From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-oi0-x22e.google.com (mail-oi0-x22e.google.com [IPv6:2607:f8b0:4003:c06::22e]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id EFFFF3B260 for ; Mon, 25 Apr 2016 12:59:12 -0400 (EDT) Received: by mail-oi0-x22e.google.com with SMTP id x201so182589240oif.3 for ; Mon, 25 Apr 2016 09:59:12 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-transfer-encoding; bh=DxATkjZDlXYprp7jbaZgZaeMNkVvx1Fz+k33yzDH7u0=; b=i3+4IGU6tc7t2TsSvgSvFCl3TmwvIPrC2QpfMYEfOjOkz1R3p4lO051rA3ov2WIz4S 6lYCi4FUVqhmwDcwLDXwEpixmz7wynH+RVyrEuhUTEpi62ThST5tgs0cOudHkIh8r+7s aE4k9PbGsqXjbBudklsO79UPO1JdoSXHOV2Ix9bNgmHi6pyUPlJaRzUdyvNYWnRU44C5 rEVII2P15L+nMR/z2SevPOINKs9YafU5tnLc0RTHKXvN1dqL9YMgJ4We0KpUNe8pTfgQ 5QQGawvvU3HkFwhnJdRo7FYHwSB5zVq5YJvuEWEO79ZAYZ2GyGOrAdW/V9dQuQSBjsjx 21Pw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:date :message-id:subject:from:to:cc:content-transfer-encoding; bh=DxATkjZDlXYprp7jbaZgZaeMNkVvx1Fz+k33yzDH7u0=; b=jeb3QHPsFkgykiFbsVi1JO6Y0O2WLe1d3CIwT9nMLxzBvF3B6Xij5vlpRY8vZnR2VS +U8DpdfNDu6KGLxZziPBAUwgOiwQRoFDdWIdwA1lTu3y9HVyssqUmECHiZLu9zyejeT+ ShlAQnFtLX11dEiro6RSKCaszG0TbmzCDi66T4C7msNkiFM9KExXvcXlyTnEpaEKxv6w IdJnI53XdhsXFS3BgzGbdEAzONkmEheVxhFbxcse/B7uqwf8i/Nik+1dO15zmc/dkHQj cQ+2SjKlNYz9VSh0cQSi87vfQojjt0XNT61hMchspSmDiIAbr1xKsG/lc27dbxx97BKd s9lQ== X-Gm-Message-State: AOPr4FUYw2p5beGbsqjHftiSja9DxhaMM5sNSgTOnxhA2Jkvkv4fCbOOwZLTvduqMkSs9e8C4ZslidcOtEhjDQ== MIME-Version: 1.0 X-Received: by 10.157.57.132 with SMTP id y4mr13444696otb.169.1461603552371; Mon, 25 Apr 2016 09:59:12 -0700 (PDT) Received: by 10.202.79.194 with HTTP; Mon, 25 Apr 2016 09:59:12 -0700 (PDT) In-Reply-To: References: <571BC61C.4070008@darbyshire-bryant.me.uk> <571BD3BC.2090405@darbyshire-bryant.me.uk> Date: Mon, 25 Apr 2016 09:59:12 -0700 Message-ID: From: Dave Taht To: Arie Cc: Alec Robertson , cake@lists.bufferbloat.net Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Cake] Beating bufferbloat 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: Mon, 25 Apr 2016 16:59:13 -0000 It turns out babeld from trunk worked, so I'm happy. I still have not the foggiest idea as to how to go about testing the triple-isolate feature.... On Mon, Apr 25, 2016 at 9:01 AM, Dave Taht wrote: > On Sun, Apr 24, 2016 at 3:37 AM, Arie wrote: >> If you want a very recent cake version, you could use my build from here= : >> http://ariekanarie.nl/openwrt/mvebu/ It's based on the very feature heav= y >> OpenWRT build by trondah ( https://forum.openwrt.org/viewtopic.php?id=3D= 50914 >> ), but using more recent cerowrt and cake stuff. >> >> You'll want to flash the shelby factory.img if still on stock firmware, = else >> shelby sysupgrade.tar. > > Thank you! I flashed the caiman build to my linksys box That has the > correct looking cake in it. > > Unfortunately babeld is not in that build, could you stick it in? (I > am heavily reliant on that to get around my insanely complex testbed) > > It is also based on the 4.1.20 kernel (thought openwrt was on 4.4). I > have had all sorts of weird issues with sending test traffic to the > prior build (flow starvation), will try this. > > It is nice to finally see the triple-isolate feature land somewhere it > could be tested. > > > root@linksys1200ac opkg# tc -s qdisc show dev eth0 > qdisc cake 800d: root refcnt 9 bandwidth 14Mbit diffserv4 > triple-isolate rtt 100.0ms raw > Sent 802270 bytes 2932 pkt (dropped 0, overlimits 403 requeues 0) > backlog 0b 0p requeues 0 > memory used: 21528b of 700000b > capacity estimate: 14Mbit > Bulk Best Effort Video Voice > thresh 14Mbit 13125Kbit 10500Kbit 3500Kbit > target 5.0ms 5.0ms 5.0ms 5.2ms > interval 100.0ms 100.0ms 100.0ms 100.2ms > pk_delay 1us 287us 14us 2us > av_delay 0us 44us 0us 1us > sp_delay 0us 1us 0us 0us > pkts 67 1302 72 1491 > bytes 10935 529189 6480 255666 > way_inds 0 0 0 3 > way_miss 40 44 72 931 > way_cols 0 0 0 0 > drops 0 0 0 0 > marks 0 0 0 0 > sp_flows 0 0 0 1 > bk_flows 0 0 0 0 > last_len 171 609 90 171 > max_len 187 3028 90 483 > > > > >> >> >> >> On 24 April 2016 at 00:22, Alec Robertson wr= ote: >>> >>> Dear All, >>> >>> I=E2=80=99ve realised that I have been responding to Kevin rather than = to the >>> mailing list - my bad! >>> >>> I think I will purchase a Linksys WRT1900ACS as it seems to be fairly w= ell >>> regarded and is easily accessible in the UK. >>> >>> How do I go about setting up Cake on it? >>> >>> -- >>> Alec Robertson >>> >>> On 23 April 2016 at 9:00:52 pm, Alec Robertson (alecrobertson13@gmail.c= om) >>> wrote: >>> >>> Dear Kevin, >>> >>> I did look at the Linksys WRT1200AC but it seems to get some rather poo= r >>> reviews on Amazon? >>> >>> -- >>> Alec Robertson >>> >>> On 23 April 2016 at 8:58:02 pm, Kevin Darbyshire-Bryant >>> (kevin@darbyshire-bryant.me.uk) wrote: >>> >>> >>> >>> On 23/04/2016 20:42, Alec Robertson wrote: >>> >>> Dear Kevin, >>> >>> That=E2=80=99s very useful thanks. >>> >>> You say the TP-LINK Archer C7 should just be okay. What could I get tha= t I >>> know will last me for a long time? What=E2=80=99s got good WiFi range t= oo? >>> >>> The honest answer is I've absolutely no idea and I'm in that dilemma >>> myself. There's a remake of the linksys WRT range (WRT1200????) that >>> apparently is very powerful, I guess the issue is how far along OpenWrt= is. >>> I might have the wrong end of hte stick but I think Dave Taht may have >>> something working. >>> >>> Apologies, I'm not really a mine of information. >>> >>> >>> I=E2=80=99m using powerline at the moment but fed up with it disconnect= ing. I >>> think it is probably the TP-LINK adaptors I am using (known issue >>> apparently) but wiring up the house is unpractical at the moment. I don= =E2=80=99t >>> think there is a better solution really. >>> >>> -- >>> Alec Robertson >>> >>> On 23 April 2016 at 8:00:00 pm, Kevin Darbyshire-Bryant >>> (kevin@darbyshire-bryant.me.uk) wrote: >>> >>> Hi Alec, >>> >>> I'm not familiar with TalkTalk but they sound like they do similar >>> things to Sky - Sky just need a 'login ID' as part of the DHCP request >>> packet (which funnily enough are the PPPoA/E login details) >>> >>> In terms of speed sacrifice, erm, none really. I've set 40mpbs incoming >>> and 9990kbps for outgoing on a 40000/9999 link as reported by the >>> modem. Probably critically I've set the packet overheads to 12, and I >>> now can't remember why... there's an on-wire vlan tag (4 bytes) but the >>> reason for the other 8 have fallen out of the brain cell. >>> >>> I've a semi-regular backup job overnight that on a bad day overruns int= o >>> the day - a week or so ago it ran for something like 2 days and I had >>> absolutely no idea - thinkbroadband's ping monitor was registering >>> something like an extra 5mS latency over the baseline, peaks were >>> something like 25mS - backup stats and openwrt's stats package were >>> registering the full 10mbps uplink in use during that time. >>> >>> Does that help? >>> >>> Kevin >>> >>> >>> >>> >>> On 23/04/16 19:40, Alec Robertson wrote: >>> > Dear Kevin, >>> > >>> > I am on TalkTalk which uses