From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from plane.gmane.org (plane.gmane.org [80.91.229.3]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (Client did not present a certificate) by huchra.bufferbloat.net (Postfix) with ESMTPS id 1307E21F0B3 for ; Fri, 15 Feb 2013 09:13:40 -0800 (PST) Received: from public by plane.gmane.org with local (Exim 4.69) (envelope-from ) id 1U6OrA-0004k9-2q for bloat@lists.bufferbloat.net; Fri, 15 Feb 2013 18:13:52 +0100 Received: from public by plane.gmane.org with local (Exim 4.69) (envelope-from ) id 1U6Or8-0004j6-QD for public-bloat-JXvr2/1DY2fm6VMwtOF2vx4hnT+Y9+D1@plane.gmane.org; Fri, 15 Feb 2013 18:13:50 +0100 Received: from mx12.netapp.com ([216.240.18.77]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1U6OpJ-0003WB-EW; Fri, 15 Feb 2013 18:11:57 +0100 X-IronPort-AV: E=Sophos;i="4.84,674,1355126400"; d="scan'208";a="21784612" Received: from smtp1.corp.netapp.com ([10.57.156.124]) by mx12-out.netapp.com with ESMTP; 15 Feb 2013 09:11:36 -0800 Received: from vmwexceht01-prd.hq.netapp.com (vmwexceht01-prd.hq.netapp.com [10.106.76.239]) by smtp1.corp.netapp.com (8.13.1/8.13.1/NTAP-1.6) with ESMTP id r1FHBYRM009861; Fri, 15 Feb 2013 09:11:35 -0800 (PST) Received: from SACEXCMBX01-PRD.hq.netapp.com ([169.254.2.54]) by vmwexceht01-prd.hq.netapp.com ([10.106.76.239]) with mapi id 14.02.0328.009; Fri, 15 Feb 2013 09:11:34 -0800 From: "Eggert, Lars" To: =?iso-8859-1?Q?Toke_H=F8iland-J=F8rgensen?= Thread-Topic: latency measurement tools? Thread-Index: AQHOChc7BbXKxTY/00qbRHpY5Kaq15h5DkVIgAKiqwA= Date: Fri, 15 Feb 2013 17:11:33 +0000 Message-ID: References: <87half8e9u.fsf@toke.dk> In-Reply-To: <87half8e9u.fsf@toke.dk> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.106.53.51] Content-Type: text/plain; charset="iso-8859-1" Content-ID: <6E593162AD3A09468B514302AC221BFB@tahoe.netapp.com> Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 X-TMDA-Confirmed: Fri, 15 Feb 2013 18:13:50 +0100 X-Mailman-Approved-At: Thu, 21 Feb 2013 10:34:23 -0800 Cc: "bloat-JXvr2/1DY2fm6VMwtOF2vx4hnT+Y9+D1@public.gmane.org" , "Eggert, Lars" Subject: Re: [Bloat] latency measurement tools? X-BeenThere: bloat@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: General list for discussing Bufferbloat List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 15 Feb 2013 17:13:41 -0000 Thanks, I'll check it out! Lars On Feb 14, 2013, at 0:56, Toke H=F8iland-J=F8rgensen wrote: > "Eggert, Lars" writes: >=20 >> what are the preferred measurement tools for latency under load? Are >> any publicly available or is everyone basically rolling their own? >=20 > Hi Lars >=20 > Usually, people employ netperf and/or iperf to load up a link, and then > do some sort of latency measurement while the link is loaded (through > netperf or just straight ping). I'm the author of a python wrapper that > handles this in a reproducible way, and which also has a prototype of > the Realtime Response Under Load (RRUL) test incorporated. The wrapper > is available here: >=20 > https://github.com/tohojo/netperf-wrapper >=20 > (Or just install it through pip with `pip install netperf-wrapper`). >=20 > Also see this page for some example test plots: >=20 > http://www.bufferbloat.net/projects/codel/wiki/RRUL_Rogues_Gallery >=20 > And this one for some general advise on benchmarking (though that is > mostly relevant at lower bandwidths): >=20 > http://www.bufferbloat.net/projects/codel/wiki/Best_practices_for_benchma= rking_Codel_and_FQ_Codel >=20 >> (I'm esp. interested in tools that can deal with 10 GbE and faster LAN >> fabrics.) >=20 > The RRUL test prototype runs four simultaneous TCP streams in each > direction which can generally load up pretty much any link on a modern > Linux kernel (though I haven't tested it personally on high-speed > networks due to lack of hardware). >=20 > Good luck with the benchmarking. :) >=20 > -Toke >=20 > --=20 > Toke H=F8iland-J=F8rgensen > toke@toke.dk