From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp126.iad3a.emailsrvr.com (smtp126.iad3a.emailsrvr.com [173.203.187.126]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 40B273B29D for ; Wed, 4 Jan 2023 17:21:02 -0500 (EST) Received: from app2.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by smtp16.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id BE5AF5876 for ; Wed, 4 Jan 2023 17:21:01 -0500 (EST) Received: from deepplum.com (localhost.localdomain [127.0.0.1]) by app2.wa-webapps.iad3a (Postfix) with ESMTP id AADEEA255A for ; Wed, 4 Jan 2023 17:21:01 -0500 (EST) Received: by apps.rackspace.com (Authenticated sender: dpreed@deepplum.com, from: dpreed@deepplum.com) with HTTP; Wed, 4 Jan 2023 17:21:01 -0500 (EST) X-Auth-ID: dpreed@deepplum.com Date: Wed, 4 Jan 2023 17:21:01 -0500 (EST) From: "David P. Reed" To: starlink@lists.bufferbloat.net MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20230104172101000000_59262" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: References: X-Client-IP: 209.6.168.128 Message-ID: <1672870861.69732760@apps.rackspace.com> X-Mailer: webmail/19.0.22-RC X-Classification-ID: f6fdd09a-2fc1-4fa0-816b-4b284301ea16-1-1 Subject: Re: [Starlink] [Rpm] the grinch meets cloudflare's christmas present X-BeenThere: starlink@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: "Starlink has bufferbloat. Bad." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 04 Jan 2023 22:21:02 -0000 ------=_20230104172101000000_59262 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0AI don't know how to debug this, but the cloudflare speed test really suc= ks on my home wired network, compared to others. And also I discovered to m= y chagrin that the DSLReports speedtest is now broken, at least in my Chrom= e browser.=0A =0AThe speeds measured by Cloudflare are essentially 1/2 of b= oth Ookla and Fast.com.=0A =0ACloudflare in my test config gives ~500/10 Mb= /sec, with a download latency of 14.6 msec, upload latency of 11.6.=0A =0AO= ookla and Fast give ~980/25 Mb/sec, with "latency" being about 11 or 12 mse= c.=0A =0AThis difference is observed over a tuned "cake" install on my Linu= x router, and the home network is 10 GigE to my workstation from the router= , and the router talks to my DOCSIS cable modem on RCN at 1 GigE, with RCN'= s product offering being its "Gig" product.=0A =0ANow, I hate using Ookla a= nd getting bombarded with ads! I don't really trust fast.com, but it used t= o be OK.=0A =0AThe real disappointment was that the DSL Reports speed test,= which I used to recommend is so broken. It claims it can't reach any of it= s 3 selected test sites, because I may have an "alien script" and suggests = my DNS might be "slow" or my Chrome might have browser malware.=0A =0ANow, = maybe Chrome has browser malware on my machine. This is troubling to a seri= ous degree to me, and I will be investigating.=0A =0AHowever, Cloudflare se= ems to be somewhat flaky to a significant degree. (It also doesn't seem to = push a fast network connection nearly hard enough to measure lag under load= , it seems to me).=0A =0A =0A =0AOn Wednesday, January 4, 2023 2:20pm, star= link-request@lists.bufferbloat.net said:=0A=0A=0A=0A> Send Starlink mailing= list submissions to=0A> starlink@lists.bufferbloat.net=0A> =0A> To subscri= be or unsubscribe via the World Wide Web, visit=0A> https://lists.bufferblo= at.net/listinfo/starlink=0A> or, via email, send a message with subject or = body 'help' to=0A> starlink-request@lists.bufferbloat.net=0A> =0A> You can = reach the person managing the list at=0A> starlink-owner@lists.bufferbloat.= net=0A> =0A> When replying, please edit your Subject line so it is more spe= cific=0A> than "Re: Contents of Starlink digest..."