Lets make wifi fast again!
 help / color / mirror / Atom feed
From: Bob McMahon <bob.mcmahon@broadcom.com>
To: Dave Taht <dave.taht@gmail.com>
Cc: Make-Wifi-fast <make-wifi-fast@lists.bufferbloat.net>
Subject: Re: [Make-wifi-fast] Fwd: a succession of networkQuality tests
Date: Sun, 19 Dec 2021 17:54:37 -0800	[thread overview]
Message-ID: <CAHb6LvpVyuK0LrFmeNvJ-vYJT2OQukZ92xo-xHKtpD2nC4763w@mail.gmail.com> (raw)
In-Reply-To: <CAHb6LvqgSNYZ6opN8S+uRObNJds0hR3-O9SpE6bUG+uJDbMUcg@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 7855 bytes --]

sorry, but BB8 is the bounceback histogram.  The OWD is on it's own line.
Also, one must build from the master branch to get this feature w/latest
bug fixes (as it's early and hasn't been thoroughly tested nor released as
part of an iperf 2 release.)

[  1] 0.00-10.09 sec  *OWD Delays (ms) Cnt=472
 To=15.290/1.068/123.623/17.616 From=5.498/4.806/23.336/1.342
Asymmetry=-9.792/-118.707/4.013/17.356*

The OWDs include the to, from and asymmetry in a mean/min/max/var format.
(uses Welford's algo
<https://nullbuffer.com/articles/welford_algorithm.html> to calculate
running mean and variance.) The asymmetry is left signed (vs an absolute
value) in order to preserve directional information.

Bob


On Sun, Dec 19, 2021 at 5:40 PM Bob McMahon <bob.mcmahon@broadcom.com>
wrote:

