
From michawe@ifi.uio.no  Thu Dec  9 11:50:07 2010
Return-Path: <michawe@ifi.uio.no>
X-Original-To: ledbat@core3.amsl.com
Delivered-To: ledbat@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 754553A6BAA for <ledbat@core3.amsl.com>; Thu,  9 Dec 2010 11:50:07 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 6lNN8U57yW3y for <ledbat@core3.amsl.com>; Thu,  9 Dec 2010 11:50:06 -0800 (PST)
Received: from mail-out2.uio.no (mail-out2.uio.no [129.240.10.58]) by core3.amsl.com (Postfix) with ESMTP id B126C3A6B4F for <ledbat@ietf.org>; Thu,  9 Dec 2010 11:50:05 -0800 (PST)
Received: from mail-mx1.uio.no ([129.240.10.29]) by mail-out2.uio.no with esmtp (Exim 4.69) (envelope-from <michawe@ifi.uio.no>) id 1PQmWc-0007e6-MY for ledbat@ietf.org; Thu, 09 Dec 2010 20:51:34 +0100
Received: from cm-84.208.175.27.getinternet.no ([84.208.175.27] helo=[192.168.0.199]) by mail-mx1.uio.no with esmtpsa (TLSv1:AES128-SHA:128) user michawe (Exim 4.69) (envelope-from <michawe@ifi.uio.no>) id 1PQmWc-0001YR-6S for ledbat@ietf.org; Thu, 09 Dec 2010 20:51:34 +0100
Message-Id: <0EA8160F-FD4C-40F5-ABE4-2958659A6632@ifi.uio.no>
From: Michael Welzl <michawe@ifi.uio.no>
To: ledbat@ietf.org
In-Reply-To: <3D7D5DD1-7B95-431B-BC76-6F2AC20EEA08@telecom-bretagne.eu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed; delsp=yes
Content-Transfer-Encoding: quoted-printable
Mime-Version: 1.0 (Apple Message framework v936)
Date: Thu, 9 Dec 2010 20:51:33 +0100
References: <20101201003014.AB9443A6BE7@core3.amsl.com> <3D7D5DD1-7B95-431B-BC76-6F2AC20EEA08@telecom-bretagne.eu>
X-Mailer: Apple Mail (2.936)
X-UiO-Ratelimit-Test: rcpts/h 2 msgs/h 2 sum rcpts/h 5 sum msgs/h 3 total rcpts 4910 max rcpts/h 33 ratelimit 0
X-UiO-Spam-info: not spam, SpamAssassin (score=-5.0, required=5.0, autolearn=disabled, UIO_MAIL_IS_INTERNAL=-5, uiobl=NO, uiouri=NO)
X-UiO-Scanned: 2092DB2B0578266285923BD8D1CA47D6A48EA73B
X-UiO-SPAM-Test: remote_host: 84.208.175.27 spam_score: -49 maxlevel 80 minaction 2 bait 0 mail/h: 2 total 262 max/h 11 blacklist 0 greylist 0 ratelimit 0
Subject: Re: [ledbat] New Version Notification for draft-ietf-ledbat-survey-02
X-BeenThere: ledbat@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Mailing list of the LEDBAT WG <ledbat.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/ledbat>
List-Post: <mailto:ledbat@ietf.org>
List-Help: <mailto:ledbat-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Dec 2010 19:50:07 -0000

Hi all,

We still haven't heard back from any volunteers... this isn't a very =20
long or complicated document, it would be great to get our third and =20
final volunteer to quickly review it.

Cheers,
Michael


On Dec 1, 2010, at 1:40 AM, David Ros wrote:

> Dear all,
>
> As promised in Beijing, we've posted an update to the survey draft, =20=

> see announcement below. Comments are welcome.
>
> Also, the chairs requested (at least) one more review; please =20
> volunteer!
>
> Thanks in advance & best regards,
>
> David.
>
>
> D=E9but du message r=E9exp=E9di=E9 :
>
>> De : IETF I-D Submission Tool <idsubmission@ietf.org>
>> Date : 1 d=E9cembre 2010 01:30:14 HNEC
>> =C0 : david.ros@telecom-bretagne.eu
>> Cc : michawe@ifi.uio.no
>> Objet : New Version Notification for draft-ietf-ledbat-survey-02
>>
>>
>> A new version of I-D, draft-ietf-ledbat-survey-02.txt has been =20
>> successfully submitted by David Ros and posted to the IETF =20
>> repository.
>>
>> Filename:	 draft-ietf-ledbat-survey
>> Revision:	 02
>> Title:		 A Survey of Lower-than-Best-Effort Transport =
Protocols
>> Creation_date:	 2010-12-01
>> WG ID:		 ledbat
>> Number_of_pages: 17
>>
>> Abstract:
>> This document provides a survey of transport protocols which are
>> designed to have a smaller bandwidth and/or delay impact on standard
>> TCP than standard TCP itself when they share a bottleneck with it.
>> Such protocols could be used for delay-insensitive "background"
>> traffic, as they provide what is sometimes called a "less than" (or
>> "lower than") best-effort service.
>>
>>
>>
>> The IETF Secretariat.
>>
>>
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> David ROS
> http://www.rennes.enst-bretagne.fr/~dros/
>
> I love deadlines. I like the whooshing sound they make as they fly =20
> by. -- Douglas Adams
>
>
>
>


From wes@mti-systems.com  Mon Dec 13 18:50:03 2010
Return-Path: <wes@mti-systems.com>
X-Original-To: ledbat@core3.amsl.com
Delivered-To: ledbat@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E8CA828C143 for <ledbat@core3.amsl.com>; Mon, 13 Dec 2010 18:50:03 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Sd-8aP5zA342 for <ledbat@core3.amsl.com>; Mon, 13 Dec 2010 18:50:02 -0800 (PST)
Received: from omr11.networksolutionsemail.com (omr11.networksolutionsemail.com [205.178.146.61]) by core3.amsl.com (Postfix) with ESMTP id B24D228C13F for <ledbat@ietf.org>; Mon, 13 Dec 2010 18:50:02 -0800 (PST)
Received: from cm-omr14 (mail.networksolutionsemail.com [205.178.146.50]) by omr11.networksolutionsemail.com (8.13.6/8.13.6) with ESMTP id oBE2pfV2010925 for <ledbat@ietf.org>; Mon, 13 Dec 2010 21:51:41 -0500
Authentication-Results: cm-omr14 smtp.user=wes@mti-systems.com; auth=pass (PLAIN)
X-Authenticated-UID: wes@mti-systems.com
Received: from [174.130.113.191] ([174.130.113.191:22111] helo=[192.168.1.104]) by cm-omr14 (envelope-from <wes@mti-systems.com>) (ecelerity 2.2.2.41 r(31179/31189)) with ESMTPA id 53/A9-29026-DBBD60D4; Mon, 13 Dec 2010 21:51:41 -0500
Message-ID: <4D06DBC0.9080503@mti-systems.com>
Date: Mon, 13 Dec 2010 21:51:44 -0500
From: Wesley Eddy <wes@mti-systems.com>
Organization: MTI Systems
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.12) Gecko/20101027 Thunderbird/3.1.6
MIME-Version: 1.0
To: ledbat@ietf.org
References: <20101201003014.AB9443A6BE7@core3.amsl.com> <3D7D5DD1-7B95-431B-BC76-6F2AC20EEA08@telecom-bretagne.eu>
In-Reply-To: <3D7D5DD1-7B95-431B-BC76-6F2AC20EEA08@telecom-bretagne.eu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [ledbat] Fwd: New Version Notification for	draft-ietf-ledbat-survey-02
X-BeenThere: ledbat@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Mailing list of the LEDBAT WG <ledbat.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/ledbat>
List-Post: <mailto:ledbat@ietf.org>
List-Help: <mailto:ledbat-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Dec 2010 02:50:04 -0000

