
From rightkicktech@gmail.com  Mon Nov 30 15:37:49 2015
Return-Path: <rightkicktech@gmail.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 4251B1B332F for <multipathtcp@ietfa.amsl.com>; Mon, 30 Nov 2015 15:37:49 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.999
X-Spam-Level: 
X-Spam-Status: No, score=-1.999 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id IvjuNAHPoMgq for <multipathtcp@ietfa.amsl.com>; Mon, 30 Nov 2015 15:37:46 -0800 (PST)
Received: from mail-wm0-x235.google.com (mail-wm0-x235.google.com [IPv6:2a00:1450:400c:c09::235]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 91A941B332C for <multipathtcp@ietf.org>; Mon, 30 Nov 2015 15:37:46 -0800 (PST)
Received: by wmec201 with SMTP id c201so160769704wme.1 for <multipathtcp@ietf.org>; Mon, 30 Nov 2015 15:37:45 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=user-agent:mime-version:content-type:content-transfer-encoding :subject:from:date:to:message-id; bh=nAoXOgj9JjEFdo63vEtjFqC+19/b2Q4HxpzOJikvI7Y=; b=ERwf9JX2yIGTPH+tQpSdp9/cc1doGHoWFSno1AYfwoW2OY118yDMboyhN9FCuMpFRi i/U+G33Rw0YupLopLkQg+oCqrz9nEhv2auA6cbJs9DGpt66zwXHhAegQSTAmos58KLZh ev55NV2eUGksiSYhbSww+9Pt9CKshgI0gn2P6+Vq1eirozpg6CKGtCxklseGH7HhVx6E uE4UgXs4HP4oREIPcohVF8ehPYZpFPsR5pC92gxh/Is6Y9S9301OWd11lsD9vOYKWr7T QoVmCTL/cRyv+D6iZjQo74AcDgO8tJZ9LhOSDpzmM5KbLv+7X2aMVVQcV2jACQE9YORc QFLQ==
X-Received: by 10.28.158.75 with SMTP id h72mr31072504wme.74.1448926665086; Mon, 30 Nov 2015 15:37:45 -0800 (PST)
Received: from android-6b1b2246a4c3940a.lan (dsl-aauu78.dyn.edudsl.gr. [37.32.166.100]) by smtp.gmail.com with ESMTPSA id z1sm48683726wje.35.2015.11.30.15.37.43 for <multipathtcp@ietf.org> (version=TLSv1/SSLv3 cipher=OTHER); Mon, 30 Nov 2015 15:37:44 -0800 (PST)
User-Agent: K-9 Mail for Android
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----402AHP9ZUXFUR0CF2FEDPUXJI1GRUO"
Content-Transfer-Encoding: 8bit
From: "rightkicktech.gmail.com" <rightkicktech@gmail.com>
Date: Tue, 01 Dec 2015 01:38:03 +0200
To: multipathtcp@ietf.org
Message-ID: <DED9695E-5D0D-429A-8F96-81A93C2FE2BA@gmail.com>
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/c6ZvcNQfs6i8cb2IzKsUgqZmTvc>
X-Mailman-Approved-At: Tue, 01 Dec 2015 02:48:21 -0800
Subject: [multipathtcp] MPTCP Usage for VPN hub.
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 30 Nov 2015 23:40:32 -0000

------402AHP9ZUXFUR0CF2FEDPUXJI1GRUO
Content-Transfer-Encoding: 8bit
Content-Type: text/plain;
 charset=UTF-8

Hi All,

I've been searching on ways to aggregate bandwidth of multiple WAN links and the mptcp project struck me as an interesting way to share TCP traffic between links.

I've followed the intructions from the site and I was able to generate subflows on multiple WAN interfaces when reaching iperf server multipath-tcp.org. Im not able though to generate the same multipath flows when testing with my server using iperf. I've installed the mptcp enabled kernel on debian wheezy by installing the linux-mptcp package from the repo, both on the server and the client side, as mentioned on the site. Am I missing sth? The end goal is to use the server as the a mptcp proxy, to proxy and forward all mptcp traffic to plain tcp. The mptcp teaffic will be generated from a remote location, which has its own mptcp enabled router (debian wheezy with linux-mptcp installed) which will terminate to the server with at least two vpn tunnels. The server then, having a fast wan connection, will route the traffic to final destination with plain tcp, then back again with mptcp.

Any ideas or hints will be appreciated.
Alex
------402AHP9ZUXFUR0CF2FEDPUXJI1GRUO
Content-Type: text/html;
 charset=utf-8
Content-Transfer-Encoding: 8bit

Hi All,<br>
<br>
I&#39;ve been searching on ways to aggregate bandwidth of multiple WAN links and the mptcp project struck me as an interesting way to share TCP traffic between links.<br>
<br>
I&#39;ve followed the intructions from the site and I was able to generate subflows on multiple WAN interfaces when reaching iperf server <a href="http://multipath-tcp.org">multipath-tcp.org</a>. Im not able though to generate the same multipath flows when testing with my server using iperf. I&#39;ve installed the mptcp enabled kernel on debian wheezy by installing the linux-mptcp package from the repo, both on the server and the client side, as mentioned on the site. Am I missing sth? The end goal is to use the server as the a mptcp proxy, to proxy and forward all mptcp traffic to plain tcp. The mptcp teaffic will be generated from a remote location, which has its own mptcp enabled router (debian wheezy with linux-mptcp installed) which will terminate to the server with at least two vpn tunnels. The server then, having a fast wan connection, will route the traffic to final destination with plain tcp, then back again with mptcp.<br>
<br>
Any ideas or hints will be appreciated.<br>
Alex
------402AHP9ZUXFUR0CF2FEDPUXJI1GRUO--


From nobody Fri Dec  4 10:41:39 2015
Return-Path: <nishida@sfc.wide.ad.jp>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 31B691B320B for <multipathtcp@ietfa.amsl.com>; Fri,  4 Dec 2015 10:41:39 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.22
X-Spam-Level: *
X-Spam-Status: No, score=1.22 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, FM_FORGED_GMAIL=0.622, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id wqSX3j8-0ZgZ for <multipathtcp@ietfa.amsl.com>; Fri,  4 Dec 2015 10:41:37 -0800 (PST)
Received: from mail.sfc.wide.ad.jp (shonan.sfc.wide.ad.jp [IPv6:2001:200:0:8803::53]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 75EA81B3215 for <multipathtcp@ietf.org>; Fri,  4 Dec 2015 10:41:36 -0800 (PST)
Received: from mail-ig0-f182.google.com (mail-ig0-f182.google.com [209.85.213.182]) by mail.sfc.wide.ad.jp (Postfix) with ESMTPSA id 2088B279237 for <multipathtcp@ietf.org>; Sat,  5 Dec 2015 03:41:35 +0900 (JST)
Received: by igbxm8 with SMTP id xm8so43304650igb.1 for <multipathtcp@ietf.org>; Fri, 04 Dec 2015 10:41:33 -0800 (PST)
MIME-Version: 1.0
X-Received: by 10.50.178.177 with SMTP id cz17mr3189956igc.94.1449254493600; Fri, 04 Dec 2015 10:41:33 -0800 (PST)
Received: by 10.79.38.79 with HTTP; Fri, 4 Dec 2015 10:41:33 -0800 (PST)
In-Reply-To: <DED9695E-5D0D-429A-8F96-81A93C2FE2BA@gmail.com>
References: <DED9695E-5D0D-429A-8F96-81A93C2FE2BA@gmail.com>
Date: Fri, 4 Dec 2015 10:41:33 -0800
X-Gmail-Original-Message-ID: <CAO249ydMZvYhzsRMsJMUVdCAyT=6bHVhuq=M8XfHWNyu1nM5pA@mail.gmail.com>
Message-ID: <CAO249ydMZvYhzsRMsJMUVdCAyT=6bHVhuq=M8XfHWNyu1nM5pA@mail.gmail.com>
From: Yoshifumi Nishida <nishida@sfc.wide.ad.jp>
To: "rightkicktech.gmail.com" <rightkicktech@gmail.com>
Content-Type: text/plain; charset=UTF-8
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/8sk3VR6dvjK_40EXACKbTSGoNdY>
Cc: multipathtcp <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] MPTCP Usage for VPN hub.
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 04 Dec 2015 18:41:39 -0000

Hi Alex,

I think what you're trying to do is a reasonable use case for mptcp.
Have you checked linux mptcp developer list? It seems that this is
linux related to stuff rather than protocol spec.
If not, please check  https://listes-2.sipr.ucl.ac.be/sympa/info/mptcp-dev

Thanks,
--
Yoshi

On Mon, Nov 30, 2015 at 3:38 PM, rightkicktech.gmail.com
<rightkicktech@gmail.com> wrote:
> Hi All,
>
> I've been searching on ways to aggregate bandwidth of multiple WAN links and
> the mptcp project struck me as an interesting way to share TCP traffic
> between links.
>
> I've followed the intructions from the site and I was able to generate
> subflows on multiple WAN interfaces when reaching iperf server
> multipath-tcp.org. Im not able though to generate the same multipath flows
> when testing with my server using iperf. I've installed the mptcp enabled
> kernel on debian wheezy by installing the linux-mptcp package from the repo,
> both on the server and the client side, as mentioned on the site. Am I
> missing sth? The end goal is to use the server as the a mptcp proxy, to
> proxy and forward all mptcp traffic to plain tcp. The mptcp teaffic will be
> generated from a remote location, which has its own mptcp enabled router
> (debian wheezy with linux-mptcp installed) which will terminate to the
> server with at least two vpn tunnels. The server then, having a fast wan
> connection, will route the traffic to final destination with plain tcp, then
> back again with mptcp.
>
> Any ideas or hints will be appreciated.
> Alex
> _______________________________________________
> multipathtcp mailing list
> multipathtcp@ietf.org
> https://www.ietf.org/mailman/listinfo/multipathtcp
>


From nobody Fri Dec  4 10:50:29 2015
Return-Path: <olivier.bonaventure@uclouvain.be>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 4F7C31B3238 for <multipathtcp@ietfa.amsl.com>; Fri,  4 Dec 2015 10:50:28 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.402
X-Spam-Level: 
X-Spam-Status: No, score=-2.402 tagged_above=-999 required=5 tests=[BAYES_40=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id g3ILGzDzkPbW for <multipathtcp@ietfa.amsl.com>; Fri,  4 Dec 2015 10:50:25 -0800 (PST)
Received: from smtp5.sgsi.ucl.ac.be (smtp.sgsi.ucl.ac.be [130.104.5.67]) by ietfa.amsl.com (Postfix) with ESMTP id 600B81B3235 for <multipathtcp@ietf.org>; Fri,  4 Dec 2015 10:50:25 -0800 (PST)
Received: from mbpobo.local (host-78-129-6-94.dynamic.voo.be [78.129.6.94]) (using TLSv1 with cipher DHE-RSA-AES128-SHA (128/128 bits)) (No client certificate requested) (Authenticated sender: obonaventure@smtp5.sgsi.ucl.ac.be) by smtp5.sgsi.ucl.ac.be (Postfix) with ESMTPSA id 140A9198023; Fri,  4 Dec 2015 19:50:19 +0100 (CET)
X-DKIM: Sendmail DKIM Filter v2.8.3 smtp5.sgsi.ucl.ac.be 140A9198023
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=uclouvain.be; s=selucl; t=1449255019; bh=Jtgf3sV3Tx11TbT5wx73u/ox7wPgUwLyxWK05+sJc60=; h=Reply-To:Subject:References:To:Cc:From:Message-ID:Date: MIME-Version:In-Reply-To:Content-Type:Content-Transfer-Encoding; b=fHljtMxm2CVApCJX0PT1XrCgsJV1rbdBMEr9iMEmi9NZOdFqMZ4pJ8mKfsQatjKdH vXPZzsgUpLAjQQO1kDEtv5I/3F7IQ0Pp1cZnuhX0szw8bGNKHMiRMhiTlZYTFWunOO UwXm1gJZSIARRpny43c3gOCWhTWikQe2Lg0fgZDM=
References: <DED9695E-5D0D-429A-8F96-81A93C2FE2BA@gmail.com> <CAO249ydMZvYhzsRMsJMUVdCAyT=6bHVhuq=M8XfHWNyu1nM5pA@mail.gmail.com>
To: Yoshifumi Nishida <nishida@sfc.wide.ad.jp>, "rightkicktech.gmail.com" <rightkicktech@gmail.com>
From: Olivier Bonaventure <Olivier.Bonaventure@uclouvain.be>
Message-ID: <5661E06A.1060606@uclouvain.be>
Date: Fri, 4 Dec 2015 19:50:18 +0100
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.11; rv:38.0) Gecko/20100101 Thunderbird/38.3.0
MIME-Version: 1.0
In-Reply-To: <CAO249ydMZvYhzsRMsJMUVdCAyT=6bHVhuq=M8XfHWNyu1nM5pA@mail.gmail.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: clamav-milter 0.99-beta1 at smtp-5.sipr-dc.ucl.ac.be
X-Virus-Status: Clean
X-Sgsi-Spamcheck: SASL authenticated, 
X-SGSI-MailScanner-ID: 140A9198023.A6BED
X-SGSI-MailScanner: Found to be clean
X-SGSI-From: olivier.bonaventure@uclouvain.be
X-SGSI-Spam-Status: No
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/Mm7rd-3t-cx6DyB8dGC-LKOEXUk>
Cc: multipathtcp <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] MPTCP Usage for VPN hub.
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
Reply-To: Olivier.Bonaventure@uclouvain.be
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 04 Dec 2015 18:50:28 -0000

Alex, Yoshifumi
>
> I think what you're trying to do is a reasonable use case for mptcp.
> Have you checked linux mptcp developer list? It seems that this is
> linux related to stuff rather than protocol spec.
> If not, please check  https://listes-2.sipr.ucl.ac.be/sympa/info/mptcp-dev

I also suggest to move this discussion to the mptcp-dev mailing list.
>>
>> I've been searching on ways to aggregate bandwidth of multiple WAN links and
>> the mptcp project struck me as an interesting way to share TCP traffic
>> between links.
>>
>> I've followed the intructions from the site and I was able to generate
>> subflows on multiple WAN interfaces when reaching iperf server
>> multipath-tcp.org. Im not able though to generate the same multipath flows
>> when testing with my server using iperf. I've installed the mptcp enabled
>> kernel on debian wheezy by installing the linux-mptcp package from the repo,
>> both on the server and the client side, as mentioned on the site. Am I
>> missing sth? The end goal is to use the server as the a mptcp proxy, to
>> proxy and forward all mptcp traffic to plain tcp. The mptcp teaffic will be
>> generated from a remote location, which has its own mptcp enabled router
>> (debian wheezy with linux-mptcp installed) which will terminate to the
>> server with at least two vpn tunnels. The server then, having a fast wan
>> connection, will route the traffic to final destination with plain tcp, then
>> back again with mptcp.
>>
>> Any ideas or hints will be appreciated.

Some people have reported to use OpenVPN in TCP mode for this use case. 
Others have deployed SOCKS based solutions for this use case. The 
OverTheBox solution that is being developed by OVH relies on encrypted 
SOCKS tunnels to aggregate DSL links. This is probably very close to 
what you have in mind. The client side of their solution is open-source 
and builds upon OpenWrt, see https://github.com/ovh/overthebox-openwrt
for the current source. This version runs on x86 servers and there is 
documentation in French on https://www.ovhtelecom.fr/overthebox/
I don't think that they have released the server side of this service.

