7GN-0002Ye-09; Thu, 26 Apr 2007 12:56:11 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh7GM-0002YY-6F
	for tsv-area@ietf.org; Thu, 26 Apr 2007 12:56:10 -0400
Received: from mailgw4.ericsson.se ([193.180.251.62])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh7GL-0007Bo-De
	for tsv-area@ietf.org; Thu, 26 Apr 2007 12:56:10 -0400
Received: from mailgw4.ericsson.se (unknown [127.0.0.1])
	by mailgw4.ericsson.se (Symantec Mail Security) with ESMTP id
	D388A20521; Thu, 26 Apr 2007 18:56:08 +0200 (CEST)
X-AuditID: c1b4fb3e-af9eebb0000061ca-ec-4630d9a8b432 
Received: from esealmw129.eemea.ericsson.se (unknown [153.88.254.124])
	by mailgw4.ericsson.se (Symantec Mail Security) with ESMTP id
	B0EA720415; Thu, 26 Apr 2007 18:56:08 +0200 (CEST)
Received: from esealmw127.eemea.ericsson.se ([153.88.254.171]) by
	esealmw129.eemea.ericsson.se with Microsoft SMTPSVC(6.0.3790.1830); 
	Thu, 26 Apr 2007 18:56:08 +0200
Received: from [147.214.30.247] ([147.214.30.247]) by
	esealmw127.eemea.ericsson.se with Microsoft SMTPSVC(6.0.3790.1830); 
	Thu, 26 Apr 2007 18:56:07 +0200
Message-ID: <4630D9A7.7010400@ericsson.com>
Date: Thu, 26 Apr 2007 18:56:07 +0200
From: Magnus Westerlund <magnus.westerlund@ericsson.com>
User-Agent: Thunderbird 1.5.0.10 (Windows/20070221)
MIME-Version: 1.0
To: Dan Wing <dwing@cisco.com>
Subject: Re: [tsv-area] ICE pacing connectivity checks
References: <07f401c7881b$1cd9aa70$05716b80@amer.cisco.com>
In-Reply-To: <07f401c7881b$1cd9aa70$05716b80@amer.cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-OriginalArrivalTime: 26 Apr 2007 16:56:07.0992 (UTC)
	FILETIME=[C8F28B80:01C78823]
X-Brightmail-Tracker: AAAAAA==
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 6cca30437e2d04f45110f2ff8dc1b1d5
Cc: tsv-area@ietf.org, "'Jonathan Rosenberg \(jdrosen\)'" <jdrosen@cisco.com>
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Hi

A few clarifications

Dan Wing skrev:
> Lars and Magnus raised an issue in MMUSIC that ICE's timers may not be TCP
> friendly
> <http://www1.ietf.org/mail-archive/web/mmusic/current/msg05571.html>.  We
> are moving this discussion to TSVWG where there is better interest in this
> specific matter.
> 
> Background:
> -----------
> 
> ICE is http://tools.ietf.org/html/draft-ietf-mmusic-ice-15, and ICE's
> primary purpose is for NAT traversal of RTP media flows.  To do this, ICE
> gathers a list of candidate addresses from both media endpoints (private
> IPv4 address, public IPv4 address, IPv6 addresses, VPN'd addresses, and so
> on), exchange those addresses with each other, and then performs
> connectivity checks.  A more detailed explanation of the steps is the
> Overview section of the ICE specification.
> 
> Details:
> --------
> 
> The primary concern is that ICE's connectivity checks are worse to the
> network than an RTP flow with RTCP Feedback (RFC4585).  RFC4585 allows an
> RTP receiver to issue an RTCP message after 500ms, and that message can tell
> the RTP sender to 'slow down'; this message might mean switch to a more
> bandwidth-efficient codec, for example.  For the first 500ms, however, the
> sender will send RTP packets towards the receiver.

Our general concern is that ICE has no congestion control and in worst 
case can consume large amounts of bandwidth. The primary desire would be 
to have ICE automatically adopt to the amount of available bandwidth on 
the different paths. However that is almost impossible as there is no 
information if the connectivity check is lost due to reaching a NAT 
without an active binding, its filtering behavior or congestion 
introduced loss. Making this into a very hard problem.

I made a comparison with RTP when using RTCP as at least being able to 
react to congestion on a medium timescale. Not TCP-friendly but at least 
it prevents congestion collapse. Also RTP flows can range from 1kb up to 
a couple of GB so there is no good comparison to do. Thus I would like 
to steer away from comparing with RTP as being good or bad.

[snip]

> 
> With more conservative RTP payload sizes or RTP packetization intervals, or
> with a lot more candidate pairs, ICE becomes quite a lot worse than RTP.  

I am especially worried about the increase in number of candidate pairs 
as this is what makes the bandwidth run away as it currently is defined. 
What I think is important for the proposals is that the maintain a 
sensible bandwidth for the whole connectivity check phase independent of 
number of candidate on any number of streams, checkboards etc.

I would also ask people to consider if there is anything that can be 
improved compared relatively to use some basic RTP bandwidth for the 
planned stream.

There are contradicting requirements here. On one side avoid causing 
serious congestion issues, on the other have a reasonable time for 
establishing working paths for all the media streams.

Cheers

Magnus Westerlund

IETF Transport Area Director & TSVWG Chair
----------------------------------------------------------------------
Multimedia Technologies, Ericsson Research EAB/TVM/M
----------------------------------------------------------------------
Ericsson AB                | Phone +46 8 4048287
Torshamsgatan 23           | Fax   +46 8 7575550
S-164 80 Stockholm, Sweden | mailto: magnus.westerlund@ericsson.com
----------------------------------------------------------------------




From tsv-area-bounces@ietf.org Thu Apr 26 13:13:03 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh7WZ-0002fe-94; Thu, 26 Apr 2007 13:12:55 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh7WX-0002fU-KA
	for tsv-area@ietf.org; Thu, 26 Apr 2007 13:12:53 -0400
Received: from sj-iport-5.cisco.com ([171.68.10.87])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh7WX-0001m8-8r
	for tsv-area@ietf.org; Thu, 26 Apr 2007 13:12:53 -0400
Received: from sj-dkim-3.cisco.com ([171.71.179.195])
	by sj-iport-5.cisco.com with ESMTP; 26 Apr 2007 10:12:53 -0700
X-IronPort-AV: i="4.14,456,1170662400"; 
	d="scan'208"; a="415978861:sNHT77885110"
Received: from sj-core-1.cisco.com (sj-core-1.cisco.com [171.71.177.237])
	by sj-dkim-3.cisco.com (8.12.11/8.12.11) with ESMTP id l3QHCqf2020224; 
	Thu, 26 Apr 2007 10:12:52 -0700
Received: from dwingwxp ([10.32.240.196])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id l3QHCqMF008784;
	Thu, 26 Apr 2007 17:12:52 GMT
From: "Dan Wing" <dwing@cisco.com>
To: "'Elwell, John'" <john.elwell@siemens.com>, <tsv-area@ietf.org>
Date: Thu, 26 Apr 2007 10:12:52 -0700
Message-ID: <084601c78826$1fc27180$05716b80@amer.cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3028
Thread-Index: AceCmOYTyAWMEsXOTcySp9+m8ZV0YgAAdgbgAWAcIdAAAbFS4AABCCtg
In-Reply-To: <0D5F89FAC29E2C41B98A6A762007F5D00AFA03@GBNTHT12009MSX.gb002.siemens.net>
DKIM-Signature: v=0.5; a=rsa-sha256; q=dns/txt; l=3352; t=1177607572;
	x=1178471572; c=relaxed/simple; s=sjdkim3002;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=dwing@cisco.com;
	z=From:=20=22Dan=20Wing=22=20<dwing@cisco.com>
	|Subject:=20RE=3A=20ICE=20pacing=20connectivity=20checks=09(wasRE=3A[MMUS
	IC]WGLC=3Adraft-ietf-mmusic-ice-15.txt) |Sender:=20;
	bh=b2BWIdLvg94yzuDt01aFWIp94qxu7yGUE7fv2SiDypQ=;
	b=tKno4z7Yox/DIWwwK/DcKVY0au0Czo7e5Y0LSl+uQCDFz1Gt/FVCZHwgXxbhklb807z865MG
	EZdHu36OLDGDwdV/8F8y13r/fFiu5XJ/0DKgWZcdFeY8czPG6fWCGUbv;
Authentication-Results: sj-dkim-3; header.From=dwing@cisco.com; dkim=pass (s
	ig from cisco.com/sjdkim3002 verified; ); 
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 14582b0692e7f70ce7111d04db3781c8
Cc: 
Subject: [tsv-area] RE: ICE pacing connectivity
	checks	(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Do you have a suggestion?

