From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp67.iad3a.emailsrvr.com (smtp67.iad3a.emailsrvr.com [173.203.187.67]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id E3FF23CB38 for ; Sat, 25 Jul 2020 13:48:03 -0400 (EDT) Received: from app47.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by smtp25.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id 6A8262520C; Sat, 25 Jul 2020 13:48:03 -0400 (EDT) Received: from deepplum.com (localhost.localdomain [127.0.0.1]) by app47.wa-webapps.iad3a (Postfix) with ESMTP id 582C2E1A62; Sat, 25 Jul 2020 13:48:03 -0400 (EDT) Received: by apps.rackspace.com (Authenticated sender: dpreed@deepplum.com, from: dpreed@deepplum.com) with HTTP; Sat, 25 Jul 2020 13:48:03 -0400 (EDT) X-Auth-ID: dpreed@deepplum.com Date: Sat, 25 Jul 2020 13:48:03 -0400 (EDT) From: "David P. Reed" To: "Kevin Darbyshire-Bryant" Cc: "Cake List" MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20200725134803000000_70776" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: <0CCA78BD-201C-4668-A013-24A3F6F4EC87@darbyshire-bryant.me.uk> References: <0CCA78BD-201C-4668-A013-24A3F6F4EC87@darbyshire-bryant.me.uk> Message-ID: <1595699283.358416190@apps.rackspace.com> X-Mailer: webmail/17.3.12-RC X-Classification-ID: c3fcdc95-ebbe-4447-a7ae-265944b46685-1-1 Subject: Re: [Cake] diffserv3 vs diffserv4 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: Sat, 25 Jul 2020 17:48:04 -0000 ------=_20200725134803000000_70776 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0AThis idea (dividing the link rate capacity, since "bandwidth" is an inco= rrect term not to be promulgated), is absurd, but typical of "bellhead" thi= nking.=0A =0APer packet latency is the key control variable, even for TCP. = That's because capacity/rate is not controllable by routers, but by routing= in a general Internet situation.=0A =0ALatency is controlled by queuing de= lay in a packet network, not bitrate. And in mixed traffic, which after all= is why traffic is classified in the first place, by its characteristics an= d response to increased latency end-to-end, is the core "control" for the i= nternetwork as a whole.=0A =0ASo, by promoting thinking about "bandwidth" a= whole sequence of misformulations of network management is embedded into t= he thinking of those designing queue management algorithms.=0A =0AAnd make = no mistake, queue management is the ONLY knob other than sending different = packets on different routes that one has for routers.=0A =0AI don't know wh= o proposed this fractional division, but it is clearly a bellhead-influence= d thinker who thinks all protocols are CBR flows like in the old phone syst= em.=0A =0ABut almost no flows in the internet are CBR flows! File transfers= are not, streaming TV is not, web ttraffic is not, game traffic is not. On= ly non-statistically multiplexed real-time telephony and *some* video confe= rencing is CBR.=0A =0AYet this bizarre idea of dividing "bandwidth" among a= ll categories of flows pops up. Probably from employees of phone companies = or phone equipment suppliers. Or folks who went to Uni and were trained in = "communications" by former phone engineers.=0A =0ALatency, latency, latency= . Queue delay, queue delay, queue delay. Not link speed! Change your brains= .=0A =0AIt's hard fo fight this bellhead crowd (or the bellheadedness in yo= ur own thinking) but think about packets and queues instead.=0A =0AMy good = friend Len Kleinrock didn't invent "Bandwidth Theory"! He invented Queueing= Theory. For a reason.=0A =0AOn Saturday, July 25, 2020 6:12am, "Kevin Darb= yshire-Bryant" said:=0A=0A=0A=0A> _________= ______________________________________=0A> Cake mailing list=0A> Cake@lists= .bufferbloat.net=0A> https://lists.bufferbloat.net/listinfo/cake=0A> =0A> = =0A> > On 24 Jul 2020, at 18:42, Kevin Darbyshire-Bryant=0A> wrote:=0A> >=0A> >=0A> > The move from diffserv4 to diffs= erv5 WAS about de-prioritization.=0A> =0A> It was also about minimum bandwi= dth allocations:=0A> =0A> LE: 1/64th=0A> BK: 1/16th=0A> BE: 1/1=0A> VI: 1/2= =0A> VO: 1/4=0A> =0A> So worst case, best effort should get 11/64ths in the= extreme case of all other=0A> tins in use.=0A> =0A> Cheers,=0A> =0A> Kevin= D-B=0A> =0A> gpg: 012C ACB2 28C6 C53E 9775 9123 B3A2 389B 9DE2 334A=0A> = =0A> ------=_20200725134803000000_70776 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

