From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mout.gmx.net (mout.gmx.net [212.227.17.21]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id D1AE03B29E for ; Thu, 5 Nov 2020 03:21:17 -0500 (EST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=gmx.net; s=badeba3b8450; t=1604564476; bh=o+MbFviwTo1sjCrNd+w5wImdf2Q8NJHV+YngQXP73AY=; h=X-UI-Sender-Class:From:To:Cc:Subject:Date:In-Reply-To:References; b=QGhmL0dYx56zHtzMG+chYQwQSiGEbkbmD2GdWIP7UBFHwazdNDle4X4awyedXcDTp 6erglNk88SE9oZEHCBPLXOGMSnodYryKhB4BuZ2Jzo7D2u1aqqviMpqjtoEKJYJoux UIskQvg7Sg6B794A/XYzPlFNS8gOuv3QVcdlQuvQ= X-UI-Sender-Class: 01bb95c1-4bf8-414a-932a-4f6e2808ef9c Received: from [134.76.241.253] ([134.76.241.253]) by web-mail.gmx.net (3c-app-gmx-bap61.server.lan [172.19.172.131]) (via HTTP); Thu, 5 Nov 2020 09:21:16 +0100 MIME-Version: 1.0 Message-ID: From: Sebastian Moeller To: Sam Westwood Cc: bloat@lists.bufferbloat.net Content-Type: text/plain; charset=UTF-8 Date: Thu, 5 Nov 2020 09:21:16 +0100 Importance: normal Sensitivity: Normal In-Reply-To: References: Content-Transfer-Encoding: quoted-printable X-UI-Message-Type: mail X-Priority: 3 X-Provags-ID: V03:K1:QkuuPsbTRMpmtD0futst8MsDNnNdhzxrjdieFOdpuxR2G+99VPYIq9IQqoq6jffXmwbhO Jf3+kvK0iMqoY9CAIVLZVpMkrSaaG46yDBQR7rkYrBzXRaNqHYDt2L6iUydgG0T7U/e12Jsr9FSp WglAFjMnyR1OrmyhpDjArpAoVw+nx9Eqbdk7+h2Ex4Jkban1wKoxOqAo9Y75i7sKJWI+XAaBFfu2 UTXy8ofM4kA2bWBVAbSbhD8l2xWxIkir6QlVcZClWw2R93ozKXkfd2V6r87IeVjjwqK3UF3uUxy5 sU= X-Spam-Flag: NO X-UI-Out-Filterresults: notjunk:1;V03:K0:wQCXIMH01AI=:k5KY/6BEa7MRsG8+Vletup fLA0p0phWDScSOtLlpY9yJiIWBwt9YCG5YYd95mcZd5OhFi6tTC+mv6DyYv48OQyCaif8Ex8T q83Y6dXxG63UJYaXc6wZcqjb8/f0CMBOteyINVhXm5gVpbvmkNepVlTVRKUG9vXe2cjX2b7vF qeKK1CHZoa9+ACfOxgukgpivgFdB6JybkDNZJG3m4PVBeH5fbbXMxYCJbcKkX7qv47cz4mUxE mN2iUJ5fchCljFximm9mcJd3OGvt/Uaccdd/uuIV9925mXLIvdP6IPjhVGGqvGxdiBKzqVBrs ycpBA2rzt4NOYgZnhpQ7wJmKxHQpp/3OabVjCnQCc1Tv1tNPtOrVCZH5J270a58mxS4LVmzYG 1RTT+4EwjyeIlVYxp38uvy3kXRm0Uhnh1V4SYbCYcgacCpFlVgK+/XveRGAf5PZkJvAaoJVHz 03h2957ikxBJBvfpCMyNXYSPUjUCQZClVQhpXgPu2DcB229CdPhuYexqeH+QOkZaXRJbxPYCF OvhwzLXkjOZWHlWXNv6VaXJv89uXFahnK4g7bCAyvsyScdwv/r2tTr5TTVdPJAz7xvcF2fE77 cyQY2tRjicrHY= Subject: Re: [Bloat] We built a new bufferbloat test and keen for feedback X-BeenThere: bloat@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: General list for discussing Bufferbloat List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 05 Nov 2020 08:21:18 -0000 Hi Sam, =C2=A0 first thanks, always good to see more dedicated tools to asses latency und= er load, especially tools that are easy to use and do not require the user = to maintain her/his own=C2=A0dedicated endpoints! More below in-line, prefixed [SM]=2E =C2=A0 =C2=A0 Gesendet:=C2=A0Mittwoch, 04=2E November 2020 um 22:30 Uhr Von:=C2=A0"Sam Westwood" An:=C2=A0bloat@lists=2Ebufferbloat=2Enet Betreff:=C2=A0[Bloat] We built a new bufferbloat test and keen for feedbac= k Hi everyone,=C2=A0 My name is Sam and I'm the co-founder and COO of Waveform=2Ecom=2E At Wave= form we provide equipment to help improve cell phone service, and being in = the industry we've always been interested in all aspects of network connect= ivity=2E Bufferbloat for us has always been interesting, and while there ar= e a few tests out there we never found one that was fantastic=2E So we thou= ght we'd try and build one! My colleague Arshan has built the test, which we based upon the Cloudflare= Speedtest template that was discussed earlier in the summer in a previous = thread=2E =C2=A0 We measure bufferbloat under two conditions: when downlink is saturated an= d when uplink is saturated=2E =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 [SM] This is a decent starting point=2E In add= ition it might be helpful to at least optionally include a test with with b= idirectional saturating load, in the past such tests typically were quite s= uccessful in detecting bufferbloat sources, that were less obvious in the u= ni-directional load tests=2E I am not sure however how well that can work w= ith a browser based test? =C2=A0 =C2=A0 The test involves three stages: Unloaded, Downlink Saturated, and Uplink S= aturated=2E In the first stage we simply