-d
 

> -----Original Message-----
> From: Elwell, John [mailto:john.elwell@siemens.com] 
> Sent: Thursday, April 26, 2007 9:49 AM
> To: Dan Wing; tsv-area@ietf.org
> Subject: RE: ICE pacing connectivity checks 
> (wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> 
> Dan,
> 
> All these calculations are a bit academic if they don't take into
> account SIP forking. Suppose a SIP INVITE request containing an ICE
> offer gets forked to 10 UASs, and each agrees to use ICE. 
> There is then
> 10 times the amount of connectivity check traffic. The UAC 
> might be on a
> low bandwidth link, yet it has no control over the degree of forking
> that might be encountered downstream.
> 
> John 
> 
> > -----Original Message-----
> > From: Dan Wing [mailto:dwing@cisco.com] 
> > Sent: 26 April 2007 17:18
> > To: mmusic@ietf.org
> > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > WGLC:draft-ietf-mmusic-ice-15.txt)
> > 
> > (Jonathan had also listed this as ICE issue #11.)
> > 
> > I have moved discussion of ICE pacing to the tsv-area mailing list
> > <https://www1.ietf.org/mailman/listinfo/tsv-area>, as the 
> > tsv-area list is
> > more appropriate for discussing TCP friendliness.  
> > 
> > My post to tsv-area
> > 
> <http://www1.ietf.org/mail-archive/web/tsv-area/current/msg00077.html>
> > proposed new calculations for ICE's Ta and STUN's RTO.  
> These changes
> > prevent ICE from exceeding the bandwidth RTP would have consumed.
> > 
> > -d
> > 
> > 
> > > -----Original Message-----
> > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > Sent: Thursday, April 19, 2007 9:13 AM
> > > To: 'Lars Eggert'; 'Magnus Westerlund'
> > > Cc: mmusic@ietf.org
> > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > WGLC: draft-ietf-mmusic-ice-15.txt)
> > > 
> > > I have created an Excel spreadsheet analyzing ICE's behavior 
> > > compared to
> > > RTP's behavior.  I am tweaking this spreadsheet and expect it 
> > > to be useful
> > > in helping us tune ICE's algorithms.  I want to make this a 
> > > more useful
> > > spreadsheet before sharing it; it doesn't quite do everything 
> > > I like yet.  I
> > > expect it will remain awkward to use, as I'm not an excel wizard.
> > > 
> > > In my initial analysis using this spreadsheet, I do agree 
> > > that ICE's default
> > > 20ms pacing and STUN's default 100ms retransmit timers are 
> > > too aggressive if
> > > we assume severely bandwidth-constrained links.  I expect 
> > > that adjusting
> > > these values based on the ptime, RTP packet size, STUN 
> > > Request packet size,
> > > and number of candidates will result in ICE behaving as well 
> > > as RTP (with
> > > RTCP feedback) in all cases, and better than RTP (with RTCP 
> > > feedback) in
> > > some cases.  The drawback of such tuning, however, is that 
> > > ICE will take
> > > longer to complete if there are many candidates.
> > > 
> > > -d
> > > 
> > > _______________________________________________
> > > mmusic mailing list
> > > mmusic@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > 
> > 
> > _______________________________________________
> > mmusic mailing list
> > mmusic@ietf.org
> > https://www1.ietf.org/mailman/listinfo/mmusic
> > 




From tsv-area-bounces@ietf.org Thu Apr 26 13:22:11 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh7fO-00062H-IB; Thu, 26 Apr 2007 13:22:02 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh7fM-0005yy-Uz
	for tsv-area@ietf.org; Thu, 26 Apr 2007 13:22:00 -0400
Received: from mailgate.siemenscomms.co.uk ([195.171.110.225]
	helo=bemg01.siemenscomms.co.uk)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh7fM-0003OH-Fq
	for tsv-area@ietf.org; Thu, 26 Apr 2007 13:22:00 -0400
Received: from GBNTHT12009MSX.gb002.siemens.net ([137.223.219.235])
	by siemenscomms.co.uk (PMDF V6.0-24 #40642)
	with ESMTP id <0JH4008AX88GDF@siemenscomms.co.uk> for tsv-area@ietf.org;
	Thu, 26 Apr 2007 18:21:52 +0100 (BST)
Date: Thu, 26 Apr 2007 18:21:58 +0100
From: "Elwell, John" <john.elwell@siemens.com>
In-reply-to: <084601c78826$1fc27180$05716b80@amer.cisco.com>
To: Dan Wing <dwing@cisco.com>, tsv-area@ietf.org
Message-id: <0D5F89FAC29E2C41B98A6A762007F5D00AFA07@GBNTHT12009MSX.gb002.siemens.net>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft Exchange V6.5
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7BIT
Thread-Topic: ICE pacing connectivity
	checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
Thread-Index: AceCmOYTyAWMEsXOTcySp9+m8ZV0YgAAdgbgAWAcIdAAAbFS4AABCCtgAABH7FA=
Content-class: urn:content-classes:message
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
References: <0D5F89FAC29E2C41B98A6A762007F5D00AFA03@GBNTHT12009MSX.gb002.siemens.net>
	<084601c78826$1fc27180$05716b80@amer.cisco.com>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: c83ccb5cc10e751496398f1233ca9c3a
Cc: 
Subject: [tsv-area] RE: ICE pacing connectivity
 checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Well, perhaps the pacing needs to be adjusted according to the number of
simultaneous ICE sessions on a given interface, if bandwidth is believed
to be limited.

John 

> -----Original Message-----
> From: Dan Wing [mailto:dwing@cisco.com] 
> Sent: 26 April 2007 18:13
> To: Elwell, John; tsv-area@ietf.org
> Subject: RE: ICE pacing connectivity 
> checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> 
> Do you have a suggestion?
> 
> -d
>  
> 
> > -----Original Message-----
> > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > Sent: Thursday, April 26, 2007 9:49 AM
> > To: Dan Wing; tsv-area@ietf.org
> > Subject: RE: ICE pacing connectivity checks 
> > (wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > 
> > Dan,
> > 
> > All these calculations are a bit academic if they don't take into
> > account SIP forking. Suppose a SIP INVITE request containing an ICE
> > offer gets forked to 10 UASs, and each agrees to use ICE. 
> > There is then
> > 10 times the amount of connectivity check traffic. The UAC 
> > might be on a
> > low bandwidth link, yet it has no control over the degree of forking
> > that might be encountered downstream.
> > 
> > John 
> > 
> > > -----Original Message-----
> > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > Sent: 26 April 2007 17:18
> > > To: mmusic@ietf.org
> > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > WGLC:draft-ietf-mmusic-ice-15.txt)
> > > 
> > > (Jonathan had also listed this as ICE issue #11.)
> > > 
> > > I have moved discussion of ICE pacing to the tsv-area mailing list
> > > <https://www1.ietf.org/mailman/listinfo/tsv-area>, as the 
> > > tsv-area list is
> > > more appropriate for discussing TCP friendliness.  
> > > 
> > > My post to tsv-area
> > > 
> > 
> <http://www1.ietf.org/mail-archive/web/tsv-area/current/msg00077.html>
> > > proposed new calculations for ICE's Ta and STUN's RTO.  
> > These changes
> > > prevent ICE from exceeding the bandwidth RTP would have consumed.
> > > 
> > > -d
> > > 
> > > 
> > > > -----Original Message-----
> > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > Sent: Thursday, April 19, 2007 9:13 AM
> > > > To: 'Lars Eggert'; 'Magnus Westerlund'
> > > > Cc: mmusic@ietf.org
> > > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > > WGLC: draft-ietf-mmusic-ice-15.txt)
> > > > 
> > > > I have created an Excel spreadsheet analyzing ICE's behavior 
> > > > compared to
> > > > RTP's behavior.  I am tweaking this spreadsheet and expect it 
> > > > to be useful
> > > > in helping us tune ICE's algorithms.  I want to make this a 
> > > > more useful
> > > > spreadsheet before sharing it; it doesn't quite do everything 
> > > > I like yet.  I
> > > > expect it will remain awkward to use, as I'm not an 
> excel wizard.
> > > > 
> > > > In my initial analysis using this spreadsheet, I do agree 
> > > > that ICE's default
> > > > 20ms pacing and STUN's default 100ms retransmit timers are 
> > > > too aggressive if
> > > > we assume severely bandwidth-constrained links.  I expect 
> > > > that adjusting
> > > > these values based on the ptime, RTP packet size, STUN 
> > > > Request packet size,
> > > > and number of candidates will result in ICE behaving as well 
> > > > as RTP (with
> > > > RTCP feedback) in all cases, and better than RTP (with RTCP 
> > > > feedback) in
> > > > some cases.  The drawback of such tuning, however, is that 
> > > > ICE will take
> > > > longer to complete if there are many candidates.
> > > > 
> > > > -d
> > > > 
> > > > _______________________________________________
> > > > mmusic mailing list
> > > > mmusic@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > 
> > > 
> > > _______________________________________________
> > > mmusic mailing list
> > > mmusic@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > 
> 




