From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from alln-iport-2.cisco.com (alln-iport-2.cisco.com [173.37.142.89]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (Client CN "alln-iport.cisco.com", Issuer "HydrantID SSL ICA G2" (verified OK)) by huchra.bufferbloat.net (Postfix) with ESMTPS id 3EFDE21F41B for ; Fri, 27 Mar 2015 14:45:14 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=7276; q=dns/txt; s=iport; t=1427492743; x=1428702343; h=from:to:subject:date:message-id:mime-version; bh=Ttqo03SrIyb/yAQ1CCRpawr+Glhm5u0f/pydTDvaIm8=; b=NSMwbPJUvPgJh5bF59g5Py+nC60tARTrMP5WzFBIETZOdzX+DD1PkdWI LZMYp0sGjQwiW7qxxVl0C9zucg/+qxB49D4OETZX0gF4uxO0fSvgcRdZf dw7uNO4WfY7m7xbEvzJhxJU7hdLZxtDjMv8j2B9ynR06U/6tu18eEFpf9 s=; X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: A0ALBQA6zhVV/5FdJa1cgkNDUl7FU4VzAoE6TAEBAQEBAX2EFgEELV4BKlYmAQQbiCcNonepbwEBAQEBAQEDAQEBAQEBAQEWBI9vLYMigRYFkFeVKIh5IoNugjN/AQEB X-IronPort-AV: E=Sophos;i="5.11,481,1422921600"; d="scan'208,217";a="136097434" Received: from rcdn-core-9.cisco.com ([173.37.93.145]) by alln-iport-2.cisco.com with ESMTP; 27 Mar 2015 21:45:13 +0000 Received: from xhc-aln-x02.cisco.com (xhc-aln-x02.cisco.com [173.36.12.76]) by rcdn-core-9.cisco.com (8.14.5/8.14.5) with ESMTP id t2RLjDB1023276 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL) for ; Fri, 27 Mar 2015 21:45:13 GMT Received: from xmb-aln-x05.cisco.com ([169.254.11.172]) by xhc-aln-x02.cisco.com ([173.36.12.76]) with mapi id 14.03.0195.001; Fri, 27 Mar 2015 16:45:12 -0500 From: "Bill Ver Steeg (versteb)" To: "bloat@lists.bufferbloat.net" Thread-Topic: setting queue depth on tail drop configurations of pfifo_fast Thread-Index: AdBo10mlMURNVJRVS0S5jn0w/qD+Iw== Date: Fri, 27 Mar 2015 21:45:11 +0000 Message-ID: Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.117.75.44] Content-Type: multipart/alternative; boundary="_000_AE7F97DB5FEE054088D82E836BD15BE9319902F7xmbalnx05ciscoc_" MIME-Version: 1.0 Subject: [Bloat] setting queue depth on tail drop configurations of pfifo_fast 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, 27 Mar 2015 21:45:43 -0000 --_000_AE7F97DB5FEE054088D82E836BD15BE9319902F7xmbalnx05ciscoc_ Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Bloaters- I am looking into how Adaptive Bitrate video algorithms interact with the v= arious queue management schemes. I have been using the netperf and netperf = wrapper tools, along with the macros to set the links states (thanks Toke a= nd Dave T). I am using HTB rather than BQL, which may have something to do = with the issues below. I am getting some interesting ABR results, which I w= ill share in detail with the group once I write them up. I need to set the transmit queue length of my Ubuntu ethernet path while ru= nning tests against the legacy pfifo_fast (tail drop) algorithm. The defau= lt value is 1000 packets, which boils down to 1.5 MBytes. At 100 Mbps, this= gives me a 120ms tail drop buffer, which is big, but somewhat reasonable. = When I then run tests at 10 Mbps, the buffer becomes a 1.2 second bloaty bu= ffer. When I run tests at 4 Mbps, the buffer becomes a 3 second extra-bloat= y buffer. This gives me some very distinct ABR results, which I am looking = into in some detail. I do want to try a few more delay values for tail drop= at 4 Mbps. https://www.bufferbloat.net/projects/codel/wiki/Best_practices_for_benchmar= king_Codel_and_FQ_Codel says to set txqueuelen to the desired size, which m= akes sense. I have tried several ways to do this on Ubuntu, with no glory. = The way that seems it should have worked was "ifconfig eth8 txqueuelen 100"= . When I then check the txqueuelen using ifconfig, it looks correct. Howeve= r, the delay measurements still stay up near 3 seconds under load. When I c= heck the queue depth using "tc -s -d qdisc ls dev ifb_eth8", it shows the v= ery large backlog in pfifo_fast under load. So, has anybody recently changed the ethernet/HTB transmit packet queue siz= e for pfifo_fast in Ubuntu? If so, any pointers? I will also try to move ov= er to BQL and see if that works better than HTB...... I am not sure that my= ethernet drivers have BQL support though, as they complain when I try to l= oad it as the queue discipline. Thanks in advance Bill VerSteeg --_000_AE7F97DB5FEE054088D82E836BD15BE9319902F7xmbalnx05ciscoc_ Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable

Bloaters-

 

I am looking into how Adaptive Bitrate video algorit= hms interact with the various queue management schemes. I have been using t= he netperf and netperf wrapper tools, along with the macros to set the link= s states (thanks Toke and Dave T). I am using HTB rather than BQL, which may have something to do with the is= sues below. I am getting some interesting ABR results, which I will share i= n detail with the group once I write them up.

 

I need to set the transmit queue length of my Ubuntu= ethernet path while running tests against the legacy pfifo_fast (tail drop= ) algorithm.  The default value is 1000 packets, which boils down to 1= .5 MBytes. At 100 Mbps, this gives me a 120ms tail drop buffer, which is big, but somewhat reasonable. When I then= run tests at 10 Mbps, the buffer becomes a 1.2 second bloaty buffer. When = I run tests at 4 Mbps, the buffer becomes a 3 second extra-bloaty buffer. T= his gives me some very distinct ABR results, which I am looking into in some detail. I do want to try a fe= w more delay values for tail drop at 4 Mbps.

 

https://www.buff= erbloat.net/projects/codel/wiki/Best_practices_for_benchmarking_Codel_and_F= Q_Codel says to set txqueuelen to the desired size, which makes sense. I have tried several ways to do this = on Ubuntu, with no glory. The way that seems it should have worked was R= 20;ifconfig eth8 txqueuelen 100”. When I then check the txqueuelen us= ing ifconfig, it looks correct. However, the delay measurements still stay up near 3 seconds under load. When I check t= he queue depth using “tc –s –d qdisc ls dev ifb_eth8̶= 1;, it shows the very large backlog in pfifo_fast under load.

 

So, has anybody recently changed the ethernet/HTB tr= ansmit packet queue size for pfifo_fast in Ubuntu? If so, any pointers? I w= ill also try to move over to BQL and see if that works better than HTB̷= 0;… I am not sure that my ethernet drivers have BQL support though, as they complain when I try to load it as the que= ue discipline.

 

Thanks in advance

Bill VerSteeg  

 

--_000_AE7F97DB5FEE054088D82E836BD15BE9319902F7xmbalnx05ciscoc_--