On 11/30/2010 7:40 PM, David Ros wrote:
> Dear all,
>
> As promised in Beijing, we've posted an update to the survey draft, see announcement below. Comments are welcome.
>
> Also, the chairs requested (at least) one more review; please volunteer!
>


Here are some review comments I had.  Aside from some general ones, 
they're all editorial or possibly minor technical clarifications that 
could be made.

General / Overall

- We should have some material early on to set context; things to 
mention include:
   - intended audience (familiarity with standard TCP, RFC2581/5681, is 
assumed)
   - this was a product of the LEDBAT WG for comparison/contrast to the 
chosen approach (this needs to be mentioned, because remember that when 
the RFC is published, a casual reader won't know it came from LEDBAT 
unless you say so)
   - most techniques discussed here are tested in limited simulations or 
experimental testbeds, but LEDBAT's algorithm is already under 
widespread deployment (i.e. we need to say that we're acknowledging 
other techniques, but not advocating them here; this does not interfere 
with advancement of the LEDBAT algorithm)
   - this is not exhaustive; that wouldn't be possible or useful, but 
the authors/editors have selected key, well-known, or otherwise 
interesting techniques for inclusion at their discretion
   - was applicability to big-I Internet deployment scenarios a 
consideration?  (I think yes, but don't actually know in some of the cases)

- Does it need to be more clearly mentioned which classes of approaches 
were mechanisms incorporated in TCP, TCP-alternatives (e.g. for DCCP or 
new transports), or other designs entirely.  For instance, do they have 
subtle (or obvious) differences in the services provided to applications 
(reliability, smoothness, etc.)?  Section 4's first paragraph does a 
good job of explaining this, but other sections had me wondering 
slightly.  It may be a personal problem on my part, so it would be good 
to check with others :).

- Are there fairly common assumptions in a lot of the existing academic 
analysis cited here which influence results?  For instance, there are 
often assumptions about flow length, RTT distributions, or other 
elements that may be relevant to mention here? (e.g. all flows are FTP 
of 10 GB files, etc.)

- Should there be a (very short) conclusions section that says where 
LEDBAT fits in and the main pros/cons versus the other classes of system 
discussed?


Section 1

- (editorial) It might be more clear to replace "end-to-end mechanisms" 
with "end-host mechanisms", since end-to-end is relative.

- (editorial) The LBE acronymn is defined twice in the first paragraph; 
the second expansion at the end of the paragraph can probably either 
just use "LBE" or put the expansion in quotation marks.


Section 2

- (editorial) The second sentence of the first paragraph could be 
clarified and shortened to just "Without ECN support, standard TCP will 
normally increase its congestion window (and effective sending rate) 
until a queue overflows, causing one or more packets to be dropped and 
the effective rate to be reduced."  I didn't like the "for some reason" 
in the current sentence because there are clear reasons like application 
limited, advertised receive window limited, kernel knob limited, etc, 
but they all seem to be beside the point rather than contributing to it.

- (editorial) At one point the text capitalizes "Congestion Avoidance", 
and in others it doesn't.  I think lower-case is best.

- (editorial) "threshold beta" might be more clear as "threshold (called 
"beta")" (same comment applies for alpha)

- (minor) "actual (measured) throughput" might be more clear as "actual 
throughput measured by recent acknowledgements"

- (minor) is the increase at the end of the 3rd paragraph every RTT or 
BaseRTT (I don't recall myself)

- (minor) "maintaining the window size" should be "maintaining the 
current congestion window"

- (editorial) "it the standard" should be "it in the standard"

- (minor) What does "remain transparent to standard TCP" really mean? 
There must be a better way to word this.


Section 2.1

- (minor) under the "RTT measurement issues", is it worth mentioning the 
case of fast or short-distance links where the delay is all in the stack 
and not on the wire?


Section 2.2

- (minor) is it worth mentioning LEDBAT's sensitivity to changes in the 
route also, when we mention Vegas?  Is it purposeful that we never 
really compare/contrast to LEDBAT itself in this document?


Section 4

- (editorial) In the last paragraph, is a ":" really the best way to 
punctuate this?


Section 5

- (minor) For the network-assisted approaches, it wasn't entirely clear 
to me if these were new transport protocols or new mechanisms for 
existing transports; if that could be clarified without the reader 
needing to check the references, I think that would be good.

- (minor) Should there be a summary/conclusion sentence here about the 
current viability of such approaches?