From tsv-area-bounces@ietf.org Thu Apr 26 13:56:47 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh8Cf-0001dK-CJ; Thu, 26 Apr 2007 13:56:25 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh8Ce-0001cv-GM
	for tsv-area@ietf.org; Thu, 26 Apr 2007 13:56:24 -0400
Received: from sj-iport-6.cisco.com ([171.71.176.117])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh8CY-0002gD-Ul
	for tsv-area@ietf.org; Thu, 26 Apr 2007 13:56:21 -0400
Received: from sj-dkim-4.cisco.com ([171.71.179.196])
	by sj-iport-6.cisco.com with ESMTP; 26 Apr 2007 10:56:18 -0700
X-IronPort-AV: i="4.14,456,1170662400"; 
	d="scan'208"; a="140993269:sNHT49821696"
Received: from sj-core-1.cisco.com (sj-core-1.cisco.com [171.71.177.237])
	by sj-dkim-4.cisco.com (8.12.11/8.12.11) with ESMTP id l3QHuISr026661; 
	Thu, 26 Apr 2007 10:56:18 -0700
Received: from dwingwxp (dhcp-128-107-163-142.cisco.com [128.107.163.142])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id l3QHuIMF013108;
	Thu, 26 Apr 2007 17:56:18 GMT
From: "Dan Wing" <dwing@cisco.com>
To: "'Elwell, John'" <john.elwell@siemens.com>, <tsv-area@ietf.org>
Date: Thu, 26 Apr 2007 10:56:18 -0700
Message-ID: <087201c7882c$31129b30$05716b80@amer.cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3028
Thread-Index: AceCmOYTyAWMEsXOTcySp9+m8ZV0YgAAdgbgAWAcIdAAAbFS4AABCCtgAABH7FAAAR8hMA==
In-Reply-To: <0D5F89FAC29E2C41B98A6A762007F5D00AFA07@GBNTHT12009MSX.gb002.siemens.net>
DKIM-Signature: v=0.5; a=rsa-sha256; q=dns/txt; l=4927; t=1177610178;
	x=1178474178; c=relaxed/simple; s=sjdkim4002;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=dwing@cisco.com;
	z=From:=20=22Dan=20Wing=22=20<dwing@cisco.com>
	|Subject:=20RE=3A=20ICE=20pacing=20connectivity=20checks(wasRE=3A[MMUSIC]
	WGLC=3Adraft-ietf-mmusic-ice-15.txt) |Sender:=20;
	bh=KcaAeWdEx3jimcaKG+RPmbpsMHsP0gZMnYVMSkhWcBc=;
	b=L0VbAojqIpM04Gyx4p0gKlzpZ98YlYMW44anSN8g80sHKjw8CWC0NLbN2+8EvZf3d8ElFBxB
	NY1IaUgUmfBbDmSWhDFrn/6R1VI2Pxxw8zjt53LDPtxEJ+01MjvjFUdv;
Authentication-Results: sj-dkim-4; header.From=dwing@cisco.com; dkim=pass (s
	ig from cisco.com/sjdkim4002 verified; ); 
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 6640e3bbe8a4d70c4469bcdcbbf0921d
Cc: 
Subject: [tsv-area] RE: ICE pacing connectivity
	checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

That could be done by calculating candidate_count to include all candidates
from all answerers - that would adjust Ta accordingly. 

That would help from the Offerer towards all of the forked endpoints, but it
wouldn't help in the opposite direction because the forked endpoints don't
know there are other forked endpoints.  It was that case (ICE checks from
answerers->offerer) I thought you were worried about.

-d


> -----Original Message-----
> From: Elwell, John [mailto:john.elwell@siemens.com] 
> Sent: Thursday, April 26, 2007 10:22 AM
> To: Dan Wing; tsv-area@ietf.org
> Subject: RE: ICE pacing connectivity 
> checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> 
> Well, perhaps the pacing needs to be adjusted according to 
> the number of
> simultaneous ICE sessions on a given interface, if bandwidth 
> is believed
> to be limited.
> 
> John 
> 
> > -----Original Message-----
> > From: Dan Wing [mailto:dwing@cisco.com] 
> > Sent: 26 April 2007 18:13
> > To: Elwell, John; tsv-area@ietf.org
> > Subject: RE: ICE pacing connectivity 
> > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > 
> > Do you have a suggestion?
> > 
> > -d
> >  
> > 
> > > -----Original Message-----
> > > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > > Sent: Thursday, April 26, 2007 9:49 AM
> > > To: Dan Wing; tsv-area@ietf.org
> > > Subject: RE: ICE pacing connectivity checks 
> > > (wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > 
> > > Dan,
> > > 
> > > All these calculations are a bit academic if they don't take into
> > > account SIP forking. Suppose a SIP INVITE request 
> containing an ICE
> > > offer gets forked to 10 UASs, and each agrees to use ICE. 
> > > There is then
> > > 10 times the amount of connectivity check traffic. The UAC 
> > > might be on a
> > > low bandwidth link, yet it has no control over the degree 
> of forking
> > > that might be encountered downstream.
> > > 
> > > John 
> > > 
> > > > -----Original Message-----
> > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > Sent: 26 April 2007 17:18
> > > > To: mmusic@ietf.org
> > > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > > WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > 
> > > > (Jonathan had also listed this as ICE issue #11.)
> > > > 
> > > > I have moved discussion of ICE pacing to the tsv-area 
> mailing list
> > > > <https://www1.ietf.org/mailman/listinfo/tsv-area>, as the 
> > > > tsv-area list is
> > > > more appropriate for discussing TCP friendliness.  
> > > > 
> > > > My post to tsv-area
> > > > 
> > > 
> > 
> <http://www1.ietf.org/mail-archive/web/tsv-area/current/msg00077.html>
> > > > proposed new calculations for ICE's Ta and STUN's RTO.  
> > > These changes
> > > > prevent ICE from exceeding the bandwidth RTP would have 
> consumed.
> > > > 
> > > > -d
> > > > 
> > > > 
> > > > > -----Original Message-----
> > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > Sent: Thursday, April 19, 2007 9:13 AM
> > > > > To: 'Lars Eggert'; 'Magnus Westerlund'
> > > > > Cc: mmusic@ietf.org
> > > > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > > > WGLC: draft-ietf-mmusic-ice-15.txt)
> > > > > 
> > > > > I have created an Excel spreadsheet analyzing ICE's behavior 
> > > > > compared to
> > > > > RTP's behavior.  I am tweaking this spreadsheet and expect it 
> > > > > to be useful
> > > > > in helping us tune ICE's algorithms.  I want to make this a 
> > > > > more useful
> > > > > spreadsheet before sharing it; it doesn't quite do everything 
> > > > > I like yet.  I
> > > > > expect it will remain awkward to use, as I'm not an 
> > excel wizard.
> > > > > 
> > > > > In my initial analysis using this spreadsheet, I do agree 
> > > > > that ICE's default
> > > > > 20ms pacing and STUN's default 100ms retransmit timers are 
> > > > > too aggressive if
> > > > > we assume severely bandwidth-constrained links.  I expect 
> > > > > that adjusting
> > > > > these values based on the ptime, RTP packet size, STUN 
> > > > > Request packet size,
> > > > > and number of candidates will result in ICE behaving as well 
> > > > > as RTP (with
> > > > > RTCP feedback) in all cases, and better than RTP (with RTCP 
> > > > > feedback) in
> > > > > some cases.  The drawback of such tuning, however, is that 
> > > > > ICE will take
> > > > > longer to complete if there are many candidates.
> > > > > 
> > > > > -d
> > > > > 
> > > > > _______________________________________________
> > > > > mmusic mailing list
> > > > > mmusic@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > > 
> > > > 
> > > > _______________________________________________
> > > > mmusic mailing list
> > > > mmusic@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > 
> > 




From tsv-area-bounces@ietf.org Thu Apr 26 15:03:11 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh9FG-0004TW-Mn; Thu, 26 Apr 2007 15:03:10 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh9FF-0004TR-DD
	for tsv-area@ietf.org; Thu, 26 Apr 2007 15:03:09 -0400
Received: from mailgate.siemenscomms.co.uk ([195.171.110.225]
	helo=bemg01.siemenscomms.co.uk)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh9FD-0003Tv-SV
	for tsv-area@ietf.org; Thu, 26 Apr 2007 15:03:09 -0400
Received: from GBNTHT12009MSX.gb002.siemens.net ([137.223.219.235])
	by siemenscomms.co.uk (PMDF V6.0-24 #40642)
	with ESMTP id <0JH40098QCX03B@siemenscomms.co.uk> for tsv-area@ietf.org;
	Thu, 26 Apr 2007 20:03:00 +0100 (BST)
Date: Thu, 26 Apr 2007 20:03:05 +0100
From: "Elwell, John" <john.elwell@siemens.com>
In-reply-to: <087201c7882c$31129b30$05716b80@amer.cisco.com>
To: Dan Wing <dwing@cisco.com>, tsv-area@ietf.org
Message-id: <0D5F89FAC29E2C41B98A6A762007F5D00AFA0C@GBNTHT12009MSX.gb002.siemens.net>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft Exchange V6.5
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7BIT
Thread-Topic: ICE pacing connectivity
	checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
Thread-Index: AceCmOYTyAWMEsXOTcySp9+m8ZV0YgAAdgbgAWAcIdAAAbFS4AABCCtgAABH7FAAAR8hMAACZkow
Content-class: urn:content-classes:message
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
References: <0D5F89FAC29E2C41B98A6A762007F5D00AFA07@GBNTHT12009MSX.gb002.siemens.net>
	<087201c7882c$31129b30$05716b80@amer.cisco.com>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 3fbd9b434023f8abfcb1532abaec7a21
Cc: 
Subject: [tsv-area] RE: ICE pacing connectivity
 checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

True - it just shows that forking beyond modest levels (2-3 branches,
say) can be problematic.