Olivier


From nobody Fri Dec  4 17:39:56 2015
Return-Path: <rightkicktech@gmail.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 36DDB1A1A82 for <multipathtcp@ietfa.amsl.com>; Fri,  4 Dec 2015 17:39:56 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.999
X-Spam-Level: 
X-Spam-Status: No, score=-1.999 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ym6bLsMv-iT9 for <multipathtcp@ietfa.amsl.com>; Fri,  4 Dec 2015 17:39:54 -0800 (PST)
Received: from mail-wm0-x22a.google.com (mail-wm0-x22a.google.com [IPv6:2a00:1450:400c:c09::22a]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 7D7291A1A7D for <multipathtcp@ietf.org>; Fri,  4 Dec 2015 17:39:53 -0800 (PST)
Received: by wmvv187 with SMTP id v187so98036452wmv.1 for <multipathtcp@ietf.org>; Fri, 04 Dec 2015 17:39:52 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=user-agent:in-reply-to:references:mime-version:content-type :content-transfer-encoding:subject:from:date:to:cc:message-id; bh=1sNKQCj+GaEVITUJSK40WvLAqLY5TTp25yloksVplDA=; b=Z2qChma88bEoGOuzwuKtEZ9NoPUOzVkA6nU9QFoNSozEt70c/TKnK4KNuDpLUClLw/ +085RsvitLsXBS5/4tpugB8/11CZ30c709aaGksNQZSdGxz+ZLaVEtTrLiPLk9eIgP8h 7hJW/1EJ7nk4pW9nKcTCSkK3h5x5mYQzady7aFZJO67W02JKzCuusfLpUmTmpkZJgExo ywGYuyykSxdVvj96ZRVSDshYQq+VMdjMe1rUV/tH30QwZsbfsvasoaFEySwUPGhigFM4 /sQnWvi2WaA93tag/67tbdUMZvtyXrg6aONou4QUV7NR/ivf/yWs+VEYdMF5ypLG+hyG C1/A==
X-Received: by 10.28.188.86 with SMTP id m83mr7662860wmf.51.1449279592146; Fri, 04 Dec 2015 17:39:52 -0800 (PST)
Received: from android-6b1b2246a4c3940a.lan (dsl-aauxur.dyn.edudsl.gr. [37.32.184.227]) by smtp.gmail.com with ESMTPSA id jo6sm14343450wjb.48.2015.12.04.17.39.50 (version=TLSv1/SSLv3 cipher=OTHER); Fri, 04 Dec 2015 17:39:51 -0800 (PST)
User-Agent: K-9 Mail for Android
In-Reply-To: <5661E06A.1060606@uclouvain.be>
References: <DED9695E-5D0D-429A-8F96-81A93C2FE2BA@gmail.com> <CAO249ydMZvYhzsRMsJMUVdCAyT=6bHVhuq=M8XfHWNyu1nM5pA@mail.gmail.com> <5661E06A.1060606@uclouvain.be>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----9LJT0FZVNB2X2DOR6LJFB2OXS446VG"
Content-Transfer-Encoding: 8bit
From: "rightkicktech.gmail.com" <rightkicktech@gmail.com>
Date: Sat, 05 Dec 2015 03:40:06 +0200
To: Olivier.Bonaventure@uclouvain.be, Yoshifumi Nishida <nishida@sfc.wide.ad.jp>
Message-ID: <9BB1FC57-389C-4310-8363-7E61F231D789@gmail.com>
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/06oOypfD2aHjKovGCjVxkUklHxg>
Cc: multipathtcp <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] MPTCP Usage for VPN hub.
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 05 Dec 2015 01:39:56 -0000

------9LJT0FZVNB2X2DOR6LJFB2OXS446VG
Content-Transfer-Encoding: 8bit
Content-Type: text/plain;
 charset=UTF-8

Thanx Olivier and Yoshifumi,

I will divert the thread to dev list.
It will be very interesting to discuss further on this topic.

Cheers,
Alex


On December 4, 2015 8:50:18 PM EET, Olivier Bonaventure <Olivier.Bonaventure@uclouvain.be> wrote:
>Alex, Yoshifumi
>>
>> I think what you're trying to do is a reasonable use case for mptcp.
>> Have you checked linux mptcp developer list? It seems that this is
>> linux related to stuff rather than protocol spec.
>> If not, please check 
>https://listes-2.sipr.ucl.ac.be/sympa/info/mptcp-dev
>
>I also suggest to move this discussion to the mptcp-dev mailing list.
>>>
>>> I've been searching on ways to aggregate bandwidth of multiple WAN
>links and
>>> the mptcp project struck me as an interesting way to share TCP
>traffic
>>> between links.
>>>
>>> I've followed the intructions from the site and I was able to
>generate
>>> subflows on multiple WAN interfaces when reaching iperf server
>>> multipath-tcp.org. Im not able though to generate the same multipath
>flows
>>> when testing with my server using iperf. I've installed the mptcp
>enabled
>>> kernel on debian wheezy by installing the linux-mptcp package from
>the repo,
>>> both on the server and the client side, as mentioned on the site. Am
>I
>>> missing sth? The end goal is to use the server as the a mptcp proxy,
>to
>>> proxy and forward all mptcp traffic to plain tcp. The mptcp teaffic
>will be
>>> generated from a remote location, which has its own mptcp enabled
>router
>>> (debian wheezy with linux-mptcp installed) which will terminate to
>the
>>> server with at least two vpn tunnels. The server then, having a fast
>wan
>>> connection, will route the traffic to final destination with plain
>tcp, then
>>> back again with mptcp.
>>>
>>> Any ideas or hints will be appreciated.
>
>Some people have reported to use OpenVPN in TCP mode for this use case.
>
>Others have deployed SOCKS based solutions for this use case. The 
>OverTheBox solution that is being developed by OVH relies on encrypted 
>SOCKS tunnels to aggregate DSL links. This is probably very close to 
>what you have in mind. The client side of their solution is open-source
>
>and builds upon OpenWrt, see https://github.com/ovh/overthebox-openwrt
>for the current source. This version runs on x86 servers and there is 
>documentation in French on https://www.ovhtelecom.fr/overthebox/
>I don't think that they have released the server side of this service.
>
>Olivier

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
------9LJT0FZVNB2X2DOR6LJFB2OXS446VG
Content-Type: text/html;
 charset=utf-8
Content-Transfer-Encoding: 8bit

<html><head></head><body>Thanx Olivier and Yoshifumi,<br>
<br>
I will divert the thread to dev list.<br>
It will be very interesting to discuss further on this topic.<br>
<br>
Cheers,<br>
Alex<br>
<br><br><div class="gmail_quote">On December 4, 2015 8:50:18 PM EET, Olivier Bonaventure &lt;Olivier.Bonaventure@uclouvain.be&gt; wrote:<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<pre class="k9mail">Alex, Yoshifumi<br /><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #729fcf; padding-left: 1ex;"><br /> I think what you're trying to do is a reasonable use case for mptcp.<br /> Have you checked linux mptcp developer list? It seems that this is<br /> linux related to stuff rather than protocol spec.<br /> If not, please check  <a href="https://listes-2.sipr.ucl.ac.be/sympa/info/mptcp-dev">https://listes-2.sipr.ucl.ac.be/sympa/info/mptcp-dev</a><br /></blockquote><br />I also suggest to move this discussion to the mptcp-dev mailing list.<br /><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #729fcf; padding-left: 1ex;"><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #ad7fa8; padding-left: 1ex;"><br /> I've been searching on ways to aggregate bandwidth of multiple WAN links and<br /> the mptcp project struck me as an interesting way to share
TCP traffic<br /> between links.<br /><br /> I've followed the intructions from the site and I was able to generate<br /> subflows on multiple WAN interfaces when reaching iperf server<br /> <a href="http://multipath-tcp.org">multipath-tcp.org</a>. Im not able though to generate the same multipath flows<br /> when testing with my server using iperf. I've installed the mptcp enabled<br /> kernel on debian wheezy by installing the linux-mptcp package from the repo,<br /> both on the server and the client side, as mentioned on the site. Am I<br /> missing sth? The end goal is to use the server as the a mptcp proxy, to<br /> proxy and forward all mptcp traffic to plain tcp. The mptcp teaffic will be<br /> generated from a remote location, which has its own mptcp enabled router<br /> (debian wheezy with linux-mptcp installed) which will terminate to the<br /> server with at least two vpn tunnels. The server then, having a fast wan<br /> connection, will route the traffic to final
destination with plain tcp, then<br /> back again with mptcp.<br /><br /> Any ideas or hints will be appreciated.<br /></blockquote></blockquote><br />Some people have reported to use OpenVPN in TCP mode for this use case. <br />Others have deployed SOCKS based solutions for this use case. The <br />OverTheBox solution that is being developed by OVH relies on encrypted <br />SOCKS tunnels to aggregate DSL links. This is probably very close to <br />what you have in mind. The client side of their solution is open-source <br />and builds upon OpenWrt, see <a href="https://github.com/ovh/overthebox-openwrt">https://github.com/ovh/overthebox-openwrt</a><br />for the current source. This version runs on x86 servers and there is <br />documentation in French on <a href="https://www.ovhtelecom.fr/overthebox">https://www.ovhtelecom.fr/overthebox</a>/<br />I don't think that they have released the server side of this service.<br /><br />Olivier<br /><br /></pre></blockquote></div><br>
-- <br>
Sent from my Android device with K-9 Mail. Please excuse my brevity.</body></html>
------9LJT0FZVNB2X2DOR6LJFB2OXS446VG--


From nobody Mon Dec 14 09:30:10 2015
Return-Path: <wwwrun@rfc-editor.org>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 9FAAE1A701C for <multipathtcp@ietfa.amsl.com>; Mon, 14 Dec 2015 07:48:03 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.912
X-Spam-Level: 
X-Spam-Status: No, score=-106.912 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_HI=-5, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_WHITELIST=-100] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id EQY6vqDU99RX for <multipathtcp@ietfa.amsl.com>; Mon, 14 Dec 2015 07:48:02 -0800 (PST)
Received: from rfc-editor.org (rfc-editor.org [4.31.198.49]) by ietfa.amsl.com (Postfix) with ESMTP id 64EEA1A854B for <multipathtcp@ietf.org>; Mon, 14 Dec 2015 07:47:48 -0800 (PST)
Received: by rfc-editor.org (Postfix, from userid 30) id 689FE180004; Mon, 14 Dec 2015 07:45:27 -0800 (PST)
To: marcelo@it.uc3m.es, christoph.paasch@gmail.com, fgont@si6networks.com, Olivier.Bonaventure@uclouvain.be, costin.raiciu@cs.pub.ro, spencerdawkins.ietf@gmail.com, mls.ietf@gmail.com, philip.eardley@bt.com, nishida@sfc.wide.ad.jp
X-PHP-Originating-Script: 30:errata_mail_lib.php
From: RFC Errata System <rfc-editor@rfc-editor.org>
Message-Id: <20151214154527.689FE180004@rfc-editor.org>
Date: Mon, 14 Dec 2015 07:45:27 -0800 (PST)
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/nGVdMZgFwUgFFXKY19bNL52mGzA>
X-Mailman-Approved-At: Mon, 14 Dec 2015 09:30:09 -0800
Cc: multipathtcp@ietf.org, rfc-editor@rfc-editor.org
Subject: [multipathtcp] [Technical Errata Reported] RFC7430 (4565)
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 14 Dec 2015 15:48:03 -0000

The following errata report has been submitted for RFC7430,
"Analysis of Residual Threats and Possible Fixes for Multipath TCP (MPTCP)".

--------------------------------------
You may review the report below and at:
http://www.rfc-editor.org/errata_search.php?rfc=7430&eid=4565

--------------------------------------
Type: Technical
Reported by: Fabrizio Demaria <fabrizio.demaria@intel.com>

Section: 6

Original Text
-------------
   Summary of the attack:

      Type of attack: An attacker that can intercept the SYN/JOIN
      message can alter the source address being added.

      Type of attacker: partial-time on-path eavesdropper

   Description:

   The attacker is present along the path when the SYN/JOIN exchange
   takes place.  This allows the attacker to add any new address it
   wants to by simply substituting the source address of the SYN/JOIN
   packet for one it chooses.  This vulnerability was readily identified
   when designing the MPTCP security solution [RFC6181], and the threat
   was considered acceptable.

Corrected Text
--------------
   Summary of the attack:

      Type of attack: An attacker that can intercept the SYN/JOIN
      message can alter the source address being added.

      Type of attacker: partial-time on-path active attacker

   Description:

   The attacker is present along the path when the SYN/JOIN exchange
   takes place.  This allows the attacker to add any new address it
   wants to by simply substituting the source address of the SYN/JOIN
   packet for one it chooses.  This vulnerability was readily identified
   when designing the MPTCP security solution [RFC6181], and the threat
   was considered acceptable.

Notes
-----
As noted in section 1, an active attacker is able to change, discard, or delay some of the packets of the MPTCP session. This coincide with the description of the SYN/JOIN attack in section 6.

Instructions:
-------------
This erratum is currently posted as "Reported". If necessary, please
use "Reply All" to discuss whether it should be verified or
rejected. When a decision is reached, the verifying party (IESG)
can log in to change the status and edit the report, if necessary. 

--------------------------------------
RFC7430 (draft-ietf-mptcp-attacks-04)
--------------------------------------
Title               : Analysis of Residual Threats and Possible Fixes for Multipath TCP (MPTCP)
Publication Date    : July 2015
Author(s)           : M. Bagnulo, C. Paasch, F. Gont, O. Bonaventure, C. Raiciu
Category            : INFORMATIONAL
Source              : Multipath TCP
Area                : Transport
Stream              : IETF
Verifying Party     : IESG


From nobody Tue Dec 15 05:10:37 2015
Return-Path: <prvs=0791f6e550=zeeshan.afzal@kau.se>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id E9C0F1A88CB for <multipathtcp@ietfa.amsl.com>; Tue, 15 Dec 2015 05:07:44 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.849
X-Spam-Level: 
X-Spam-Status: No, score=-3.849 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HELO_EQ_SE=0.35, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-2.3] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Fo0foZxKmRMk for <multipathtcp@ietfa.amsl.com>; Tue, 15 Dec 2015 05:07:41 -0800 (PST)
Received: from tiger.dc.kau.se (smtp.kau.se [193.10.220.38]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 5481A1A88DB for <multipathtcp@ietf.org>; Tue, 15 Dec 2015 05:07:20 -0800 (PST)
X-Spam-Processed: mail.kau.se, Tue, 15 Dec 2015 14:07:14 +0100 (not processed: spam filter heuristic analysis disabled)
X-MDHelo: WorldClient
X-MDArrival-Date: Tue, 15 Dec 2015 14:07:14 +0100
X-Authenticated-Sender: zeesafza@kau.se
X-Return-Path: zeeshan.afzal@kau.se
X-Envelope-From: zeeshan.afzal@kau.se
Date: Tue, 15 Dec 2015 14:07:14 +0100
From: "Zeeshan Afzal" <zeeshan.afzal@kau.se>
To: marcelo@it.uc3m.es, christoph.paasch@gmail.com, fgont@si6networks.com, Olivier.Bonaventure@uclouvain.be, costin.raiciu@cs.pub.ro, spencerdawkins.ietf@gmail.com, mls.ietf@gmail.com, philip.eardley@bt.com, nishida@sfc.wide.ad.jp
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="1215-1307-14-03-PART_BREAK"
Message-ID: <WC20151215130714.101AB4@kau.se>
X-Mailer: WorldClient 15.5.2
In-Reply-To: <20151214154527.689FE180004@rfc-editor.org>
References: <20151214154527.689FE180004@rfc-editor.org>
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/LwwJe2Foja4o0_gsAI3TEYb5SB0>
X-Mailman-Approved-At: Tue, 15 Dec 2015 05:10:36 -0800
Cc: multipathtcp@ietf.org
Subject: Re: [multipathtcp] [Technical Errata Reported] RFC7430 (4565)
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 15 Dec 2015 13:07:45 -0000

--1215-1307-14-03-PART_BREAK
Content-Type: text/plain; charset="us-ascii"

Hi all,

After seeing the errata report, I observed something in Section 3 (DoS 
Attack on MP_JOIN). 

Current text:

The attacker needs to know the token TB in order to perform the
described attack.  This can be achieved if it is a partial-time on-
path eavesdropper observing the 3-way handshake of the establishment
of an additional subflow between Host A and Host B.  If the attacker
is never on-path, it has to guess the 32-bit token.

Can't the attacker instead eavesdrop on the keys exchanged in the initial 
connection establishment handshake and then generate the same token?  
Eavesdropping on the subflow establishment handshake is not the only way to 
know the token I guess. I am not sure if it needs to be clarified though.