From Rolf.Winter@neclab.eu  Thu Dec 16 00:53:45 2010
Return-Path: <Rolf.Winter@neclab.eu>
X-Original-To: ledbat@core3.amsl.com
Delivered-To: ledbat@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0057F3A7091 for <ledbat@core3.amsl.com>; Thu, 16 Dec 2010 00:53:45 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.405
X-Spam-Level: 
X-Spam-Status: No, score=-102.405 tagged_above=-999 required=5 tests=[AWL=0.194, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id e60clJ3NUTto for <ledbat@core3.amsl.com>; Thu, 16 Dec 2010 00:53:43 -0800 (PST)
Received: from smtp0.neclab.eu (smtp0.neclab.eu [195.37.70.41]) by core3.amsl.com (Postfix) with ESMTP id 56CF73A6E71 for <ledbat@ietf.org>; Thu, 16 Dec 2010 00:53:43 -0800 (PST)
Received: from localhost (localhost.localdomain [127.0.0.1]) by smtp0.neclab.eu (Postfix) with ESMTP id 8FEA22C0003F8 for <ledbat@ietf.org>; Thu, 16 Dec 2010 09:56:32 +0100 (CET)
X-Virus-Scanned: Amavisd on Debian GNU/Linux (atlas2.office.hd)
Received: from smtp0.neclab.eu ([127.0.0.1]) by localhost (atlas2.office.hd [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id PeqYKDNCXBh6 for <ledbat@ietf.org>; Thu, 16 Dec 2010 09:56:32 +0100 (CET)
Received: from ENCELADUS.office.hd (ENCELADUS.office.hd [192.168.24.52]) by smtp0.neclab.eu (Postfix) with ESMTP id 71D282C0002E0 for <ledbat@ietf.org>; Thu, 16 Dec 2010 09:56:27 +0100 (CET)
Received: from PALLENE.office.hd ([169.254.1.113]) by ENCELADUS.office.hd ([192.168.24.52]) with mapi id 14.01.0255.000; Thu, 16 Dec 2010 09:55:22 +0100
From: Rolf Winter <Rolf.Winter@neclab.eu>
To: "ledbat@ietf.org" <ledbat@ietf.org>
Thread-Topic: [ledbat] New Version Notification for draft-ietf-ledbat-survey-02
Thread-Index: AQHLnP74DomFK8/3XU6+Fjjgq5ub7A==
Date: Thu, 16 Dec 2010 08:55:21 +0000
Message-ID: <791AD3077F94194BB2BDD13565B6295D86B527@PALLENE.office.hd>
References: <20101201003014.AB9443A6BE7@core3.amsl.com> <3D7D5DD1-7B95-431B-BC76-6F2AC20EEA08@telecom-bretagne.eu> <0EA8160F-FD4C-40F5-ABE4-2958659A6632@ifi.uio.no>
In-Reply-To: <0EA8160F-FD4C-40F5-ABE4-2958659A6632@ifi.uio.no>
Accept-Language: en-US, de-DE
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [10.1.6.28]
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [ledbat] New Version Notification for draft-ietf-ledbat-survey-02
X-BeenThere: ledbat@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Mailing list of the LEDBAT WG <ledbat.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/ledbat>
List-Post: <mailto:ledbat@ietf.org>
List-Help: <mailto:ledbat-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Dec 2010 08:53:45 -0000

Hi,

I would also be good if the previous reviewers could assess whether their c=
omments have been addressed or not.

Best,

	Rolf


NEC Europe Limited | Registered Office: NEC House, 1 Victoria Road, London =
W3 6BL | Registered in England 2832014=20


> -----Original Message-----
> From: ledbat-bounces@ietf.org [mailto:ledbat-bounces@ietf.org] On
> Behalf Of Michael Welzl
> Sent: Donnerstag, 9. Dezember 2010 20:52
> To: ledbat@ietf.org
> Subject: Re: [ledbat] New Version Notification for draft-ietf-ledbat-
> survey-02
>=20
> Hi all,
>=20
> We still haven't heard back from any volunteers... this isn't a very
> long or complicated document, it would be great to get our third and
> final volunteer to quickly review it.
>=20
> Cheers,
> Michael
>=20
>=20
> On Dec 1, 2010, at 1:40 AM, David Ros wrote:
>=20
> > Dear all,
> >
> > As promised in Beijing, we've posted an update to the survey draft,
> > see announcement below. Comments are welcome.
> >
> > Also, the chairs requested (at least) one more review; please
> > volunteer!
> >
> > Thanks in advance & best regards,
> >
> > David.
> >
> >
> > D=E9but du message r=E9exp=E9di=E9 :
> >
> >> De : IETF I-D Submission Tool <idsubmission@ietf.org>
> >> Date : 1 d=E9cembre 2010 01:30:14 HNEC
> >> =C0 : david.ros@telecom-bretagne.eu
> >> Cc : michawe@ifi.uio.no
> >> Objet : New Version Notification for draft-ietf-ledbat-survey-02
> >>
> >>
> >> A new version of I-D, draft-ietf-ledbat-survey-02.txt has been
> >> successfully submitted by David Ros and posted to the IETF
> >> repository.
> >>
> >> Filename:	 draft-ietf-ledbat-survey
> >> Revision:	 02
> >> Title:		 A Survey of Lower-than-Best-Effort Transport
> Protocols
> >> Creation_date:	 2010-12-01
> >> WG ID:		 ledbat
> >> Number_of_pages: 17
> >>
> >> Abstract:
> >> This document provides a survey of transport protocols which are
> >> designed to have a smaller bandwidth and/or delay impact on standard
> >> TCP than standard TCP itself when they share a bottleneck with it.
> >> Such protocols could be used for delay-insensitive "background"
> >> traffic, as they provide what is sometimes called a "less than" (or
> >> "lower than") best-effort service.
> >>
> >>
> >>
> >> The IETF Secretariat.
> >>
> >>
> >
> > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> > David ROS
> > http://www.rennes.enst-bretagne.fr/~dros/
> >
> > I love deadlines. I like the whooshing sound they make as they fly
> > by. -- Douglas Adams
> >
> >
> >
> >
>=20
> _______________________________________________
> ledbat mailing list
> ledbat@ietf.org
> https://www.ietf.org/mailman/listinfo/ledbat

From michawe@ifi.uio.no  Fri Dec 17 13:01:28 2010
Return-Path: <michawe@ifi.uio.no>
X-Original-To: ledbat@core3.amsl.com
Delivered-To: ledbat@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8A51A3A6C24 for <ledbat@core3.amsl.com>; Fri, 17 Dec 2010 13:01:28 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UcIba9L5dDKt for <ledbat@core3.amsl.com>; Fri, 17 Dec 2010 13:01:27 -0800 (PST)
Received: from mail-out2.uio.no (mail-out2.uio.no [129.240.10.58]) by core3.amsl.com (Postfix) with ESMTP id C82D73A695F for <ledbat@ietf.org>; Fri, 17 Dec 2010 13:01:26 -0800 (PST)
Received: from mail-mx5.uio.no ([129.240.10.46]) by mail-out2.uio.no with esmtp (Exim 4.72) (envelope-from <michawe@ifi.uio.no>) id 1PThSL-0006KV-Ko for ledbat@ietf.org; Fri, 17 Dec 2010 22:03:13 +0100
Received: from cm-84.208.175.27.getinternet.no ([84.208.175.27] helo=[192.168.0.199]) by mail-mx5.uio.no with esmtpsa (TLSv1:AES128-SHA:128) user michawe (Exim 4.72) (envelope-from <michawe@ifi.uio.no>) id 1PThSK-0001O5-Qb for ledbat@ietf.org; Fri, 17 Dec 2010 22:03:13 +0100
Message-Id: <9FF43AB0-FA71-4536-ADE7-78CA2CEEEF12@ifi.uio.no>
From: Michael Welzl <michawe@ifi.uio.no>
To: ledbat@ietf.org
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v936)
Date: Fri, 17 Dec 2010 22:02:51 +0100
X-Mailer: Apple Mail (2.936)
X-UiO-Ratelimit-Test: rcpts/h 1 msgs/h 1 sum rcpts/h 2 sum msgs/h 2 total rcpts 5109 max rcpts/h 33 ratelimit 0
X-UiO-Spam-info: not spam, SpamAssassin (score=-5.0, required=5.0, autolearn=disabled, UIO_MAIL_IS_INTERNAL=-5, uiobl=NO, uiouri=NO)
X-UiO-Scanned: 6521FA1F5538D5A905AE1F335E43C4ABA0173B88
X-UiO-SPAM-Test: remote_host: 84.208.175.27 spam_score: -49 maxlevel 80 minaction 2 bait 0 mail/h: 1 total 338 max/h 11 blacklist 0 greylist 0 ratelimit 0
Subject: [ledbat] draft-ietf-ledbat-survey-03.txt
X-BeenThere: ledbat@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Mailing list of the LEDBAT WG <ledbat.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/ledbat>
List-Post: <mailto:ledbat@ietf.org>
List-Help: <mailto:ledbat-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 17 Dec 2010 21:01:28 -0000

