<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:st1="urn:schemas-microsoft-com:office:smarttags" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 11 (filtered medium)">
<o:SmartTagType namespaceuri="urn:schemas-microsoft-com:office:smarttags"
name="country-region"/>
<o:SmartTagType namespaceuri="urn:schemas-microsoft-com:office:smarttags"
name="place"/>
<o:SmartTagType namespaceuri="urn:schemas-microsoft-com:office:smarttags"
name="PersonName"/>
<!--[if !mso]>
<style>
st1\:*{behavior:url(#default#ieooui) }
</style>
<![endif]-->
<style>
<!--
/* Font Definitions */
@font-face
{font-family:Wingdings;
panose-1:5 0 0 0 0 0 0 0 0 0;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:12.0pt;
font-family:"Times New Roman";}
a:link, span.MsoHyperlink
{color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{color:purple;
text-decoration:underline;}
p.MsoPlainText, li.MsoPlainText, div.MsoPlainText
{margin:0in;
margin-bottom:.0001pt;
font-size:10.0pt;
font-family:"Courier New";}
@page Section1
{size:8.5in 11.0in;
margin:1.0in 77.95pt 1.0in 77.95pt;}
div.Section1
{page:Section1;}
-->
</style>
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>-----Original Message-----<br>
From: Starlink [mailto:starlink-bounces@lists.bufferbloat.net] On Behalf Of rjmcmahon
via Starlink<br>
Sent: Monday, January 9, 2023 12:47 PM<br>
To: Dave Taht<br>
Cc: starlink@lists.bufferbloat.net; mike.reynolds@netforecast.com; libreqos; <st1:PersonName
w:st="on">David P. Reed</st1:PersonName>; Rpm; bloat<br>
Subject: Re: [Starlink] [Rpm] Researchers Seeking Probe Volunteers in <st1:country-region
w:st="on"><st1:place w:st="on">USA</st1:place></st1:country-region></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>The write to read latencies (OWD) are on the server side in CLT form. <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>Use --histograms on the server side to enable them.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>Your client side sampled TCP RTT is 6ms with less than a 1 ms of <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>variance (or sqrt of variance as variance is typically squared)<o:p></o:p></span></font></p>
<p class=MsoPlainText><b><i><font size=2 face="Courier New"><span
style='font-size:10.0pt;font-weight:bold;font-style:italic'>[RR] or standard
deviation (std for short) </span></font></i></b><b><i><font face=Wingdings><span
style='font-family:Wingdings;font-weight:bold;font-style:italic'>J</span></font><o:p></o:p></i></b></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'> No <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>retries suggest the network isn't dropping packets.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>All the newer bounceback code is only master and requires a compile from
<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>source. It will be released in 2.1.9 after testing cycles. Hopefully,
in <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>early March 2023<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>Bob<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>https://sourceforge.net/projects/iperf2/<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'><o:p> </o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> The DC that so graciously loaned us 3 machines for the testbed
(thx<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> equinix!), does support ptp, but we have not configured it yet. In
ntp<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> tests between these hosts we seem to be within 500us, and
certainly<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 50us would be great, in the future.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> I note that in all my kvetching about the new tests' needing<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> validation today... I kind of elided that I'm pretty happy with<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> iperf2's new tests that landed last august, and are now appearing
in<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> linux package managers around the world. I hope more folk use
them.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> (sorry robert, it's been a long time since last august!)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> Our new testbed has multiple setups. In one setup - basically the<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> machine name is equal to a given ISP plan, and a key testing point
is<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> looking at the differences between the FCC 25-3 and 100/20 plans
in<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> the real world. However at our scale (25gbit) it turned out that<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> emulating the delay realistically has problematic.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> Anyway, here's a 25/3 result for iperf (other results and iperf
test<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> type requests gladly accepted)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> root@lqos:~# iperf -6 --trip-times -c c25-3 -e -i 1<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> ------------------------------------------------------------<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> Client connecting to c25-3, TCP port 5001 with pid 2146556 (1
flows)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> Write buffer size: 131072 Byte<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> TOS set to 0x0 (Nagle on)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> TCP window size: 85.3 KByte (default)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> ------------------------------------------------------------<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] local fd77::3%bond0.4 port 59396 connected with fd77::1:2
port<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 5001 (trip-times) (sock=3) (icwnd/mss/irtt=13/1428/948) (ct=1.10
ms)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> on 2023-01-09 20:13:37 (UTC)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ ID] Interval Transfer Bandwidth Write/Err
Rtry<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> Cwnd/RTT(var) NetPwr<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 0.0000-1.0000 sec 3.25 MBytes 27.3 Mbits/sec
26/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 19K/6066(262) us 562<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 1.0000-2.0000 sec 3.00 MBytes 25.2 Mbits/sec
24/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 15K/4671(207) us 673<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 2.0000-3.0000 sec 3.00 MBytes 25.2 Mbits/sec
24/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 13K/5538(280) us 568<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 3.0000-4.0000 sec 3.12 MBytes 26.2 Mbits/sec
