From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from homiemail-a54.g.dreamhost.com (sub4.mail.dreamhost.com [69.163.253.135]) by huchra.bufferbloat.net (Postfix) with ESMTP id 513EB21F359 for ; Sat, 11 Apr 2015 08:53:12 -0700 (PDT) Received: from homiemail-a54.g.dreamhost.com (localhost [127.0.0.1]) by homiemail-a54.g.dreamhost.com (Postfix) with ESMTP id 479FB400F8E13; Sat, 11 Apr 2015 08:53:12 -0700 (PDT) Received: from kmnimac.local (c-50-156-111-45.hsd1.ca.comcast.net [50.156.111.45]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: nichols@pollere.net) by homiemail-a54.g.dreamhost.com (Postfix) with ESMTPSA id 2C22D400F8E11; Sat, 11 Apr 2015 08:53:12 -0700 (PDT) Message-ID: <55294367.4000101@pollere.com> Date: Sat, 11 Apr 2015 08:53:11 -0700 From: Kathleen Nichols User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.10; rv:31.0) Gecko/20100101 Thunderbird/31.6.0 MIME-Version: 1.0 To: bloat@lists.bufferbloat.net References: <2E760E83-7224-4B9B-B918-B3EAD578E831@gmail.com> <67BE685A-724C-45D7-A619-EC49498AF165@gmail.com> <9A8B9375-C683-4A85-9D65-564AEECE1300@gmail.com> In-Reply-To: Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: quoted-printable Subject: Re: [Bloat] Questions about the use of HTB & fq_codel in simple.qos, simplest.qos 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: Sat, 11 Apr 2015 15:53:41 -0000 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Just to clarify...I wrote the sfqcodel code in ns using existing sfq code so I could put something together quickly to show the benefits, most specifically, of binning acks and other short packets separately from longer data transfer packets. At the time, we had been explaining that AQM does one job (controlling queue) but that some form of separation/isolation was needed to deal with the well known ack issue. The work we do to understand issues and algorithms is often different from the work we do to deploy something. However, both are important. On 4/10/15 5:14 AM, Rich Brown wrote: > Hi Jonathan, >=20 > Thanks for the response - this is really helpful. As I said in my=20 > original note, I'm casting about for usable advice that we can > offer to people today. I want it to be unarguably true, and I want > to understand the nuances so that I don't get tripped up by > Gotcha's. >=20 > I'm currently focussed on OpenWrt, as that's a platform where it's=20 > easy to roll out fq_codel in current builds simply by downloading > a couple packages. I'll work on a draft note with recommendations > for OpenWrt and send it past the list for comments. Thanks again. >=20 > Rich >=20 >=20 > On Apr 9, 2015, at 6:02 PM, Jonathan Morton > wrote: >=20 >>=20 >>> On 10 Apr, 2015, at 00:35, Rich Brown >>> wrote: >>>=20 >>> - Why do we provide an HTB-based shaper in simple.qos and=20 >>> simplest.qos? - Do the shapers in these sqm-scripts actually=20 >>> limit bandwidth for various kinds of traffic? Might that not=20 >>> leave unused bandwidth? - Or do they just shunt certain >>> packets to higher or lower priority fq_codel >>> tiers/bands/levels (terminology used in Dave's note below)? - >>> And if the latter, how does the "link" (I'm not sure of the >>> proper term) know which of the tiers/bands/levels to dequeue >>> next? >>=20 >> The short answer is: because cake isn=92t out in the real world >> yet. We=92re working on it. >>=20 >> HTB and IFB as used in those scripts is a stopgap solution, to >> take control of the bottleneck queue so that fq_codel can work on >> it. Cake includes a shaper which does the job more effectively >> and more efficiently. >>=20 >> Ultimately, what we=92d like is for fq_codel (or even something as=20 >> sophisticated as cake) to be implemented in the *actual* >> bottleneck queues, so that artificially taking control of the >> bottleneck isn=92t necessary. >>=20 >>> I'll state up front that I'm not entirely clear on the=20 >>> distinction between shapers, qdisc's, IFBs, etc. But I'm >>> groping around for a simple, clear recommendation for what we >>> should tell people to do so they can: a) Make their router work >>> very well, with minimal latency b) Spend their time more >>> usefully than tweaking QoS/priority settings (for example, by >>> actually playing the game that whose lag you're trying to >>> minimize :-) >>=20 >> If they=92ve got a router with the sqm-scripts installed, use >> those and follow the directions. The implementation is a little >> messy, but it works and it keeps things simple for the user. >>=20 >> When cake arrives, the implementation will get simpler and more=20 >> efficient. >>=20 >> - Jonathan Morton >>=20 >=20 >=20 >=20 > _______________________________________________ Bloat mailing list > Bloat@lists.bufferbloat.net=20 > https://lists.bufferbloat.net/listinfo/bloat >=20 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlUpQ2cACgkQ+dkULylClLbLZwCePPF//nCLuOGKo6HUzYdx7k57 9y4AoORL7oNWNpOOUli9eWcBY3R3Px1Q =3Dj1yO -----END PGP SIGNATURE-----