Hi all,

We have now submitted draft-ietf-ledbat-survey-03.txt, in which we  
tried to address Wes' comments:
http://www.ietf.org/id/draft-ietf-ledbat-survey-03.txt
See below for his list of comments, with our answers in line.

We would like to raise a question to the chairs regarding one matter.  
Wes, and Mirja before him, have suggested to include a concluding  
section where we discuss pro's and con's of LEDBAT in comparison to  
the related work here. It seems clear to us that such a section would  
not make any sense without citing the LEDBAT document, but such a  
citation would put this document on hold until LEDBAT is published. It  
seems to us that this just goes to show that this is in fact the wrong  
sequence, and the wrong place for this discussion: in our opinion, the  
LEDBAT document should contain a short paragraph where it is explained  
how it relates to the state of the art, as documented in our survey.  
This would seem like a much nicer structure to us (e.g. in academia,  
first you look at the state of the art, then you propose your  
mechanism in relation to it).

We would appreciate if the chairs could let us know what they think  
about this matter, and if the reviewers could let us know if they are  
happy with how we addressed all their other comments.

Thanks!!

Michael and David


----

ANSWERS TO WES' COMMENTS:

Here are some review comments I had.  Aside from some general ones,  
they're all editorial or possibly minor technical clarifications that  
could be made.

General / Overall

- We should have some material early on to set context; things to  
mention include:
- intended audience (familiarity with standard TCP, RFC2581/5681, is  
assumed)
- this was a product of the LEDBAT WG for comparison/contrast to the  
chosen approach (this needs to be mentioned, because remember that  
when the RFC is published, a casual reader won't know it came from  
LEDBAT unless you say so)
- most techniques discussed here are tested in limited simulations or  
experimental testbeds, but LEDBAT's algorithm is already under  
widespread deployment (i.e. we need to say that we're acknowledging  
other techniques, but not advocating them here; this does not  
interfere with advancement of the LEDBAT algorithm)
- this is not exhaustive; that wouldn't be possible or useful, but the  
authors/editors have selected key, well-known, or otherwise  
interesting techniques for inclusion at their discretion
- was applicability to big-I Internet deployment scenarios a  
consideration?  (I think yes, but don't actually know in some of the  
cases)

Done. Regarding the "big-I" question, it is clear that almost all of  
the presented mechanisms were designed for that; if there are one or  
two exceptions, then stressing that here just for these mechanisms  
would not contribute much to readability, we believe - the general  
picture, as presented and introduced here, is about the Internet,  
throughout the document.


- Does it need to be more clearly mentioned which classes of  
approaches were mechanisms incorporated in TCP, TCP-alternatives (e.g.  
for DCCP or new transports), or other designs entirely.  For instance,  
do they have subtle (or obvious) differences in the services provided  
to applications (reliability, smoothness, etc.)?  Section 4's first  
paragraph does a good job of explaining this, but other sections had  
me wondering slightly.  It may be a personal problem on my part, so it  
would be good to check with others :).

We tried to state whether mechanisms are incorporated in TCP or  
entirely new protocols upon their appearance in our document (as you  
mention below, we have failed to make this clear enough in section 5,  
but this has now been fixed). Generally, the new protocol / TCP-change  
classification is orthogonal to ours, i.e. both variants are scattered  
throughout the document. Regarding the service provided to  
applications, there is no simple answer here. LBE is, as we define, "a  
service which results in smaller bandwidth and/or delay impact on  
standard TCP than standard TCP itself, when sharing a bottleneck with  
it". This is fairly network-centric; a more user-centric perspective  
is not commonly taken in this literature, and no user-centric  
comparison exists to the best of our knowledge. Yet, the user  
experience will be different, depending on the dynamics of the  
underlying mechanism -- much like you could get a different user  
experience from CTCP, HTCP and CUBIC.


- Are there fairly common assumptions in a lot of the existing  
academic analysis cited here which influence results?  For instance,  
there are often assumptions about flow length, RTT distributions, or  
other elements that may be relevant to mention here? (e.g. all flows  
are FTP of 10 GB files, etc.)

We think not.


- Should there be a (very short) conclusions section that says where  
LEDBAT fits in and the main pros/cons versus the other classes of  
system discussed?

Maybe; this was also requested by Mirja. See our question to the  
chairs regarding this matter.


Section 1

- (editorial) It might be more clear to replace "end-to-end  
mechanisms" with "end-host mechanisms", since end-to-end is relative.

Done.


- (editorial) The LBE acronymn is defined twice in the first  
paragraph; the second expansion at the end of the paragraph can  
probably either just use "LBE" or put the expansion in quotation marks.

Done (just used "LBE").


Section 2

- (editorial) The second sentence of the first paragraph could be  
clarified and shortened to just "Without ECN support, standard TCP  
will normally increase its congestion window (and effective sending  
rate) until a queue overflows, causing one or more packets to be  
dropped and the effective rate to be reduced."  I didn't like the "for  
some reason" in the current sentence because there are clear reasons  
like application limited, advertised receive window limited, kernel  
knob limited, etc, but they all seem to be beside the point rather  
than contributing to it.

Done.


- (editorial) At one point the text capitalizes "Congestion  
Avoidance", and in others it doesn't.  I think lower-case is best.

Done.


- (editorial) "threshold beta" might be more clear as "threshold  
(called "beta")" (same comment applies for alpha)

Done (but without brackets because they don't go well with another  
bracket in the "alpha" part).


- (minor) "actual (measured) throughput" might be more clear as  
"actual throughput measured by recent acknowledgements"

Done.


- (minor) is the increase at the end of the 3rd paragraph every RTT or  
BaseRTT (I don't recall myself)

We checked: the original [Bra94] paper only talks about increasing  
linearly and decreasing linearly. I have therefore removed the "by one  
segment per RTT" bit.


- (minor) "maintaining the window size" should be "maintaining the  
current congestion window"

Done.


- (editorial) "it the standard" should be "it in the standard"

Done.


- (minor) What does "remain transparent to standard TCP" really mean?  
There must be a better way to word this.

