From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp64.iad3a.emailsrvr.com (smtp64.iad3a.emailsrvr.com [173.203.187.64]) (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 4816F3B2A4 for ; Tue, 14 May 2019 18:05:33 -0400 (EDT) Received: from smtp33.relay.iad3a.emailsrvr.com (localhost [127.0.0.1]) by smtp33.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id 178C054EF; Tue, 14 May 2019 18:05:33 -0400 (EDT) X-SMTPDoctor-Processed: csmtpprox beta Received: from smtp33.relay.iad3a.emailsrvr.com (localhost [127.0.0.1]) by smtp33.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id 1066854F4; Tue, 14 May 2019 18:05:33 -0400 (EDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=g001.emailsrvr.com; s=20190322-9u7zjiwi; t=1557871533; bh=pVerozl7UOqk4dqdPqHBPXROroRo/Dwo0M5PcDWYWiU=; h=Date:Subject:From:To:From; b=t2RfmMBLhAW74NU06/OTfknzoccbLkbB3nQ/nTJSg9Rx4lsluEMkC36Yw4MHQynAG g6MUXk/wETHIJPwJb1L8h93xdSVSf6//VGKZcxtUm1cm3XIHWEwQgdJq7OB9TyEe5m cHy//04jM5hgU55pmp7283j3k+of859bRRtnUDZI= Received: from app11.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by smtp33.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id CD81B54EF; Tue, 14 May 2019 18:05:32 -0400 (EDT) X-Sender-Id: dpreed@deepplum.com Received: from app11.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by 0.0.0.0:25 (trex/5.7.12); Tue, 14 May 2019 18:05:32 -0400 Received: from deepplum.com (localhost.localdomain [127.0.0.1]) by app11.wa-webapps.iad3a (Postfix) with ESMTP id B8B3DA0075; Tue, 14 May 2019 18:05:32 -0400 (EDT) Received: by apps.rackspace.com (Authenticated sender: dpreed@deepplum.com, from: dpreed@deepplum.com) with HTTP; Tue, 14 May 2019 18:05:32 -0400 (EDT) X-Auth-ID: dpreed@deepplum.com Date: Tue, 14 May 2019 18:05:32 -0400 (EDT) From: "David P. Reed" To: "David P. Reed" Cc: "=?utf-8?Q?Valdis_Kl=C4=93tnieks?=" , "Rich Brown" , "cerowrt-devel" , "bloat" MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20190514180532000000_55912" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: <1557859131.759530583@apps.rackspace.com> References: <2936.1557856670@turing-police> <1557859131.759530583@apps.rackspace.com> Message-ID: <1557871532.754117608@apps.rackspace.com> X-Mailer: webmail/16.4.2-RC Subject: Re: [Cerowrt-devel] =?utf-8?q?fq=5Fcodel_is_SEVEN_years_old_today=2E?= =?utf-8?b?Li4=?= X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.20 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: Tue, 14 May 2019 22:05:33 -0000 ------=_20190514180532000000_55912 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0AI wonder if an interesting project to design and pitch for CrowdSupply t= o fund would be a little board that packages sch_cake or something in the m= inimal hardware package that could sit between a 1 GigE symmetric port and = either an asymmetric GigE or a symmetric 1 GigE connection into a 10 GigE s= witch.=0AThe key point is that it needs to support wire-rate forwarding wit= h small packets of Gigabit throughput. Ideally, it also supports a dnsmasq = NAT and wireguard optionally.=0A =0AI know a Celeron with 2 GB of RAM can e= asily do it (because that is what I use). We know (well that's what you guy= s tell me) that the dinky MIPS processors are underpowered to handle sch_ca= ke at such packet rates. The Linksys and Netgear and TP-link guys seem to s= ee no market at all for any such thing. But I see it as a useful jellybean = device if it could be cheap and simple.