Zeeshan



-----Original Message-----
From: RFC Errata System <rfc-editor@rfc-editor.org>
To: marcelo@it.uc3m.es, christoph.paasch@gmail.com, fgont@si6networks.com, 
Olivier.Bonaventure@uclouvain.be, costin.raiciu@cs.pub.ro, 
spencerdawkins.ietf@gmail.com, mls.ietf@gmail.com, philip.eardley@bt.com, 
nishida@sfc.wide.ad.jp
Cc: multipathtcp@ietf.org, rfc-editor@rfc-editor.org
Date: Mon, 14 Dec 2015 07:45:27 -0800 (PST)
Subject: [multipathtcp] [Technical Errata Reported] RFC7430 (4565)

The following errata report has been submitted for RFC7430,
"Analysis of Residual Threats and Possible Fixes for Multipath TCP (MPTCP)".

--------------------------------------
You may review the report below and at:
http://www.rfc-editor.org/errata_search.php?rfc=7430&eid=4565

--------------------------------------
Type: Technical
Reported by: Fabrizio Demaria <fabrizio.demaria@intel.com>

Section: 6

Original Text
-------------
   Summary of the attack:

      Type of attack: An attacker that can intercept the SYN/JOIN
      message can alter the source address being added.

      Type of attacker: partial-time on-path eavesdropper

   Description:

   The attacker is present along the path when the SYN/JOIN exchange
   takes place.  This allows the attacker to add any new address it
   wants to by simply substituting the source address of the SYN/JOIN
   packet for one it chooses.  This vulnerability was readily identified
   when designing the MPTCP security solution [RFC6181], and the threat
   was considered acceptable.

Corrected Text
--------------
   Summary of the attack:

      Type of attack: An attacker that can intercept the SYN/JOIN
      message can alter the source address being added.

      Type of attacker: partial-time on-path active attacker

   Description:

   The attacker is present along the path when the SYN/JOIN exchange
   takes place.  This allows the attacker to add any new address it
   wants to by simply substituting the source address of the SYN/JOIN
   packet for one it chooses.  This vulnerability was readily identified
   when designing the MPTCP security solution [RFC6181], and the threat
   was considered acceptable.

Notes
-----
As noted in section 1, an active attacker is able to change, discard, or 
delay some of the packets of the MPTCP session. This coincide with the 
description of the SYN/JOIN attack in section 6.

Instructions:
-------------
This erratum is currently posted as "Reported". If necessary, please
use "Reply All" to discuss whether it should be verified or
rejected. When a decision is reached, the verifying party (IESG)
can log in to change the status and edit the report, if necessary.

--------------------------------------
RFC7430 (draft-ietf-mptcp-attacks-04)
--------------------------------------
Title               : Analysis of Residual Threats and Possible Fixes for 
Multipath TCP (MPTCP)
Publication Date    : July 2015
Author(s)           : M. Bagnulo, C. Paasch, F. Gont, O. Bonaventure, C. 
Raiciu
Category            : INFORMATIONAL
Source              : Multipath TCP
Area                : Transport
Stream              : IETF
Verifying Party     : IESG

_______________________________________________
multipathtcp mailing list
multipathtcp@ietf.org
https://www.ietf.org/mailman/listinfo/multipathtcp
--1215-1307-14-03-PART_BREAK
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html><body>
<div style=3D"font-size: 13.3333px; font-family: tahoma; color: rgb(0, 0, 0=
); font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">Hi all,</div>

<div style=3D"font-size: 13.3333px; font-family: tahoma; color: rgb(0, 0, 0=
); font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">&nbsp;</div>

<div style=3D"font-size: 13.3333px; font-family: tahoma; color: rgb(0, 0, 0=
); font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">After seeing the errata=
&nbsp;report,&nbsp;I observed something in Section 3 (DoS Attack on MP_JOIN=
).&nbsp;</div>

<div style=3D"font-size: 13.3333px; font-family: tahoma; color: rgb(0, 0, 0=
); font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">&nbsp;</div>

<div style=3D"font-size: 13.3333px; font-family: tahoma; color: rgb(0, 0, 0=
); font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">Current text:</div>

<div style=3D"font-size: 13.3333px; font-family: tahoma; color: rgb(0, 0, 0=
); font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">&nbsp;</div>

<div style=3D"font-size: 13.3333px; font-family: tahoma; color: rgb(0, 0, 0=
); font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">
<pre class=3D"newpage" style=3D"font-size: 13.3333px; margin-top: 0px; marg=
in-bottom: 0px; page-break-before: always;">
The attacker needs to know the token TB in order to perform the
described attack.  This can be achieved if it is a partial-time on-
path eavesdropper observing the 3-way handshake of the establishment
of an additional subflow between Host A and Host B.  If the attacker
is never on-path, it has to guess the 32-bit token.</pre>
</div>

<div style=3D"font-family: tahoma; color: rgb(0, 0, 0); font-size: 13.3333p=
x; font-weight: 400; font-style: normal; background: none 0% 0% / auto repe=
at scroll padding-box border-box rgba(0, 0, 0, 0);">
<div>&nbsp;</div>

<div>Can&#39;t the attacker instead&nbsp;eavesdrop on the keys exchanged in=
 the initial connection establishment handshake and then generate the same =
token? &nbsp;Eavesdropping on the subflow establishment handshake is not th=
e only way to know the token I guess. I am not sure if it needs to be clari=
fied though.</div>

<div>&nbsp;</div>

<div>Zeeshan</div>

<div>&nbsp;</div>

<div>&nbsp;</div>

<div>&nbsp;</div>

<blockquote style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5p=
x; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">-----Original Message=
-----<br />
From: RFC Errata System &lt;rfc-editor@rfc-editor.org&gt;<br />
To: marcelo@it.uc3m.es, christoph.paasch@gmail.com, fgont@si6networks.com, =
Olivier.Bonaventure@uclouvain.be, costin.raiciu@cs.pub.ro, spencerdawkins.i=
etf@gmail.com, mls.ietf@gmail.com, philip.eardley@bt.com, nishida@sfc.wide.=
ad.jp<br />
Cc: multipathtcp@ietf.org, rfc-editor@rfc-editor.org<br />
Date: Mon, 14 Dec 2015 07:45:27 -0800 (PST)<br />
Subject: [multipathtcp] [Technical Errata Reported] RFC7430 (4565)<br />
&nbsp;
<div style=3D"font-family: monospace, courier new, courier">The following e=
rrata report has been submitted for RFC7430,<br />
&quot;Analysis of Residual Threats and Possible Fixes for Multipath TCP (MP=
TCP)&quot;.<br />
<br />
--------------------------------------<br />
You may review the report below and at:<br />
<a href=3D"http://www.rfc-editor.org/errata_search.php?rfc=3D7430&amp;eid=
=3D4565" target=3D"_blank">http://www.rfc-editor.org/errata_search.php?rfc=
=3D7430&amp;eid=3D4565</a><br />
<br />
--------------------------------------<br />
Type: Technical<br />
Reported by: Fabrizio Demaria &lt;<a href=3D"mailto:fabrizio.demaria%40inte=
l.com">fabrizio.demaria@intel.com</a>&gt;<br />
<br />
Section: 6<br />
<br />
Original Text<br />
-------------<br />
&nbsp; &nbsp;Summary of the attack:<br />
<br />
&nbsp; &nbsp; &nbsp; Type of attack: An attacker that can intercept the SYN=
/JOIN<br />
&nbsp; &nbsp; &nbsp; message can alter the source address being added.<br />
<br />
&nbsp; &nbsp; &nbsp; Type of attacker: partial-time on-path eavesdropper<br=
 />
<br />
&nbsp; &nbsp;Description:<br />
<br />
&nbsp; &nbsp;The attacker is present along the path when the SYN/JOIN excha=
nge<br />
&nbsp; &nbsp;takes place. &nbsp;This allows the attacker to add any new add=
ress it<br />
&nbsp; &nbsp;wants to by simply substituting the source address of the SYN/=
JOIN<br />
&nbsp; &nbsp;packet for one it chooses. &nbsp;This vulnerability was readil=
y identified<br />
&nbsp; &nbsp;when designing the MPTCP security solution [RFC6181], and the =
threat<br />
&nbsp; &nbsp;was considered acceptable.<br />
<br />
Corrected Text<br />
--------------<br />
&nbsp; &nbsp;Summary of the attack:<br />
<br />
&nbsp; &nbsp; &nbsp; Type of attack: An attacker that can intercept the SYN=
/JOIN<br />
&nbsp; &nbsp; &nbsp; message can alter the source address being added.<br />
<br />
&nbsp; &nbsp; &nbsp; Type of attacker: partial-time on-path active attacker=
<br />
<br />
&nbsp; &nbsp;Description:<br />
<br />
&nbsp; &nbsp;The attacker is present along the path when the SYN/JOIN excha=
nge<br />
&nbsp; &nbsp;takes place. &nbsp;This allows the attacker to add any new add=
ress it<br />
&nbsp; &nbsp;wants to by simply substituting the source address of the SYN/=
JOIN<br />
&nbsp; &nbsp;packet for one it chooses. &nbsp;This vulnerability was readil=
y identified<br />
&nbsp; &nbsp;when designing the MPTCP security solution [RFC6181], and the =
threat<br />
&nbsp; &nbsp;was considered acceptable.<br />
<br />
Notes<br />
-----<br />
As noted in section 1, an active attacker is able to change, discard, or de=
lay some of the packets of the MPTCP session. This coincide with the descri=
ption of the SYN/JOIN attack in section 6.<br />
<br />
Instructions:<br />
-------------<br />
This erratum is currently posted as &quot;Reported&quot;. If necessary, ple=
ase<br />
use &quot;Reply All&quot; to discuss whether it should be verified or<br />
rejected. When a decision is reached, the verifying party (IESG)<br />
can log in to change the status and edit the report, if necessary.<br />
<br />
--------------------------------------<br />
RFC7430 (draft-ietf-mptcp-attacks-04)<br />
--------------------------------------<br />
Title &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; : Analysis of Residu=
al Threats and Possible Fixes for Multipath TCP (MPTCP)<br />
Publication Date &nbsp; &nbsp;: July 2015<br />
Author(s) &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; : M. Bagnulo, C. Paasch, F. Go=
nt, O. Bonaventure, C. Raiciu<br />
Category &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;: INFORMATIONAL<br />
Source &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;: Multipath TCP<br />
Area &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;: Transport<br =
/>
Stream &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;: IETF<br />
Verifying Party &nbsp; &nbsp; : IESG<br />
<br />
_______________________________________________<br />
multipathtcp mailing list<br />
<a href=3D"mailto:multipathtcp%40ietf.org">multipathtcp@ietf.org</a><br />
<a href=3D"https://www.ietf.org/mailman/listinfo/multipathtcp" target=3D"_b=
lank">https://www.ietf.org/mailman/listinfo/multipathtcp</a></div>
</blockquote>
</div>
</body></html>

--1215-1307-14-03-PART_BREAK--



From nobody Tue Dec 15 17:21:33 2015
Return-Path: <cpaasch@apple.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 55DDA1A013F for <multipathtcp@ietfa.amsl.com>; Tue, 15 Dec 2015 16:58:22 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.311
X-Spam-Level: 
X-Spam-Status: No, score=-4.311 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id P4lX_mhcddMU for <multipathtcp@ietfa.amsl.com>; Tue, 15 Dec 2015 16:58:20 -0800 (PST)
Received: from mail-in5.apple.com (mail-out5.apple.com [17.151.62.27]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 01F851A010B for <multipathtcp@ietf.org>; Tue, 15 Dec 2015 16:58:19 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; d=apple.com; s=mailout2048s; c=relaxed/simple;  q=dns/txt; i=@apple.com; t=1450227499; x=2314141099; h=From:Sender:Reply-To:Subject:Date:Message-id:To:Cc:MIME-version:Content-type: Content-Transfer-Encoding:Content-ID:Content-Description:Resent-Date:Resent-From: Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:In-reply-to:References:List-Id: List-Help:List-Unsubscribe:List-Subscribe:List-Post:List-Owner:List-Archive; bh=JK7W3ZVcbh+Fh/kdBdqSqMAi3I8TiQep/Wlf5VMowzY=; b=ISXFabW0KcLuNwjZJxLwWu//I9WX7nMXDFXdw0IHrbrT5tVFZwTmPRqw6TNcJkEG JYn4+RPeJ7PI6Dv/ae4g9XRlh8iam0TifXsX406KAB8dUDR9PTQIYE4/NN9lqFoJ +cNOnmrqWl/7Hiie4ZLPRUwpRbJUHunYS0M3JOE1r9/hGUbNa4DqAE5Qimzg4hmv cTTWrbpf2V1dO3wSVC6pugICVO8InHUzbJbG8p5EVXEUp5m5hSekXoN0apE8I9XD 3WVr0eB5qgbiCYGtRka3QYQ6BeB3gUHXA2l56NKYsNjutg2dyQ8rWArwWcZG3+dR NFzhWPalAeIC+tVAnaxGBA==;
Received: from relay6.apple.com (relay6.apple.com [17.128.113.90]) by mail-in5.apple.com (Apple Secure Mail Relay) with SMTP id 36.B2.13397.B27B0765; Tue, 15 Dec 2015 16:58:19 -0800 (PST)
X-AuditID: 11973e13-f798b6d000003455-be-5670b72bf0f8
Received: from chive.apple.com (chive.apple.com [17.128.115.15]) (using TLS with cipher DHE-RSA-AES128-SHA (128/128 bits)) (Client did not present a certificate) by relay6.apple.com (Apple SCV relay) with SMTP id 32.F7.29392.B27B0765; Tue, 15 Dec 2015 16:58:19 -0800 (PST)
Received: from localhost ([17.114.152.118]) by chive.apple.com (Oracle Communications Messaging Server 7.0.5.35.0 64bit (built Mar 31 2015)) with ESMTPSA id <0NZF00LX0EP6SV00@chive.apple.com> for multipathtcp@ietf.org; Tue, 15 Dec 2015 16:58:19 -0800 (PST)
Sender: cpaasch@apple.com
Date: Tue, 15 Dec 2015 16:58:19 -0800
From: Christoph Paasch <cpaasch@apple.com>
To: Zeeshan Afzal <zeeshan.afzal@kau.se>
Message-id: <20151216005819.GJ84348@Chimay.local>
References: <20151214154527.689FE180004@rfc-editor.org> <WC20151215130714.101AB4@kau.se>
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii
Content-disposition: inline
In-reply-to: <WC20151215130714.101AB4@kau.se>
User-Agent: Mutt/1.5.24 (08e81162482f) (2015-08-30)
X-Brightmail-Tracker: H4sIAAAAAAAAA+NgFjrALMWRmVeSWpSXmKPExsUi2FAYpau9vSDM4OBWeYvPq6+zOTB6LFny kymAMYrLJiU1J7MstUjfLoEr48Cd24wFh1Qrvs3cwNjAeFK2i5GTQ0LARGLCv39MELaYxIV7 69m6GLk4hAT2Mkrs29rDDFO0/VY7M0RiGpPE64ftrBBOG5PEt4ffWEGqhAUkJbrv3AHrYBFQ lbh2+RfYWDYBLYm3t9vBakQENCTWXF8NtoJZoJlJYt7za8wQzW4SXSf/s4HYvAKGEheOvWIB sYUEoiT+floOFReU+DH5HlicGWjo+p3HmSBsaYlHf2ewg9icAroSP/a1M4LYogLGEk1H5oCd LSGwgE3i5oJVbBMYRWYhmTULyaxZSGYtYGRexSiUm5iZo5uZZ6qXWFCQk6qXnJ+7iREU5NPt hHcwnl5ldYhRgINRiYf3QmRBmBBrYllxZe4hRmkOFiVx3jdhQCGB9MSS1OzU1ILUovii0pzU 4kOMTBycUg2Mh9Z3hnG++FXOsODAfK0zBaZ2QSzvynaZz19ZUpRT5TN1p8jrr8vaMht/t3LN eyDDzCKT/bAlwM1t/3dvvyPSUzbZTGK1OWH/s/Cjzu4HzOVXbyWnrK3b1cm9L22WVmnn0vZH Gmxz3y3qMVvkZmD5PK36yQNNR5V2ZY6K68nr72zt7r/iLvteiaU4I9FQi7moOBEAA+C0+VMC AAA=
X-Brightmail-Tracker: H4sIAAAAAAAAA+NgFvrNLMWRmVeSWpSXmKPExsUi2FDMr6u9vSDM4OESKYvPq6+zOTB6LFny kymAMYrLJiU1J7MstUjfLoEr48Cd24wFh1Qrvs3cwNjAeFK2i5GTQ0LARGL7rXZmCFtM4sK9 9WxdjFwcQgLTmCReP2xnhXDamCS+PfzGClIlLCAp0X3nDlgHi4CqxLXLv5hAbDYBLYm3t9vB akQENCTWXF8NNolZoJlJYt7za8wQzW4SXSf/s4HYvAKGEheOvWIBsYUEoiT+floOFReU+DH5 HlicGWjo+p3HmSBsaYlHf2ewg9icAroSP/a1M4LYogLGEk1H5jBPYBSchaR9FpL2WUjaFzAy r2IUKErNSaw000ssKMhJ1UvOz93ECA7KwqgdjA3LrQ4xCnAwKvHwXogsCBNiTSwrrsw9xCjB wawkwtsdDxTiTUmsrEotyo8vKs1JLT7EmAwMiYnMUqLJ+cCIySuJNzQxMTAxNjYzNjY3MSdN WEmcV8IaaIVAemJJanZqakFqEcwWJg5OqQZGOdXFt6Vm9H4ua1LlZBdVenvrne7baVEBhTf0 9h/YfqT9TC53zRPWDeuE8j76zbXbVKGkvOrsoZSGpRYb359MFThqU+WneHPDd/eZn0r7P/rM PlO38I11kfj/mGAdpae/rS5v9BezM+HWepb28F/Fxtjd2iVbpGP2yn69UMk+neG7v/3jTyf2 KbEUZyQaajEXFScCAAI5FyWOAgAA
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/r9lEYLyBye00BHj8HmxjTd8533M>
X-Mailman-Approved-At: Tue, 15 Dec 2015 17:21:32 -0800
Cc: costin.raiciu@cs.pub.ro, christoph.paasch@gmail.com, multipathtcp@ietf.org, mls.ietf@gmail.com, fgont@si6networks.com
Subject: Re: [multipathtcp] [Technical Errata Reported] RFC7430 (4565)
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 16 Dec 2015 00:58:22 -0000

