<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">No one responded to Luca's Sept 1 comment (on the bloat list) that the new code seems to do tail drop rather than longest queue drop. <div><br></div><div>If this is so, bandwidth sharing will not be fair since FQ alone is not enough. This was shown in the previously cited Bell Labs paper : <a href="http://ect.bell-labs.com/who/stiliadi/papers/jsac99.pdf">http://ect.bell-labs.com/who/stiliadi/papers/jsac99.pdf</a>. The following table is copied from the paper.</div><div><img id="6b2d6a14-c085-411b-b4b5-6e7798575e8c" height="164" width="374" apple-width="yes" apple-height="yes" src="cid:3A266A72-5934-4A47-8AA6-218C0E9A3675"></div><div><br></div><div>These results are a kind of benchmark showing that fairness needs both FQ and LQD. The paper also shows that a more easily implemented, approximate LQD works OK and avoids the need to sort flow queues in size order.</div><div><br></div><div>Jim Roberts</div><div><br></div><div><br></div><div><br></div><div>They showed the <br><div><div>On 21 Sep 2013, at 01:18, Dave Taht wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div>The best writeup of Eric Dumazet's (and team's) latest assault on<br>bloat was in lwn last month. The article just came out from behind the<br>paywall for general reading and is here:<br><br><a href="http://lwn.net/Articles/564978/">http://lwn.net/Articles/564978/</a><br><br>(I haven't fiddled with these new features because I'm in the final<br>throes of getting a decent beta of cerowrt out on 3.10.12 with some<br>annoy-the-nsa features in it. )<br><br>The only benchmark I've seen on the new scheduler so far was:<br><br>http://thread.gmane.org/gmane.linux.network/281023<br><br>Anybody got relevant benchmarks?<br><br>I am very happy to see the TSO/GSO sizing stuff as well. I hope that<br>this will result in GSO doing sane things on 100Mbit and slower<br>devices.<br><br>-- <br>Dave Täht<br>_______________________________________________<br>aqm mailing list<br>aqm@ietf.org<br>https://www.ietf.org/mailman/listinfo/aqm<br></div></blockquote></div><br></div></body></html>