John 

> -----Original Message-----
> From: Dan Wing [mailto:dwing@cisco.com] 
> Sent: 26 April 2007 18:56
> To: Elwell, John; tsv-area@ietf.org
> Subject: RE: ICE pacing connectivity 
> checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> 
> That could be done by calculating candidate_count to include 
> all candidates
> from all answerers - that would adjust Ta accordingly. 
> 
> That would help from the Offerer towards all of the forked 
> endpoints, but it
> wouldn't help in the opposite direction because the forked 
> endpoints don't
> know there are other forked endpoints.  It was that case (ICE 
> checks from
> answerers->offerer) I thought you were worried about.
> 
> -d
> 
> 
> > -----Original Message-----
> > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > Sent: Thursday, April 26, 2007 10:22 AM
> > To: Dan Wing; tsv-area@ietf.org
> > Subject: RE: ICE pacing connectivity 
> > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > 
> > Well, perhaps the pacing needs to be adjusted according to 
> > the number of
> > simultaneous ICE sessions on a given interface, if bandwidth 
> > is believed
> > to be limited.
> > 
> > John 
> > 
> > > -----Original Message-----
> > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > Sent: 26 April 2007 18:13
> > > To: Elwell, John; tsv-area@ietf.org
> > > Subject: RE: ICE pacing connectivity 
> > > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > 
> > > Do you have a suggestion?
> > > 
> > > -d
> > >  
> > > 
> > > > -----Original Message-----
> > > > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > > > Sent: Thursday, April 26, 2007 9:49 AM
> > > > To: Dan Wing; tsv-area@ietf.org
> > > > Subject: RE: ICE pacing connectivity checks 
> > > > (wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > 
> > > > Dan,
> > > > 
> > > > All these calculations are a bit academic if they don't 
> take into
> > > > account SIP forking. Suppose a SIP INVITE request 
> > containing an ICE
> > > > offer gets forked to 10 UASs, and each agrees to use ICE. 
> > > > There is then
> > > > 10 times the amount of connectivity check traffic. The UAC 
> > > > might be on a
> > > > low bandwidth link, yet it has no control over the degree 
> > of forking
> > > > that might be encountered downstream.
> > > > 
> > > > John 
> > > > 
> > > > > -----Original Message-----
> > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > Sent: 26 April 2007 17:18
> > > > > To: mmusic@ietf.org
> > > > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > > > WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > > 
> > > > > (Jonathan had also listed this as ICE issue #11.)
> > > > > 
> > > > > I have moved discussion of ICE pacing to the tsv-area 
> > mailing list
> > > > > <https://www1.ietf.org/mailman/listinfo/tsv-area>, as the 
> > > > > tsv-area list is
> > > > > more appropriate for discussing TCP friendliness.  
> > > > > 
> > > > > My post to tsv-area
> > > > > 
> > > > 
> > > 
> > 
> <http://www1.ietf.org/mail-archive/web/tsv-area/current/msg00077.html>
> > > > > proposed new calculations for ICE's Ta and STUN's RTO.  
> > > > These changes
> > > > > prevent ICE from exceeding the bandwidth RTP would have 
> > consumed.
> > > > > 
> > > > > -d
> > > > > 
> > > > > 
> > > > > > -----Original Message-----
> > > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > > Sent: Thursday, April 19, 2007 9:13 AM
> > > > > > To: 'Lars Eggert'; 'Magnus Westerlund'
> > > > > > Cc: mmusic@ietf.org
> > > > > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > > > > WGLC: draft-ietf-mmusic-ice-15.txt)
> > > > > > 
> > > > > > I have created an Excel spreadsheet analyzing ICE's 
> behavior 
> > > > > > compared to
> > > > > > RTP's behavior.  I am tweaking this spreadsheet and 
> expect it 
> > > > > > to be useful
> > > > > > in helping us tune ICE's algorithms.  I want to make this a 
> > > > > > more useful
> > > > > > spreadsheet before sharing it; it doesn't quite do 
> everything 
> > > > > > I like yet.  I
> > > > > > expect it will remain awkward to use, as I'm not an 
> > > excel wizard.
> > > > > > 
> > > > > > In my initial analysis using this spreadsheet, I do agree 
> > > > > > that ICE's default
> > > > > > 20ms pacing and STUN's default 100ms retransmit timers are 
> > > > > > too aggressive if
> > > > > > we assume severely bandwidth-constrained links.  I expect 
> > > > > > that adjusting
> > > > > > these values based on the ptime, RTP packet size, STUN 
> > > > > > Request packet size,
> > > > > > and number of candidates will result in ICE 
> behaving as well 
> > > > > > as RTP (with
> > > > > > RTCP feedback) in all cases, and better than RTP (with RTCP 
> > > > > > feedback) in
> > > > > > some cases.  The drawback of such tuning, however, is that 
> > > > > > ICE will take
> > > > > > longer to complete if there are many candidates.
> > > > > > 
> > > > > > -d
> > > > > > 
> > > > > > _______________________________________________
> > > > > > mmusic mailing list
> > > > > > mmusic@ietf.org
> > > > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > > > 
> > > > > 
> > > > > _______________________________________________
> > > > > mmusic mailing list
> > > > > mmusic@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > > 
> > > 
> 




From tsv-area-bounces@ietf.org Thu Apr 26 15:40:45 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1Hh9pd-0004k0-15; Thu, 26 Apr 2007 15:40:45 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1Hh9pZ-0004jr-8h
	for tsv-area@ietf.org; Thu, 26 Apr 2007 15:40:41 -0400
Received: from sj-iport-6.cisco.com ([171.71.176.117])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1Hh9pX-0003G6-LB
	for tsv-area@ietf.org; Thu, 26 Apr 2007 15:40:41 -0400
Received: from sj-dkim-3.cisco.com ([171.71.179.195])
	by sj-iport-6.cisco.com with ESMTP; 26 Apr 2007 12:40:39 -0700
X-IronPort-AV: i="4.14,457,1170662400"; 
	d="scan'208"; a="141049610:sNHT52132419"
Received: from sj-core-1.cisco.com (sj-core-1.cisco.com [171.71.177.237])
	by sj-dkim-3.cisco.com (8.12.11/8.12.11) with ESMTP id l3QJedRA029372; 
	Thu, 26 Apr 2007 12:40:39 -0700
Received: from dwingwxp (dhcp-128-107-114-150.cisco.com [128.107.114.150])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id l3QJedMF007773;
	Thu, 26 Apr 2007 19:40:39 GMT
From: "Dan Wing" <dwing@cisco.com>
To: "'Elwell, John'" <john.elwell@siemens.com>, <tsv-area@ietf.org>
Date: Thu, 26 Apr 2007 12:40:39 -0700
Message-ID: <08a401c7883a$c4da25a0$05716b80@amer.cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3028
Thread-Index: AceINYgEVmZU4VcWQc6Irfi9fJyNjAABSSNg
In-Reply-To: <0D5F89FAC29E2C41B98A6A762007F5D00AFA0C@GBNTHT12009MSX.gb002.siemens.net>
DKIM-Signature: v=0.5; a=rsa-sha256; q=dns/txt; l=6303; t=1177616439;
	x=1178480439; c=relaxed/simple; s=sjdkim3002;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=dwing@cisco.com;
	z=From:=20=22Dan=20Wing=22=20<dwing@cisco.com>
	|Subject:=20RE=3A=20ICE=20pacing=20connectivity=20checks(wasRE=3A[MMUSIC]
	WGLC=3Adraft-ietf-mmusic-ice-15.txt) |Sender:=20;
	bh=VNtsnLme1AyoE5g1xDLLvu2OUiWuMYI8pZ2Ez+hi6Uk=;
	b=rUjU9UQiJnEEiB2BvUbh3Dms2BimzMYXeOFGmcggPA1QMwKAT4qzV4Mt0m81Z+PRYsYSV8xJ
	iiLAUxQ/Nvv49KpscN/Y5YrkSLgk54JgOt+0WeKlgQ5jt7CSPHZHoINk;
Authentication-Results: sj-dkim-3; header.From=dwing@cisco.com; dkim=pass (s
	ig from cisco.com/sjdkim3002 verified; ); 
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9af087f15dbdd4c64ae6bbcdbc5b1d44
Cc: 
Subject: [tsv-area] RE: ICE pacing connectivity
	checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Yep; but it's problematic for RTP, too, if early media is sent by the
branches.