Here, it means "avoid interfering with standard TCP traffic", and  
that's what we've now written there.


Section 2.1

- (minor) under the "RTT measurement issues", is it worth mentioning  
the case of fast or short-distance links where the delay is all in the  
stack and not on the wire?

Absolutely! I am surprised that we missed this, thanks for catching  
it! I put that there, now.


Section 2.2

- (minor) is it worth mentioning LEDBAT's sensitivity to changes in  
the route also, when we mention Vegas?  Is it purposeful that we never  
really compare/contrast to LEDBAT itself in this document?

Yes, this might be worth mentioning regarding LEDBAT, but rather than  
scattering LEDBAT comments throughout the document, we have opted for  
putting all such LEDBAT (except for the introductory statement that  
you suggested) in a "Conclusion and relationship to LEDBAT" section if  
we indeed end up writing one (see our question to the chairs).


Section 4

- (editorial) In the last paragraph, is a ":" really the best way to  
punctuate this?

Done.


Section 5

- (minor) For the network-assisted approaches, it wasn't entirely  
clear to me if these were new transport protocols or new mechanisms  
for existing transports; if that could be clarified without the reader  
needing to check the references, I think that would be good.

Done (clarified for every reference where it occurs).


- (minor) Should there be a summary/conclusion sentence here about the  
current viability of such approaches?

We think not, as these proposals are so heterogeneous: e.g., one  
involves changing TCP's AIMD parameters, the other involves no changes  
to TCP at all but essentially putting a multipath-LBE-PEP in the  
network. Anyway, as you said, we're acknowledging other techniques,  
but not advocating them here; we would prefer to present them as  
neutrally as possible.

From Rolf.Winter@neclab.eu  Tue Dec 21 07:01:50 2010
Return-Path: <Rolf.Winter@neclab.eu>
X-Original-To: ledbat@core3.amsl.com
Delivered-To: ledbat@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2FD823A6AFC for <ledbat@core3.amsl.com>; Tue, 21 Dec 2010 07:01:50 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.409
X-Spam-Level: 
X-Spam-Status: No, score=-102.409 tagged_above=-999 required=5 tests=[AWL=0.190, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 10z5IQcHHIQm for <ledbat@core3.amsl.com>; Tue, 21 Dec 2010 07:01:48 -0800 (PST)
Received: from smtp0.neclab.eu (smtp0.neclab.eu [195.37.70.41]) by core3.amsl.com (Postfix) with ESMTP id 8886A3A6AF8 for <ledbat@ietf.org>; Tue, 21 Dec 2010 07:01:48 -0800 (PST)
Received: from localhost (localhost.localdomain [127.0.0.1]) by smtp0.neclab.eu (Postfix) with ESMTP id 9817F2C000406; Tue, 21 Dec 2010 16:05:12 +0100 (CET)
X-Virus-Scanned: Amavisd on Debian GNU/Linux (atlas2.office.hd)
Received: from smtp0.neclab.eu ([127.0.0.1]) by localhost (atlas2.office.hd [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yg8+BEvqM6+v; Tue, 21 Dec 2010 16:05:12 +0100 (CET)
Received: from METHONE.office.hd (Methone.office.hd [192.168.24.54]) by smtp0.neclab.eu (Postfix) with ESMTP id 7631F2C000403; Tue, 21 Dec 2010 16:05:02 +0100 (CET)
Received: from PALLENE.office.hd ([169.254.1.113]) by METHONE.office.hd ([192.168.24.54]) with mapi id 14.01.0255.000; Tue, 21 Dec 2010 16:03:34 +0100
From: Rolf Winter <Rolf.Winter@neclab.eu>
To: Michael Welzl <michawe@ifi.uio.no>, "ledbat@ietf.org" <ledbat@ietf.org>
Thread-Topic: [ledbat] draft-ietf-ledbat-survey-03.txt
Thread-Index: AQHLni3c3MCU2z+4kUKd7mWHHOAW8JOrAybg
Date: Tue, 21 Dec 2010 15:03:33 +0000
Message-ID: <791AD3077F94194BB2BDD13565B6295D86D2C5@PALLENE.office.hd>
References: <9FF43AB0-FA71-4536-ADE7-78CA2CEEEF12@ifi.uio.no>
In-Reply-To: <9FF43AB0-FA71-4536-ADE7-78CA2CEEEF12@ifi.uio.no>
Accept-Language: en-US, de-DE
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [10.1.6.28]
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [ledbat] draft-ietf-ledbat-survey-03.txt
X-BeenThere: ledbat@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Mailing list of the LEDBAT WG <ledbat.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/ledbat>
List-Post: <mailto:ledbat@ietf.org>
List-Help: <mailto:ledbat-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 21 Dec 2010 15:01:50 -0000

Micheal,

you say that the order is wrong, however, when you look at our original cha=
rter, the last item to finish actually is the LEDBAT survey. So the request=
 that both Wes and Mirja have made make complete sense to me. You can alrea=
dy write that section now as I do not see a reason why you would need to wa=
it.=20

Best,

	Rolf


NEC Europe Limited | Registered Office: NEC House, 1 Victoria Road, London =
W3 6BL | Registered in England 2832014=20