> FYI, I've added a --bounceback option to iperf 2
> <https://iperf2.sourceforge.io/iperf-manpage.html>.  One can use the BB
> cnt as analogous for messages per unit time. It by default turns on
> TCP_QUICKACK so one can compare the TCP RTT vs the app level bounceback (-i
> is recommended for tcpinfo sampling.) There is an option to add a full
> duplex TCP stream to the same dest along with the bounceback traffic.  If
> you use --trip-times you'll get a histogram (BB8) of the one way delays too
> (clocks must be sync'ed and assumed so)
>
> This is designed to be a simple socket level test. Please do feel free to
> comment and suggest improvements.
>
> [rjmcmahon@ryzen3950 iperf2-code]$ src/iperf - hidden.com --bounceback -t
> 10 --trip-times -e --hide-ips
> ------------------------------------------------------------
> Client connecting to (**hidden**), TCP port 5001 with pid 240864 (1 flows)
> Write buffer size:  100 Byte
> Bounce-back test (size= 100 Byte) (hold=0 usecs)
> TOS set to 0x0 and nodelay (Nagle off)
> TCP window size: 85.0 KByte (default)
> ------------------------------------------------------------
> [  1] local *.*.*.133%enp4s0 port 55634 connected with *.*.*.123 port 5001
> (MSS=1448) (bb len/hold=100/0) (trip-times) (sock=3) (irtt/icwnd=14469/14)
> (ct=14.52 ms) on 2021-12-19 17:29:23 (PST)
> [ ID] Interval        Transfer    Bandwidth         *BB cnt*=avg/min/max/stdev
>         Rtry  Cwnd/RTT
> [  1] 0.00-10.03 sec   178 KBytes   145 Kbits/sec    *911*=10.982/6.187/21.876/1.883
> ms    0   14K/13160 us
> [  1] 0.00-10.03 sec  OWD Delays (ms) Cnt=911  To=5.222/0.383/16.277/1.875
> From=5.760/5.500/6.091/0.092 Asymmetry=0.538/-10.678/5.421/1.871
> [  1] 0.00-10.03 sec BB8(f)-PDF:
> bin(w=100us):cnt(911)=62:1,79:1,97:13,98:23,99:65,100:92,101:115,102:135,103:85,104:70,105:43,106:30,107:17,108:15,109:11,110:6,111:3,112:4,113:5,114:5,115:5,116:2,117:4,118:2,119:4,120:5,121:4,122:2,123:3,124:2,125:1,126:3,127:3,128:1,131:1,132:1,133:7,134:1,135:4,138:6,139:6,141:4,142:2,143:4,144:2,145:7,146:6,147:5,148:6,149:6,150:12,151:1,153:3,154:2,155:1,156:4,157:6,158:1,159:4,160:2,162:3,163:2,165:2,166:2,167:3,168:1,169:1,171:1,172:1,177:2,179:1,183:1,187:1,188:1,191:2,197:1,208:1,219:1
> (5.00/95.00/99.7%=99/154/197,Outliers=0,obl/obu=0/0) (0.000 ms/0.0)
>
>
>
> [rjmcmahon@ryzen3950 iperf2-code]$ src/iperf -c hidden.com --bounceback
> -t 10 --trip-times -e --hide-ips -*-bounceback-congest*
> [  2] local *.*.*.133 port 55640 connected with *.*.*.123 port 5001
> (full-duplex) (trip-times)
> [  1] local *.*.*.133%enp4s0 port 55642 connected with *.*.*.123 port 5001
> (MSS=1448) (bb len/hold=100/0) (trip-times) (sock=3) (irtt/icwnd=10580/14)
> (ct=10.67 ms) on 2021-12-19 17:29:48 (PST)
> ------------------------------------------------------------
> Client connecting to (**hidden**), TCP port 5001 with pid 240911 (1 flows)
> Write buffer size:  100 Byte
> Bounce-back test (size= 100 Byte) (hold=0 usecs)
> TOS set to 0x0 and nodelay (Nagle off)
> TCP window size: 85.0 KByte (default)
> ------------------------------------------------------------
> [ ID] Interval       Transfer     Bandwidth
> [  2] 0.00-10.02 sec  7.82 MBytes  6.55 Mbits/sec
> [ ID] Interval        Transfer    Bandwidth         BB
> cnt=avg/min/max/stdev         Rtry  Cwnd/RTT
> [  1] 0.00-10.09 sec  92.2 KBytes  74.9 Kbits/sec    *472*=20.788/6.149/128.539/17.973
> ms    1   14K/42283 us
> [  1] 0.00-10.09 sec  OWD Delays (ms) Cnt=472
>  To=15.290/1.068/123.623/17.616 From=5.498/4.806/23.336/1.342
> Asymmetry=-9.792/-118.707/4.013/17.356
> [  1] 0.00-10.09 sec BB8(f)-PDF:
> bin(w=100us):cnt(472)=62:1,82:1,91:1,94:5,95:1,96:5,97:6,98:6,99:3,100:6,101:4,102:4,103:3,104:3,105:7,106:6,107:4,108:6,109:2,110:5,111:5,112:2,113:3,114:3,115:4,116:3,117:4,118:1,119:3,120:4,121:3,122:3,123:1,124:2,126:2,127:2,128:1,129:3,132:3,133:1,135:2,136:2,137:5,138:1,139:3,140:8,141:8,142:8,143:9,144:12,145:10,146:14,147:6,148:7,149:4,150:6,151:7,152:3,153:5,154:5,155:4,156:2,157:6,158:8,159:2,160:6,161:11,162:3,163:2,164:6,165:2,166:5,167:3,168:2,169:2,170:7,171:6,172:1,174:2,175:3,176:1,177:1,178:1,179:2,180:1,181:1,182:1,183:3,184:2,186:1,187:3,189:2,190:1,193:2,194:1,195:1,196:2,197:1,199:2,200:2,201:1,202:1,204:4,206:3,208:1,209:1,213:1,216:1,218:1,219:1,220:1,221:1,225:1,226:1,228:1,240:1,248:1,252:1,259:1,262:1,269:1,275:1,276:1,279:1,281:2,282:1,283:1,290:1,291:2,293:2,295:1,297:1,298:1,301:1,304:1,313:1,317:1,319:1,325:1,327:1,330:1,344:1,359:1,360:1,365:1,372:1,375:1,396:1,406:1,412:1,433:1,440:1,442:1,469:1,480:2,485:1,515:1,521:1,533:1,535:1,536:1,543:1,544:1,547:1,559:1,571:1,597:1,600:1,604:1,608:1,663:1,668:1,689:1,735:1,750:1,784:1,810:1,818:1,832:1,840:1,847:1,850:1,884:1,892:1,897:1,924:1,997:1,1010:1,1039:1,1046:1,1177:1,1286:1
> (5.00/95.00/99.7%=98/604/1177,Outliers=2,obl/obu=0/0) (0.000 ms/0.0)
> [ ID] Interval        Transfer    Bandwidth    Burst Latency
> avg/min/max/stdev (cnt/size) inP NetPwr  Reads=Dist
> [ *2] 0.00-10.40 sec   149 MBytes   120 Mbits/sec
>  243.512/21.708/616.257/106.136 ms (1190/131072) 3.48 MByte 61.58
>  1567094=900:2894:3359:2773:1327:1282:1624:1760
> [SUM] 0.00-10.41 sec   157 MBytes   126 Mbits/sec
>
>
> CLIENT SPECIFIC OPTIONS*-b*, *--bandwidth* *n*[kmgKMG][,*n*[kmgKMG]] | *n*
> [kmgKMG]ppsset target bandwidth to *n* bits/sec (default 1 Mbit/sec) or
> *n* packets per sec. This may be used with TCP or UDP. Optionally, for
> variable loads, use format of mean,standard deviation*--bounceback*run a
> tcp bounceback test (set size with -l or --len, defaults to 100 bytes)
> *--bounceback-congest*request a concurrent full-duplex TCP stream
> *--bounceback-hold **n*request the server to insert a delay of n
> milliseconds between its read and write (default is no delay)
> *--bounceback-period **n*request the client schedule a send every n
> milliseconds (w/o option bouncebacks are immediately scheduled. With
> command option and no optional value given, the interval will be one
> second.)*--bounceback-no-quickack*request the server not set the
> TCP_QUICKACK socket option (disabling TCP ACK delays) during a bounceback
> test (see NOTES)*--burst-period **n*
> Bob
>