IPoE, so no PPPoE use at all, as far as I >>> > know. I certainly haven=E2=80=99t ever configured login details. >>> > >>> > How much speed do you have to sacrifice on your connection to >>> > eliminate bufferbloat? >>> > >>> > -- >>> > Alec Robertson >>> > >>> > On 23 April 2016 at 10:46:35 am, Kevin Darbyshire-Bryant >>> > (kevin@darbyshire-bryant.me.uk = ) >>> >>> > wrote: >>> > >>> >> Hi Alec, >>> >> >>> >> A brief appearance from me whilst I have a spare few seconds. >>> >> >>> >> Not stupid! >>> >> >>> >> There=E2=80=99s an obvious question: Why are Billion still shipping = buffer >>> >> bloated devices? Have they been sent graphs/demos/logs of how their >>> >> kit is faulty? >>> >> >>> >> To offer some hopefully constructive pointers: For FTTC service I=E2= =80=99m >>> >> guessing you=E2=80=99ll be using the Billion as a vdsl modem. Who=E2= =80=99s the ISP? >>> >> AFAIK anyone other than Sky will need to run PPPoE and hence hit the >>> >> 1492 MTU restriction *unless* the Billion supports mini jumbo frames >>> >> on the ethernet side and the PPPoE MTU extension (the rfc number >>> >> escapes the brain at the moment) Sky don=E2=80=99t use PPP and just = run >>> >> ethernet frames over PTM=E2=80=A6..the way it should be. The incomin= g >>> >> packets from ISP to you are policed at something close to sync rate, >>> >> this is part of the BT specification. The uplink of course can be as >>> >> bloated as hell ;-) >>> >> >>> >> I use an Archer C7 with BT=E2=80=99s equally horrendously bloated HG= 612 vdsl >>> >> modem on a 40/10 link with sky as my isp. In terms of CPU usage it= =E2=80=99s >>> >> about 1% per megabit so a full 40/10 uses around 55% cpu, I think >>> >> there=E2=80=99s enough for your 60/20=E2=80=A6just. >>> >> https://middling.me.uk/blog/2015/03/customising-openwrt-to-my-needs/ >>> >> offers further advice which I found useful. >>> >> >>> >> Kevin >>> >> >>> >> >>> >> >>> >> >>> >>> On 22 Apr 2016, at 23:01, Alec Robertson >> >>> > wrote: >>> >>> >>> >>> Hi all, >>> >>> >>> >>> I=E2=80=99ve been out of the bufferbloat game for a while and want = to try >>> >>> and beat it once again. >>> >>> >>> >>> I=E2=80=99ve got an FTTC connection (UK) which I get around 60Mbps = on but >>> >>> with horrible bufferbloat on my Billion 8800NL. What router should = I >>> >>> get that can run OpenWRT and handle this connection? Do the newest >>> >>> builds of OpenWRT have cake built-in now via sqm-scripts or would I >>> >>> need to install this manually. If so, how would I do this? >>> >>> >>> >>> Would appreciate any help and apologies if I come off in any way >>> >>> stupid. >>> >>> >>> >>> -- >>> >>> Alec Robertson >>> >>> _______________________________________________ >>> >>> Cake mailing list >>> >>> Cake@lists.bufferbloat.net >>> >>> https://lists.bufferbloat.net/listinfo/cake >>> >> >>> >>> >>> ________________________________ >>> >>> >>> -- >>> Thanks, >>> >>> Kevin@Darbyshire-Bryant.me.uk >>> M: +44 7947 355344 H: +44 1256 478597 >>> >>> >>> _______________________________________________ >>> Cake mailing list >>> Cake@lists.bufferbloat.net >>> https://lists.bufferbloat.net/listinfo/cake >>> >> >> >> _______________________________________________ >> Cake mailing list >> Cake@lists.bufferbloat.net >> https://lists.bufferbloat.net/listinfo/cake >> > > > > -- > Dave T=C3=A4ht > Let's go make home routers and wifi faster! With better software! > http://blog.cerowrt.org --=20 Dave T=C3=A4ht Let's go make home routers and wifi faster! With better software! http://blog.cerowrt.org