measure latency to a file hosted o= n a CDN=2E This is usually around 5ms and could vary a bit based on the use= r's location=2E We use the webTiming API to find the time-to-first-byte, an= d consider that as the latency=2E=20 [SM] Mmmh, I like that this is a relevant latency measure, it migh= t make sense though to make sure users realize that this is not the eqivale= nt number to runing a ICMP eche request against the same endpoint? In the second stage we run a download, while simultaneously measuring late= ncy=2E In the third stage we do the same but for upload=2E Both download an= d upload usually take around 5 seconds=2E=20 [SM] On heavily bufferbloated links it often takes a considerable = amount of time for the bottleneck buffers to drain after a uni-directional = test, so it might make sense to separate the two direction test with an add= itional phase of idle latency measurements=2E If that latency is like the i= nitial unloaded latency, all is well, but if latency slowly ramps down in t= hat phase you have a smoking gun for bad bufferbloat=2E Also, there are link technologies and scheduler techniques that c= an prioritise relative short flows (e=2Eg=2E Comcast's powerboost) to avoid= just measuring the properties of these short duration special modes, it mi= ght make sense to optionally and considerably lengthen the duration of the = test durations to say 30 seconds (empirically powerboost does not engage fo= r a full 30second perid at full rate, but that might be arms race)=2E Also = to assess possible root causes for latency and rate issues, it is very help= ful to show time resolved plots, that show the development of rate and late= ncy over the duration of all phases of the test=2E For example, using longe= r running flent tests I could pinpoint the cyclic channel scanning of my la= ptop's wifi as a source of repeated bufferbloat with a period of ~10 second= s, by seeing evenly spaced latency spikes and rate dips every 10 seconds th= en went away when switching to wired ethernet=2E=2E=2E We show the median, first quartile and the third quartile on distribution = charts corresponding to each stage to provide a visual representation of th= e latency variations=2E For download and upload we have used Cloudflare's s= peedtest backend=2E [SM] This is great, it would be nice though to also add a graphica= l representation, be it a histogram or a cumulative density plot of latenci= es (split out for idle, download, upload and the idle period between down- = and upload)=2E Best Regards Sebastian =C2=A0 You can find the test here: https://www=2Ewaveform=2Ecom/apps/dev-arshan =C2=A0 We built testing it on Chrome, but it works on Firefox and mobile too=2E O= n mobile results may be a little different, as the APIs aren't available an= d so instead we implemented a more manual method, which can be a little noi= sier=2E =C2=A0 This is a really early alpha, and so we are keen to get any and all feedba= ck you have :-)=2E Things that we would particularly like feedback on: How does the bufferbloat measure compare to other tests you may have run o= n the same connection (e=2Eg=2E dslreports, fast=2Ecom[http://fast=2Ecom])H= ow the throughput results (download/upload/latency) look compared to other = toolsAny feedback on the user interface of the test itself? We know that be= fore releasing more widely we will put more effort into explaining bufferbl= oat than we have so far=2E =C2=A0Anything else you would like to give feedb= ack on?We have added a feature to share results via a URL, so please feel f= ree to share these if you have specific feedback=2E=C2=A0 Thanks! Sam and Arshan =C2=A0 ************************* Sam Westwood Co-Founder & COO | RSRF & Waveform E =C2=A0 sam@waveform=2Ecom[mailto:sam@waveform=2Ecom] D =C2=A0=C2=A0(949) 207-3175=C2=A0=C2=A0 T =C2=A0=C2=A01-800-761-3041 Ext=2E 100 W =C2=A0=C2=A0www=2Ersrf=2Ecom[http://www=2Ersrf=2Ecom] & www=2Ewaveform= =2Ecom[http://www=2Ewaveform=2Ecom] =C2=A0_______________________________________________ Bloat mailing list B= loat@lists=2Ebufferbloat=2Enet https://lists=2Ebufferbloat=2Enet/listinfo/b= loat[https://lists=2Ebufferbloat=2Enet/listinfo/bloat]