> -----Original Message-----
> From: ledbat-bounces@ietf.org [mailto:ledbat-bounces@ietf.org] On
> Behalf Of Michael Welzl
> Sent: Freitag, 17. Dezember 2010 22:03
> To: ledbat@ietf.org
> Subject: [ledbat] draft-ietf-ledbat-survey-03.txt
>=20
> Hi all,
>=20
> We have now submitted draft-ietf-ledbat-survey-03.txt, in which we
> tried to address Wes' comments:
> http://www.ietf.org/id/draft-ietf-ledbat-survey-03.txt
> See below for his list of comments, with our answers in line.
>=20
> We would like to raise a question to the chairs regarding one matter.
> Wes, and Mirja before him, have suggested to include a concluding
> section where we discuss pro's and con's of LEDBAT in comparison to
> the related work here. It seems clear to us that such a section would
> not make any sense without citing the LEDBAT document, but such a
> citation would put this document on hold until LEDBAT is published. It
> seems to us that this just goes to show that this is in fact the wrong
> sequence, and the wrong place for this discussion: in our opinion, the
> LEDBAT document should contain a short paragraph where it is explained
> how it relates to the state of the art, as documented in our survey.
> This would seem like a much nicer structure to us (e.g. in academia,
> first you look at the state of the art, then you propose your
> mechanism in relation to it).
>=20
> We would appreciate if the chairs could let us know what they think
> about this matter, and if the reviewers could let us know if they are
> happy with how we addressed all their other comments.
>=20
> Thanks!!
>=20
> Michael and David
>=20
>=20
> ----
>=20
> ANSWERS TO WES' COMMENTS:
>=20
> Here are some review comments I had.  Aside from some general ones,
> they're all editorial or possibly minor technical clarifications that
> could be made.
>=20
> General / Overall
>=20
> - We should have some material early on to set context; things to
> mention include:
> - intended audience (familiarity with standard TCP, RFC2581/5681, is
> assumed)
> - this was a product of the LEDBAT WG for comparison/contrast to the
> chosen approach (this needs to be mentioned, because remember that
> when the RFC is published, a casual reader won't know it came from
> LEDBAT unless you say so)
> - most techniques discussed here are tested in limited simulations or
> experimental testbeds, but LEDBAT's algorithm is already under
> widespread deployment (i.e. we need to say that we're acknowledging
> other techniques, but not advocating them here; this does not
> interfere with advancement of the LEDBAT algorithm)
> - this is not exhaustive; that wouldn't be possible or useful, but the
> authors/editors have selected key, well-known, or otherwise
> interesting techniques for inclusion at their discretion
> - was applicability to big-I Internet deployment scenarios a
> consideration?  (I think yes, but don't actually know in some of the
> cases)
>=20
> Done. Regarding the "big-I" question, it is clear that almost all of
> the presented mechanisms were designed for that; if there are one or
> two exceptions, then stressing that here just for these mechanisms
> would not contribute much to readability, we believe - the general
> picture, as presented and introduced here, is about the Internet,
> throughout the document.
>=20
>=20
> - Does it need to be more clearly mentioned which classes of
> approaches were mechanisms incorporated in TCP, TCP-alternatives (e.g.
> for DCCP or new transports), or other designs entirely.  For instance,
> do they have subtle (or obvious) differences in the services provided
> to applications (reliability, smoothness, etc.)?  Section 4's first
> paragraph does a good job of explaining this, but other sections had
> me wondering slightly.  It may be a personal problem on my part, so it
> would be good to check with others :).
>=20
> We tried to state whether mechanisms are incorporated in TCP or
> entirely new protocols upon their appearance in our document (as you
> mention below, we have failed to make this clear enough in section 5,
> but this has now been fixed). Generally, the new protocol / TCP-change
> classification is orthogonal to ours, i.e. both variants are scattered
> throughout the document. Regarding the service provided to
> applications, there is no simple answer here. LBE is, as we define, "a
> service which results in smaller bandwidth and/or delay impact on
> standard TCP than standard TCP itself, when sharing a bottleneck with
> it". This is fairly network-centric; a more user-centric perspective
> is not commonly taken in this literature, and no user-centric
> comparison exists to the best of our knowledge. Yet, the user
> experience will be different, depending on the dynamics of the
> underlying mechanism -- much like you could get a different user
> experience from CTCP, HTCP and CUBIC.
>=20
>=20
> - Are there fairly common assumptions in a lot of the existing
> academic analysis cited here which influence results?  For instance,
> there are often assumptions about flow length, RTT distributions, or
> other elements that may be relevant to mention here? (e.g. all flows
> are FTP of 10 GB files, etc.)
>=20
> We think not.
>=20
>=20
> - Should there be a (very short) conclusions section that says where
> LEDBAT fits in and the main pros/cons versus the other classes of
> system discussed?
>=20
> Maybe; this was also requested by Mirja. See our question to the
> chairs regarding this matter.
>=20
>=20
> Section 1
>=20
> - (editorial) It might be more clear to replace "end-to-end
> mechanisms" with "end-host mechanisms", since end-to-end is relative.
>=20
> Done.
>=20
>=20
> - (editorial) The LBE acronymn is defined twice in the first
> paragraph; the second expansion at the end of the paragraph can
> probably either just use "LBE" or put the expansion in quotation marks.
>=20
> Done (just used "LBE").
>=20
>=20
> Section 2
>=20
> - (editorial) The second sentence of the first paragraph could be
> clarified and shortened to just "Without ECN support, standard TCP
> will normally increase its congestion window (and effective sending
> rate) until a queue overflows, causing one or more packets to be
> dropped and the effective rate to be reduced."  I didn't like the "for
> some reason" in the current sentence because there are clear reasons
> like application limited, advertised receive window limited, kernel
> knob limited, etc, but they all seem to be beside the point rather
> than contributing to it.
>=20
> Done.
>=20
>=20
> - (editorial) At one point the text capitalizes "Congestion
> Avoidance", and in others it doesn't.  I think lower-case is best.
>=20
> Done.
>=20
>=20
> - (editorial) "threshold beta" might be more clear as "threshold
> (called "beta")" (same comment applies for alpha)
>=20
> Done (but without brackets because they don't go well with another
> bracket in the "alpha" part).
>=20
>=20
> - (minor) "actual (measured) throughput" might be more clear as
> "actual throughput measured by recent acknowledgements"
>=20
> Done.
>=20
>=20
> - (minor) is the increase at the end of the 3rd paragraph every RTT or
> BaseRTT (I don't recall myself)
>=20
> We checked: the original [Bra94] paper only talks about increasing
> linearly and decreasing linearly. I have therefore removed the "by one
> segment per RTT" bit.
>=20
>=20
> - (minor) "maintaining the window size" should be "maintaining the
> current congestion window"
>=20
> Done.
>=20
>=20
> - (editorial) "it the standard" should be "it in the standard"
>=20
> Done.
>=20
>=20
> - (minor) What does "remain transparent to standard TCP" really mean?
> There must be a better way to word this.
>=20
> Here, it means "avoid interfering with standard TCP traffic", and
> that's what we've now written there.
>=20
>=20
> Section 2.1
>=20
> - (minor) under the "RTT measurement issues", is it worth mentioning
> the case of fast or short-distance links where the delay is all in the
> stack and not on the wire?
>=20
> Absolutely! I am surprised that we missed this, thanks for catching
> it! I put that there, now.
>=20
>=20
> Section 2.2
>=20
> - (minor) is it worth mentioning LEDBAT's sensitivity to changes in
> the route also, when we mention Vegas?  Is it purposeful that we never
> really compare/contrast to LEDBAT itself in this document?
>=20
> Yes, this might be worth mentioning regarding LEDBAT, but rather than
> scattering LEDBAT comments throughout the document, we have opted for
> putting all such LEDBAT (except for the introductory statement that
> you suggested) in a "Conclusion and relationship to LEDBAT" section if
> we indeed end up writing one (see our question to the chairs).
>=20
>=20
> Section 4
>=20
> - (editorial) In the last paragraph, is a ":" really the best way to
> punctuate this?
>=20
> Done.
>=20
>=20
> Section 5
>=20
> - (minor) For the network-assisted approaches, it wasn't entirely
> clear to me if these were new transport protocols or new mechanisms
> for existing transports; if that could be clarified without the reader
> needing to check the references, I think that would be good.
>=20
> Done (clarified for every reference where it occurs).
>=20
>=20
> - (minor) Should there be a summary/conclusion sentence here about the
> current viability of such approaches?
>=20
> We think not, as these proposals are so heterogeneous: e.g., one
> involves changing TCP's AIMD parameters, the other involves no changes
> to TCP at all but essentially putting a multipath-LBE-PEP in the
> network. Anyway, as you said, we're acknowledging other techniques,
> but not advocating them here; we would prefer to present them as
> neutrally as possible.
> _______________________________________________
> ledbat mailing list
> ledbat@ietf.org
> https://www.ietf.org/mailman/listinfo/ledbat

