From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp105.iad3a.emailsrvr.com (smtp105.iad3a.emailsrvr.com [173.203.187.105]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 2D9673B29E for ; Sat, 25 Apr 2020 16:56:28 -0400 (EDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=g001.emailsrvr.com; s=20190322-9u7zjiwi; t=1587848187; bh=xgYP5h12/5n9b3wfax5ufOJzfGEAsQrivMOdxODUW0Y=; h=Date:Subject:From:To:From; b=uF3pQ1fM2uArlid3qY/MTNh3PwSmXi+IdiCa5WhcwSyeUuhLhezZP+/VwEkrKWcfB ohDCWv7BT2FEbK2b04dunRCICcRzsZxeJGi/UhvuRD0kdlLbZvTptkSykoCdwXGwsE euqqHjXd0vQw1E7jBrkUa+AsLOexOn9KV44bCwZo= Received: from app19.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by smtp22.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id BE8C15EF6; Sat, 25 Apr 2020 16:56:27 -0400 (EDT) X-Sender-Id: dpreed@deepplum.com Received: from app19.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by 0.0.0.0:25 (trex/5.7.12); Sat, 25 Apr 2020 16:56:27 -0400 Received: from deepplum.com (localhost.localdomain [127.0.0.1]) by app19.wa-webapps.iad3a (Postfix) with ESMTP id D483CE0047; Sat, 25 Apr 2020 16:56:26 -0400 (EDT) Received: by apps.rackspace.com (Authenticated sender: dpreed@deepplum.com, from: dpreed@deepplum.com) with HTTP; Sat, 25 Apr 2020 16:56:26 -0400 (EDT) X-Auth-ID: dpreed@deepplum.com Date: Sat, 25 Apr 2020 16:56:26 -0400 (EDT) From: "David P. Reed" To: "Kevin Darbyshire-Bryant" Cc: "Jonathan Morton" , "Cake List" MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20200425165626000000_56464" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: <32DE972A-3359-462A-A12C-77714B2563F6@darbyshire-bryant.me.uk> References: <62228545-5DE2-4600-A9BB-52D891FF5AD4@darbyshire-bryant.me.uk> <4D896254-FFB2-4CEB-B596-A6D2E510243C@gmail.com> <32DE972A-3359-462A-A12C-77714B2563F6@darbyshire-bryant.me.uk> Message-ID: <1587848186.866926178@apps.rackspace.com> X-Mailer: webmail/17.3.7-RC X-Classification-ID: d06e8b88-1de1-4259-b498-e01543f608d3-1-1 Subject: Re: [Cake] Cake tin behaviour - discuss.... 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 Apr 2020 20:56:28 -0000 ------=_20200425165626000000_56464 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0AQuestion: what's the "lag under load" experienced when these two loads a= re filling the capacity of the bottleneck router (the DSL link)?=0AI'm wond= ering whether your cake setup is deliberately building up a big queue withi= n the router for any of the 10 bulk/best efforts flows.=0A =0AOn Saturday, = April 25, 2020 4:34pm, "Kevin Darbyshire-Bryant" said:=0A=0A=0A=0A> _______________________________________________=0A= > Cake mailing list=0A> Cake@lists.bufferbloat.net=0A> https://lists.buffer= bloat.net/listinfo/cake=0A> =0A> =0A> > On 25 Apr 2020, at 16:25, Jonathan = Morton =0A> wrote:=0A> >=0A> >> On 25 Apr, 2020, at = 2:07 pm, Kevin Darbyshire-Bryant=0A> wrote:= =0A> >>=0A> >> Download from =E2=80=98onedrive=E2=80=99 from 1 box, using 5= flows,=0A> classified as Bulk. Little other traffic going on, sits there a= t circa 70Mbit, no=0A> problem.=0A> >>=0A> >> If I started another download= on another box, say 5 flows, classified as=0A> Best Effort, what rates wou= ld you expect the Bulk & Best effort tins to flow at?=0A> >=0A> > Approxima= tely speaking, Cake should give the Best Effort traffic priority=0A> over B= ulk, until the latter is squashed down to its tin's capacity. So you may=0A= > see 5-10Mbps of Bulk and 65-70Mbps of Best Effort, depending on some shor= t-term=0A> effects.=0A> >=0A> > This assumes that the Diffserv marking actu= ally reaches Cake, of course.=0A> =0A> Thanks Jonathan. I can assure you di= ffserv markings are reaching cake both egress=0A> & ingress due to my pet = =E2=80=98act_ctinfo/connmark -savedscp=E2=80=99 project. =0A> Amongst other= monitoring methods a simple 'watch -t tc -s qdisc show dev $1=E2=80=99=0A>= albeit with a slightly modified cake module & tc to report per tin traffic= as a=0A> percentage of total & per tin % of threshold is used.=0A> =0A> eg= :=0A> Bulk Best Effort Video Voice=0A> thresh 4812Kbit 77Mbit 38500Kbit 192= 50Kbit=0A> target 5.0ms 5.0ms 5.0ms 5.0ms=0A> interval 100.0ms 100.0ms 100.= 0ms 100.0ms=0A> pk_delay 961us 167us 311us 164us=0A> av_delay 453us 78us 14= 1us 75us=0A> sp_delay 51us 12us 17us 9us=0A> backlog 9084b 0b 0b 0b=0A> pkt= s 60618617 2006708 460725 11129=0A> bytes 91414263264 2453185010 636385583 = 5205008=0A> traffic% 89 0 0 0=0A> traftin% 1435 0 0 0=0A> way_inds 2703134 = 8957 169 111=0A> way_miss 922 6192 104 525=0A> way_cols 0 0 0 0=0A> drops 8= 442 230 37 0=0A> marks 5 0 0 0=0A> ack_drop 0 0 0 0=0A> sp_flows 2 3 1 3=0A= > bk_flows 1 0 0 0=0A> un_flows 0 0 0 0=0A> max_len 66616 12112 9084 3360= =0A> quantum 300 1514 1174 587=0A> =0A> Your expectation is that Best Effor= t would exert downward pressure on Bulk traffic=0A> reducing bulk traffic t= o about bulk threshold level which is my expectation also. =0A> Tin priorit= y then host (fairness), then flow.=0A> =0A> As you may have guessed, that= =E2=80=99s not quite what I=E2=80=99m seeing but as=0A> I=E2=80=99ve manage= d to see the issue when using =E2=80=98flowblind=E2=80=99 am now much=0A> l= ess inclined to point the finger at host fairness & friends. I remain confu= sed=0A> why =E2=80=98bulk=E2=80=99 is exceeding its allocation though in wh= at should be=0A> pressure from best effort but it ends up going all over th= e place and being a bit=0A> unstable. Odd.=0A> =0A> BTW: The =E2=80=98onedr= ive=E2=80=99 client box is actually running linux.=0A> =0A> ------=_20200425165626000000_56464 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