25/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 16K/6244(355) us 525<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 4.0000-5.0000 sec 3.00 MBytes 25.2 Mbits/sec
24/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 19K/6152(216) us 511<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 5.0000-6.0000 sec 3.00 MBytes 25.2 Mbits/sec
24/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 22K/6764(529) us 465<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 6.0000-7.0000 sec 3.12 MBytes 26.2 Mbits/sec
25/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 15K/5918(605) us 554<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 7.0000-8.0000 sec 3.00 MBytes 25.2 Mbits/sec
24/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 18K/5178(327) us 608<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 8.0000-9.0000 sec 3.00 MBytes 25.2 Mbits/sec
24/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 19K/5758(473) us 546<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 9.0000-10.0000 sec 3.00 MBytes 25.2 Mbits/sec
24/0 0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 16K/6141(280) us 512<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> [ 1] 0.0000-10.0952 sec 30.6 MBytes 25.4 Mbits/sec 245/0<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> 0 19K/5924(491) us 537<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> On Mon, Jan 9, 2023 at 11:13 AM rjmcmahon <rjmcmahon@rjmcmahon.com>
<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>> wrote:<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> My biggest barrier is the lack of clock sync by the devices,
i.e. very<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> limited support for PTP in data centers and in end devices.
This <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> limits<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> the ability to measure one way delays (OWD) and most assume
that OWD <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> is<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> 1/2 and RTT which typically is a mistake. We know this
intuitively <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> with<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> airplane flight times or even car commute times where the one
way time<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> is not 1/2 a round trip time. Google maps & directions
provide a time<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> estimate for the one way link. It doesn't compute a round trip
and<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> divide by two.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> For those that can get clock sync working, the iperf 2
--trip-times<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> options is useful.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> --trip-times<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> enable the measurement of end to end write to read
latencies <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> (client<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> and server clocks must be synchronized)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> Bob<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > I have many kvetches about the new latency under load
tests being<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > designed and distributed over the past year. I am
delighted! that they<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > are happening, but most really need third party evaluation,
and<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > calibration, and a solid explanation of what network
pathologies they<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > do and don't cover. Also a RED team attitude towards
them, as well as<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > thinking hard about what you are not measuring
(operations research).<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > I actually rather love the new cloudflare speedtest,
because it tests<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > a single TCP connection, rather than dozens, and at the
same time folk<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > are complaining that it doesn't find the actual
"speed!". yet... the<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > test itself more closely emulates a user experience than
speedtest.net<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > does. I am personally pretty convinced that the fewer
numbers of flows<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > that a web page opens improves the likelihood of a good
user<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > experience, but lack data on it.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > To try to tackle the evaluation and calibration part,
I've reached out<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > to all the new test designers in the hope that we could
get together<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > and produce a report of what each new test is actually
doing. I've<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > tweeted, linked in, emailed, and spammed every
measurement list I know<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > of, and only to some response, please reach out to other
test designer<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > folks and have them join the rpm email list?<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > My principal kvetches in the new tests so far are:<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > 0) None of the tests last long enough.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > Ideally there should be a mode where they at least run to
"time of<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > first loss", or periodically, just run longer than
the<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > industry-stupid^H^H^H^H^H^Hstandard 20 seconds. There be
dragons<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > there! It's really bad science to optimize the internet
for 20<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > seconds. It's like optimizing a car, to handle well, for
just 20<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > seconds.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > 1) Not testing up + down + ping at the same time<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > None of the new tests actually test the same thing that
the infamous<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > rrul test does - all the others still test up, then down,
