From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from bifrost.lang.hm (mail.lang.hm [64.81.33.126]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by huchra.bufferbloat.net (Postfix) with ESMTPS id 9B59E21F629; Fri, 25 Jul 2014 14:20:59 -0700 (PDT) Received: from asgard.lang.hm (asgard.lang.hm [10.0.0.100]) by bifrost.lang.hm (8.13.4/8.13.4/Debian-3) with ESMTP id s6PLKrvK005954; Fri, 25 Jul 2014 14:20:53 -0700 Date: Fri, 25 Jul 2014 14:20:53 -0700 (PDT) From: David Lang X-X-Sender: dlang@asgard.lang.hm To: Neil Davies In-Reply-To: Message-ID: References: <03292B76-5273-4912-BB18-90E95C16A9F5@pnsol.com> <66FF8435-C8A5-4596-B43A-EC12D537D49E@gmx.de> User-Agent: Alpine 2.02 (DEB 1266 2009-07-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Cc: cerowrt-devel , bloat Subject: Re: [Cerowrt-devel] [Bloat] Check out www.speedof.me - no Flash 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: Fri, 25 Jul 2014 21:20:59 -0000 On Fri, 25 Jul 2014, Neil Davies wrote: > Sebastian > > On 25 Jul 2014, at 15:17, Sebastian Moeller wrote: > >> But how do you propose to measure the (bottleneck) link capacity then? >> It turns out for current CPE and CMTS/DSLAM equipment one typically can not >> relay on good QoE out of the box, since typically these devices do not use >> their (largish) buffers wisely. Instead the current remedy is to take back >> control over the bottleneck link by shaping the actually sent traffic to stay >> below the hardware link capacity thereby avoiding feeling the consequences of >> the over-buffering. But to do this is is quite helpful to get an educated >> guess what the bottleneck links capacity actually is. And for that purpose a >> speediest seems useful. > > > I totally agree that what you are trying to do is to take control "back" for > the upstream delay and loss (which is the network level activity that directly > influences QoE). Observationally the "constraining link" is the point at which > the delay and loss start to grow as the the offered load is increased (there > are interesting interactions with the scheduling in the CMTS/3GPP node B - but > they are tractable) if we don't have direct access to the constraint (which in > the CPE, for ADSL you have) we track that "quality attenuation" inflection > point. Saturating the path is a bit of a sledgehammer (and has nasty > cost/scaling implications). The thing is that there is little effect on latency until the congestion starts, so we can only measure the problem when there is congestion. Saturating the link is a bit of a sledgehammer, but there really isn't any other way to get to the worst case situation. In terms of scaling, have the server detect that all the requests have combined to saturate it's link, and have it tell the clients that it's overloaded, wait a random amount of time and retry (or try another location) cost of bandwidth for this is just something to get someone to pay for (ideally someone with tons of bandwidth already who won't notice this sort of test, even if there are a few going on at once.) David Lang