Question: what's the "= lag under load" experienced when these two loads are filling the capacity o= f the bottleneck router (the DSL link)?

=0A

I'm wond= ering whether your cake setup is deliberately building up a big queue withi= n the router for any of the 10 bulk/best efforts flows.

=0A

 

=0A

On Saturday, April 25, 2020 4:34p= m, "Kevin Darbyshire-Bryant" <kevin@darbyshire-bryant.me.uk> said:

=0A
=0A

&g= t; _______________________________________________
> Cake mailing l= ist
> Cake@lists.bufferbloat.net
> https://lists.bufferbloa= t.net/listinfo/cake
>
>
> > On 25 Apr 2020, at= 16:25, Jonathan Morton <chromatix99@gmail.com>
> wrote:
> >
> >> On 25 Apr, 2020, at 2:07 pm, Kevin Darbyshire= -Bryant
> <kevin@darbyshire-bryant.me.uk> wrote:
> &g= t;>
> >> Download from =E2=80=98onedrive=E2=80=99 from 1 b= ox, using 5 flows,
> classified as Bulk. Little other traffic going= on, sits there at circa 70Mbit, no
> problem.
> >>> >> If I started another download on another box, say 5 flows= , classified as
> Best Effort, what rates would you expect the Bulk= & Best effort tins to flow at?
> >
> > Approxima= tely speaking, Cake should give the Best Effort traffic priority
> = over Bulk, until the latter is squashed down to its tin's capacity. So you = may
> see 5-10Mbps of Bulk and 65-70Mbps of Best Effort, depending = on some short-term
> effects.
> >
> > This as= sumes that the Diffserv marking actually reaches Cake, of course.
>=
> Thanks Jonathan. I can assure you diffserv markings are reachin= g cake both egress
> & ingress due to my pet =E2=80=98act_ctinf= o/connmark -savedscp=E2=80=99 project.
> Amongst other monitoring = methods a simple 'watch -t tc -s qdisc show dev $1=E2=80=99
> albei= t with a slightly modified cake module & tc to report per tin traffic a= s a
> percentage of total & per tin % of threshold is used.
>
> eg:
> Bulk Best Effort Video Voice
> thre= sh 4812Kbit 77Mbit 38500Kbit 19250Kbit
> target 5.0ms 5.0ms 5.0ms 5= .0ms
> interval 100.0ms 100.0ms 100.0ms 100.0ms
> pk_delay = 961us 167us 311us 164us
> av_delay 453us 78us 141us 75us
> = sp_delay 51us 12us 17us 9us
> backlog 9084b 0b 0b 0b
> pkts= 60618617 2006708 460725 11129
> bytes 91414263264 2453185010 63638= 5583 5205008
> traffic% 89 0 0 0
> traftin% 1435 0 0 0
> way_inds 2703134 8957 169 111
> way_miss 922 6192 104 525
> way_cols 0 0 0 0
> drops 8442 230 37 0
> marks 5 0 0= 0
> ack_drop 0 0 0 0
> sp_flows 2 3 1 3
> bk_flows= 1 0 0 0
> un_flows 0 0 0 0
> max_len 66616 12112 9084 3360=
> quantum 300 1514 1174 587
>
> Your expectation = is that Best Effort would exert downward pressure on Bulk traffic
>= reducing bulk traffic to about bulk threshold level which is my expectatio= n also.
> Tin priority then host (fairness), then flow.
> =
> As you may have guessed, that=E2=80=99s not quite what I=E2=80= =99m seeing but as
> I=E2=80=99ve managed to see the issue when usi= ng =E2=80=98flowblind=E2=80=99 am now much
> less inclined to point= the finger at host fairness & friends. I remain confused
> why= =E2=80=98bulk=E2=80=99 is exceeding its allocation though in what should b= e
> pressure from best effort but it ends up going all over the pla= ce and being a bit
> unstable. Odd.
>
> BTW: The = =E2=80=98onedrive=E2=80=99 client box is actually running linux.
> =
>

=0A
 ------=_20200425165626000000_56464--