Hello,

On 15/12/15 - 14:07:14, Zeeshan Afzal wrote:
> After seeing the errata report, I observed something in Section 3 (DoS 
> Attack on MP_JOIN). 
> 
> Current text:
> 
> The attacker needs to know the token TB in order to perform the
> described attack.  This can be achieved if it is a partial-time on-
> path eavesdropper observing the 3-way handshake of the establishment
> of an additional subflow between Host A and Host B.  If the attacker
> is never on-path, it has to guess the 32-bit token.
> 
> Can't the attacker instead eavesdrop on the keys exchanged in the initial 
> connection establishment handshake and then generate the same token?  
> Eavesdropping on the subflow establishment handshake is not the only way to 
> know the token I guess. I am not sure if it needs to be clarified though.

I think this would rather be the "Eavesdropper in the Initial
Handshake"-attack. As the attacker's possibilities to do harm are much higher
here, I don't think that we should combine this with the DoS Attack on
MP_JOIN.


Christoph


> -----Original Message-----
> From: RFC Errata System <rfc-editor@rfc-editor.org>
> To: marcelo@it.uc3m.es, christoph.paasch@gmail.com, fgont@si6networks.com, 
> Olivier.Bonaventure@uclouvain.be, costin.raiciu@cs.pub.ro, 
> spencerdawkins.ietf@gmail.com, mls.ietf@gmail.com, philip.eardley@bt.com, 
> nishida@sfc.wide.ad.jp
> Cc: multipathtcp@ietf.org, rfc-editor@rfc-editor.org
> Date: Mon, 14 Dec 2015 07:45:27 -0800 (PST)
> Subject: [multipathtcp] [Technical Errata Reported] RFC7430 (4565)
> 
> The following errata report has been submitted for RFC7430,
> "Analysis of Residual Threats and Possible Fixes for Multipath TCP (MPTCP)".
> 
> --------------------------------------
> You may review the report below and at:
> http://www.rfc-editor.org/errata_search.php?rfc=7430&eid=4565
> 
> --------------------------------------
> Type: Technical
> Reported by: Fabrizio Demaria <fabrizio.demaria@intel.com>
> 
> Section: 6
> 
> Original Text
> -------------
>    Summary of the attack:
> 
>       Type of attack: An attacker that can intercept the SYN/JOIN
>       message can alter the source address being added.
> 
>       Type of attacker: partial-time on-path eavesdropper
> 
>    Description:
> 
>    The attacker is present along the path when the SYN/JOIN exchange
>    takes place.  This allows the attacker to add any new address it
>    wants to by simply substituting the source address of the SYN/JOIN
>    packet for one it chooses.  This vulnerability was readily identified
>    when designing the MPTCP security solution [RFC6181], and the threat
>    was considered acceptable.
> 
> Corrected Text
> --------------
>    Summary of the attack:
> 
>       Type of attack: An attacker that can intercept the SYN/JOIN
>       message can alter the source address being added.
> 
>       Type of attacker: partial-time on-path active attacker
> 
>    Description:
> 
>    The attacker is present along the path when the SYN/JOIN exchange
>    takes place.  This allows the attacker to add any new address it
>    wants to by simply substituting the source address of the SYN/JOIN
>    packet for one it chooses.  This vulnerability was readily identified
>    when designing the MPTCP security solution [RFC6181], and the threat
>    was considered acceptable.
> 
> Notes
> -----
> As noted in section 1, an active attacker is able to change, discard, or 
> delay some of the packets of the MPTCP session. This coincide with the 
> description of the SYN/JOIN attack in section 6.
> 
> Instructions:
> -------------
> This erratum is currently posted as "Reported". If necessary, please
> use "Reply All" to discuss whether it should be verified or
> rejected. When a decision is reached, the verifying party (IESG)
> can log in to change the status and edit the report, if necessary.
> 
> --------------------------------------
> RFC7430 (draft-ietf-mptcp-attacks-04)
> --------------------------------------
> Title               : Analysis of Residual Threats and Possible Fixes for 
> Multipath TCP (MPTCP)
> Publication Date    : July 2015
> Author(s)           : M. Bagnulo, C. Paasch, F. Gont, O. Bonaventure, C. 
> Raiciu
> Category            : INFORMATIONAL
> Source              : Multipath TCP
> Area                : Transport
> Stream              : IETF
> Verifying Party     : IESG
> 
> _______________________________________________
> multipathtcp mailing list
> multipathtcp@ietf.org
> https://www.ietf.org/mailman/listinfo/multipathtcp

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


From nobody Thu Dec 17 04:42:17 2015
Return-Path: <abhishekv.verma@gmail.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id BE9921B2B3F for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 04:42:11 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.701
X-Spam-Level: 
X-Spam-Status: No, score=0.701 tagged_above=-999 required=5 tests=[BAYES_50=0.8, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9_BUSupE6BoJ for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 04:42:10 -0800 (PST)
Received: from mail-ig0-x234.google.com (mail-ig0-x234.google.com [IPv6:2607:f8b0:4001:c05::234]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id EE4ED1ACE11 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 04:42:09 -0800 (PST)
Received: by mail-ig0-x234.google.com with SMTP id m11so11982069igk.1 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 04:42:09 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:date:message-id:subject:from:to:content-type; bh=QpH+9+WdZmx3dNvlQzjAeFGiYMTx5el1Tje6aJh3dms=; b=iUlHyj81xZ1W/tyEEmJuFCuoXOrfuhWCgi53TBiIJwis0ZXPwhtld58JRRXxEHGLk+ YI05RxfDvSvfwXaxNTOiJsXAbM4lRIE9wukin2vh2aG/RpR2TvPRDfTEPqlhr6D2x6aL VnMACy9AYyN5emNEIE7nHQGwfvECFd4Ux5P1pA+r1OaFrKMgQuK9YJJYQRy9t/wmsFKT +BLlSHusu0PNb+wHEwMO0Q0saLw3kb9evi3hdyRY5/x+OgH7WkUgJzfabICSaT3wep9q WEW/k2iNttX/PCgHmjztXcOf1gQ14GoOthv7UEhmNFX5cCwN935CNRWK1Ce4RFBXZAaD ivEg==
MIME-Version: 1.0
X-Received: by 10.50.65.39 with SMTP id u7mr3448467igs.85.1450356129357; Thu, 17 Dec 2015 04:42:09 -0800 (PST)
Received: by 10.64.62.101 with HTTP; Thu, 17 Dec 2015 04:42:09 -0800 (PST)
Date: Thu, 17 Dec 2015 18:12:09 +0530
Message-ID: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com>
From: Abhishek Verma <abhishekv.verma@gmail.com>
To: multipathtcp@ietf.org
Content-Type: multipart/alternative; boundary=047d7b2e1455e366100527175909
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/iH5ARjSGLn28oW_gp40nIAmGMnU>
Subject: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 17 Dec 2015 12:42:12 -0000

--047d7b2e1455e366100527175909
Content-Type: text/plain; charset=UTF-8

Hi TCP experts,

I have an observation that i am unable to explain and understand. Would
appreciate if folks here can make some sense out of this:

I have a fixed size file, 100 Mb.

I use multiple TCP streams to transfer the file over my network. I have a
10Mbps broadband connection.

There are times when i am able to fill up the pipe with very few TCP
streams (like 2-4) but there are times when i need more than 20 to fill up
my connection.

My understanding is as follows:

TCP Throughput = k (MSS/RTT)*(1/sqrt(Loss)) where k is some constant

In a file transfer between two end points the MSS and RTT are constant, so
the equation reduces to:

Throughput = k'/sqrt(Loss)

This is the throughput that i get with 1 TCP stream.
Hence, when am pushing N chunks, the throughput that i will get is
N*k/sqrt(Loss)

I hence assume that when i need more TCP streams to push my file out, the
Loss in the network is higher than when i need fewer TCP streams. This is
the part that i understand very well.

What i dont understand is this:

I figure out that i need, say 7, TCP streams to fill my 10Mbps pipe when
pushing out my file.

So, i if i do this experiment back to back 10 times, then while 80% of the
times i am able to fill the pipe with 7 streams. There are always these 20%
runs, where 7 streams gives me a BW thats lower than 10, say around 7-8
Mbps. This is something that i cannot explain.

Can somebody help me with this? Just give me some clues and i will take it
on from there.

I could rephrase my question like this:
Is it possible that for certain TCP sessions (the 20%) the Loss that i get
is more than the Loss that 80% of my TCP streams experience.

Thanks, Abhishek

--047d7b2e1455e366100527175909
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi TCP experts,<br><br>I have an observation that i am una=
ble to explain and understand. Would appreciate if folks here can make some=
 sense out of this:<br><br>I have a fixed size file, 100 Mb.<br><br>I use m=
ultiple TCP streams to transfer the file over my network. I have a 10Mbps b=
roadband connection.<br><br>There are times when i am able to fill up the p=
ipe with very few TCP streams (like 2-4) but there are times when i need mo=
re than 20 to fill up my connection.<br><br>My understanding is as follows:=
<br><br>TCP Throughput =3D k (MSS/RTT)*(1/sqrt(Loss)) where k is some const=
ant<div><br></div><div>In a file transfer between two end points the MSS an=
d RTT are constant, so the equation reduces to:</div><div><br></div><div>Th=
roughput =3D k&#39;/sqrt(Loss)</div><div><br></div><div>This is the through=
put that i get with 1 TCP stream.</div><div>Hence, when am pushing N chunks=
, the throughput that i will get is N*k/sqrt(Loss)</div><div><br></div><div=
>I hence assume that when i need more TCP streams to push my file out, the =
Loss in the network is higher than when i need fewer TCP streams. This is t=
he part that i understand very well.</div><div><br></div><div>What i dont u=
nderstand is this:</div><div><br></div><div>I figure out that i need, say 7=
, TCP streams to fill my 10Mbps pipe when pushing out my file.</div><div><b=
r></div><div>So, i if i do this experiment back to back 10 times, then whil=
e 80% of the times i am able to fill the pipe with 7 streams. There are alw=
ays these 20% runs, where 7 streams gives me a BW thats lower than 10, say =
around 7-8 Mbps. This is something that i cannot explain.</div><div><br></d=
iv><div>Can somebody help me with this? Just give me some clues and i will =
take it on from there.</div><div><br></div><div>I could rephrase my questio=
n like this:</div><div>Is it possible that for certain TCP sessions (the 20=
%) the Loss that i get is more than the Loss that 80% of my TCP streams exp=
erience.</div><div><br></div><div>Thanks, Abhishek</div></div>

--047d7b2e1455e366100527175909--


From nobody Thu Dec 17 04:58:44 2015
Return-Path: <c.raiciu@cs.ucl.ac.uk>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id E61061B2C34 for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 04:58:43 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.21
X-Spam-Level: 
X-Spam-Status: No, score=-4.21 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3lw_rcegBO9Z for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 04:58:41 -0800 (PST)
Received: from bells2.cs.ucl.ac.uk (bells2.cs.ucl.ac.uk [128.16.5.33]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 9AF8B1B2C28 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 04:58:41 -0800 (PST)
Received: from [141.85.225.202] (helo=[10.0.0.102]) by bells2.cs.ucl.ac.uk with esmtpsa (TLSv1:AES256-SHA:256) (C.Raiciu authenticated) (Exim 4.54) id 1a9Y8j-000LFK-Df; Thu, 17 Dec 2015 12:58:37 +0000
Content-Type: text/plain; charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2070.6\))
From: Costin Raiciu <c.raiciu@cs.ucl.ac.uk>
In-Reply-To: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com>
Date: Thu, 17 Dec 2015 14:58:35 +0200
Content-Transfer-Encoding: quoted-printable
Message-Id: <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com>
To: Abhishek Verma <abhishekv.verma@gmail.com>
X-Mailer: Apple Mail (2.2070.6)
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/PCgK_a3_hFMUlmiq4OhHsBtj3Gs>
Cc: MultiPath WG <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 17 Dec 2015 12:58:44 -0000

Hi,

Most likely your results show congestion upstream of your access link.

One TCP Cubic should be able to the 10Mbps of your link, provided there =
is no other bottleneck.

If there is another bottleneck, with many flows fighting for it, by =
increasing the number of TCP flows you are increasing your share of that =
bottleneck link, until you go to 10Mbps at which point your access link =
is becoming the bottleneck.

The variability you are seeing in the number of TCP connections is =
likely due to variability of upstream bottleneck capacity (e.g. how many =
other users=E2=80=99 flows you are fighting with).