-d


> -----Original Message-----
> From: Elwell, John [mailto:john.elwell@siemens.com] 
> Sent: Thursday, April 26, 2007 12:03 PM
> To: Dan Wing; tsv-area@ietf.org
> Subject: RE: ICE pacing connectivity 
> checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> 
> True - it just shows that forking beyond modest levels (2-3 branches,
> say) can be problematic.
> 
> John 
> 
> > -----Original Message-----
> > From: Dan Wing [mailto:dwing@cisco.com] 
> > Sent: 26 April 2007 18:56
> > To: Elwell, John; tsv-area@ietf.org
> > Subject: RE: ICE pacing connectivity 
> > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > 
> > That could be done by calculating candidate_count to include 
> > all candidates
> > from all answerers - that would adjust Ta accordingly. 
> > 
> > That would help from the Offerer towards all of the forked 
> > endpoints, but it
> > wouldn't help in the opposite direction because the forked 
> > endpoints don't
> > know there are other forked endpoints.  It was that case (ICE 
> > checks from
> > answerers->offerer) I thought you were worried about.
> > 
> > -d
> > 
> > 
> > > -----Original Message-----
> > > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > > Sent: Thursday, April 26, 2007 10:22 AM
> > > To: Dan Wing; tsv-area@ietf.org
> > > Subject: RE: ICE pacing connectivity 
> > > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > 
> > > Well, perhaps the pacing needs to be adjusted according to 
> > > the number of
> > > simultaneous ICE sessions on a given interface, if bandwidth 
> > > is believed
> > > to be limited.
> > > 
> > > John 
> > > 
> > > > -----Original Message-----
> > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > Sent: 26 April 2007 18:13
> > > > To: Elwell, John; tsv-area@ietf.org
> > > > Subject: RE: ICE pacing connectivity 
> > > > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > 
> > > > Do you have a suggestion?
> > > > 
> > > > -d
> > > >  
> > > > 
> > > > > -----Original Message-----
> > > > > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > > > > Sent: Thursday, April 26, 2007 9:49 AM
> > > > > To: Dan Wing; tsv-area@ietf.org
> > > > > Subject: RE: ICE pacing connectivity checks 
> > > > > (wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > > 
> > > > > Dan,
> > > > > 
> > > > > All these calculations are a bit academic if they don't 
> > take into
> > > > > account SIP forking. Suppose a SIP INVITE request 
> > > containing an ICE
> > > > > offer gets forked to 10 UASs, and each agrees to use ICE. 
> > > > > There is then
> > > > > 10 times the amount of connectivity check traffic. The UAC 
> > > > > might be on a
> > > > > low bandwidth link, yet it has no control over the degree 
> > > of forking
> > > > > that might be encountered downstream.
> > > > > 
> > > > > John 
> > > > > 
> > > > > > -----Original Message-----
> > > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > > Sent: 26 April 2007 17:18
> > > > > > To: mmusic@ietf.org
> > > > > > Subject: RE: ICE pacing connectivity checks (wasRE:[MMUSIC] 
> > > > > > WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > > > 
> > > > > > (Jonathan had also listed this as ICE issue #11.)
> > > > > > 
> > > > > > I have moved discussion of ICE pacing to the tsv-area 
> > > mailing list
> > > > > > <https://www1.ietf.org/mailman/listinfo/tsv-area>, as the 
> > > > > > tsv-area list is
> > > > > > more appropriate for discussing TCP friendliness.  
> > > > > > 
> > > > > > My post to tsv-area
> > > > > > 
> > > > > 
> > > > 
> > > 
> > 
> <http://www1.ietf.org/mail-archive/web/tsv-area/current/msg00077.html>
> > > > > > proposed new calculations for ICE's Ta and STUN's RTO.  
> > > > > These changes
> > > > > > prevent ICE from exceeding the bandwidth RTP would have 
> > > consumed.
> > > > > > 
> > > > > > -d
> > > > > > 
> > > > > > 
> > > > > > > -----Original Message-----
> > > > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > > > Sent: Thursday, April 19, 2007 9:13 AM
> > > > > > > To: 'Lars Eggert'; 'Magnus Westerlund'
> > > > > > > Cc: mmusic@ietf.org
> > > > > > > Subject: RE: ICE pacing connectivity checks 
> (wasRE:[MMUSIC] 
> > > > > > > WGLC: draft-ietf-mmusic-ice-15.txt)
> > > > > > > 
> > > > > > > I have created an Excel spreadsheet analyzing ICE's 
> > behavior 
> > > > > > > compared to
> > > > > > > RTP's behavior.  I am tweaking this spreadsheet and 
> > expect it 
> > > > > > > to be useful
> > > > > > > in helping us tune ICE's algorithms.  I want to 
> make this a 
> > > > > > > more useful
> > > > > > > spreadsheet before sharing it; it doesn't quite do 
> > everything 
> > > > > > > I like yet.  I
> > > > > > > expect it will remain awkward to use, as I'm not an 
> > > > excel wizard.
> > > > > > > 
> > > > > > > In my initial analysis using this spreadsheet, I do agree 
> > > > > > > that ICE's default
> > > > > > > 20ms pacing and STUN's default 100ms retransmit 
> timers are 
> > > > > > > too aggressive if
> > > > > > > we assume severely bandwidth-constrained links.  I expect 
> > > > > > > that adjusting
> > > > > > > these values based on the ptime, RTP packet size, STUN 
> > > > > > > Request packet size,
> > > > > > > and number of candidates will result in ICE 
> > behaving as well 
> > > > > > > as RTP (with
> > > > > > > RTCP feedback) in all cases, and better than RTP 
> (with RTCP 
> > > > > > > feedback) in
> > > > > > > some cases.  The drawback of such tuning, 
> however, is that 
> > > > > > > ICE will take
> > > > > > > longer to complete if there are many candidates.
> > > > > > > 
> > > > > > > -d
> > > > > > > 
> > > > > > > _______________________________________________
> > > > > > > mmusic mailing list
> > > > > > > mmusic@ietf.org
> > > > > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > > > > 
> > > > > > 
> > > > > > _______________________________________________
> > > > > > mmusic mailing list
> > > > > > mmusic@ietf.org
> > > > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > > > 
> > > > 
> > 




From tsv-area-bounces@ietf.org Thu Apr 26 18:29:46 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HhCTA-0002vC-0W; Thu, 26 Apr 2007 18:29:44 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1HhCT8-0002v7-9o
	for tsv-area@ietf.org; Thu, 26 Apr 2007 18:29:42 -0400
Received: from smtpout.mac.com ([17.250.248.184])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1HhCT5-0004CS-T0
	for tsv-area@ietf.org; Thu, 26 Apr 2007 18:29:42 -0400
Received: from mac.com (smtpin07-en2 [10.13.10.152])
	by smtpout.mac.com (Xserve/smtpout14/MantshX 4.0) with ESMTP id
	l3QMTQJZ010526; Thu, 26 Apr 2007 15:29:26 -0700 (PDT)
Received: from [192.150.186.254] (laptop254.icsi.berkeley.edu
	[192.150.186.254]) (authenticated bits=0)
	by mac.com (Xserve/smtpin07/MantshX 4.0) with ESMTP id l3QMTOOl000336; 
	Thu, 26 Apr 2007 15:29:25 -0700 (PDT)
In-Reply-To: <4630D9A7.7010400@ericsson.com>
References: <07f401c7881b$1cd9aa70$05716b80@amer.cisco.com>
	<4630D9A7.7010400@ericsson.com>
Mime-Version: 1.0 (Apple Message framework v624)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <841f8e24eab33b98ef0cb84b5e716bb3@mac.com>
Content-Transfer-Encoding: 7bit
From: Sally Floyd <sallyfloyd@mac.com>
Subject: Re: [tsv-area] ICE pacing connectivity checks
Date: Thu, 26 Apr 2007 15:29:24 -0700
To: Magnus Westerlund <magnus.westerlund@ericsson.com>
X-Mailer: Apple Mail (2.624)
X-Brightmail-Tracker: AAAAAA==
X-Brightmail-scanned: yes
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 0fa76816851382eb71b0a882ccdc29ac
Cc: tsv-area@ietf.org, "'Jonathan Rosenberg \(jdrosen\)'" <jdrosen@cisco.com>,
	Dan Wing <dwing@cisco.com>
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

> Our general concern is that ICE has no congestion control and in worst 
> case can consume large amounts of bandwidth. The primary desire would 
> be to have ICE automatically adopt to the amount of available 
> bandwidth on the different paths. However that is almost impossible as 
> there is no information if the connectivity check is lost due to 
> reaching a NAT without an active binding, its filtering behavior or 
> congestion introduced loss. Making this into a very hard problem.
>
> I made a comparison with RTP when using RTCP as at least being able to 
> react to congestion on a medium timescale. Not TCP-friendly but at 
> least it prevents congestion collapse. Also RTP flows can range from 
> 1kb up to a couple of GB so there is no good comparison to do. Thus I 
> would like to steer away from comparing with RTP as being good or bad.

