Re: [tcpm] Discussion -- Pause/Resume mechanism on TCP ?

Alexander Zimmermann <alexander.zimmermann@nets.rwth-aachen.de> Wed, 21 April 2010 12:01 UTC

Return-Path: <alexander.zimmermann@nets.rwth-aachen.de>
X-Original-To: tcpm@core3.amsl.com
Delivered-To: tcpm@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A76F028C14A for <tcpm@core3.amsl.com>; Wed, 21 Apr 2010 05:01:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.801
X-Spam-Level:
X-Spam-Status: No, score=-4.801 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_DE=0.35, HELO_MISMATCH_DE=1.448, RCVD_IN_DNSWL_MED=-4]
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 MKBw6ZnTJV1z for <tcpm@core3.amsl.com>; Wed, 21 Apr 2010 05:01:15 -0700 (PDT)
Received: from mta-2.ms.rz.rwth-aachen.de (mta-2.ms.rz.RWTH-Aachen.DE [134.130.7.73]) by core3.amsl.com (Postfix) with ESMTP id 041D13A67E9 for <tcpm@ietf.org>; Wed, 21 Apr 2010 05:01:12 -0700 (PDT)
MIME-version: 1.0
Received: from ironport-out-2.rz.rwth-aachen.de ([134.130.5.41]) by mta-2.ms.rz.RWTH-Aachen.de (Sun Java(tm) System Messaging Server 6.3-7.04 (built Sep 26 2008)) with ESMTP id <0L1800GG36PPDJ50@mta-2.ms.rz.RWTH-Aachen.de> for tcpm@ietf.org; Wed, 21 Apr 2010 14:01:01 +0200 (CEST)
X-IronPort-AV: E=Sophos; i="4.52,250,1270418400"; d="sig'?scan'208"; a="28791156"
Received: from relay-auth-2.ms.rz.rwth-aachen.de (HELO relay-auth-2) ([134.130.7.79]) by ironport-in-2.rz.rwth-aachen.de with ESMTP; Wed, 21 Apr 2010 14:01:02 +0200
Received: from [137.226.12.56] ([unknown] [137.226.12.56]) by relay-auth-2.ms.rz.rwth-aachen.de (Sun Java(tm) System Messaging Server 7.0-3.01 64bit (built Dec 9 2008)) with ESMTPA id <0L180015S6PP7Q50@relay-auth-2.ms.rz.rwth-aachen.de> for tcpm@ietf.org; Wed, 21 Apr 2010 14:01:01 +0200 (CEST)
Content-type: multipart/signed; protocol="application/pgp-signature"; micalg=pgp-sha1; boundary=Apple-Mail-37--661193714
From: Alexander Zimmermann <alexander.zimmermann@nets.rwth-aachen.de>
In-reply-to: <816145D8-3C62-4F28-AC93-FC7B1051B833@nokia.com>
Date: Wed, 21 Apr 2010 14:01:05 +0200
Content-transfer-encoding: 7bit
Message-id: <129BD33F-1E6B-43FD-9769-B0075044ED70@nets.rwth-aachen.de>
References: <00d601cae06c$5a4324c0$96106f0a@china.huawei.com> <05F0AC17-703E-4F25-8730-A88ACA4D8E6B@nokia.com> <01ac01cae12d$81d50640$96106f0a@china.huawei.com> <816145D8-3C62-4F28-AC93-FC7B1051B833@nokia.com>
To: Lars Eggert <lars.eggert@nokia.com>
X-Pgp-Agent: GPGMail 1.2.3
X-Mailer: Apple Mail (2.1078)
Cc: "tcpm@ietf.org" <tcpm@ietf.org>
Subject: Re: [tcpm] Discussion -- Pause/Resume mechanism on TCP ?
X-BeenThere: tcpm@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: TCP Maintenance and Minor Extensions Working Group <tcpm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/tcpm>, <mailto:tcpm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/tcpm>
List-Post: <mailto:tcpm@ietf.org>
List-Help: <mailto:tcpm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/tcpm>, <mailto:tcpm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 21 Apr 2010 12:01:16 -0000

Hi,

Am 21.04.2010 um 12:00 schrieb Lars Eggert:

> Hi,
> 
> ...
> 
> Even proposals that are agnostic to link technology have issues, e.g., Freeze TCP, which I cited since it sounded similar to your proposal. For example, there is an implicit assumption that path characteristics (available bandwidth, RTT, etc.) do not change during the freeze period - when they do, the thawed connection can significantly overshoot along the new path, or it can underperform (and will take a long time to probe for extra capacity if it is in congestion avoidance mode). And obviously, there is an assumption that you somehow know when to freeze and thaw. 
> 

Exactly!!! This IMO the reason why TCP freeze would not work (in general).

Copied from RFC 2861:

3. Description
"When a TCP sender has sufficient data available to fill the available network capacity for that flow, cwnd and ssthresh get set to appropriate values for the network conditions. When a TCP sender stops sending, the flow stops sampling the network conditions, and so the value of the congestion window may become inaccurate."

BTW: Read the introduction of this RFC. It gave me an excellent insight why TCP freeze is a bad idea...

Alex