From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mout.gmx.net (mout.gmx.net [212.227.15.19]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "mout.gmx.net", Issuer "TeleSec ServerPass DE-1" (verified OK)) by huchra.bufferbloat.net (Postfix) with ESMTPS id DC25021F129 for ; Tue, 24 Mar 2015 00:47:54 -0700 (PDT) Received: from u-089-d060.biologie.uni-tuebingen.de ([134.2.89.60]) by mail.gmx.com (mrgmx002) with ESMTPSA (Nemesis) id 0M4WRI-1ZQCja2FuW-00yjUi; Tue, 24 Mar 2015 08:47:50 +0100 Content-Type: text/plain; charset=windows-1252 Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.6\)) From: Sebastian Moeller In-Reply-To: Date: Tue, 24 Mar 2015 08:47:48 +0100 Content-Transfer-Encoding: quoted-printable Message-Id: References: <08BAF198-87C5-42B8-8899-53F34E47156E@gmail.com> <896FAE61-B45A-4F34-9449-5ADB82194DD9@gmx.de> <48350C2E-C33A-4534-84BC-5D56F4AAF0EA@gmail.com> <8AC58249-8199-405B-997A-E8F7285A34FB@gmx.de> To: Jonathan Morton X-Mailer: Apple Mail (2.1878.6) X-Provags-ID: V03:K0:MhHM/wgR87t7mix0ALdrEzzGyd6vEV1dCAp67bLIBlmBjqDHgwa BKE7bageDon0BqccFt8gkYY9zahsuXDNnOG+mYDj98SrDakGAk/F8zrl6VKcxNppbGEjpRM rcsRWbHnPjFZehiYawXOnlzNeRhSrVAwEqhB/ffa/2/iDtaegE2u8LQtuEaUcN+kfUTvBID pWNR5s6wO4rynPIbKlUNA== X-UI-Out-Filterresults: notjunk:1; Cc: "cerowrt-devel@lists.bufferbloat.net" Subject: Re: [Cerowrt-devel] archer c7 v2, policing, hostapd, test openwrt build X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: Development issues regarding the cerowrt test router project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 24 Mar 2015 07:48:23 -0000 Hi Jonathan, On Mar 24, 2015, at 04:16 , Jonathan Morton = wrote: >=20 >> On 24 Mar, 2015, at 02:00, Sebastian Moeller wrote: >>=20 >> So I got around to a bit of rrul testing of the dual egress idea to = asses the cost of IFB, but the results are complicated (so most likely I = screwed up). >=20 > IFB is normally used on the download direction (as a substitute for a = lack of AQM at the ISP), so that=92s the one which matters. Can you try = a unidirectional test which exercises only the download direction? =20 I will try to get around to this later this week, not sure = whether I manage though. > This should get the clearest signal - without CPU-load interference = from the upload direction. I agree, but if IFB redirection truly is costly enough to bother = with fixing/avoiding it should also cause a noticeable effect on the = full ingress-egress stress test, I would assume. But at least for my = limited tests it did not=85 Or to put it differently, if avoiding the = IFB does not increase bandwidth use under full load it is not going to = help with getting a router=92s combined shaping performance improve, or = do I see something wrong. Now maybe it is a critical building block for = better performance that is masked at full load by something else, that = is why I tried the reduced bandwidth loads (35000 bidirectional) but = even there the effect was rather mild=85 That said, I will retry with = download shaping only (vie se00 egress) and simplest.qos (instead of = simple.qos) to move the heavy filtering out of the way. I wonder whether = anybody has a good idea of how to measure the router=92s cpu usage = during a rrul test (maybe the main effect of avoiding IFB is not to = increase bandwidth usage, but to free up cpu cycles for performing other = task, which still would be quite valuable, I guess) Best Regards Sebastian >=20 > - Jonathan Morton >=20