Hope this helps.
Costin

> On 17 Dec 2015, at 14:42, Abhishek Verma <abhishekv.verma@gmail.com> =
wrote:
>=20
> Hi TCP experts,
>=20
> I have an observation that i am unable to explain and understand. =
Would appreciate if folks here can make some sense out of this:
>=20
> I have a fixed size file, 100 Mb.
>=20
> I use multiple TCP streams to transfer the file over my network. I =
have a 10Mbps broadband connection.
>=20
> There are times when i am able to fill up the pipe with very few TCP =
streams (like 2-4) but there are times when i need more than 20 to fill =
up my connection.
>=20
> My understanding is as follows:
>=20
> TCP Throughput =3D k (MSS/RTT)*(1/sqrt(Loss)) where k is some constant
>=20
> In a file transfer between two end points the MSS and RTT are =
constant, so the equation reduces to:
>=20
> Throughput =3D k'/sqrt(Loss)
>=20
> This is the throughput that i get with 1 TCP stream.
> Hence, when am pushing N chunks, the throughput that i will get is =
N*k/sqrt(Loss)
>=20
> I hence assume that when i need more TCP streams to push my file out, =
the Loss in the network is higher than when i need fewer TCP streams. =
This is the part that i understand very well.
>=20
> What i dont understand is this:
>=20
> I figure out that i need, say 7, TCP streams to fill my 10Mbps pipe =
when pushing out my file.
>=20
> So, i if i do this experiment back to back 10 times, then while 80% of =
the times i am able to fill the pipe with 7 streams. There are always =
these 20% runs, where 7 streams gives me a BW thats lower than 10, say =
around 7-8 Mbps. This is something that i cannot explain.
>=20
> Can somebody help me with this? Just give me some clues and i will =
take it on from there.
>=20
> I could rephrase my question like this:
> Is it possible that for certain TCP sessions (the 20%) the Loss that i =
get is more than the Loss that 80% of my TCP streams experience.
>=20
> Thanks, Abhishek
> _______________________________________________
> multipathtcp mailing list
> multipathtcp@ietf.org
> https://www.ietf.org/mailman/listinfo/multipathtcp


From nobody Thu Dec 17 06:09:15 2015
Return-Path: <abhishekv.verma@gmail.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 61E241B2DF0 for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 06:09:13 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.999
X-Spam-Level: 
X-Spam-Status: No, score=-1.999 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id eCCMApoHquYu for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 06:09:11 -0800 (PST)
Received: from mail-io0-x22f.google.com (mail-io0-x22f.google.com [IPv6:2607:f8b0:4001:c06::22f]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 086FC1B2DEC for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 06:09:11 -0800 (PST)
Received: by mail-io0-x22f.google.com with SMTP id q126so55857444iof.2 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 06:09:10 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=oCtuJXydMNNLHGV0nZCHKCFzM5puNWABlc4viPrafFc=; b=i2AfMToiRE6PFCh3vMueqcDk81tAUHTp7MMT2kJ7YOBtOMI+U9Lr3iOdkEiO50O1ym P7xROQz/myEaHnrXI0sWnXxHiV17y5R5Ub/PsIECK+fwDWnaCOgvXCQbLlxU5nEiu10m XPdI/UfXlxnQ9QxrdMb7s7TvbdoyVRK17dm10qRz9/K7DaMrkRZBQP6cT+sBtj9z5bn/ HZpy/9IWOUwu5yUIKkgjhI8KKk9/JCbubOlvJJXmbPK4h7gHiRZd+Yd8SxgBDuvWEJAi NFKUBgXIvo7CpBwpzi08IoZ0pf63TaMqC5Qcuz+0gS4LQMRtpprWjOakKSpPLAiLj68m 2z+A==
MIME-Version: 1.0
X-Received: by 10.107.134.166 with SMTP id q38mr47892081ioi.25.1450361350340;  Thu, 17 Dec 2015 06:09:10 -0800 (PST)
Received: by 10.64.62.101 with HTTP; Thu, 17 Dec 2015 06:09:10 -0800 (PST)
In-Reply-To: <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com> <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk>
Date: Thu, 17 Dec 2015 19:39:10 +0530
Message-ID: <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com>
From: Abhishek Verma <abhishekv.verma@gmail.com>
To: Costin Raiciu <c.raiciu@cs.ucl.ac.uk>
Content-Type: multipart/alternative; boundary=001a113f9c88154700052718914a
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/P-Wn1HYJq6WEEvxsB6A3FTVlIVo>
Cc: MultiPath WG <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 17 Dec 2015 14:09:13 -0000

--001a113f9c88154700052718914a
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Hi Costin,

Thanks a lot for the response. It was indeed very helpful.


> Most likely your results show congestion upstream of your access link.
>
> One TCP Cubic should be able to the 10Mbps of your link, provided there i=
s
> no other bottleneck.
>

It usually doesnt. I would guess then its because of the losses in the
network.

>
> If there is another bottleneck, with many flows fighting for it, by
> increasing the number of TCP flows you are increasing your share of that
> bottleneck link, until you go to 10Mbps at which point your access link i=
s
> becoming the bottleneck.
>

Yes, i agree.

I believe the throughput equation does not hold good when the losses are
induced because of us congesting the link. So, when i increase the # of TCP
streams to fill up the pipe, then that equation may not hold good.

Is this correct -- that, loss in that equation is the general network loss
and not the loss induced by congesting the link artifically?


>
> The variability you are seeing in the number of TCP connections is likely
> due to variability of upstream bottleneck capacity (e.g. how many other
> users=E2=80=99 flows you are fighting with).
>

Wow. Thats strange.

I assumed that in a 10Mbps broadband connection, i will always get 10Mbps
and that will be the bottle neck. Did not imagine that its possible that i
could get choked further upstream.

I have another observation that cannot get reconciled with getting choked
upstream:

I send two 100Mb files at the same time.

On one file i get a throughput of 7Mbps, while on the other i get 3Mbps.

Because of TCP fairness, shouldnt each file get 5Mbps?

Thanks, Abhishek



>
> Hope this helps.
> Costin
>
> > On 17 Dec 2015, at 14:42, Abhishek Verma <abhishekv.verma@gmail.com>
> wrote:
> >
> > Hi TCP experts,
> >
> > I have an observation that i am unable to explain and understand. Would
> appreciate if folks here can make some sense out of this:
> >
> > I have a fixed size file, 100 Mb.
> >
> > I use multiple TCP streams to transfer the file over my network. I have
> a 10Mbps broadband connection.
> >
> > There are times when i am able to fill up the pipe with very few TCP
> streams (like 2-4) but there are times when i need more than 20 to fill u=
p
> my connection.
> >
> > My understanding is as follows:
> >
> > TCP Throughput =3D k (MSS/RTT)*(1/sqrt(Loss)) where k is some constant
> >
> > In a file transfer between two end points the MSS and RTT are constant,
> so the equation reduces to:
> >
> > Throughput =3D k'/sqrt(Loss)
> >
> > This is the throughput that i get with 1 TCP stream.
> > Hence, when am pushing N chunks, the throughput that i will get is
> N*k/sqrt(Loss)
> >
> > I hence assume that when i need more TCP streams to push my file out,
> the Loss in the network is higher than when i need fewer TCP streams. Thi=
s
> is the part that i understand very well.
> >
> > What i dont understand is this:
> >
> > I figure out that i need, say 7, TCP streams to fill my 10Mbps pipe whe=
n
> pushing out my file.
> >
> > So, i if i do this experiment back to back 10 times, then while 80% of
> the times i am able to fill the pipe with 7 streams. There are always the=
se
> 20% runs, where 7 streams gives me a BW thats lower than 10, say around 7=
-8
> Mbps. This is something that i cannot explain.
> >
> > Can somebody help me with this? Just give me some clues and i will take
> it on from there.
> >
> > I could rephrase my question like this:
> > Is it possible that for certain TCP sessions (the 20%) the Loss that i
> get is more than the Loss that 80% of my TCP streams experience.
> >
> > Thanks, Abhishek
> > _______________________________________________
> > multipathtcp mailing list
> > multipathtcp@ietf.org
> > https://www.ietf.org/mailman/listinfo/multipathtcp
>
>

--001a113f9c88154700052718914a
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi Costin,<div><br></div><div>Thanks a lot for the respons=
e. It was indeed very helpful.</div><div><br></div><div class=3D"gmail_extr=
a"><div class=3D"gmail_quote"><blockquote class=3D"gmail_quote" style=3D"ma=
rgin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,=
204);border-left-style:solid;padding-left:1ex">
<br>
Most likely your results show congestion upstream of your access link.<br>
<br>
One TCP Cubic should be able to the 10Mbps of your link, provided there is =
no other bottleneck.<br></blockquote><div><br></div><div>It usually doesnt.=
 I would guess then its because of the losses in the network.</div><blockqu=
ote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-wid=
th:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-l=
eft:1ex"><br>
If there is another bottleneck, with many flows fighting for it, by increas=
ing the number of TCP flows you are increasing your share of that bottlenec=
k link, until you go to 10Mbps at which point your access link is becoming =
the bottleneck.<br></blockquote><div><br></div><div>Yes, i agree.</div><div=
><br></div><div>I believe the throughput equation does not hold good when t=
he losses are induced because of us congesting the link. So, when i increas=
e the # of TCP streams to fill up the pipe, then that equation may not hold=
 good.</div><div><br></div><div>Is this correct -- that, loss in that equat=
ion is the general network loss and not the loss induced by congesting the =
link artifically?</div><div>=C2=A0</div><blockquote class=3D"gmail_quote" s=
tyle=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rg=
b(204,204,204);border-left-style:solid;padding-left:1ex">
<br>
The variability you are seeing in the number of TCP connections is likely d=
ue to variability of upstream bottleneck capacity (e.g. how many other user=
s=E2=80=99 flows you are fighting with).<br></blockquote><div><br></div><di=
v>Wow. Thats strange.</div><div><br></div><div>I assumed that in a 10Mbps b=
roadband connection, i will always get 10Mbps and that will be the bottle n=
eck. Did not imagine that its possible that i could get choked further upst=
ream.</div><div><br></div><div>I have another observation that cannot get r=
econciled with getting choked upstream:</div><div><br></div><div>I send two=
 100Mb files at the same time.</div><div><br></div><div>On one file i get a=
 throughput of 7Mbps, while on the other i get 3Mbps.</div><div><br></div><=
div>Because of TCP fairness, shouldnt each file get 5Mbps?</div><div><br></=
div><div>Thanks, Abhishek</div><div><br></div><div>=C2=A0</div><blockquote =
class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1=
px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:=
1ex">
<br>
Hope this helps.<br>
Costin<br>
<div><div class=3D"h5"><br>
&gt; On 17 Dec 2015, at 14:42, Abhishek Verma &lt;<a href=3D"mailto:abhishe=
kv.verma@gmail.com">abhishekv.verma@gmail.com</a>&gt; wrote:<br>
&gt;<br>
&gt; Hi TCP experts,<br>
&gt;<br>
&gt; I have an observation that i am unable to explain and understand. Woul=
d appreciate if folks here can make some sense out of this:<br>
&gt;<br>
&gt; I have a fixed size file, 100 Mb.<br>
&gt;<br>
&gt; I use multiple TCP streams to transfer the file over my network. I hav=
e a 10Mbps broadband connection.<br>
&gt;<br>
&gt; There are times when i am able to fill up the pipe with very few TCP s=
treams (like 2-4) but there are times when i need more than 20 to fill up m=
y connection.<br>
&gt;<br>
&gt; My understanding is as follows:<br>
&gt;<br>
&gt; TCP Throughput =3D k (MSS/RTT)*(1/sqrt(Loss)) where k is some constant=
<br>
&gt;<br>
&gt; In a file transfer between two end points the MSS and RTT are constant=
, so the equation reduces to:<br>
&gt;<br>
&gt; Throughput =3D k&#39;/sqrt(Loss)<br>
&gt;<br>
&gt; This is the throughput that i get with 1 TCP stream.<br>
&gt; Hence, when am pushing N chunks, the throughput that i will get is N*k=
/sqrt(Loss)<br>
&gt;<br>
&gt; I hence assume that when i need more TCP streams to push my file out, =
the Loss in the network is higher than when i need fewer TCP streams. This =
is the part that i understand very well.<br>
&gt;<br>
&gt; What i dont understand is this:<br>
&gt;<br>
&gt; I figure out that i need, say 7, TCP streams to fill my 10Mbps pipe wh=
en pushing out my file.<br>
&gt;<br>
&gt; So, i if i do this experiment back to back 10 times, then while 80% of=
 the times i am able to fill the pipe with 7 streams. There are always thes=
e 20% runs, where 7 streams gives me a BW thats lower than 10, say around 7=
-8 Mbps. This is something that i cannot explain.<br>
&gt;<br>
&gt; Can somebody help me with this? Just give me some clues and i will tak=
e it on from there.<br>
&gt;<br>
&gt; I could rephrase my question like this:<br>
&gt; Is it possible that for certain TCP sessions (the 20%) the Loss that i=
 get is more than the Loss that 80% of my TCP streams experience.<br>
&gt;<br>
&gt; Thanks, Abhishek<br>
</div></div>&gt; _______________________________________________<br>
&gt; multipathtcp mailing list<br>
&gt; <a href=3D"mailto:multipathtcp@ietf.org">multipathtcp@ietf.org</a><br>
&gt; <a href=3D"https://www.ietf.org/mailman/listinfo/multipathtcp" rel=3D"=
noreferrer" target=3D"_blank">https://www.ietf.org/mailman/listinfo/multipa=
thtcp</a><br>
<br>
</blockquote></div><br></div></div>

--001a113f9c88154700052718914a--


From nobody Thu Dec 17 07:42:14 2015
Return-Path: <philip.eardley@bt.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 61F371B2CF1 for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 07:42:13 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.111
X-Spam-Level: 
X-Spam-Status: No, score=-0.111 tagged_above=-999 required=5 tests=[BAYES_20=-0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_21=0.6, RCVD_IN_DNSWL_LOW=-0.7, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id JTfDAuKEnKaO for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 07:42:10 -0800 (PST)
Received: from smtpb1.bt.com (smtpb1.bt.com [62.7.242.142]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 58BD21B2EAB for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 07:42:09 -0800 (PST)
Received: from EPDAG01D-UKBR.domain1.systemhost.net (193.113.197.205) by EVMED06-UKBR.bt.com (10.216.161.38) with Microsoft SMTP Server (TLS) id 14.3.195.1; Thu, 17 Dec 2015 15:41:57 +0000
Received: from rew09926dag03c.domain1.systemhost.net (10.55.202.26) by EPDAG01D-UKBR.domain1.systemhost.net (193.113.197.205) with Microsoft SMTP Server (TLS) id 15.0.1104.5; Thu, 17 Dec 2015 15:42:01 +0000
Received: from rew09926dag03b.domain1.systemhost.net (10.55.202.22) by rew09926dag03c.domain1.systemhost.net (10.55.202.26) with Microsoft SMTP Server (TLS) id 15.0.1104.5; Thu, 17 Dec 2015 15:26:57 +0000
Received: from rew09926dag03b.domain1.systemhost.net ([fe80::d514:fe50:560c:401e]) by rew09926dag03b.domain1.systemhost.net ([fe80::d514:fe50:560c:401e%12]) with mapi id 15.00.1104.000; Thu, 17 Dec 2015 15:26:57 +0000
From: <philip.eardley@bt.com>
To: <abhishekv.verma@gmail.com>, <c.raiciu@cs.ucl.ac.uk>
Thread-Topic: [multipathtcp] Loss and TCP Throughput
Thread-Index: AQHROMhiNRINhLRVSUmCkOloifxj757PI7aAgAATuACAABSDkA==
Date: Thu, 17 Dec 2015 15:26:57 +0000
Message-ID: <a2ed60a6d4d34747974b4ee7d39b675d@rew09926dag03b.domain1.systemhost.net>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com> <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk> <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com>
In-Reply-To: <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com>
Accept-Language: en-GB, en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-ms-exchange-transport-fromentityheader: Hosted
x-originating-ip: [10.55.202.233]
Content-Type: multipart/alternative; boundary="_000_a2ed60a6d4d34747974b4ee7d39b675drew09926dag03bdomain1sy_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/QldY2RMpur4Vc3ybIuQiE0ZWkrI>
Cc: multipathtcp@ietf.org
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 17 Dec 2015 15:42:13 -0000

--_000_a2ed60a6d4d34747974b4ee7d39b675drew09926dag03bdomain1sy_
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
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--_000_a2ed60a6d4d34747974b4ee7d39b675drew09926dag03bdomain1sy_
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: base64
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--_000_a2ed60a6d4d34747974b4ee7d39b675drew09926dag03bdomain1sy_--


From nobody Thu Dec 17 07:52:56 2015
Return-Path: <abhishekv.verma@gmail.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id A6FAD1B2F0D for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 07:52:55 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.399
X-Spam-Level: 
X-Spam-Status: No, score=-1.399 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_21=0.6, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id l0Dq1dE5bqxE for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 07:52:54 -0800 (PST)
Received: from mail-io0-x232.google.com (mail-io0-x232.google.com [IPv6:2607:f8b0:4001:c06::232]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 6FB751B2EEF for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 07:52:31 -0800 (PST)
Received: by mail-io0-x232.google.com with SMTP id e126so59299629ioa.1 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 07:52:31 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=rCa2Tzf2G6172QDY3h+BUGjql7cFBPjfRL0rPfZZBb0=; b=F7bXQ/onKdzXEk0XT/aXYDhcoA+MHs8z8Bgw7pSMqw0zAIMnDeW1bMa0iYN1z4JJuV CdXDTtsPHo+5apcPdu19vEX1f3LQPsQPLncbwgqLhH/V0b20+9GgTGgu7/Nv0heBzmI9 CikzcQF0o7lUtN43cs4N04ZCB3UR4KDHHAL/DoYVP7bmUXhhpud+YNkcZ7fISzsTd6Yf rK8sVWuXBGIr+IULqW/oljWbHQbntEEND5CreuaxDiGkbbXUuCJ8SibI8q23AFhMNgKp 0whJWJX/0vcDjGgMpK9OIZrD6rJeCo4R9tk1Af6gG2w+PWvRSR35LaJt55AoVoV2tfK8 xhSw==
MIME-Version: 1.0
X-Received: by 10.107.160.7 with SMTP id j7mr55150939ioe.52.1450367550801; Thu, 17 Dec 2015 07:52:30 -0800 (PST)
Received: by 10.64.62.101 with HTTP; Thu, 17 Dec 2015 07:52:30 -0800 (PST)
In-Reply-To: <a2ed60a6d4d34747974b4ee7d39b675d@rew09926dag03b.domain1.systemhost.net>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com> <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk> <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com> <a2ed60a6d4d34747974b4ee7d39b675d@rew09926dag03b.domain1.systemhost.net>
Date: Thu, 17 Dec 2015 21:22:30 +0530
Message-ID: <CADDmorRthrcZGE68YBsmRJF2epT0o_k4Lm5c8GCuzGMuSw6eKg@mail.gmail.com>
From: Abhishek Verma <abhishekv.verma@gmail.com>
To: philip.eardley@bt.com
Content-Type: multipart/alternative; boundary=001a1140efb6a8fcfd05271a0239
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/tvtRgj3dHMw05_6wWot9jHu5YSQ>
Cc: MultiPath WG <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 17 Dec 2015 15:52:55 -0000

--001a1140efb6a8fcfd05271a0239
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Thanks Philip. This was very helpful.