=0A> =0A> =0A> Today's T= opics:=0A> =0A> 1. Re: [Rpm] the grinch meets cloudflare's christmas presen= t=0A> (jf@jonathanfoulkes.com)=0A> =0A> =0A> ------------------------------= ----------------------------------------=0A> =0A> Message: 1=0A> Date: Wed,= 4 Jan 2023 14:20:15 -0500=0A> From: "jf@jonathanfoulkes.com" =0A> To: Dave Taht =0A> Cc: bloat , libreqos=0A> , Cake L= ist=0A> , Dave Taht via Starlink=0A> , Rpm ,=0A> IETF IPPM WG <= ippm@ietf.org>=0A> Subject: Re: [Starlink] [Rpm] the grinch meets cloudflar= e's christmas=0A> present=0A> Message-ID: <845161E4-474C-44A9-92D4-1702748A= 3DA1@jonathanfoulkes.com>=0A> Content-Type: text/plain; charset=3D"utf-8"= =0A> =0A> HNY Dave and all the rest,=0A> =0A> Great to see yet another capa= city test add latency metrics to the results. This=0A> one looks like a goo= d start.=0A> =0A> Results from my Windstream DOCSIS 3.1 line (3.1 on downlo= ad only, up is 3.0)=0A> Gigabit down / 35Mbps up provisioning. Using an IQr= outer Pro (an i5 x86) with Cake=0A> set for 710/31 as this ISP can=E2=80=99= t deliver reliable low-latency unless you=0A> shave a good bit off the targ= ets. My local loop is pretty congested.=0A> =0A> Here=E2=80=99s the latest = Cloudflare test:=0A> =0A> -------------- next part --------------=0A> A non= -text attachment was scrubbed...=0A> Name: CFSpeedTest_Gig35_20230104.png= =0A> Type: image/png=0A> Size: 379539 bytes=0A> Desc: not available=0A> URL= :=0A> =0A> -------------- next part --------------=0A> = =0A> =0A> And an Ookla test run just afterward:=0A> =0A> -------------- nex= t part --------------=0A> A non-text attachment was scrubbed...=0A> Name: S= peedtest_net_Gig35_20230104.png=0A> Type: image/png=0A> Size: 40589 bytes= =0A> Desc: not available=0A> URL:=0A> =0A> -------= ------- next part --------------=0A> =0A> =0A> They are definitely both in = the ballpark and correspond to other tests run from=0A> the router itself o= r my (wired) MacBook Pro.=0A> =0A> Cheers,=0A> =0A> Jonathan=0A> =0A> =0A> = > On Jan 4, 2023, at 12:26 PM, Dave Taht via Rpm=0A> wrote:=0A> >=0A> > Please try the new, the shiny, the really wonderfu= l test here:=0A> > https://speed.cloudflare.com/=0A> >=0A> > I would really= appreciate some independent verification of=0A> > measurements using this = tool. In my brief experiments it appears - as=0A> > all the commercial tool= s to date - to dramatically understate the=0A> > bufferbloat, on my LTE, (a= nd my starlink terminal is out being=0A> > hacked^H^H^H^H^H^Hworked on, so = I can't measure that)=0A> >=0A> > My test of their test reports 223ms 5G la= tency under load , where=0A> > flent reports over 2seconds. See comparison = attached.=0A> >=0A> > My guess is that this otherwise lovely new tool, like= too many,=0A> > doesn't run for long enough. Admittedly, most web objects = (their=0A> > target market) are small, and so long as they remain small and= not=0A> > heavily pipelined this test is a very good start... but I'm pret= ty=0A> > sure cloudflare is used for bigger uploads and downloads than that= .=0A> > There's no way to change the test to run longer either.=0A> >=0A> >= I'd love to get some results from other networks (compared as usual to=0A>= > flent), especially ones with cake on it. I'd love to know if they=0A> > = measured more minimum rtts that can be obtained with fq_codel or cake,=0A> = > correctly.=0A> >=0A> > Love Always,=0A> > The Grinch=0A> >=0A> > --=0A> >= This song goes out to all the folk that thought Stadia would work:=0A> >= =0A> https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-698136= 6665607352320-FXtz=0A> > Dave T=C3=A4ht CEO, TekLibre, LLC=0A> >=0A> _______________________= ________________________=0A> > Rpm mailing list=0A> > Rpm@lists.bufferbloat= .net=0A> > https://lists.bufferbloat.net/listinfo/rpm=0A> =0A> =0A> -------= -----------------------=0A> =0A> Subject: Digest Footer=0A> =0A> __________= _____________________________________=0A> Starlink mailing list=0A> Starlin= k@lists.bufferbloat.net=0A> https://lists.bufferbloat.net/listinfo/starlink= =0A> =0A> =0A> ------------------------------=0A> =0A> End of Starlink Dige= st, Vol 22, Issue 10=0A> ****************************************=0A> ------=_20230104172101000000_59262 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