From michawe@ulrik.uio.no  Tue Dec 21 13:18:01 2010
Return-Path: <michawe@ulrik.uio.no>
X-Original-To: ledbat@core3.amsl.com
Delivered-To: ledbat@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2976E3A6A92 for <ledbat@core3.amsl.com>; Tue, 21 Dec 2010 13:18:01 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id APDos9EaP1XX for <ledbat@core3.amsl.com>; Tue, 21 Dec 2010 13:17:58 -0800 (PST)
Received: from mail-out2.uio.no (mail-out2.uio.no [129.240.10.58]) by core3.amsl.com (Postfix) with ESMTP id 5190B3A6A6E for <ledbat@ietf.org>; Tue, 21 Dec 2010 13:17:58 -0800 (PST)
Received: from mail-mx2.uio.no ([129.240.10.30]) by mail-out2.uio.no with esmtp (Exim 4.72) (envelope-from <michawe@ulrik.uio.no>) id 1PV9cg-0003s4-07; Tue, 21 Dec 2010 22:19:54 +0100
Received: from w3prod-wm01.uio.no ([129.240.4.214] helo=webmail.uio.no) by mail-mx2.uio.no with esmtpsa (TLSv1:AES256-SHA:256) user michawe (Exim 4.72) (envelope-from <michawe@ulrik.uio.no>) id 1PV9cf-0002OY-3l; Tue, 21 Dec 2010 22:19:53 +0100
Received: from 194.48.133.8 (SquirrelMail authenticated user michawe) by webmail.uio.no with HTTP; Tue, 21 Dec 2010 22:19:53 +0100
Message-ID: <226a95bd02babc35623ffb9d2da54189.squirrel@webmail.uio.no>
In-Reply-To: <791AD3077F94194BB2BDD13565B6295D86D2C5@PALLENE.office.hd>
References: <9FF43AB0-FA71-4536-ADE7-78CA2CEEEF12@ifi.uio.no> <791AD3077F94194BB2BDD13565B6295D86D2C5@PALLENE.office.hd>
Date: Tue, 21 Dec 2010 22:19:53 +0100
From: michawe@ifi.uio.no
To: "Rolf Winter" <Rolf.Winter@neclab.eu>
User-Agent: SquirrelMail/1.4.19
MIME-Version: 1.0
Content-Type: text/plain;charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-UiO-Ratelimit-Test: rcpts/h 2 msgs/h 1 sum rcpts/h 2 sum msgs/h 1 total rcpts 5162 max rcpts/h 33 ratelimit 0
X-UiO-Spam-info: not spam, SpamAssassin (score=-5.0, required=5.0, autolearn=disabled, T_RP_MATCHES_RCVD=-0.01, UIO_MAIL_IS_INTERNAL=-5, uiobl=NO, uiouri=NO)
X-UiO-Scanned: 83788B12D878DB3364EFB8DEECE4E5F3460B991F
X-UiO-SPAM-Test: remote_host: 129.240.4.214 spam_score: -49 maxlevel 80 minaction 2 bait 0 mail/h: 57 total 1021986 max/h 663 blacklist 0 greylist 0 ratelimit 0
Cc: "ledbat@ietf.org" <ledbat@ietf.org>
Subject: Re: [ledbat] draft-ietf-ledbat-survey-03.txt
X-BeenThere: ledbat@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Mailing list of the LEDBAT WG <ledbat.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/ledbat>
List-Post: <mailto:ledbat@ietf.org>
List-Help: <mailto:ledbat-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/ledbat>, <mailto:ledbat-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 21 Dec 2010 21:18:01 -0000

ok!