This idea (dividing th= e link rate capacity, since "bandwidth" is an incorrect term not to be prom= ulgated), is absurd, but typical of "bellhead" thinking.

=0A

 

=0A

Per packet latency is the key co= ntrol variable, even for TCP. That's because capacity/rate is not controlla= ble by routers, but by routing in a general Internet situation.

=0A

 

=0A

Latency is controlled by = queuing delay in a packet network, not bitrate. And in mixed traffic, which= after all is why traffic is classified in the first place, by its characte= ristics and response to increased latency end-to-end, is the core "control"= for the internetwork as a whole.

=0A

 

=0A<= p style=3D"margin:0;padding:0;font-family: arial; font-size: 10pt; overflow= -wrap: break-word;">So, by promoting thinking about "bandwidth" a whole seq= uence of misformulations of network management is embedded into the thinkin= g of those designing queue management algorithms.

=0A

And make no mistake, queue management i= s the ONLY knob other than sending different packets on different routes th= at one has for routers.

=0A

 

=0A

I don't know who proposed this fractional division, but it is cle= arly a bellhead-influenced thinker who thinks all protocols are CBR flows l= ike in the old phone system.

=0A

 

=0A

But almost no flows in the internet are CBR flows! File tran= sfers are not, streaming TV is not, web ttraffic is not, game traffic is no= t. Only non-statistically multiplexed real-time telephony and *some* video = conferencing is CBR.

=0A

 

=0A

Yet this bizarre idea of dividing "bandwidth" among all categories o= f flows pops up. Probably from employees of phone companies or phone equipm= ent suppliers. Or folks who went to Uni and were trained in "communications= " by former phone engineers.

=0A

 

=0A

Latency, latency, latency. Queue delay, queue delay, queue d= elay. Not link speed! Change your brains.

=0A

 =

=0A

It's hard fo fight this bellhead crowd (or the = bellheadedness in your own thinking) but think about packets and queues ins= tead.

=0A

 

=0A

My good = friend Len Kleinrock didn't invent "Bandwidth Theory"! He invented Queueing= Theory. For a reason.

=0A

 

=0A

On Saturday, July 25, 2020 6:12am, "Kevin Darbyshire-Bryant" <k= evin@darbyshire-bryant.me.uk> said:

=0A
=0A

> _______________________________= ________________
> Cake mailing list
> Cake@lists.bufferblo= at.net
> https://lists.bufferbloat.net/listinfo/cake
>
>
> > On 24 Jul 2020, at 18:42, Kevin Darbyshire-Bryant> <kevin@darbyshire-bryant.me.uk> wrote:
> >
&g= t; >
> > The move from diffserv4 to diffserv5 WAS about de-pr= ioritization.
>
> It was also about minimum bandwidth allo= cations:
>
> LE: 1/64th
> BK: 1/16th
> BE:= 1/1
> VI: 1/2
> VO: 1/4
>
> So worst case= , best effort should get 11/64ths in the extreme case of all other
>= ; tins in use.
>
> Cheers,
>
> Kevin D-B<= br />>
> gpg: 012C ACB2 28C6 C53E 9775 9123 B3A2 389B 9DE2 334A=
>
>

=0A
 ------=_20200725134803000000_70776--