-- 
This electronic communication and the information and any files transmitted 
with it, or attached to it, are confidential and are intended solely for 
the use of the individual or entity to whom it is addressed and may contain 
information that is confidential, legally privileged, protected by privacy 
laws, or otherwise restricted from disclosure to anyone else. If you are 
not the intended recipient or the person responsible for delivering the 
e-mail to the intended recipient, you are hereby notified that any use, 
copying, distributing, dissemination, forwarding, printing, or copying of 
this e-mail is strictly prohibited. If you received this e-mail in error, 
please return the e-mail to the sender, delete it from your computer, and 
destroy any printed copy of it.

[-- Attachment #2: Type: text/html, Size: 9326 bytes --]

      reply	other threads:[~2021-12-20  1:54 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CAA93jw4N97k1QYY5oBWKeRc=-AgxBEy3r6K7mtz-NKccJn9Gwg@mail.gmail.com>
2021-12-19 16:52 ` Dave Taht
2021-12-20  1:40   ` Bob McMahon
2021-12-20  1:54     ` Bob McMahon [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://lists.bufferbloat.net/postorius/lists/make-wifi-fast.lists.bufferbloat.net/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAHb6LvpVyuK0LrFmeNvJ-vYJT2OQukZ92xo-xHKtpD2nC4763w@mail.gmail.com \
    --to=bob.mcmahon@broadcom.com \
    --cc=dave.taht@gmail.com \
    --cc=make-wifi-fast@lists.bufferbloat.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox