From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mout.gmx.net (mout.gmx.net [212.227.15.15]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 8CC9A3B2BC; Sat, 4 Jun 2016 10:01:50 -0400 (EDT) Received: from hms-beagle.lan ([93.237.69.26]) by mail.gmx.com (mrgmx001) with ESMTPSA (Nemesis) id 0MTjqS-1b0MRQ19Lk-00QPhx; Sat, 04 Jun 2016 16:01:48 +0200 Content-Type: text/plain; charset=utf-8 Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2104\)) From: moeller0 In-Reply-To: <01BEA343-7C07-46FA-8DC4-07BF26309FC8@gmail.com> Date: Sat, 4 Jun 2016 16:01:47 +0200 Cc: Andrew McGregor , cake@lists.bufferbloat.net, "codel@lists.bufferbloat.net" Content-Transfer-Encoding: quoted-printable Message-Id: References: <22371476-B45C-4E81-93C0-D39A67639EA0@gmx.de> <857AEE56-E7DB-4981-B32E-82473F877139@gmail.com> <8AB0D25D-C1CA-45F1-889E-2F73CF8C44F7@gmail.com> <323AFC22-A092-4F59-8197-AF21EF73FD58@gmail.com> <274D3A0FA900FD47AA6B56991AAA32FDC5529FC8@wtl-exchp-1.sandvine.com> <574478B4.7080103@taht.net> <39F38477-A877-4C1B-9B7F-BB3358425F17@gmail.com> <0eb223f9-2873-7f53-c2ce-c6867ddec17c@gmail.com> <48A25043-19E2-4BB7-B634-A4003F7BE6F8@gmail.com> <01BEA343-7C07-46FA-8DC4-07BF26309FC8@gmail.com> To: Jonathan Morton X-Mailer: Apple Mail (2.2104) X-Provags-ID: V03:K0:zomm88oYPBTryF0w0zaZGfZl8hohZ2kuMJizZYzQIhzVcZpMWXD UoR7XtxvsXg+qfcFR9svS/vtdU8TVmNiSMl0VDr/s9SYLUDd+0CYLU4PMaDhUwxEmjc2HCj pJHh7L4FJjeVgG3Dc0Dwu9McHS1AgD8omlpJvNAE2QOr332wBrYSA+NIgKtyQbPblPcUKXE PiPZxvURUr4u1DqsngURA== X-UI-Out-Filterresults: notjunk:1;V01:K0:iTy6I+J1USw=:eaDfrA036gJ8iMOzBXE2Hf YQmOXvkRfyUauaCYw9i8plOf/8lMLKroRTT8SmHjJwanBnK0ujc3+zAg84ytTTmrBSECS8gkU 7u+hbChVrKKMUjvL1NuVEg6ptKc602W7UIdwvPQAjvtsg+k2EUq/HKDijy3Ld15MFvTKBNL/i YCXep6Nv6z0mT6Ef3ED/Nc6Urntq/6AF9i2PtvEj2aI+v3iLVACLVePY0uLUk07KatcjdkOtq aioRsyoTK7en6HtTN6i+m/G0BW6JQ5JMOU/7Aak0M5D9uE2O3EMJ1p0nQcvxYo/YY837I2cCu KuNUlhOIXjeU4KAU1yItvScQDIUa/2uA+A9tbYze3VUgnYFxstSE2VKbrBtflpbegtHurRe1D /yS1BUQJ9p9WYuKhiyGvNp3FBYzKn/stY1+2+KwecTcLMXVWWBkoK5UfvE+hxw1mMh1lUpbpp fP/1t7I3K8rH317g3hwvjoQmBfD+kTjXw1FPKqpsCbkYED6GX4OEcYT/zWsiVs8LPxHNTOuyW AkQVQ6kzupID8ATAwBTbQmBSOoQdE/9Qxwe7a+D23g8Mpb/HOLQ7cdTPYkdAFwWffIYyIb36q rwHjazQ/6+JTAggS1QY22IvkdhEu87K7cqIS5j8mp02tRwUbr5NFU+L0MzGZ7OaxWt5qza/Vn xb99vi5ay5EtgVVC/AhK4s85DR8lZrf5BCWbxtrUwR390puwzaJ9cPAJAROft/OFrA/iHXqfj TWTBz4pQu6OH6KSxCPEU2iPgc4i1lhJlkDl8Kxn0Myk4ZoLJBEusc2Pc1jIVt3jPiTqhK2i/b 6JaatsQ Subject: Re: [Cake] [Codel] Proposing COBALT 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: Sat, 04 Jun 2016 14:01:50 -0000 Hi Jonathan, > On Jun 4, 2016, at 15:55 , Jonathan Morton = wrote: >=20 >=20 >> On 4 Jun, 2016, at 04:01, Andrew McGregor = wrote: >>=20 >> ...servers with ECN response turned off even though they negotiate = ECN. >=20 > It appears that I=E2=80=99m looking at precisely that scenario. >=20 > A random selection of connections from a packet dump show very high = marking rates, which are apparently acknowledged using CWR, but a = subsequent dropped packet (probably due to queue overflow) takes many = seconds to be retransmitted (I=E2=80=99m using a rather high memory = limit for observation purposes). >=20 > Overall the TCP behaviour is approximately normal for NewReno on a = dumb FIFO, and the ECN signalling is completely ignored. This doesn=E2=80= =99t rule out the possibility that it=E2=80=99s a different Reno = relative, such as Westwood+ or Compound. >=20 > There=E2=80=99s often more than one CWR per RTT. This isn=E2=80=99t a = consistent characteristic; some connections have normal-looking CWRs = while others issue them every three packets, as if they=E2=80=99re = fishing for =E2=80=9Cmore accurate=E2=80=9D ECN feedback. It might vary = by host; I didn=E2=80=99t keep track of that. But this can=E2=80=99t be = DCTCP; even that should back off in the face of a 100% marking rate, = which is often achieved at my low bandwidth and with very persistent = queues. >=20 > Other servers respond normally to ECN signals, ruling out interference = by my ISP. It=E2=80=99s possible the ECE flag is wiped and the CWRs are = faked, but there=E2=80=99s no legitimate reason to do that. The CWRs = ultimately make no difference, since at 100% CE marks, every ack has ECE = set anyway. >=20 > Turning off ECN negotiation at the client results in a much better = managed queue with similar throughput. It=E2=80=99s not immediately = obvious whether that=E2=80=99s due to a functioning congestion response = or simply the AQM clearing out the queue the hard way. It=E2=80=99ll be = interesting to see what effect COBALT has here, when I get it to = actually work. >=20 > As for who these servers are: Valve Software=E2=80=99s Steam platform. = I did say they were large and popular. Maybe cake should allow to switch from the default mark by ECN = policy to mark by drop per command line argument? At least that would = allow much easier in the field testing=E2=80=A6 As is there is only the = option of disabling ECN at the endpoint(s)=E2=80=A6 Best Regards Sebastian >=20 > - Jonathan Morton >=20 > _______________________________________________ > Cake mailing list > Cake@lists.bufferbloat.net > https://lists.bufferbloat.net/listinfo/cake