<div dir="ltr"><div>I also found this to sample tcp latency performance: <a href="https://code.google.com/p/paping/">https://code.google.com/p/paping/</a></div><div><a href="https://code.google.com/p/paping/"><br></a>Does fq_codel's flow classification supports external classifier filter? Say I only want to hash on dst port in my case.</div>
<div><br></div><div>Thanks</div><div>Mo</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">2013/12/5 Toke Høiland-Jørgensen <span dir="ltr"><<a href="mailto:toke@toke.dk" target="_blank">toke@toke.dk</a>></span><br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">Dong Mo <<a href="mailto:montedong@gmail.com">montedong@gmail.com</a>> writes:<br>
<br>
> What I am trying to do is measure the delay of packets of the same<br>
> flow experienced in A under different queuing policies (sfq or<br>
> fq_codel).<br>
><br>
> Is there a simple way to do this?<br>
<br>
</div>I'm currently in the early planning stages for having something like<br>
this incorporated into the netperf-wrapper [1] testing tool. In the<br>
meantime, you could look at either thrulay [2] (old and apparently<br>
abandoned, but may work in one of its variants) as a measurement tool,<br>
or tcpprobe [3] to output statistics that can (probably) tell you.<br>
<br>
-Toke<br>
<br>
[1] <a href="https://github.com/tohojo/netperf-wrapper" target="_blank">https://github.com/tohojo/netperf-wrapper</a><br>
<br>
[2] <a href="http://e2epi.internet2.edu/thrulay/" target="_blank">http://e2epi.internet2.edu/thrulay/</a><br>
<a href="http://thrulay-hd.sourceforge.net/" target="_blank">http://thrulay-hd.sourceforge.net/</a><br>
<a href="http://thrulay-ng.sourceforge.net/" target="_blank">http://thrulay-ng.sourceforge.net/</a><br>
<br>
[3] <a href="http://www.linuxfoundation.org/collaborate/workgroups/networking/tcpprobe" target="_blank">http://www.linuxfoundation.org/collaborate/workgroups/networking/tcpprobe</a><br>
</blockquote></div><br></div>