From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mout.gmx.net (mout.gmx.net [212.227.17.21]) by lists.bufferbloat.net (Postfix) with ESMTPS id DC3C33B2A3 for ; Mon, 18 Jan 2016 11:20:49 -0500 (EST) Received: from [10.24.224.45] ([80.187.100.101]) by mail.gmx.com (mrgmx102) with ESMTPSA (Nemesis) id 0MZOan-1aZJ9H1oZJ-00LFAM; Mon, 18 Jan 2016 17:20:46 +0100 User-Agent: K-9 Mail for Android In-Reply-To: <72B1E8E5-F42D-4B13-A404-5DE82F8658F9@gmail.com> References: <5693E8FA.4000803@darbyshire-bryant.me.uk> <56941191.1010601@darbyshire-bryant.me.uk> <452D0F47-931B-4412-AC59-C308388AA1E4@gmail.com> <02A10F37-145C-4BF9-B428-BC1BDF700135@gmx.de> <079A86FB-D2A6-4787-BC42-D2E200AD3290@gmx.de> <72B1E8E5-F42D-4B13-A404-5DE82F8658F9@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable From: Sebastian Moeller Date: Mon, 18 Jan 2016 17:20:39 +0100 To: Jonathan Morton CC: cake@lists.bufferbloat.net Message-ID: <7B1A2A4E-B5BB-4BD0-A5E0-A36A0C168A3A@gmx.de> X-Provags-ID: V03:K0:QnlkaULgQsHOBjoV2/2LXkjJfY2yoW1Ht5KOGfYZ8g6tYgz7j75 SORAG2dlvbJTdjkFYK9P9Xj+QxiS1lROzhQgCtZN2BzRtjY0ibDI0EH/yVBnp5uAZFeP1XZ 8Sn/rfWxutRqcgB3Uuxdps2Pm5KrdHNB0hz0fI2ThDv556W5jn4qhrX/XwIdz3NK/lbr6P5 mc27FmUFoKlQfdGRXOIvA== X-UI-Out-Filterresults: notjunk:1;V01:K0:U1uzNgl8KAc=:L6lDLd57fIEKjmrl4Oa/t/ E6NVZMq7Sl2CXp5gVVYV9MHsuETJWV3aCWLJofuEk+/mYfJqwlEMl2BKtbQoMC49zTOxryVJ+ NipW+UDUA7u4H/tiJ7Pv636wG/DC8FUV+r0ck+u7sMDsAZ0qWsKr6Im7sTwiBI7cAkakfvHG3 9Ig26MehJFp7+RG+rzJpazQcqUfIhw0KpZvHrDvPb7UcxuPu8Wucr8Gm2ieidKZxpSpkiQti+ VPLffFvQoq9uFrKPk/z8RKCulqsm+cfKKGvXCgt1/rZRq46tf0FBTHOhzWPvsr2XJhOzkGyE/ UfnMyUD0/HsmGxTuyZDmrV3jcl4238EvIYK1OTCA+8Pcv4uNZLYvr1gnS5rSe9A+E0JLDV5TM 7/Bc/9Jl02Ntg1A0dC/17nR0yaPK2OWCEJzwsRFZoFgEBotMR4jHFQ4BGtN59nqjlF2CAiUhk Toh5219qzOf2i/2O4k1VtzpzWRgP6dfRh8Jjxd3v2lWxItulIw2k/BaorP52sAl0HC9GG1hgR tlYhQGQkvmHVnuKxhN5MONMQqBx5tuaxIgXg3Z6MYdVxZBJ8r49pL++UmiKvYJ8hqu4MHvxa3 +uNEJIW0nmc+WiEKd1wAeVPAb5Huk9i/FDAL0X3tsFOKL9aZeyV6H+ntXagf3rzmeKzBGDhkI kqiwgHgXsXIPDnp9/cah6v72qDlH2TBJecBCn6BYr7+X3pEFfB47n0Mv+Xq2b0xqblnCNS1Xy ZfjTgSiaZil1GTl5wd/NKRqaFxlfccKgYwTwkqVpAeckt5n+cBKcwRHldJpsjyV0sKdP8VlgX UFusMbH Subject: Re: [Cake] triple flow isolation X-BeenThere: cake@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: Cake - FQ_codel the next generation List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 18 Jan 2016 16:20:50 -0000 Hi Jonathan, On January 18, 2016 10:37:35 AM GMT+01:00, Jonathan Morton wrote: > >> On 18 Jan, 2016, at 11:21, moeller0 wrote: >>=20 >> Am I right to assume that dust and src host isolation works with the >same counters but simply ignores one of them? > >Yes=2E That=E2=80=99s explicit in the code=2E Thanks, as I am on the road since Sunday morning, I could not chec= k the code easily=2E > >> So if all internal hosts talk to one external host, does this scheme >then equal pure per-flow fairness? I am trying to understand how robust >triple-iso is going to be against attempt at shenanigans by unruly >machines on the internal/external networks=E2=80=A6 > >No=2E If there is only one host on one side (whichever side that happens >to be), then maintaining per-host fairness for that side is trivial=2E=20 >The algorithm will instead maintain per-host fairness for the other >side=2E This derives from the fact that it also maintains per-host >fairness within traffic to each individual host=2E Per-flow fairness is >also still maintained within individual host-pairs=2E > >My measurements show that this aspect of the isolation is not perfect=2E= =20 >There is still some influence from the number of flows, which biases >the actual throughput slightly from ideal per-host fairness=2E This bias >is however *much* smaller than pure per-flow fairness would be, and >I=E2=80=99ve been unable to come up with a robust way of eliminating it >entirely=2E Question: does this slight imperfection only exist for the two ,*_= host settings as well?=20 > >Hence I think it=E2=80=99s reasonable to simply switch on triple isolatio= n by >default, in the near future=2E It does approximately the right thing, >without further configuration, in the great majority of practical cases >(that I can think of), and to a greater extent than the existing >=E2=80=9Cflows=E2=80=9D mode does=2E Not that it matters much, but I agree, if triple-iso does work goo= d enough and does not have any too nasty corner cases it should be the defa= ult=2E=2E=2E People requiring stricter control can always use src_ host or = dst_host if they are not happy with the default I would argue=2E=2E=2E > >I think that might also be a good time to overhaul the documentation >and do some other overdue cleanup=2E +1 especially for the documentation overhaul ;) Best Regards Sebastian > > - Jonathan Morton --=20 Sent from my Android device with K-9 Mail=2E Please excuse my brevity=2E