> Micheal,
>
> you say that the order is wrong, however, when you look at our original
> charter, the last item to finish actually is the LEDBAT survey. So the
> request that both Wes and Mirja have made make complete sense to me. You
> can already write that section now as I do not see a reason why you would
> need to wait.
>
> Best,
>
> 	Rolf
>
>
> NEC Europe Limited | Registered Office: NEC House, 1 Victoria Road, London
> W3 6BL | Registered in England 2832014
>
>
>> -----Original Message-----
>> From: ledbat-bounces@ietf.org [mailto:ledbat-bounces@ietf.org] On
>> Behalf Of Michael Welzl
>> Sent: Freitag, 17. Dezember 2010 22:03
>> To: ledbat@ietf.org
>> Subject: [ledbat] draft-ietf-ledbat-survey-03.txt
>>
>> Hi all,
>>
>> We have now submitted draft-ietf-ledbat-survey-03.txt, in which we
>> tried to address Wes' comments:
>> http://www.ietf.org/id/draft-ietf-ledbat-survey-03.txt
>> See below for his list of comments, with our answers in line.
>>
>> We would like to raise a question to the chairs regarding one matter.
>> Wes, and Mirja before him, have suggested to include a concluding
>> section where we discuss pro's and con's of LEDBAT in comparison to
>> the related work here. It seems clear to us that such a section would
>> not make any sense without citing the LEDBAT document, but such a
>> citation would put this document on hold until LEDBAT is published. It
>> seems to us that this just goes to show that this is in fact the wrong
>> sequence, and the wrong place for this discussion: in our opinion, the
>> LEDBAT document should contain a short paragraph where it is explained
>> how it relates to the state of the art, as documented in our survey.
>> This would seem like a much nicer structure to us (e.g. in academia,
>> first you look at the state of the art, then you propose your
>> mechanism in relation to it).
>>
>> We would appreciate if the chairs could let us know what they think
>> about this matter, and if the reviewers could let us know if they are
>> happy with how we addressed all their other comments.
>>
>> Thanks!!
>>
>> Michael and David
>>
>>
>> ----
>>
>> ANSWERS TO WES' COMMENTS:
>>
>> Here are some review comments I had.  Aside from some general ones,
>> they're all editorial or possibly minor technical clarifications that
>> could be made.
>>
>> General / Overall
>>
>> - We should have some material early on to set context; things to
>> mention include:
>> - intended audience (familiarity with standard TCP, RFC2581/5681, is
>> assumed)
>> - this was a product of the LEDBAT WG for comparison/contrast to the
>> chosen approach (this needs to be mentioned, because remember that
>> when the RFC is published, a casual reader won't know it came from
>> LEDBAT unless you say so)
>> - most techniques discussed here are tested in limited simulations or
>> experimental testbeds, but LEDBAT's algorithm is already under
>> widespread deployment (i.e. we need to say that we're acknowledging
>> other techniques, but not advocating them here; this does not
>> interfere with advancement of the LEDBAT algorithm)
>> - this is not exhaustive; that wouldn't be possible or useful, but the
>> authors/editors have selected key, well-known, or otherwise
>> interesting techniques for inclusion at their discretion
>> - was applicability to big-I Internet deployment scenarios a
>> consideration?  (I think yes, but don't actually know in some of the
>> cases)
>>
>> Done. Regarding the "big-I" question, it is clear that almost all of
>> the presented mechanisms were designed for that; if there are one or
>> two exceptions, then stressing that here just for these mechanisms
>> would not contribute much to readability, we believe - the general
>> picture, as presented and introduced here, is about the Internet,
>> throughout the document.
>>
>>
>> - Does it need to be more clearly mentioned which classes of
>> approaches were mechanisms incorporated in TCP, TCP-alternatives (e.g.
>> for DCCP or new transports), or other designs entirely.  For instance,
>> do they have subtle (or obvious) differences in the services provided
>> to applications (reliability, smoothness, etc.)?  Section 4's first
>> paragraph does a good job of explaining this, but other sections had
>> me wondering slightly.  It may be a personal problem on my part, so it
>> would be good to check with others :).
>>
>> We tried to state whether mechanisms are incorporated in TCP or
>> entirely new protocols upon their appearance in our document (as you
>> mention below, we have failed to make this clear enough in section 5,
>> but this has now been fixed). Generally, the new protocol / TCP-change
>> classification is orthogonal to ours, i.e. both variants are scattered
>> throughout the document. Regarding the service provided to
>> applications, there is no simple answer here. LBE is, as we define, "a
>> service which results in smaller bandwidth and/or delay impact on
>> standard TCP than standard TCP itself, when sharing a bottleneck with
>> it". This is fairly network-centric; a more user-centric perspective
>> is not commonly taken in this literature, and no user-centric
>> comparison exists to the best of our knowledge. Yet, the user
>> experience will be different, depending on the dynamics of the
>> underlying mechanism -- much like you could get a different user
>> experience from CTCP, HTCP and CUBIC.
>>
>>
>> - Are there fairly common assumptions in a lot of the existing
>> academic analysis cited here which influence results?  For instance,
>> there are often assumptions about flow length, RTT distributions, or
>> other elements that may be relevant to mention here? (e.g. all flows
>> are FTP of 10 GB files, etc.)
>>
>> We think not.
>>
>>
>> - Should there be a (very short) conclusions section that says where
>> LEDBAT fits in and the main pros/cons versus the other classes of
>> system discussed?
>>
>> Maybe; this was also requested by Mirja. See our question to the
>> chairs regarding this matter.
>>
>>
>> Section 1
>>
>> - (editorial) It might be more clear to replace "end-to-end
>> mechanisms" with "end-host mechanisms", since end-to-end is relative.
>>
>> Done.
>>
>>
>> - (editorial) The LBE acronymn is defined twice in the first
>> paragraph; the second expansion at the end of the paragraph can
>> probably either just use "LBE" or put the expansion in quotation marks.
>>
>> Done (just used "LBE").
>>
>>
>> Section 2
>>
>> - (editorial) The second sentence of the first paragraph could be
>> clarified and shortened to just "Without ECN support, standard TCP
>> will normally increase its congestion window (and effective sending
>> rate) until a queue overflows, causing one or more packets to be
>> dropped and the effective rate to be reduced."  I didn't like the "for
>> some reason" in the current sentence because there are clear reasons
>> like application limited, advertised receive window limited, kernel
>> knob limited, etc, but they all seem to be beside the point rather
>> than contributing to it.
>>
>> Done.
>>
>>
>> - (editorial) At one point the text capitalizes "Congestion
>> Avoidance", and in others it doesn't.  I think lower-case is best.
>>
>> Done.
>>
>>
>> - (editorial) "threshold beta" might be more clear as "threshold
>> (called "beta")" (same comment applies for alpha)
>>
>> Done (but without brackets because they don't go well with another
>> bracket in the "alpha" part).
>>
>>
>> - (minor) "actual (measured) throughput" might be more clear as
>> "actual throughput measured by recent acknowledgements"
>>
>> Done.
>>
>>
>> - (minor) is the increase at the end of the 3rd paragraph every RTT or
>> BaseRTT (I don't recall myself)
>>
>> We checked: the original [Bra94] paper only talks about increasing
>> linearly and decreasing linearly. I have therefore removed the "by one
>> segment per RTT" bit.
>>
>>
>> - (minor) "maintaining the window size" should be "maintaining the
>> current congestion window"
>>
>> Done.
>>
>>
>> - (editorial) "it the standard" should be "it in the standard"
>>
>> Done.
>>
>>
>> - (minor) What does "remain transparent to standard TCP" really mean?
>> There must be a better way to word this.
>>
>> Here, it means "avoid interfering with standard TCP traffic", and
>> that's what we've now written there.
>>
>>
>> Section 2.1
>>
>> - (minor) under the "RTT measurement issues", is it worth mentioning
>> the case of fast or short-distance links where the delay is all in the
>> stack and not on the wire?
>>
>> Absolutely! I am surprised that we missed this, thanks for catching
>> it! I put that there, now.
>>
>>
>> Section 2.2
>>
>> - (minor) is it worth mentioning LEDBAT's sensitivity to changes in
>> the route also, when we mention Vegas?  Is it purposeful that we never
>> really compare/contrast to LEDBAT itself in this document?
>>
>> Yes, this might be worth mentioning regarding LEDBAT, but rather than
>> scattering LEDBAT comments throughout the document, we have opted for
>> putting all such LEDBAT (except for the introductory statement that
>> you suggested) in a "Conclusion and relationship to LEDBAT" section if
>> we indeed end up writing one (see our question to the chairs).
>>
>>
>> Section 4
>>
>> - (editorial) In the last paragraph, is a ":" really the best way to
>> punctuate this?
>>
>> Done.
>>
>>
>> Section 5
>>
>> - (minor) For the network-assisted approaches, it wasn't entirely
>> clear to me if these were new transport protocols or new mechanisms
>> for existing transports; if that could be clarified without the reader
>> needing to check the references, I think that would be good.
>>
>> Done (clarified for every reference where it occurs).
>>
>>
>> - (minor) Should there be a summary/conclusion sentence here about the
>> current viability of such approaches?
>>
>> We think not, as these proposals are so heterogeneous: e.g., one
>> involves changing TCP's AIMD parameters, the other involves no changes
>> to TCP at all but essentially putting a multipath-LBE-PEP in the
>> network. Anyway, as you said, we're acknowledging other techniques,
>> but not advocating them here; we would prefer to present them as
>> neutrally as possible.
>> _______________________________________________
>> ledbat mailing list
>> ledbat@ietf.org
>> https://www.ietf.org/mailman/listinfo/ledbat
>