We have a DSL connection. I see that i need fewer TCP streams to fill up my
pipe during my day time vis-a-vis the night hours.

Cheers, Abhishek

On Thu, Dec 17, 2015 at 8:56 PM, <philip.eardley@bt.com> wrote:

>
>
> *From:* multipathtcp [mailto:multipathtcp-bounces@ietf.org] *On Behalf Of
> *Abhishek Verma
> *Sent:* 17 December 2015 14:09
> *To:* Costin Raiciu <c.raiciu@cs.ucl.ac.uk>
> *Cc:* MultiPath WG <multipathtcp@ietf.org>
> *Subject:* Re: [multipathtcp] Loss and TCP Throughput
>
>
>
>
> The variability you are seeing in the number of TCP connections is likely
> due to variability of upstream bottleneck capacity (e.g. how many other
> users=E2=80=99 flows you are fighting with).
>
>
>
> Wow. Thats strange.
>
>
>
> I assumed that in a 10Mbps broadband connection, i will always get 10Mbps
> and that will be the bottle neck. Did not imagine that its possible that =
i
> could get choked further upstream.
>
> [phil] are you on DSL or cable? Cable is shared access. DSL networks are
> designed to have a bottleneck link at the BRAS /BNG =E2=80=93 shared poin=
t between
> typically a few tens of users. One way of finding out whether this is the
> issue is to run your tests at different times of the day =E2=80=93 off pe=
ak, esp
> middle of the night, you won=E2=80=99t be contended, in the evening peak =
you will
> be sometimes.
>
>
>
> Best wishes
>
> Phil
>
>
>
> Philip Eardley
>
> Research and Innovation
>
> This email contains BT information, which may be privileged or
> confidential. It's meant only for the individual(s) or entity named above=
.
> If you're not the intended recipient, note that disclosing, copying,
> distributing or using this information is prohibited. If you've received
> this email in error, please let me know immediately on the email address
> above. Thank you.
> We monitor our email system, and may record your emails.
>
> British Telecommunications plc
> Registered office: 81 Newgate Street London EC1A 7AJ
> Registered in England no: 1800000
>
>
>
>
>

--001a1140efb6a8fcfd05271a0239
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Thanks Philip. This was very helpful.<div><br></div><div>W=
e have a DSL connection. I see that i need fewer TCP streams to fill up my =
pipe during my day time vis-a-vis the night hours.</div><div><br></div><div=
>Cheers, Abhishek</div></div><div class=3D"gmail_extra"><br><div class=3D"g=
mail_quote">On Thu, Dec 17, 2015 at 8:56 PM,  <span dir=3D"ltr">&lt;<a href=
=3D"mailto:philip.eardley@bt.com" target=3D"_blank">philip.eardley@bt.com</=
a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"margin:0=
 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">





<div lang=3D"EN-GB" link=3D"blue" vlink=3D"purple">
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
<p class=3D"MsoNormal"><b><span lang=3D"EN-US" style=3D"font-size:11.0pt;fo=
nt-family:&quot;Calibri&quot;,sans-serif">From:</span></b><span lang=3D"EN-=
US" style=3D"font-size:11.0pt;font-family:&quot;Calibri&quot;,sans-serif"> =
multipathtcp [mailto:<a href=3D"mailto:multipathtcp-bounces@ietf.org" targe=
t=3D"_blank">multipathtcp-bounces@ietf.org</a>]
<b>On Behalf Of </b>Abhishek Verma<br>
<b>Sent:</b> 17 December 2015 14:09<br>
<b>To:</b> Costin Raiciu &lt;<a href=3D"mailto:c.raiciu@cs.ucl.ac.uk" targe=
t=3D"_blank">c.raiciu@cs.ucl.ac.uk</a>&gt;<br>
<b>Cc:</b> MultiPath WG &lt;<a href=3D"mailto:multipathtcp@ietf.org" target=
=3D"_blank">multipathtcp@ietf.org</a>&gt;<br>
<b>Subject:</b> Re: [multipathtcp] Loss and TCP Throughput<u></u><u></u></s=
pan></p>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
<div>
<div>
<div><span class=3D"">
<blockquote style=3D"border:none;border-left:solid #cccccc 1.0pt;padding:0c=
m 0cm 0cm 6.0pt;margin-left:4.8pt;margin-right:0cm">
<p class=3D"MsoNormal"><br>
The variability you are seeing in the number of TCP connections is likely d=
ue to variability of upstream bottleneck capacity (e.g. how many other user=
s=E2=80=99 flows you are fighting with).<u></u><u></u></p>
</blockquote>
<div>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
</div>
<div>
<p class=3D"MsoNormal">Wow. Thats strange.<u></u><u></u></p>
</div>
<div>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
</div>
</span><div><span class=3D"">
<p class=3D"MsoNormal">I assumed that in a 10Mbps broadband connection, i w=
ill always get 10Mbps and that will be the bottle neck. Did not imagine tha=
t its possible that i could get choked further upstream.<u></u><u></u></p>
</span><p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&=
quot;Calibri&quot;,sans-serif;color:#1f497d">[phil] are you on DSL or cable=
? Cable is shared access. DSL networks are designed to have a bottleneck li=
nk at the BRAS /BNG =E2=80=93 shared point between typically
 a few tens of users. One way of finding out whether this is the issue is t=
o run your tests at different times of the day =E2=80=93 off peak, esp midd=
le of the night, you won=E2=80=99t be contended, in the evening peak you wi=
ll be sometimes.</span>
<u></u><u></u></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d">Best wishes<u></u><u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d">Phil<u></u><u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">Philip Eardley<u></u><u><=
/u></span></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">Research and Innovation<u=
></u><u></u></span></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">This email contains BT in=
formation, which may be privileged or confidential. It&#39;s meant only
 for the individual(s) or entity named above. If you&#39;re not the intende=
d recipient, note that disclosing, copying, distributing or using this info=
rmation is prohibited. If you&#39;ve received this email in error, please l=
et me know immediately on the email address
 above. Thank you.<br>
We monitor our email system, and may record your emails.<u></u><u></u></spa=
n></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">British Telecommunication=
s plc<br>
Registered office: 81 Newgate Street London EC1A 7AJ<br>
Registered in England no: 1800000<u></u><u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
</div>
</div>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
</div>
</div>
</div>
</div>

</blockquote></div><br></div>

--001a1140efb6a8fcfd05271a0239--


From nobody Thu Dec 17 13:51:53 2015
Return-Path: <c.raiciu@cs.ucl.ac.uk>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id E272E1B30CD for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 13:51:52 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.21
X-Spam-Level: 
X-Spam-Status: No, score=-4.21 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id KIu7I9A3mJT0 for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 13:51:50 -0800 (PST)
Received: from bells2.cs.ucl.ac.uk (bells2.cs.ucl.ac.uk [128.16.5.33]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 419C51B30DD for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 13:51:50 -0800 (PST)
Received: from [86.127.154.136] (helo=192-168-0-100.rdsnet.ro) by bells2.cs.ucl.ac.uk with esmtpsa (TLSv1:AES256-SHA:256) (C.Raiciu authenticated) (Exim 4.54) id 1a9gSV-000LiP-0j; Thu, 17 Dec 2015 21:51:35 +0000
Content-Type: text/plain; charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2070.6\))
From: Costin Raiciu <c.raiciu@cs.ucl.ac.uk>
In-Reply-To: <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com>
Date: Thu, 17 Dec 2015 23:51:35 +0200
Content-Transfer-Encoding: quoted-printable
Message-Id: <C5E10922-E3BD-4BE6-9CFD-31B76B4D0744@cs.ucl.ac.uk>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com> <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk> <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com>
To: Abhishek Verma <abhishekv.verma@gmail.com>
X-Mailer: Apple Mail (2.2070.6)
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/IJ4NOfRxfRqy70NCEc18Sp62CTY>
Cc: MultiPath WG <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 17 Dec 2015 21:51:53 -0000

> I believe the throughput equation does not hold good when the losses =
are induced because of us congesting the link. So, when i increase the # =
of TCP streams to fill up the pipe, then that equation may not hold =
good.
>=20
> Is this correct -- that, loss in that equation is the general network =
loss and not the loss induced by congesting the link artifically?

If you are a single flow at the bottleneck, it really does not matter =
what the throughput equation says; as long as you have enough buffering =
(at least BDP at the bottleneck), and an aggressive enough congestion =
control (think problems of Reno long fat pipes), you will always fill =
the pipe. If you don=E2=80=99t, then either the bottleneck is not where =
you think it is, or you have some sender/receiver buffer sizing issues =
(I don=E2=80=99t think this is the case).

The Reno throughput equations that exist are probably slightly off for =
Cubic (assuming that=E2=80=99s what you are using) - but, as I said, it =
shouldn=E2=80=99t matter.

As for fairness - if you increase the file size, throughput should  =
equalize more across the two connections.
Costin


> =20
>=20
> The variability you are seeing in the number of TCP connections is =
likely due to variability of upstream bottleneck capacity (e.g. how many =
other users=E2=80=99 flows you are fighting with).
>=20
> Wow. Thats strange.
>=20
> I assumed that in a 10Mbps broadband connection, i will always get =
10Mbps and that will be the bottle neck. Did not imagine that its =
possible that i could get choked further upstream.
>=20
> I have another observation that cannot get reconciled with getting =
choked upstream:
>=20
> I send two 100Mb files at the same time.
>=20
> On one file i get a throughput of 7Mbps, while on the other i get =
3Mbps.
>=20
> Because of TCP fairness, shouldnt each file get 5Mbps?
>=20
> Thanks, Abhishek
>=20
> =20
>=20
> Hope this helps.
> Costin
>=20
> > On 17 Dec 2015, at 14:42, Abhishek Verma <abhishekv.verma@gmail.com> =
wrote:
> >
> > Hi TCP experts,
> >
> > I have an observation that i am unable to explain and understand. =
Would appreciate if folks here can make some sense out of this:
> >
> > I have a fixed size file, 100 Mb.
> >
> > I use multiple TCP streams to transfer the file over my network. I =
have a 10Mbps broadband connection.
> >
> > There are times when i am able to fill up the pipe with very few TCP =
streams (like 2-4) but there are times when i need more than 20 to fill =
up my connection.
> >
> > My understanding is as follows:
> >
> > TCP Throughput =3D k (MSS/RTT)*(1/sqrt(Loss)) where k is some =
constant
> >
> > In a file transfer between two end points the MSS and RTT are =
constant, so the equation reduces to:
> >
> > Throughput =3D k'/sqrt(Loss)
> >
> > This is the throughput that i get with 1 TCP stream.
> > Hence, when am pushing N chunks, the throughput that i will get is =
N*k/sqrt(Loss)
> >
> > I hence assume that when i need more TCP streams to push my file =
out, the Loss in the network is higher than when i need fewer TCP =
streams. This is the part that i understand very well.
> >
> > What i dont understand is this:
> >
> > I figure out that i need, say 7, TCP streams to fill my 10Mbps pipe =
when pushing out my file.
> >
> > So, i if i do this experiment back to back 10 times, then while 80% =
of the times i am able to fill the pipe with 7 streams. There are always =
these 20% runs, where 7 streams gives me a BW thats lower than 10, say =
around 7-8 Mbps. This is something that i cannot explain.
> >
> > Can somebody help me with this? Just give me some clues and i will =
take it on from there.
> >
> > I could rephrase my question like this:
> > Is it possible that for certain TCP sessions (the 20%) the Loss that =
i get is more than the Loss that 80% of my TCP streams experience.
> >
> > Thanks, Abhishek
> > _______________________________________________
> > multipathtcp mailing list
> > multipathtcp@ietf.org
> > https://www.ietf.org/mailman/listinfo/multipathtcp
>=20
>=20


From nobody Thu Dec 17 16:57:34 2015
Return-Path: <abhishekv.verma@gmail.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 8C6F91B3188 for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 16:57:33 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.999
X-Spam-Level: 
X-Spam-Status: No, score=-1.999 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 09lY6SXRMzmk for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 16:57:31 -0800 (PST)
Received: from mail-ig0-x230.google.com (mail-ig0-x230.google.com [IPv6:2607:f8b0:4001:c05::230]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 2BE571B3183 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 16:57:30 -0800 (PST)
Received: by mail-ig0-x230.google.com with SMTP id ph11so25485185igc.1 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 16:57:30 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=i8CA/861os0lPsb0z/dNZM0Tio3wEOv7lJ/Fnjlag8M=; b=mOR8b9vrxCdX+H+6Fb0YPRwwfaUQCPAUKcxjiKe0P4m7DeZbEIVujz2FXwKZIySHrH 2CgE1Se/1Oq6Z2bsEqLYMERJgpzZ+dZQt6hwP2ByN+BJv8e4nl24hQvlszNjdSkpx5p6 gtYgB0E80iftDyGgoS6Wq8Ihq2VGux6QO7qza4mkqDzTDkAXuazFpSQHpF6CoQrqOsSC kUDKHRDjqi9dJh3n/cYYnR127X+MBlaJGdv1qAslUli2KROPq6qwfKJkSyHtlqWk3YtQ je2BHJqgh/RLt71+7SYCdbLqlHAOJvFsKJKmI2KQC5i1SQGthd0BPUyr6Mv1w81GG9k9 KnzA==
MIME-Version: 1.0
X-Received: by 10.50.154.5 with SMTP id vk5mr433609igb.85.1450400250267; Thu, 17 Dec 2015 16:57:30 -0800 (PST)
Received: by 10.64.62.101 with HTTP; Thu, 17 Dec 2015 16:57:30 -0800 (PST)
In-Reply-To: <C5E10922-E3BD-4BE6-9CFD-31B76B4D0744@cs.ucl.ac.uk>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com> <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk> <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com> <C5E10922-E3BD-4BE6-9CFD-31B76B4D0744@cs.ucl.ac.uk>
Date: Fri, 18 Dec 2015 06:27:30 +0530
Message-ID: <CADDmorR2qP4dH+f7rdK8mai-y_UrO+GsxGo_jAngO7+q+r8KHw@mail.gmail.com>
From: Abhishek Verma <abhishekv.verma@gmail.com>
To: Costin Raiciu <c.raiciu@cs.ucl.ac.uk>
Content-Type: multipart/alternative; boundary=047d7bd76b98b30c170527219f72
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/fOml0Zdfcx1VJrmTjDZle3j-AKM>
Cc: MultiPath WG <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 18 Dec 2015 00:57:33 -0000

--047d7bd76b98b30c170527219f72
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

>
>
> As for fairness - if you increase the file size, throughput should
> equalize more across the two connections.
> Costin
>

And this is precisely that i am unable to understand.