and ping. It<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > was/remains my hope that the simpler parts of the flent
test suite -<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > such as the tcp_up_squarewave tests, the rrul test, and
the rtt_fair<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > tests would provide calibration to the test designers.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > we've got zillions of flent results in the archive
published here:<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > https://blog.cerowrt.org/post/found_in_flent/<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > ps. Misinformation about iperf 2 impacts my ability to do
this.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > The new tests have all added up + ping and down + ping,
but not up +<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > down + ping. Why??<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > The behaviors of what happens in that case are really
non-intuitive, I<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > know, but... it's just one more phase to add to any one
of those new<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > tests. I'd be deliriously happy if someone(s) new to the
field<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > started doing that, even optionally, and boggled at how
it defeated<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > their assumptions.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > Among other things that would show...<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > It's the home router industry's dirty secret than darn
few "gigabit"<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > home routers can actually forward in both directions at a
gigabit. I'd<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > like to smash that perception thoroughly, but given our
starting point<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > is a gigabit router was a "gigabit switch" -
and historically been<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > something that couldn't even forward at 200Mbit - we have
a long way<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > to go there.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > Only in the past year have non-x86 home routers appeared
that could<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > actually do a gbit in both directions.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > 2) Few are actually testing within-stream latency<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > Apple's rpm project is making a stab in that direction.
It looks<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > highly likely, that with a little more work, crusader and<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > go-responsiveness can finally start sampling the tcp RTT,
loss and<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > markings, more directly. As for the rest... sampling
TCP_INFO on<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > windows, and Linux, at least, always appeared simple to
me, but I'm<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > discovering how hard it is by delving deep into the rust
behind<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > crusader.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > the goresponsiveness thing is also IMHO running WAY too
many streams<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > at the same time, I guess motivated by an attempt to have
the test<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > complete quickly?<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > B) To try and tackle the validation problem:ps.
Misinformation about<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > iperf 2 impacts my ability to do this.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> <o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > In the libreqos.io project we've established a testbed
where tests can<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > be plunked through various ISP plan network emulations.
It's here:<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > https://payne.taht.net (run bandwidth test for what's
currently hooked<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > up)<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > We could rather use an AS number and at least a ipv4/24
and ipv6/48 to<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > leverage with that, so I don't have to nat the various
emulations.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > (and funding, anyone got funding?) Or, as the code is
GPLv2 licensed,<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > to see more test designers setup a testbed like this to
calibrate<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > their own stuff.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > Presently we're able to test:<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > flent<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > netperf<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > iperf2<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > iperf3<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > speedtest-cli<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > crusader<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > the broadband forum udp based test:<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > https://github.com/BroadbandForum/obudpst<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > trexx<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> ><o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > There's also a virtual machine setup that we can remotely
drive a web<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > browser from (but I didn't want to nat the results to the
world) to<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > test other web services.<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > _______________________________________________<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > Rpm mailing list<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > Rpm@lists.bufferbloat.net<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>>> > https://lists.bufferbloat.net/listinfo/rpm<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>_______________________________________________<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>Starlink mailing list<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>Starlink@lists.bufferbloat.net<o:p></o:p></span></font></p>
<p class=MsoPlainText><font size=2 face="Courier New"><span style='font-size:
10.0pt'>https://lists.bufferbloat.net/listinfo/starlink<o:p></o:p></span></font></p>
</div>
</body>
</html>