...
> I am especially worried about the increase in number of candidate 
> pairs as this is what makes the bandwidth run away as it currently is 
> defined. What I think is important for the proposals is that the 
> maintain a sensible bandwidth for the whole connectivity check phase 
> independent of number of candidate on any number of streams, 
> checkboards etc.

I would agree with this.  (I was assigned a few days ago to review
the ICE draft for congestion control issues, and so far I have read
the relevant parts once-over-lightly, so I am not coming in
with any particular expertise in this.)

A first step would be to have some text like Section 17.4.2 on the
STUN Amplification Attack.  First for the conformant case (no one
misbehaving), with a worst-case ICE scenario with many candidate
pairs and such, describing the amount of bandwidth that ICE could
easily end up using, and the time interval that this bandwidth could
be used for.

And then a mention of the security issues already discussed, or
other issues if there are any, of things going wrong that could
make the bandwidth used much worse.  Just for completeness.

Given this situation where loss cannot be used as an indication of
congestion, and where the "round-trip time" isn't really a meaningful
concept, controlling the delay Ta and the STUN retransmit interval
RTO, from Dan's email, certainly seem like the right knobs to play
with.  But instead of allowing ICE to use at most the bandwidth
RTP would have used for the first 500 ms, I would suggest allowing
ICE to use at most X Kbps, for some value of X.  50, 100, 200 Kbps?
I don't know.

When the appropriate guidelines for Ta and RTO have been figured out,
I would recommend saying that "Y" SHOULD be the value or equation used,
whatever "Y" turns out to be, but that ICE MUST NOT use a value
more aggressive than Y.  So that some hosts might know
that they want to go slower than the recommended approach, but
that they MAY NOT go faster.  (Or they MAY NOT go faster unless
some really explicit conditions have been fulfulled.)

As a comparison of how bad TCP can be, after the SYN/ACK exchange
TCP can send an initial window of roughly 4K bytes/RTT (~32 Kbits/RTT)
(from RFC 3390).  And because presumably end nodes won't be starting
ICE sessions as frequently as they start TCP connections, it would
seem reasonable to me to allow ICE to be more aggressive that TCP,
in terms of the bandwidth used in the short initial period roughly
equivalent to an RTT.  (And if there was some way to *ensure* that
end nodes didn't open many concurrent or subsequent ICE sessions,
this argument would be even more convincing...)

Another issue in ICE's favor is that it seems acceptable (to me)
to evaluate the bandwidth used by ICE in terms of bytes per second,
not in terms of packets per second.  E.g., RFC 3714, RFC 4828.
(This is based on the assumption, which is not necessarily always
correct, that the bottleneck is likely to be the bandwidth in
bits per second, not the CPU processing in packets per second.)

Take these as first thoughts...

- Sally
http://www.icir.org/floyd/





From tsv-area-bounces@ietf.org Thu Apr 26 19:26:27 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HhDM2-0000mh-Qd; Thu, 26 Apr 2007 19:26:26 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1HhDM1-0000mb-HR
	for tsv-area@ietf.org; Thu, 26 Apr 2007 19:26:25 -0400
Received: from smtpout.mac.com ([17.250.248.174])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1HhDM0-0007BV-4l
	for tsv-area@ietf.org; Thu, 26 Apr 2007 19:26:25 -0400
Received: from mac.com (smtpin07-en2 [10.13.10.152])
	by smtpout.mac.com (Xserve/smtpout04/MantshX 4.0) with ESMTP id
	l3QNQF3x016053; Thu, 26 Apr 2007 16:26:15 -0700 (PDT)
Received: from [192.150.186.254] (laptop254.icsi.berkeley.edu
	[192.150.186.254]) (authenticated bits=0)
	by mac.com (Xserve/smtpin07/MantshX 4.0) with ESMTP id l3QNQDO8002065; 
	Thu, 26 Apr 2007 16:26:14 -0700 (PDT)
In-Reply-To: <4630D9A7.7010400@ericsson.com>
References: <07f401c7881b$1cd9aa70$05716b80@amer.cisco.com>
	<4630D9A7.7010400@ericsson.com>
Mime-Version: 1.0 (Apple Message framework v624)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <05fd8cac557b85a703d23c6c3f1e18e4@mac.com>
Content-Transfer-Encoding: 7bit
From: Sally Floyd <sallyfloyd@mac.com>
Subject: Re: [tsv-area] ICE pacing connectivity checks
Date: Thu, 26 Apr 2007 16:26:13 -0700
To: Magnus Westerlund <magnus.westerlund@ericsson.com>
X-Mailer: Apple Mail (2.624)
X-Brightmail-Tracker: AAAAAA==
X-Brightmail-scanned: yes
X-Spam-Score: 0.0 (/)
X-Scan-Signature: b19722fc8d3865b147c75ae2495625f2
Cc: tsv-area@ietf.org, "'Jonathan Rosenberg \(jdrosen\)'" <jdrosen@cisco.com>,
	Dan Wing <dwing@cisco.com>
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Another thought is that congestion control for ICE could be made
somewhat more intelligent if ECN was used, and the router for the
congested link was ECN-Capable.  In this case, ICE packets could be
sent as ECN-capable, and if an end-node received an ICE packet
that was ECN-marked, that end-node would tell the other end, and
the other end would know that the path was congested.

(If the ECN-Capable ICE sending end-node heard nothing, on the other
hand, it could not use this as an indication that the path was
uncongested; even if the end-node has some way to know that all the
routers in its universe are ECN-capabke, an ECN-capable router can
still have a buffer overflow and can drop packets due to congestion.)

I am not suggested this as something to figure out and add to the
current draft - it could take a while to figure out if it would or
would not work, and it could take some time.  But if some of the
ICE people were ever motivated to take it on and add it in a
new internet-draft, it might be worth the trouble...

(I would note that using ECN for ICE would always be a bit problematic.
For TCP, if an ECN-marked data packet is dropped downstream, that
is taken by TCP as an indication of congestion, so when a packet
is ECN-marked, that indication of congestion makes its way back to
the sending node one way or anoother.  In contrast, with ICE a
packet could be ECN-marked, and then dropped downstream, and this
would not be taken by the end nodes as an indication of congestion.
 From the router's point of view, this might not be a strong
recommendation for using ECN for ICE packets.  But I think it could
be argued that the tradeoffs would make it worth it, all in all.)

- Sally
http://www.icir.org/floyd/





From tsv-area-bounces@ietf.org Fri Apr 27 02:45:26 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HhKCr-0007NB-CY; Fri, 27 Apr 2007 02:45:25 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1HhKCq-0007N3-Ba
	for tsv-area@ietf.org; Fri, 27 Apr 2007 02:45:24 -0400
Received: from mailgate.siemenscomms.co.uk ([195.171.110.225]
	helo=bemg01.siemenscomms.co.uk)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1HhKCp-0002bK-Qa
	for tsv-area@ietf.org; Fri, 27 Apr 2007 02:45:24 -0400
Received: from GBNTHT12009MSX.gb002.siemens.net ([137.223.219.235])
	by siemenscomms.co.uk (PMDF V6.0-24 #40642)
	with ESMTP id <0JH500J689FGS5@siemenscomms.co.uk> for tsv-area@ietf.org;
	Fri, 27 Apr 2007 07:45:16 +0100 (BST)
Date: Fri, 27 Apr 2007 07:45:21 +0100
From: "Elwell, John" <john.elwell@siemens.com>
In-reply-to: <08a401c7883a$c4da25a0$05716b80@amer.cisco.com>
To: Dan Wing <dwing@cisco.com>, tsv-area@ietf.org
Message-id: <0D5F89FAC29E2C41B98A6A762007F5D00AFA32@GBNTHT12009MSX.gb002.siemens.net>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft Exchange V6.5
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7BIT
Thread-Topic: ICE pacing connectivity
	checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
Thread-Index: AceINYgEVmZU4VcWQc6Irfi9fJyNjAABSSNgABXLg8A=
Content-class: urn:content-classes:message
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
References: <0D5F89FAC29E2C41B98A6A762007F5D00AFA0C@GBNTHT12009MSX.gb002.siemens.net>
	<08a401c7883a$c4da25a0$05716b80@amer.cisco.com>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: a8041eca2a724d631b098c15e9048ce9
Cc: 
Subject: [tsv-area] RE: ICE pacing connectivity
 checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Yes, true. In fact ICE is probably less severe than early media, because
tests by UASs on those candidate-pairs that fail will not result in
traffic on the final hop to the UAC. Once STUN requests start reaching
the UAC, we should be close to terminating ICE for that particular
branch.