Why isnt TCP fair for small file sizes. And when i say smaller, i mean
100Mb files on a 25Mbps broadband connection. Sometimes when i am pushing
the two files togethar, i get a BW of 19Mbps for one, and 6 for the other.
Most of the times its equitably distributed, but there are often these
outliers where one connection completely dominates the bulk of the pipe
capacity.

Cheers, Abhishek

>
>
> >
> >
> > The variability you are seeing in the number of TCP connections is
> likely due to variability of upstream bottleneck capacity (e.g. how many
> other users=E2=80=99 flows you are fighting with).
> >
> > Wow. Thats strange.
> >
> > I assumed that in a 10Mbps broadband connection, i will always get
> 10Mbps and that will be the bottle neck. Did not imagine that its possibl=
e
> that i could get choked further upstream.
> >
> > I have another observation that cannot get reconciled with getting
> choked upstream:
> >
> > I send two 100Mb files at the same time.
> >
> > On one file i get a throughput of 7Mbps, while on the other i get 3Mbps=
.
> >
> > Because of TCP fairness, shouldnt each file get 5Mbps?
> >
> > Thanks, Abhishek
> >
> >
> >
> > Hope this helps.
> > Costin
> >
> > > On 17 Dec 2015, at 14:42, Abhishek Verma <abhishekv.verma@gmail.com>
> wrote:
> > >
> > > Hi TCP experts,
> > >
> > > I have an observation that i am unable to explain and understand.
> Would appreciate if folks here can make some sense out of this:
> > >
> > > I have a fixed size file, 100 Mb.
> > >
> > > I use multiple TCP streams to transfer the file over my network. I
> have a 10Mbps broadband connection.
> > >
> > > There are times when i am able to fill up the pipe with very few TCP
> streams (like 2-4) but there are times when i need more than 20 to fill u=
p
> my connection.
> > >
> > > My understanding is as follows:
> > >
> > > TCP Throughput =3D k (MSS/RTT)*(1/sqrt(Loss)) where k is some constan=
t
> > >
> > > In a file transfer between two end points the MSS and RTT are
> constant, so the equation reduces to:
> > >
> > > Throughput =3D k'/sqrt(Loss)
> > >
> > > This is the throughput that i get with 1 TCP stream.
> > > Hence, when am pushing N chunks, the throughput that i will get is
> N*k/sqrt(Loss)
> > >
> > > I hence assume that when i need more TCP streams to push my file out,
> the Loss in the network is higher than when i need fewer TCP streams. Thi=
s
> is the part that i understand very well.
> > >
> > > What i dont understand is this:
> > >
> > > I figure out that i need, say 7, TCP streams to fill my 10Mbps pipe
> when pushing out my file.
> > >
> > > So, i if i do this experiment back to back 10 times, then while 80% o=
f
> the times i am able to fill the pipe with 7 streams. There are always the=
se
> 20% runs, where 7 streams gives me a BW thats lower than 10, say around 7=
-8
> Mbps. This is something that i cannot explain.
> > >
> > > Can somebody help me with this? Just give me some clues and i will
> take it on from there.
> > >
> > > I could rephrase my question like this:
> > > Is it possible that for certain TCP sessions (the 20%) the Loss that =
i
> get is more than the Loss that 80% of my TCP streams experience.
> > >
> > > Thanks, Abhishek
> > > _______________________________________________
> > > multipathtcp mailing list
> > > multipathtcp@ietf.org
> > > https://www.ietf.org/mailman/listinfo/multipathtcp
> >
> >
>
>

--047d7bd76b98b30c170527219f72
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div class=3D"gmail_extra"><div class=3D"gmail_quote"><blo=
ckquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #c=
cc solid;padding-left:1ex"><br>
As for fairness - if you increase the file size, throughput should=C2=A0 eq=
ualize more across the two connections.<br>
<span class=3D"HOEnZb"><font color=3D"#888888">Costin<br></font></span></bl=
ockquote><div><br></div><div>And this is precisely that i am unable to unde=
rstand.</div><div><br></div><div>Why isnt TCP fair for small file sizes. An=
d when i say smaller, i mean 100Mb files on a 25Mbps broadband connection. =
Sometimes when i am pushing the two files togethar, i get a BW of 19Mbps fo=
r one, and 6 for the other. Most of the times its equitably distributed, bu=
t there are often these outliers where one connection completely dominates =
the bulk of the pipe capacity.</div><div><br></div><div>Cheers, Abhishek=C2=
=A0</div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;borde=
r-left:1px #ccc solid;padding-left:1ex"><span class=3D"HOEnZb"><font color=
=3D"#888888">
</font></span><div class=3D"HOEnZb"><div class=3D"h5"><br>
<br>
&gt;<br>
&gt;<br>
&gt; The variability you are seeing in the number of TCP connections is lik=
ely due to variability of upstream bottleneck capacity (e.g. how many other=
 users=E2=80=99 flows you are fighting with).<br>
&gt;<br>
&gt; Wow. Thats strange.<br>
&gt;<br>
&gt; I assumed that in a 10Mbps broadband connection, i will always get 10M=
bps and that will be the bottle neck. Did not imagine that its possible tha=
t i could get choked further upstream.<br>
&gt;<br>
&gt; I have another observation that cannot get reconciled with getting cho=
ked upstream:<br>
&gt;<br>
&gt; I send two 100Mb files at the same time.<br>
&gt;<br>
&gt; On one file i get a throughput of 7Mbps, while on the other i get 3Mbp=
s.<br>
&gt;<br>
&gt; Because of TCP fairness, shouldnt each file get 5Mbps?<br>
&gt;<br>
&gt; Thanks, Abhishek<br>
&gt;<br>
&gt;<br>
&gt;<br>
&gt; Hope this helps.<br>
&gt; Costin<br>
&gt;<br>
&gt; &gt; On 17 Dec 2015, at 14:42, Abhishek Verma &lt;<a href=3D"mailto:ab=
hishekv.verma@gmail.com">abhishekv.verma@gmail.com</a>&gt; wrote:<br>
&gt; &gt;<br>
&gt; &gt; Hi TCP experts,<br>
&gt; &gt;<br>
&gt; &gt; I have an observation that i am unable to explain and understand.=
 Would appreciate if folks here can make some sense out of this:<br>
&gt; &gt;<br>
&gt; &gt; I have a fixed size file, 100 Mb.<br>
&gt; &gt;<br>
&gt; &gt; I use multiple TCP streams to transfer the file over my network. =
I have a 10Mbps broadband connection.<br>
&gt; &gt;<br>
&gt; &gt; There are times when i am able to fill up the pipe with very few =
TCP streams (like 2-4) but there are times when i need more than 20 to fill=
 up my connection.<br>
&gt; &gt;<br>
&gt; &gt; My understanding is as follows:<br>
&gt; &gt;<br>
&gt; &gt; TCP Throughput =3D k (MSS/RTT)*(1/sqrt(Loss)) where k is some con=
stant<br>
&gt; &gt;<br>
&gt; &gt; In a file transfer between two end points the MSS and RTT are con=
stant, so the equation reduces to:<br>
&gt; &gt;<br>
&gt; &gt; Throughput =3D k&#39;/sqrt(Loss)<br>
&gt; &gt;<br>
&gt; &gt; This is the throughput that i get with 1 TCP stream.<br>
&gt; &gt; Hence, when am pushing N chunks, the throughput that i will get i=
s N*k/sqrt(Loss)<br>
&gt; &gt;<br>
&gt; &gt; I hence assume that when i need more TCP streams to push my file =
out, the Loss in the network is higher than when i need fewer TCP streams. =
This is the part that i understand very well.<br>
&gt; &gt;<br>
&gt; &gt; What i dont understand is this:<br>
&gt; &gt;<br>
&gt; &gt; I figure out that i need, say 7, TCP streams to fill my 10Mbps pi=
pe when pushing out my file.<br>
&gt; &gt;<br>
&gt; &gt; So, i if i do this experiment back to back 10 times, then while 8=
0% of the times i am able to fill the pipe with 7 streams. There are always=
 these 20% runs, where 7 streams gives me a BW thats lower than 10, say aro=
und 7-8 Mbps. This is something that i cannot explain.<br>
&gt; &gt;<br>
&gt; &gt; Can somebody help me with this? Just give me some clues and i wil=
l take it on from there.<br>
&gt; &gt;<br>
&gt; &gt; I could rephrase my question like this:<br>
&gt; &gt; Is it possible that for certain TCP sessions (the 20%) the Loss t=
hat i get is more than the Loss that 80% of my TCP streams experience.<br>
&gt; &gt;<br>
&gt; &gt; Thanks, Abhishek<br>
&gt; &gt; _______________________________________________<br>
&gt; &gt; multipathtcp mailing list<br>
&gt; &gt; <a href=3D"mailto:multipathtcp@ietf.org">multipathtcp@ietf.org</a=
><br>
&gt; &gt; <a href=3D"https://www.ietf.org/mailman/listinfo/multipathtcp" re=
l=3D"noreferrer" target=3D"_blank">https://www.ietf.org/mailman/listinfo/mu=
ltipathtcp</a><br>
&gt;<br>
&gt;<br>
<br>
</div></div></blockquote></div><br></div></div>

--047d7bd76b98b30c170527219f72--


From nobody Thu Dec 17 20:48:45 2015
Return-Path: <nishida@sfc.wide.ad.jp>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 1334A1B3307 for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 20:48:42 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.115
X-Spam-Level: **
X-Spam-Status: No, score=2.115 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, FM_FORGED_GMAIL=0.622, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, HTML_MESSAGE=0.001, J_CHICKENPOX_21=0.6, RCVD_IN_DNSWL_LOW=-0.7, RELAY_IS_203=0.994, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HVomP6o4MN8R for <multipathtcp@ietfa.amsl.com>; Thu, 17 Dec 2015 20:48:40 -0800 (PST)
Received: from mail.sfc.wide.ad.jp (shonan.sfc.wide.ad.jp [203.178.142.130]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id CDCD51B3305 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 20:48:39 -0800 (PST)
Received: from mail-ob0-f182.google.com (mail-ob0-f182.google.com [209.85.214.182]) by mail.sfc.wide.ad.jp (Postfix) with ESMTPSA id 2EE122790F0 for <multipathtcp@ietf.org>; Fri, 18 Dec 2015 13:48:37 +0900 (JST)
Received: by mail-ob0-f182.google.com with SMTP id sd4so73279146obb.0 for <multipathtcp@ietf.org>; Thu, 17 Dec 2015 20:48:37 -0800 (PST)
MIME-Version: 1.0
X-Received: by 10.60.58.166 with SMTP id s6mr647010oeq.80.1450414115536; Thu, 17 Dec 2015 20:48:35 -0800 (PST)
Received: by 10.202.190.6 with HTTP; Thu, 17 Dec 2015 20:48:35 -0800 (PST)
In-Reply-To: <CADDmorRthrcZGE68YBsmRJF2epT0o_k4Lm5c8GCuzGMuSw6eKg@mail.gmail.com>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com> <1CA40276-1B28-4483-9B0C-E95283BC1C53@cs.ucl.ac.uk> <CADDmorS6cb0gmFALyEcN9swyf1-Q4qm52QyJJFFJFviEtarrVg@mail.gmail.com> <a2ed60a6d4d34747974b4ee7d39b675d@rew09926dag03b.domain1.systemhost.net> <CADDmorRthrcZGE68YBsmRJF2epT0o_k4Lm5c8GCuzGMuSw6eKg@mail.gmail.com>
Date: Thu, 17 Dec 2015 20:48:35 -0800
X-Gmail-Original-Message-ID: <CAO249yeO2brVfQQbcGSR7q+SMi3a8bo6iB0hE4Zg3eWtGXJA-Q@mail.gmail.com>
Message-ID: <CAO249yeO2brVfQQbcGSR7q+SMi3a8bo6iB0hE4Zg3eWtGXJA-Q@mail.gmail.com>
From: Yoshifumi Nishida <nishida@sfc.wide.ad.jp>
To: Abhishek Verma <abhishekv.verma@gmail.com>
Content-Type: multipart/alternative; boundary=089e0139faa8224e52052724da43
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/4c-21y9600pythMJ-XKXd5Gu_l4>
Cc: MultiPath WG <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 18 Dec 2015 04:48:42 -0000

--089e0139faa8224e52052724da43
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Hi Abhishek,

I think It would be good if you can test it in a local network where
there's no competitive traffic.
--
Yoshi


On Thu, Dec 17, 2015 at 7:52 AM, Abhishek Verma <abhishekv.verma@gmail.com>
wrote:

> Thanks Philip. This was very helpful.
>
> We have a DSL connection. I see that i need fewer TCP streams to fill up
> my pipe during my day time vis-a-vis the night hours.
>
> Cheers, Abhishek
>
> On Thu, Dec 17, 2015 at 8:56 PM, <philip.eardley@bt.com> wrote:
>
>>
>>
>> *From:* multipathtcp [mailto:multipathtcp-bounces@ietf.org] *On Behalf
>> Of *Abhishek Verma
>> *Sent:* 17 December 2015 14:09
>> *To:* Costin Raiciu <c.raiciu@cs.ucl.ac.uk>
>> *Cc:* MultiPath WG <multipathtcp@ietf.org>
>> *Subject:* Re: [multipathtcp] Loss and TCP Throughput
>>
>>
>>
>>
>> The variability you are seeing in the number of TCP connections is likel=
y
>> due to variability of upstream bottleneck capacity (e.g. how many other
>> users=E2=80=99 flows you are fighting with).
>>
>>
>>
>> Wow. Thats strange.
>>
>>
>>
>> I assumed that in a 10Mbps broadband connection, i will always get 10Mbp=
s
>> and that will be the bottle neck. Did not imagine that its possible that=
 i
>> could get choked further upstream.
>>
>> [phil] are you on DSL or cable? Cable is shared access. DSL networks are
>> designed to have a bottleneck link at the BRAS /BNG =E2=80=93 shared poi=
nt between
>> typically a few tens of users. One way of finding out whether this is th=
e
>> issue is to run your tests at different times of the day =E2=80=93 off p=
eak, esp
>> middle of the night, you won=E2=80=99t be contended, in the evening peak=
 you will
>> be sometimes.
>>
>>
>>
>> Best wishes
>>
>> Phil
>>
>>
>>
>> Philip Eardley
>>
>> Research and Innovation
>>
>> This email contains BT information, which may be privileged or
>> confidential. It's meant only for the individual(s) or entity named abov=
e.
>> If you're not the intended recipient, note that disclosing, copying,
>> distributing or using this information is prohibited. If you've received
>> this email in error, please let me know immediately on the email address
>> above. Thank you.
>> We monitor our email system, and may record your emails.
>>
>> British Telecommunications plc
>> Registered office: 81 Newgate Street London EC1A 7AJ
>> Registered in England no: 1800000
>>
>>
>>
>>
>>
>
>
> _______________________________________________
> multipathtcp mailing list
> multipathtcp@ietf.org
> https://www.ietf.org/mailman/listinfo/multipathtcp
>
>

--089e0139faa8224e52052724da43
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi Abhishek,<div><br></div><div>I think It would be good i=
f you can test it in a local network where there&#39;s no competitive traff=
ic.</div><div>--</div><div>Yoshi</div><div><br><div class=3D"gmail_extra"><=
br><div class=3D"gmail_quote">On Thu, Dec 17, 2015 at 7:52 AM, Abhishek Ver=
ma <span dir=3D"ltr">&lt;<a href=3D"mailto:abhishekv.verma@gmail.com" targe=
t=3D"_blank">abhishekv.verma@gmail.com</a>&gt;</span> wrote:<br><blockquote=
 class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc soli=
d;padding-left:1ex"><div dir=3D"ltr">Thanks Philip. This was very helpful.<=
div><br></div><div>We have a DSL connection. I see that i need fewer TCP st=
reams to fill up my pipe during my day time vis-a-vis the night hours.</div=
><div><br></div><div>Cheers, Abhishek</div></div><div class=3D"HOEnZb"><div=
 class=3D"h5"><div class=3D"gmail_extra"><br><div class=3D"gmail_quote">On =
Thu, Dec 17, 2015 at 8:56 PM,  <span dir=3D"ltr">&lt;<a href=3D"mailto:phil=
ip.eardley@bt.com" target=3D"_blank">philip.eardley@bt.com</a>&gt;</span> w=
rote:<br><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;borde=
r-left:1px #ccc solid;padding-left:1ex">