=0A =0ACould maybe design, produce, = and sell this for $100? No one else seems to want to make such a thing. I c= ould just barely design and implement the board and get it made, but to be = honest I'm better at spec'ing and prototyping than making manufacturable ha= rdware designs. I suspect I could find someone to do the PCB design, layout= and parts selection as a project.=0A =0AThe idea for this hardware "produc= t" is to decouple this buffer management from the WiFi compatibility and dr= iver mess, and make it easy for people, maybe to demonstrate that it could = be a great product. Forget designing the packaging, negotiating a sales cha= nnel, etc. Just do what is needed to make a few thousand for the CrowdSuppl= y market.=0A =0AThoughts?=0A =0A-----Original Message-----=0AFrom: "David P= . Reed" =0ASent: Tuesday, May 14, 2019 2:38pm=0ATo: "V= aldis Kl=C4=93tnieks" =0ACc: "Rich Brown" , "cerowrt-devel" , "= bloat" =0ASubject: Re: [Cerowrt-devel] fq_code= l is SEVEN years old today...=0A=0A=0A=0AWell, of all the devices in my hou= se (maybe 100), only the router attached to the cable modem (which is a 2x = GigE Intel Linux board based on Fedora 29 server with sch_cake configured) = is running fq_codel. And setting that up was a labor of love. But it works = a charm for my asymmetric Gigabit cable service.=0A =0AMy home's backbone i= s 10 GigE fiber, so I suppose fq_codel would be helpful for devices that ru= n on 1 GigE subnets like my 2 802.11ac access points when talking to my NAS= 's.=0AHowever, the 802.11ac access point high speed functionality doesn't s= eem to be supportable by LEDE. So what can I do? =0A =0AI suppose I could s= tick some little custom Intel Linux 2x GigE devices between access points a= nd the 10 GigE backbone, and put fq_codel in there.=0A =0AMy point is, to g= et the primary benefit of bufferbloat reduction, one has to stick little Li= nux boxes everywhere, because fq_codel is not supported except via DIY hack= ing.=0A =0AAnd indeed, 10 GigE->1 GigE buffering does affect storage access= latency in bad ways.=0A =0AWe see the same problem in datacenter networks = that have excessive buffering - a famous switch company backed by Andy Bech= tolsheim is really problematic because they claim building up huge buffers = is a "feature" not a bug.=0A-----Original Message-----=0AFrom: "Valdis Kl= =C4=93tnieks" =0ASent: Tuesday, May 14, 2019 1:57p= m=0ATo: "Rich Brown" =0ACc: "cerowrt-devel" , "bloat" =0ASu= bject: Re: [Cerowrt-devel] fq_codel is SEVEN years old today...=0A=0A=0A=0A= _______________________________________________=0ACerowrt-devel mailing lis= t=0ACerowrt-devel@lists.bufferbloat.net=0Ahttps://lists.bufferbloat.net/lis= tinfo/cerowrt-devel=0AOn Tue, 14 May 2019 08:16:06 -0400, Rich Brown said:= =0A=0A> Let's all pat ourselves on the back for this good work!=0A=0ADo we = have an estimate of what percent of connected devices=0Aare actually using = fq_codel or other modern anti-bloat methods?=0AI'm reasonably sure my TV, m= y PS3, and my PS4 are still=0Abehind the curve. ------=_20190514180532000000_55912 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

I wonder if an interes= ting project to design and pitch for CrowdSupply to fund would be a little = board that packages sch_cake or something in the minimal hardware package t= hat could sit between a 1 GigE symmetric port and either an asymmetric GigE= or a symmetric 1 GigE connection into a 10 GigE switch.

=0A

The key point is that it needs to support wire-rate forwarding wit= h small packets of Gigabit throughput. Ideally, it also supports a dnsmasq = NAT and wireguard optionally.

=0A

 

=0A

I know a Celeron with 2 GB of RAM can easily do it (because= that is what I use). We know (well that's what you guys tell me) that the = dinky MIPS processors are underpowered to handle sch_cake at such packet ra= tes. The Linksys and Netgear and TP-link guys seem to see no market at all = for any such thing. But I see it as a useful jellybean device if it could b= e cheap and simple.

=0A

 

=0A

Could maybe design, produce, and sell this for $100? No one else seem= s to want to make such a thing. I could just barely design and implement th= e board and get it made, but to be honest I'm better at spec'ing and protot= yping than making manufacturable hardware designs. I suspect I could find s= omeone to do the PCB design, layout and parts selection as a project.

= =0A

 

=0A

The idea for this= hardware "product" is to decouple this buffer management from the WiFi com= patibility and driver mess, and make it easy for people, maybe to demonstra= te that it could be a great product. Forget designing the packaging, negoti= ating a sales channel, etc. Just do what is needed to make a few thousand f= or the CrowdSupply market.

=0A

 

=0A

Thoughts?

=0A

 

=0A

-----Original Message-----
From: "David P. Reed" <dpreed@d= eepplum.com>
Sent: Tuesday, May 14, 2019 2:38pm
To: "Valdis Kl= =C4=93tnieks" <valdis.kletnieks@vt.edu>
Cc: "Rich Brown" <ric= hb.hanover@gmail.com>, "cerowrt-devel" <cerowrt-devel@lists.bufferblo= at.net>, "bloat" <bloat@lists.bufferbloat.net>
Subject: Re: [= Cerowrt-devel] fq_codel is SEVEN years old today...

=0A
=0A

= Well, of all the devices in my house (maybe 100), only the router attached = to the cable modem (which is a 2x GigE Intel Linux board based on Fedora 29= server with sch_cake configured) is running fq_codel. And setting that up = was a labor of love. But it works a charm for my asymmetric Gigabit cable s= ervice.

=0A

 

=0A

My home's backbone is 10 GigE fiber= , so I suppose fq_codel would be helpful for devices that run on 1 GigE sub= nets like my 2 802.11ac access points when talking to my NAS's.

=0A

However, the 802.11ac access point h= igh speed functionality doesn't seem to be supportable by LEDE. So what can= I do? 

=0A

 

= =0A

I suppose I could stick some= little custom Intel Linux 2x GigE devices between access points and the 10= GigE backbone, and put fq_codel in there.

=0A

 

=0A

= My point is, to get the primary benefit of bufferbloat reduction, one has t= o stick little Linux boxes everywhere, because fq_codel is not supported ex= cept via DIY hacking.

=0A

&nb= sp;

=0A

And indeed, 10 GigE-&= gt;1 GigE buffering does affect storage access latency in bad ways.

=0A<= p style=3D"margin:0;padding:0;margin: 0; padding: 0; font-family: arial; fo= nt-size: 12pt; overflow-wrap: break-word;"> 

=0A

We see the same problem in datacenter networks tha= t have excessive buffering - a famous switch company backed by Andy Bechtol= sheim is really problematic because they claim building up huge buffers is = a "feature" not a bug.

=0A

--= ---Original Message-----
From: "Valdis Kl=C4=93tnieks" <valdis.klet= nieks@vt.edu>
Sent: Tuesday, May 14, 2019 1:57pm
To: "Rich Bro= wn" <richb.hanover@gmail.com>
Cc: "cerowrt-devel" <cerowrt-de= vel@lists.bufferbloat.net>, "bloat" <bloat@lists.bufferbloat.net><= br />Subject: Re: [Cerowrt-devel] fq_codel is SEVEN years old today...

=0A
=0A

_______________________________________________
Cer= owrt-devel mailing list
Cerowrt-devel@lists.bufferbloat.net
https= ://lists.bufferbloat.net/listinfo/cerowrt-devel
On Tue, 14 May 2019 08= :16:06 -0400, Rich Brown said:

> Let's all pat ourselves on t= he back for this good work!

Do we have an estimate of what perce= nt of connected devices
are actually using fq_codel or other modern an= ti-bloat methods?
I'm reasonably sure my TV, my PS3, and my PS4 are st= ill
behind the curve.

=0A
=0A
 ------=_20190514180532000000_55912--