John 

> -----Original Message-----
> From: Dan Wing [mailto:dwing@cisco.com] 
> Sent: 26 April 2007 20:41
> To: Elwell, John; tsv-area@ietf.org
> Subject: RE: ICE pacing connectivity 
> checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> 
> Yep; but it's problematic for RTP, too, if early media is sent by the
> branches.
> 
> -d
> 
> 
> > -----Original Message-----
> > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > Sent: Thursday, April 26, 2007 12:03 PM
> > To: Dan Wing; tsv-area@ietf.org
> > Subject: RE: ICE pacing connectivity 
> > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > 
> > True - it just shows that forking beyond modest levels (2-3 
> branches,
> > say) can be problematic.
> > 
> > John 
> > 
> > > -----Original Message-----
> > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > Sent: 26 April 2007 18:56
> > > To: Elwell, John; tsv-area@ietf.org
> > > Subject: RE: ICE pacing connectivity 
> > > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > 
> > > That could be done by calculating candidate_count to include 
> > > all candidates
> > > from all answerers - that would adjust Ta accordingly. 
> > > 
> > > That would help from the Offerer towards all of the forked 
> > > endpoints, but it
> > > wouldn't help in the opposite direction because the forked 
> > > endpoints don't
> > > know there are other forked endpoints.  It was that case (ICE 
> > > checks from
> > > answerers->offerer) I thought you were worried about.
> > > 
> > > -d
> > > 
> > > 
> > > > -----Original Message-----
> > > > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > > > Sent: Thursday, April 26, 2007 10:22 AM
> > > > To: Dan Wing; tsv-area@ietf.org
> > > > Subject: RE: ICE pacing connectivity 
> > > > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > 
> > > > Well, perhaps the pacing needs to be adjusted according to 
> > > > the number of
> > > > simultaneous ICE sessions on a given interface, if bandwidth 
> > > > is believed
> > > > to be limited.
> > > > 
> > > > John 
> > > > 
> > > > > -----Original Message-----
> > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > Sent: 26 April 2007 18:13
> > > > > To: Elwell, John; tsv-area@ietf.org
> > > > > Subject: RE: ICE pacing connectivity 
> > > > > checks(wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > > 
> > > > > Do you have a suggestion?
> > > > > 
> > > > > -d
> > > > >  
> > > > > 
> > > > > > -----Original Message-----
> > > > > > From: Elwell, John [mailto:john.elwell@siemens.com] 
> > > > > > Sent: Thursday, April 26, 2007 9:49 AM
> > > > > > To: Dan Wing; tsv-area@ietf.org
> > > > > > Subject: RE: ICE pacing connectivity checks 
> > > > > > (wasRE:[MMUSIC]WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > > > 
> > > > > > Dan,
> > > > > > 
> > > > > > All these calculations are a bit academic if they don't 
> > > take into
> > > > > > account SIP forking. Suppose a SIP INVITE request 
> > > > containing an ICE
> > > > > > offer gets forked to 10 UASs, and each agrees to use ICE. 
> > > > > > There is then
> > > > > > 10 times the amount of connectivity check traffic. The UAC 
> > > > > > might be on a
> > > > > > low bandwidth link, yet it has no control over the degree 
> > > > of forking
> > > > > > that might be encountered downstream.
> > > > > > 
> > > > > > John 
> > > > > > 
> > > > > > > -----Original Message-----
> > > > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > > > Sent: 26 April 2007 17:18
> > > > > > > To: mmusic@ietf.org
> > > > > > > Subject: RE: ICE pacing connectivity checks 
> (wasRE:[MMUSIC] 
> > > > > > > WGLC:draft-ietf-mmusic-ice-15.txt)
> > > > > > > 
> > > > > > > (Jonathan had also listed this as ICE issue #11.)
> > > > > > > 
> > > > > > > I have moved discussion of ICE pacing to the tsv-area 
> > > > mailing list
> > > > > > > <https://www1.ietf.org/mailman/listinfo/tsv-area>, as the 
> > > > > > > tsv-area list is
> > > > > > > more appropriate for discussing TCP friendliness.  
> > > > > > > 
> > > > > > > My post to tsv-area
> > > > > > > 
> > > > > > 
> > > > > 
> > > > 
> > > 
> > 
> <http://www1.ietf.org/mail-archive/web/tsv-area/current/msg00077.html>
> > > > > > > proposed new calculations for ICE's Ta and STUN's RTO.  
> > > > > > These changes
> > > > > > > prevent ICE from exceeding the bandwidth RTP would have 
> > > > consumed.
> > > > > > > 
> > > > > > > -d
> > > > > > > 
> > > > > > > 
> > > > > > > > -----Original Message-----
> > > > > > > > From: Dan Wing [mailto:dwing@cisco.com] 
> > > > > > > > Sent: Thursday, April 19, 2007 9:13 AM
> > > > > > > > To: 'Lars Eggert'; 'Magnus Westerlund'
> > > > > > > > Cc: mmusic@ietf.org
> > > > > > > > Subject: RE: ICE pacing connectivity checks 
> > (wasRE:[MMUSIC] 
> > > > > > > > WGLC: draft-ietf-mmusic-ice-15.txt)
> > > > > > > > 
> > > > > > > > I have created an Excel spreadsheet analyzing ICE's 
> > > behavior 
> > > > > > > > compared to
> > > > > > > > RTP's behavior.  I am tweaking this spreadsheet and 
> > > expect it 
> > > > > > > > to be useful
> > > > > > > > in helping us tune ICE's algorithms.  I want to 
> > make this a 
> > > > > > > > more useful
> > > > > > > > spreadsheet before sharing it; it doesn't quite do 
> > > everything 
> > > > > > > > I like yet.  I
> > > > > > > > expect it will remain awkward to use, as I'm not an 
> > > > > excel wizard.
> > > > > > > > 
> > > > > > > > In my initial analysis using this spreadsheet, 
> I do agree 
> > > > > > > > that ICE's default
> > > > > > > > 20ms pacing and STUN's default 100ms retransmit 
> > timers are 
> > > > > > > > too aggressive if
> > > > > > > > we assume severely bandwidth-constrained links. 
>  I expect 
> > > > > > > > that adjusting
> > > > > > > > these values based on the ptime, RTP packet size, STUN 
> > > > > > > > Request packet size,
> > > > > > > > and number of candidates will result in ICE 
> > > behaving as well 
> > > > > > > > as RTP (with
> > > > > > > > RTCP feedback) in all cases, and better than RTP 
> > (with RTCP 
> > > > > > > > feedback) in
> > > > > > > > some cases.  The drawback of such tuning, 
> > however, is that 
> > > > > > > > ICE will take
> > > > > > > > longer to complete if there are many candidates.
> > > > > > > > 
> > > > > > > > -d
> > > > > > > > 
> > > > > > > > _______________________________________________
> > > > > > > > mmusic mailing list
> > > > > > > > mmusic@ietf.org
> > > > > > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > > > > > 
> > > > > > > 
> > > > > > > _______________________________________________
> > > > > > > mmusic mailing list
> > > > > > > mmusic@ietf.org
> > > > > > > https://www1.ietf.org/mailman/listinfo/mmusic
> > > > > > > 
> > > > > 
> > > 
> 




From tsv-area-bounces@ietf.org Fri Apr 27 10:01:22 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HhR0j-0002gB-Pu; Fri, 27 Apr 2007 10:01:21 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1HhR0i-0002g1-As
	for tsv-area@ietf.org; Fri, 27 Apr 2007 10:01:20 -0400
Received: from mailgw4.ericsson.se ([193.180.251.62])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1HhR0h-0002ZY-Hx
	for tsv-area@ietf.org; Fri, 27 Apr 2007 10:01:20 -0400
Received: from mailgw4.ericsson.se (unknown [127.0.0.1])
	by mailgw4.ericsson.se (Symantec Mail Security) with ESMTP id
	0A21B216B9; Fri, 27 Apr 2007 16:01:19 +0200 (CEST)
X-AuditID: c1b4fb3e-b1a1dbb0000061ca-f1-4632022eb08e 
Received: from esealmw128.eemea.ericsson.se (unknown [153.88.254.121])
	by mailgw4.ericsson.se (Symantec Mail Security) with ESMTP id
	E549621011; Fri, 27 Apr 2007 16:01:18 +0200 (CEST)
Received: from esealmw126.eemea.ericsson.se ([153.88.254.170]) by
	esealmw128.eemea.ericsson.se with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 27 Apr 2007 16:01:18 +0200
Received: from [147.214.30.247] ([147.214.30.247]) by
	esealmw126.eemea.ericsson.se with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 27 Apr 2007 16:01:18 +0200
Message-ID: <4632022E.5080204@ericsson.com>
Date: Fri, 27 Apr 2007 16:01:18 +0200
From: Magnus Westerlund <magnus.westerlund@ericsson.com>
User-Agent: Thunderbird 1.5.0.10 (Windows/20070221)
MIME-Version: 1.0
To: Sally Floyd <sallyfloyd@mac.com>
Subject: Re: [tsv-area] ICE pacing connectivity checks
References: <07f401c7881b$1cd9aa70$05716b80@amer.cisco.com>
	<4630D9A7.7010400@ericsson.com>
	<05fd8cac557b85a703d23c6c3f1e18e4@mac.com>
In-Reply-To: <05fd8cac557b85a703d23c6c3f1e18e4@mac.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-OriginalArrivalTime: 27 Apr 2007 14:01:18.0379 (UTC)
	FILETIME=[87105FB0:01C788D4]
X-Brightmail-Tracker: AAAAAA==
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 0ddefe323dd869ab027dbfff7eff0465
Cc: tsv-area@ietf.org, "'Jonathan Rosenberg \(jdrosen\)'" <jdrosen@cisco.com>,
	Dan Wing <dwing@cisco.com>
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Sally Floyd skrev:
> Another thought is that congestion control for ICE could be made
> somewhat more intelligent if ECN was used, and the router for the
> congested link was ECN-Capable.  In this case, ICE packets could be
> sent as ECN-capable, and if an end-node received an ICE packet
> that was ECN-marked, that end-node would tell the other end, and
> the other end would know that the path was congested.
> 
> (If the ECN-Capable ICE sending end-node heard nothing, on the other
> hand, it could not use this as an indication that the path was
> uncongested; even if the end-node has some way to know that all the
> routers in its universe are ECN-capabke, an ECN-capable router can
> still have a buffer overflow and can drop packets due to congestion.)
> 
> I am not suggested this as something to figure out and add to the
> current draft - it could take a while to figure out if it would or
> would not work, and it could take some time.  But if some of the
> ICE people were ever motivated to take it on and add it in a
> new internet-draft, it might be worth the trouble...
> 
> (I would note that using ECN for ICE would always be a bit problematic.
> For TCP, if an ECN-marked data packet is dropped downstream, that
> is taken by TCP as an indication of congestion, so when a packet
> is ECN-marked, that indication of congestion makes its way back to
> the sending node one way or anoother.  In contrast, with ICE a
> packet could be ECN-marked, and then dropped downstream, and this
> would not be taken by the end nodes as an indication of congestion.
>  From the router's point of view, this might not be a strong
> recommendation for using ECN for ICE packets.  But I think it could
> be argued that the tradeoffs would make it worth it, all in all.)
> 

The problem with using ECN is that most of the sent connectivity checks 
will be dropped by the external side of NATs or hosts that don't have an 
open port or for which the arrived packet is garbage and discarded. Only 
on packets where STUN requests are answered are there any possibilities 
to use ECN to notice that there are some issues on the path.

Cheers

Magnus Westerlund

IETF Transport Area Director & TSVWG Chair
----------------------------------------------------------------------
Multimedia Technologies, Ericsson Research EAB/TVM/M
----------------------------------------------------------------------
Ericsson AB                | Phone +46 8 4048287
Torshamsgatan 23           | Fax   +46 8 7575550
S-164 80 Stockholm, Sweden | mailto: magnus.westerlund@ericsson.com
----------------------------------------------------------------------




From tsv-area-bounces@ietf.org Fri Apr 27 11:42:25 2007
Return-path: <tsv-area-bounces@ietf.org>
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1HhSaW-0002dw-4y; Fri, 27 Apr 2007 11:42:24 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1HhSaU-0002di-Kn
	for tsv-area@ietf.org; Fri, 27 Apr 2007 11:42:22 -0400
Received: from sj-iport-3-in.cisco.com ([171.71.176.72]
	helo=sj-iport-3.cisco.com) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1HhSaU-0006RX-AV
	for tsv-area@ietf.org; Fri, 27 Apr 2007 11:42:22 -0400
Received: from sj-dkim-1.cisco.com ([171.71.179.21])
	by sj-iport-3.cisco.com with ESMTP; 27 Apr 2007 08:42:22 -0700
X-IronPort-AV: i="4.14,462,1170662400"; 
	d="scan'208"; a="482000244:sNHT49913132"
Received: from sj-core-5.cisco.com (sj-core-5.cisco.com [171.71.177.238])
	by sj-dkim-1.cisco.com (8.12.11/8.12.11) with ESMTP id l3RFgLg7017115; 
	Fri, 27 Apr 2007 08:42:21 -0700
Received: from dwingwxp ([10.32.240.196])
	by sj-core-5.cisco.com (8.12.10/8.12.6) with ESMTP id l3RFgLEi010012;
	Fri, 27 Apr 2007 15:42:21 GMT
From: "Dan Wing" <dwing@cisco.com>
To: "'Magnus Westerlund'" <magnus.westerlund@ericsson.com>,
	"'Sally Floyd'" <sallyfloyd@mac.com>
Subject: RE: [tsv-area] ICE pacing connectivity checks
Date: Fri, 27 Apr 2007 08:42:20 -0700
Message-ID: <0ac501c788e2$a4dbe790$05716b80@amer.cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 11
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3028
Thread-Index: AceI1InRRKNPEPi1ReCTQUoqLwe2AQADM8ZQ
In-Reply-To: <4632022E.5080204@ericsson.com>
DKIM-Signature: v=0.5; a=rsa-sha256; q=dns/txt; l=2995; t=1177688541;
	x=1178552541; c=relaxed/simple; s=sjdkim1004;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=dwing@cisco.com;
	z=From:=20=22Dan=20Wing=22=20<dwing@cisco.com>
	|Subject:=20RE=3A=20[tsv-area]=20ICE=20pacing=20connectivity=20checks
	|Sender:=20; bh=vOFqySAW6CfvsgRhu6VFckE7LVGAsXsAlsQrsy9Yv6U=;
	b=Uai44+ka8wNaG1jmi+XIzCAbfmnWyYsF3+OQoA2cZseJ6aCVc+c1EgIXk9lOHKklFv7wpUCn
	SKYFRPgnMpFu8PwPR3w+LN3ZQBYDB6y9fjOqIp4lh78l2hC3vHUbvQSAL4CIzP/UXKJ93GQyca
	PDv1Yw64YmQd5eZc/9ePxFV3U=;
Authentication-Results: sj-dkim-1; header.From=dwing@cisco.com; dkim=pass (s
	ig from cisco.com/sjdkim1004 verified; ); 
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 0a7aa2e6e558383d84476dc338324fab
Cc: tsv-area@ietf.org, "'Jonathan Rosenberg \(jdrosen\)'" <jdrosen@cisco.com>
X-BeenThere: tsv-area@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: IETF Transport Area Mailing List <tsv-area.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www1.ietf.org/pipermail/tsv-area>
List-Post: <mailto:tsv-area@ietf.org>
List-Help: <mailto:tsv-area-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsv-area>,
	<mailto:tsv-area-request@ietf.org?subject=subscribe>
Errors-To: tsv-area-bounces@ietf.org

Magnus Westerlund wrote:
> Sally Floyd skrev:
> > Another thought is that congestion control for ICE could be made
> > somewhat more intelligent if ECN was used, and the router for the
> > congested link was ECN-Capable.  In this case, ICE packets could be
> > sent as ECN-capable, and if an end-node received an ICE packet
> > that was ECN-marked, that end-node would tell the other end, and
> > the other end would know that the path was congested.
> > 
> > (If the ECN-Capable ICE sending end-node heard nothing, on the other
> > hand, it could not use this as an indication that the path was
> > uncongested; even if the end-node has some way to know that all the
> > routers in its universe are ECN-capabke, an ECN-capable router can
> > still have a buffer overflow and can drop packets due to 
> > congestion.)
> > 
> > I am not suggested this as something to figure out and add to the
> > current draft - it could take a while to figure out if it would or
> > would not work, and it could take some time.  But if some of the
> > ICE people were ever motivated to take it on and add it in a
> > new internet-draft, it might be worth the trouble...
> > 
> > (I would note that using ECN for ICE would always be a bit 
> > problematic.
> > For TCP, if an ECN-marked data packet is dropped downstream, that
> > is taken by TCP as an indication of congestion, so when a packet
> > is ECN-marked, that indication of congestion makes its way back to
> > the sending node one way or anoother.  In contrast, with ICE a
> > packet could be ECN-marked, and then dropped downstream, and this
> > would not be taken by the end nodes as an indication of congestion.
> >  From the router's point of view, this might not be a strong
> > recommendation for using ECN for ICE packets.  But I think it could
> > be argued that the tradeoffs would make it worth it, all in all.)
> > 
> 
> The problem with using ECN is that most of the sent 
> connectivity checks will be dropped by the external side of 
> NATs or hosts that don't have an open port or for which 
> the arrived packet is garbage and discarded.  Only 
> on packets where STUN requests are answered are there any 
> possibilities to use ECN to notice that there are some 
> issues on the path.

Right.  But if there are multiple paths which /do/ work, 
ECN could help provide additional information about which of
those paths is better.  ICE does allow each end to do some
optimization of its decisions, and communicate its decision
to the other end via the USE-CANDIDATE attribute.  Right
now about all the controlling agent could learn from the
network is the round-trip time of a Request/Response.  ECN
would improve that.

I could imagine that if ECN support were added to ICE, we 
would define another STUN attribute that is conveyed in 
the STUN Response.  This new attribute would communicate
the ECN bits of the received STUN Request back to the
controlling (ICE) agent.

-d




