From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp81.iad3a.emailsrvr.com (smtp81.iad3a.emailsrvr.com [173.203.187.81]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by huchra.bufferbloat.net (Postfix) with ESMTPS id 3C6EA21F55E for ; Sat, 24 Oct 2015 09:34:14 -0700 (PDT) Received: from smtp19.relay.iad3a.emailsrvr.com (localhost.localdomain [127.0.0.1]) by smtp19.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id 6FF8618039B; Sat, 24 Oct 2015 12:34:13 -0400 (EDT) Received: by smtp19.relay.iad3a.emailsrvr.com (Authenticated sender: dpreed-AT-reed.com) with ESMTPSA id 6CAD6180394; Sat, 24 Oct 2015 12:34:12 -0400 (EDT) X-Sender-Id: dpreed@reed.com Received: from [100.118.164.155] (55.sub-70-199-67.myvzw.com [70.199.67.55]) (using TLSv1.2 with cipher DHE-RSA-AES256-SHA) by 0.0.0.0:465 (trex/5.4.2); Sat, 24 Oct 2015 16:34:13 GMT User-Agent: K-@ Mail X-Priority: 3 In-Reply-To: References: <562A5BE5.6010101@gmail.com> <12883.1445625688@sandelman.ca> <562A9611.4050403@gmail.com> MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----36HOEFSO6ACXR30GPI1NP70VC7PG91" Content-Transfer-Encoding: 7bit From: "David P. Reed" Date: Sat, 24 Oct 2015 12:34:09 -0400 To: Sebastian Moeller Message-ID: <931be41d-68b7-4fea-9d4e-14bf1636e9bc@reed.com> Cc: cerowrt-devel@lists.bufferbloat.net Subject: Re: [Cerowrt-devel] Problems testing sqm X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.13 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: Sat, 24 Oct 2015 16:34:37 -0000 ------36HOEFSO6ACXR30GPI1NP70VC7PG91 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Not trying to haggle=2E Just pointing out that this test configuration has = a very short RTT=2E maybe too short for our SQM to adjust to=2E On Oct 24,= 2015, Sebastian Moeller wrote: >Hi David, > >On Oct 24= , 2015, at 00:53 , David P=2E Reed wrote: > >> In parti= cular, the DUT should probably have no more than 2 packets >of outbound qu= eueing given the very small RTT=2E 2xRTT is the most >buffering you want in= the loop=2E > > Let=E2=80=99s not haggle about the precise amount of queue= ing we deem >acceptable, as long as we all agree that >=3D 2 seconds is sim= ply not >acceptable ;) (the default sqm will approximately limit the latenc= y >under load increase (LULI) to roughly twice the target or typically 10 >= ms; note that this LULI only applies to unrelated flows)=2E The exact >numb= er of queued packets seems to correlate with the beefiness of the >DUT, the= beefier the fewer packets should work, wimpier devices might >need to batc= h some processing up, resulting in higher LULI=E2=80=A6 > >Best Regards > = Sebastian > >> >> On Oct 23, 2015, Richard Smith w= rote: >> On 10/23/2015 02:41 PM, Michael Richardson wrote: >> Richard Smith= wrote: >> My test setup: >> >> Laptop<--1000BaseT= -->DUT<--1000baseT-->Server >> >> So, given that the DUT is the only real = constraint in the network, >what >> do you expect to see from this setup? >= > >> Given that the probably DUT can't forward at Gb/s, and it certainly >= can't >> shape anything, it's gonna drop packets, and it's probably gonna d= rop >them in >> Rx, having overrun the Rx-queue (so tail-drop)=2E If there = is too much >ram >> (bufferbloated), then you'll see different results=2E= =2E=2E >> >> Setting ingress/egress to 10Mbit/s I expected to see the spee= d >> measurements bounce around those limits with the ping times staying >= in >> the low double digits of ms=2E What I saw however, was the data rate= s >> going well past 10Mbit limit and pings up to 2000 ms=2E >> >> This i= s what I've seen in prior rrul testing using a the 50/10 cable >> link at = our office and my 25(ish)/6 link at my apartment and a well >> connected s= erver on the net=2E That however was using QoS and not SQM=2E >> >> Its th= at a reasonable expectation? >> >> -- Sent with K-@ Mail - the evolution o= f emailing=2E >_______________________________________________ >> Cerowrt-d= evel mailing list >> Cerowrt-devel@lists=2Ebufferbloat=2Enet >> https://lis= ts=2Ebufferbloat=2Enet/listinfo/cerowrt-devel -- Sent with K-@ Mail - the = evolution of emailing=2E ------36HOEFSO6ACXR30GPI1NP70VC7PG91 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable Not trying to haggle=2E Just pointing out that thi= s test configuration has a very short RTT=2E maybe too short for our SQM to= adjust to=2E

On Oct 24, 2015, Sebastian Moeller <moeller0@gmx=2Ede> wrote: 
Hi David,

On Oct 24, 2015, at 00:5=
3 , David P=2E Reed <dpreed@reed=2Ecom> wrote:

In pa= rticular, the DUT should probably have no more than 2 packets of outbound = queueing given the very small RTT=2E 2xRTT is the most buffering you want i= n the loop=2E

Let’s not haggle about = the precise amount of queueing we deem acceptable, as long as we all agree = that >=3D 2 seconds is simply not acceptable ;) (the default sqm will ap= proximately limit the latency under load increase (LULI) to roughly twice t= he target or typically 10 ms; note that this LULI only applies to unrelated= flows)=2E The exact number of queued packets seems to correlate with the b= eefiness of the DUT, the beefier the fewer packets should work, wimpier dev= ices might need to batch some processing up, resulting in higher LULI&hell= ip;

Best Regards
Se= bastian


On Oct 23, 2015, Richard Smith <sm= ithbone@gmail=2Ecom> wrote:
On 10/23/2015 02:41 PM, Mi= chael Richardson wrote:
Richard Smith <smithbone@gmail= =2Ecom> wrote:
My test setup:

Laptop<--1000BaseT-->DUT<--1000baseT-->Server

So, given that the DUT is the only real co= nstraint in the network, what
do you expect to see from t= his setup?

Given that the probably DUT= can't forward at Gb/s, and it certainly can't
shape anyt= hing, it's gonna drop packets, and it's probably gonna drop them in
Rx, having overrun the Rx-queue (so tail-drop)=2E If there is to= o much ram
(bufferbloated), then you'll see different res= ults=2E=2E=2E

Setting ingress/egress t= o 10Mbit/s I expected to see the speed
measurements boun= ce around those limits with the ping times staying in
th= e low double digits of ms=2E What I saw however, was the data rates
going well past 10Mbit limit and pings up to 2000 ms=2E

This is what I've seen in prior rrul testing = using a the 50/10 cable
link at our office and my 25(ish= )/6 link at my apartment and a well
connected server on = the net=2E That however was using QoS and not SQM=2E

Its that a reasonable expectation?

-- Sent with K-@ Mail - the evolution of emailing=2E

Cerowrt-devel mailing list
Cerowrt-devel@list= s=2Ebufferbloat=2Enet
https://lists=2Ebufferbloa= t=2Enet/listinfo/cerowrt-devel

-- Sent with K-@ Mail - the evolution of emailing=2E ------36HOEFSO6ACXR30GPI1NP70VC7PG91--