<div lang=3D"EN-GB" link=3D"blue" vlink=3D"purple">
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
<p class=3D"MsoNormal"><b><span lang=3D"EN-US" style=3D"font-size:11.0pt;fo=
nt-family:&quot;Calibri&quot;,sans-serif">From:</span></b><span lang=3D"EN-=
US" style=3D"font-size:11.0pt;font-family:&quot;Calibri&quot;,sans-serif"> =
multipathtcp [mailto:<a href=3D"mailto:multipathtcp-bounces@ietf.org" targe=
t=3D"_blank">multipathtcp-bounces@ietf.org</a>]
<b>On Behalf Of </b>Abhishek Verma<br>
<b>Sent:</b> 17 December 2015 14:09<br>
<b>To:</b> Costin Raiciu &lt;<a href=3D"mailto:c.raiciu@cs.ucl.ac.uk" targe=
t=3D"_blank">c.raiciu@cs.ucl.ac.uk</a>&gt;<br>
<b>Cc:</b> MultiPath WG &lt;<a href=3D"mailto:multipathtcp@ietf.org" target=
=3D"_blank">multipathtcp@ietf.org</a>&gt;<br>
<b>Subject:</b> Re: [multipathtcp] Loss and TCP Throughput<u></u><u></u></s=
pan></p>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
<div>
<div>
<div><span>
<blockquote style=3D"border:none;border-left:solid #cccccc 1.0pt;padding:0c=
m 0cm 0cm 6.0pt;margin-left:4.8pt;margin-right:0cm">
<p class=3D"MsoNormal"><br>
The variability you are seeing in the number of TCP connections is likely d=
ue to variability of upstream bottleneck capacity (e.g. how many other user=
s=E2=80=99 flows you are fighting with).<u></u><u></u></p>
</blockquote>
<div>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
</div>
<div>
<p class=3D"MsoNormal">Wow. Thats strange.<u></u><u></u></p>
</div>
<div>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
</div>
</span><div><span>
<p class=3D"MsoNormal">I assumed that in a 10Mbps broadband connection, i w=
ill always get 10Mbps and that will be the bottle neck. Did not imagine tha=
t its possible that i could get choked further upstream.<u></u><u></u></p>
</span><p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&=
quot;Calibri&quot;,sans-serif;color:#1f497d">[phil] are you on DSL or cable=
? Cable is shared access. DSL networks are designed to have a bottleneck li=
nk at the BRAS /BNG =E2=80=93 shared point between typically
 a few tens of users. One way of finding out whether this is the issue is t=
o run your tests at different times of the day =E2=80=93 off peak, esp midd=
le of the night, you won=E2=80=99t be contended, in the evening peak you wi=
ll be sometimes.</span>
<u></u><u></u></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d">Best wishes<u></u><u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d">Phil<u></u><u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">Philip Eardley<u></u><u><=
/u></span></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">Research and Innovation<u=
></u><u></u></span></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">This email contains BT in=
formation, which may be privileged or confidential. It&#39;s meant only
 for the individual(s) or entity named above. If you&#39;re not the intende=
d recipient, note that disclosing, copying, distributing or using this info=
rmation is prohibited. If you&#39;ve received this email in error, please l=
et me know immediately on the email address
 above. Thank you.<br>
We monitor our email system, and may record your emails.<u></u><u></u></spa=
n></p>
<p class=3D"MsoNormal"><span lang=3D"EN" style=3D"font-size:11.0pt;font-fam=
ily:&quot;Calibri&quot;,sans-serif;color:#1f497d">British Telecommunication=
s plc<br>
Registered office: 81 Newgate Street London EC1A 7AJ<br>
Registered in England no: 1800000<u></u><u></u></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:11.0pt;font-family:&quot;Ca=
libri&quot;,sans-serif;color:#1f497d"><u></u>=C2=A0<u></u></span></p>
</div>
</div>
<p class=3D"MsoNormal"><u></u>=C2=A0<u></u></p>
</div>
</div>
</div>
</div>

</blockquote></div><br></div>
</div></div><br>_______________________________________________<br>
multipathtcp mailing list<br>
<a href=3D"mailto:multipathtcp@ietf.org">multipathtcp@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/multipathtcp" rel=3D"noref=
errer" target=3D"_blank">https://www.ietf.org/mailman/listinfo/multipathtcp=
</a><br>
<br></blockquote></div><br></div></div></div>

--089e0139faa8224e52052724da43--


From nobody Fri Dec 18 00:38:16 2015
Return-Path: <olivier.bonaventure@uclouvain.be>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id AF8031B3467 for <multipathtcp@ietfa.amsl.com>; Fri, 18 Dec 2015 00:38:14 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.402
X-Spam-Level: 
X-Spam-Status: No, score=-2.402 tagged_above=-999 required=5 tests=[BAYES_20=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id QuHtEhgO2Xa9 for <multipathtcp@ietfa.amsl.com>; Fri, 18 Dec 2015 00:38:12 -0800 (PST)
Received: from smtp5.sgsi.ucl.ac.be (smtp.sgsi.ucl.ac.be [130.104.5.67]) by ietfa.amsl.com (Postfix) with ESMTP id 259521B3459 for <multipathtcp@ietf.org>; Fri, 18 Dec 2015 00:38:12 -0800 (PST)
Received: from mbpobo.lan (unknown [5.149.142.22]) (using TLSv1 with cipher DHE-RSA-AES128-SHA (128/128 bits)) (No client certificate requested) (Authenticated sender: obonaventure@smtp5.sgsi.ucl.ac.be) by smtp5.sgsi.ucl.ac.be (Postfix) with ESMTPSA id CFE731980C1; Fri, 18 Dec 2015 09:37:53 +0100 (CET)
X-DKIM: Sendmail DKIM Filter v2.8.3 smtp5.sgsi.ucl.ac.be CFE731980C1
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=uclouvain.be; s=selucl; t=1450427873; bh=a1owT5cV9gsrIZLsUXxDAGSRzUhCVlm/M6yYksykT1o=; h=Reply-To:Subject:References:To:From:Message-ID:Date:MIME-Version: In-Reply-To:Content-Type:Content-Transfer-Encoding; b=zK8KQMLW7lDTodowy/fIIkIiGVnrU48Fec+GfK2r0tX39vgNxJfIiSDkPMlxUxPbW HA26v88jmrQYWIq/WF3ZNcX2HAsSuYpVctWqXVIZy4LdPmBUN7BwFFsyrF5ir0M5Sn LGxNYuDjlZQrwk494KTZfvuAKgbi320SEQIRZnCs=
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com>
To: Abhishek Verma <abhishekv.verma@gmail.com>, multipathtcp@ietf.org
From: Olivier Bonaventure <Olivier.Bonaventure@uclouvain.be>
Message-ID: <5673C5E1.1050909@uclouvain.be>
Date: Fri, 18 Dec 2015 09:37:53 +0100
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.11; rv:38.0) Gecko/20100101 Thunderbird/38.4.0
MIME-Version: 1.0
In-Reply-To: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: clamav-milter 0.99-beta1 at smtp-5.sipr-dc.ucl.ac.be
X-Virus-Status: Clean
X-Sgsi-Spamcheck: SASL authenticated, 
X-SGSI-MailScanner-ID: CFE731980C1.A3131
X-SGSI-MailScanner: Found to be clean
X-SGSI-From: olivier.bonaventure@uclouvain.be
X-SGSI-Spam-Status: No
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/LdtT-pzGhH4v6k44w3bhZ1enPUI>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
Reply-To: Olivier.Bonaventure@uclouvain.be
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 18 Dec 2015 08:38:14 -0000

Abhishek,
>
> I have an observation that i am unable to explain and understand. Would
> appreciate if folks here can make some sense out of this:
>
> I have a fixed size file, 100 Mb.
>
> I use multiple TCP streams to transfer the file over my network. I have
> a 10Mbps broadband connection.
>
> There are times when i am able to fill up the pipe with very few TCP
> streams (like 2-4) but there are times when i need more than 20 to fill
> up my connection.
>
> My understanding is as follows:
>
> TCP Throughput = k (MSS/RTT)*(1/sqrt(Loss)) where k is some constant
>
> In a file transfer between two end points the MSS and RTT are constant,
> so the equation reduces to:
>
> Throughput = k'/sqrt(Loss)
>
> This is the throughput that i get with 1 TCP stream.
> Hence, when am pushing N chunks, the throughput that i will get is
> N*k/sqrt(Loss)

This is only valid if the loss ratio is independant of the number of TCP 
connections that you have. If you increase congestion by adding more 
subflows, then Loss will also depend on N (and could increase with N).

Your measurement scenario is

client --- DSL network ---- server

I guess that the server has much more capacity than the client and you 
want to optimise the utilisation of the client DSL link.

The DSL network cannot always be considered as a blackbox. When you use 
multiple TCP subflows, you use different source and destination ports. 
If your network uses equal cost multipath, you will end up in using 
different paths for different subflows :

client --- R1---R2 ---R3 ---- server
             +         +
            R4---------R5


In this example, R1 can load balance the flows on the R1-R2-R3 and 
R1-R4-R5 paths that might have different levels of congestion.

You might use paris traceroute to check the load balancing in the DSL 
network that you use. I would also encourage you to collect packet 
traces of the different transfers and analyse the low performing flows


Olivier




From nobody Fri Dec 18 06:08:38 2015
Return-Path: <abhishekv.verma@gmail.com>
X-Original-To: multipathtcp@ietfa.amsl.com
Delivered-To: multipathtcp@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 7DB8B1B3652 for <multipathtcp@ietfa.amsl.com>; Fri, 18 Dec 2015 06:08:37 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.999
X-Spam-Level: 
X-Spam-Status: No, score=-1.999 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id CLvc5R0rrv6u for <multipathtcp@ietfa.amsl.com>; Fri, 18 Dec 2015 06:08:36 -0800 (PST)
Received: from mail-io0-x22c.google.com (mail-io0-x22c.google.com [IPv6:2607:f8b0:4001:c06::22c]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 140A11B3650 for <multipathtcp@ietf.org>; Fri, 18 Dec 2015 06:08:36 -0800 (PST)
Received: by mail-io0-x22c.google.com with SMTP id o67so90277796iof.3 for <multipathtcp@ietf.org>; Fri, 18 Dec 2015 06:08:36 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=sW1HYJMtUUcIFImCSiwd++OvpHCO42lKMq0iN8iIuno=; b=0F8uvTU6+S6ptfp42z+W1pZk5tK8uJpWsEKTMMjY6MiSS+Ri8M2E0jPQwkF19Mw2ZK 3RTl0dB3cQm9g5Zjl5BxtODGQGBRiJcrFOJdWRptmBy59RVgfJ0rCMVkKskliR02mzRP w4pD2Tq0QgixjHFRfG8nymguy98tRn4Boujp5EiGkXRNV9KvXN9MTFP2GHDnLo39h4n/ qM58AAOX+HaQN0/T92L7SAJV0aeioABSuPKoDBbk60WmWUiMSmnv6D3NxRWa0ei1qGz1 UkLfwbuG95W1aAxUcr3hLYPLV/s7vzoJy3aI4bIHDB56W7aYUmqthkNhKiJ9GU9mVIcB p0xw==
MIME-Version: 1.0
X-Received: by 10.107.156.21 with SMTP id f21mr22666ioe.54.1450447715544; Fri, 18 Dec 2015 06:08:35 -0800 (PST)
Received: by 10.64.0.233 with HTTP; Fri, 18 Dec 2015 06:08:35 -0800 (PST)
In-Reply-To: <5673C5E1.1050909@uclouvain.be>
References: <CADDmorSTCJx8RAwdiw8v+4Mm+j+OhwGGBHA0W3W9k2zPzimr6w@mail.gmail.com> <5673C5E1.1050909@uclouvain.be>
Date: Fri, 18 Dec 2015 19:38:35 +0530
Message-ID: <CADDmorQhfATxn6sY+UCvxxUjOOH4yU20GdcW=wjG1WKO6NW2tw@mail.gmail.com>
From: Abhishek Verma <abhishekv.verma@gmail.com>
To: Olivier.Bonaventure@uclouvain.be
Content-Type: multipart/alternative; boundary=001a1141bba6d9b6c105272cac86
Archived-At: <http://mailarchive.ietf.org/arch/msg/multipathtcp/KZTChc4LmKCz5eSN7XuW9kNW4SA>
Cc: MultiPath WG <multipathtcp@ietf.org>
Subject: Re: [multipathtcp] Loss and TCP Throughput
X-BeenThere: multipathtcp@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-path extensions for TCP <multipathtcp.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/multipathtcp/>
List-Post: <mailto:multipathtcp@ietf.org>
List-Help: <mailto:multipathtcp-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/multipathtcp>, <mailto:multipathtcp-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 18 Dec 2015 14:08:37 -0000

--001a1141bba6d9b6c105272cac86
Content-Type: text/plain; charset=UTF-8

Hi Olivier,


> I guess that the server has much more capacity than the client and you
> want to optimise the utilisation of the client DSL link.
>
> The DSL network cannot always be considered as a blackbox. When you use
> multiple TCP subflows, you use different source and destination ports. If
> your network uses equal cost multipath, you will end up in using different
> paths for different subflows :
>
> client --- R1---R2 ---R3 ---- server
>             +         +
>            R4---------R5
>
>
> In this example, R1 can load balance the flows on the R1-R2-R3 and
> R1-R4-R5 paths that might have different levels of congestion.
>
> You might use paris traceroute to check the load balancing in the DSL
> network that you use. I would also encourage you to collect packet traces
> of the different transfers and analyse the low performing flows


Very valid point and I did think about it. However, i was unsure if there
is ECMP at the access networks or at this level.

I was under the impression that ECMP or hashing because of link aggregation
(LAGs) usually happens at the edge or the core. However, i hear you and
this would definitely explain why certain TCP streams give lower throughput
while others work well.

Let me try paris traceroute -- I wasnt aware of this.

Thanks for the lead!

Abhishek



>
>
> Olivier
>
>
>
>

--001a1141bba6d9b6c105272cac86
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi Olivier,<div><br></div><div><div class=3D"gmail_extra">=
<div class=3D"gmail_quote"><blockquote class=3D"gmail_quote" style=3D"margi=
n:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
I guess that the server has much more capacity than the client and you want=
 to optimise the utilisation of the client DSL link.<br>
<br>
The DSL network cannot always be considered as a blackbox. When you use mul=
tiple TCP subflows, you use different source and destination ports. If your=
 network uses equal cost multipath, you will end up in using different path=
s for different subflows :<br>
<br>
client --- R1---R2 ---R3 ---- server<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 +=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0+<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0R4---------R5<br>
<br>
<br>
In this example, R1 can load balance the flows on the R1-R2-R3 and R1-R4-R5=
 paths that might have different levels of congestion.<br>
<br>
You might use paris traceroute to check the load balancing in the DSL netwo=
rk that you use. I would also encourage you to collect packet traces of the=
 different transfers and analyse the low performing flows</blockquote><div>=
<br></div><div>Very valid point and I did think about it. However, i was un=
sure if there is ECMP at the access networks or at this level.</div><div><b=
r></div><div>I was under the impression that ECMP or hashing because of lin=
k aggregation (LAGs) usually happens at the edge or the core. However, i he=
ar you and this would definitely explain why certain TCP streams give lower=
 throughput while others work well.</div><div><br></div><div>Let me try par=
is traceroute -- I wasnt aware of this.</div><div><br></div><div>Thanks for=
 the lead!</div><div><br></div><div>Abhishek</div><div><br></div><div><br><=
/div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-le=
ft:1px #ccc solid;padding-left:1ex"><span class=3D"HOEnZb"><font color=3D"#=
888888"><br>
<br>
<br>
Olivier<br>
<br>
<br>
<br>
</font></span></blockquote></div><br></div></div></div>

--001a1141bba6d9b6c105272cac86--