I don't know how to de= bug this, but the cloudflare speed test really sucks on my home wired netwo= rk, compared to others. And also I discovered to my chagrin that the DSLRep= orts speedtest is now broken, at least in my Chrome browser.

=0A

 

=0A

The speeds measured by Cloud= flare are essentially 1/2 of both Ookla and Fast.com.

=0A

 

=0A

Cloudflare in my test config gives = ~500/10 Mb/sec, with a download latency of 14.6 msec, upload latency of 11.= 6.

=0A

 

=0A

Oookla and = Fast give ~980/25 Mb/sec, with "latency" being about 11 or 12 msec.

=0A<= p style=3D"margin:0;padding:0;font-family: arial; font-size: 10pt; overflow= -wrap: break-word;"> 

=0A

This difference is ob= served over a tuned "cake" install on my Linux router, and the home network= is 10 GigE to my workstation from the router, and the router talks to my D= OCSIS cable modem on RCN at 1 GigE, with RCN's product offering being its "= Gig" product.

=0A

 

=0A

= Now, I hate using Ookla and getting bombarded with ads! I don't really trus= t fast.com, but it used to be OK.

=0A

 

=0A<= p style=3D"margin:0;padding:0;font-family: arial; font-size: 10pt; overflow= -wrap: break-word;">The real disappointment was that the DSL Reports speed = test, which I used to recommend is so broken. It claims it can't reach any = of its 3 selected test sites, because I may have an "alien script" and sugg= ests my DNS might be "slow" or my Chrome might have browser malware.

=0A=

 

=0A

Now, maybe Chrome ha= s browser malware on my machine. This is troubling to a serious degree to m= e, and I will be investigating.

=0A

 

=0A

However, Cloudflare seems to be somewhat flaky to a signi= ficant degree. (It also doesn't seem to push a fast network connection near= ly hard enough to measure lag under load, it seems to me).

=0A

 

=0A

 

=0A

 

=0A

On Wednesday, January 4, 2023 2:20p= m, starlink-request@lists.bufferbloat.net said:

=0A
=0A

> Send Starlink mailin= g list submissions to
> starlink@lists.bufferbloat.net
> > To subscribe or unsubscribe via the World Wide Web, visit
>= ; https://lists.bufferbloat.net/listinfo/starlink
> or, via email, = send a message with subject or body 'help' to
> starlink-request@li= sts.bufferbloat.net
>
> You can reach the person managing = the list at
> starlink-owner@lists.bufferbloat.net
>
= > When replying, please edit your Subject line so it is more specific> than "Re: Contents of Starlink digest..."
>
>
> Today's Topics:
>
> 1. Re: [Rpm] the grinch meets c= loudflare's christmas present
> (jf@jonathanfoulkes.com)
> =
>
> -----------------------------------------------------= -----------------
>
> Message: 1
> Date: Wed, 4 Ja= n 2023 14:20:15 -0500
> From: "jf@jonathanfoulkes.com" <jf@jonat= hanfoulkes.com>
> To: Dave Taht <dave.taht@gmail.com>
> Cc: bloat <bloat@lists.bufferbloat.net>, libreqos
> <= ;libreqos@lists.bufferbloat.net>, Cake List
> <cake@lists.buf= ferbloat.net>, Dave Taht via Starlink
> <starlink@lists.buffe= rbloat.net>, Rpm <rpm@lists.bufferbloat.net>,
> IETF IPPM = WG <ippm@ietf.org>
> Subject: Re: [Starlink] [Rpm] the grinch= meets cloudflare's christmas
> present
> Message-ID: <8= 45161E4-474C-44A9-92D4-1702748A3DA1@jonathanfoulkes.com>
> Conte= nt-Type: text/plain; charset=3D"utf-8"
>
> HNY Dave and al= l the rest,
>
> Great to see yet another capacity test add= latency metrics to the results. This
> one looks like a good start= .
>
> Results from my Windstream DOCSIS 3.1 line (3.1 on d= ownload only, up is 3.0)
> Gigabit down / 35Mbps up provisioning. U= sing an IQrouter Pro (an i5 x86) with Cake
> set for 710/31 as this= ISP can=E2=80=99t deliver reliable low-latency unless you
> shave = a good bit off the targets. My local loop is pretty congested.
> > Here=E2=80=99s the latest Cloudflare test:
>
> --= ------------ next part --------------
> A non-text attachment was s= crubbed...
> Name: CFSpeedTest_Gig35_20230104.png
> Type: i= mage/png
> Size: 379539 bytes
> Desc: not available
&g= t; URL:
> <https://lists.bufferbloat.net/pipermail/starlink/atta= chments/20230104/d89867fe/attachment.png>
> -------------- next = part --------------
>
>
> And an Ookla test run j= ust afterward:
>
> -------------- next part --------------=
> A non-text attachment was scrubbed...
> Name: Speedtest_= net_Gig35_20230104.png
> Type: image/png
> Size: 40589 byte= s
> Desc: not available
> URL:
> <https://lists.= bufferbloat.net/pipermail/starlink/attachments/20230104/d89867fe/attachment= -0001.png>
> -------------- next part --------------
> <= br />>
> They are definitely both in the ballpark and correspon= d to other tests run from
> the router itself or my (wired) MacBook= Pro.
>
> Cheers,
>
> Jonathan
> =
>
> > On Jan 4, 2023, at 12:26 PM, Dave Taht via Rpm> <rpm@lists.bufferbloat.net> wrote:
> >
> = > Please try the new, the shiny, the really wonderful test here:
&g= t; > https://speed.cloudflare.com/
> >
> > I would= really appreciate some independent verification of
> > measurem= ents using this tool. In my brief experiments it appears - as
> >= ; all the commercial tools to date - to dramatically understate the
&g= t; > bufferbloat, on my LTE, (and my starlink terminal is out being
> > hacked^H^H^H^H^H^Hworked on, so I can't measure that)
> = >
> > My test of their test reports 223ms 5G latency under lo= ad , where
> > flent reports over 2seconds. See comparison attac= hed.
> >
> > My guess is that this otherwise lovely n= ew tool, like too many,
> > doesn't run for long enough. Admitte= dly, most web objects (their
> > target market) are small, and s= o long as they remain small and not
> > heavily pipelined this t= est is a very good start... but I'm pretty
> > sure cloudflare i= s used for bigger uploads and downloads than that.
> > There's n= o way to change the test to run longer either.
> >
> >= ; I'd love to get some results from other networks (compared as usual to> > flent), especially ones with cake on it. I'd love to know if t= hey
> > measured more minimum rtts that can be obtained with fq_= codel or cake,
> > correctly.
> >
> > Love= Always,
> > The Grinch
> >
> > --
&g= t; > This song goes out to all the folk that thought Stadia would work:<= br />> >
> https://www.linkedin.com/posts/dtaht_the-mushroom-= song-activity-6981366665607352320-FXtz
> > Dave T=C3=A4ht CEO, T= ekLibre, LLC
> >
> <image.png><tcp_nup-2023-01-= 04T090937.211620.LTE.flent.gz>__________________________________________= _____
> > Rpm mailing list
> > Rpm@lists.bufferbloat.= net
> > https://lists.bufferbloat.net/listinfo/rpm
> >
> ------------------------------
>
> Subj= ect: Digest Footer
>
> ___________________________________= ____________
> Starlink mailing list
> Starlink@lists.buffe= rbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
&= gt;
>
> ------------------------------
>
&g= t; End of Starlink Digest, Vol 22, Issue 10
> *********************= *******************
>

=0A
 ------=_20230104172101000000_59262--