http://lwn.net/Articles/128681/ This is a special feed containing comments posted to the individual LWN article titled "Pluggable congestion avoidance modules". hourly 2 http://lwn.net/Articles/129039/rss 2005-03-24T15:20:01+00:00 hmh <p>Agreed. The explanations are very very good, and this is what sets LWN appart from the other publications.</p> <p>As a side note, after reading the article, I tried to select Westwood instead of BIC (which the kernel selects by default as I compiled all of them in), and got instant higher TCP performance on a network with an WiFi backbone link.</p> <p><em>Thank you LWN!</em></p> http://lwn.net/Articles/129038/rss 2005-03-24T15:05:13+00:00 Duncan Indeed. I had in mind to post something very similar. Unlike some of the <br> other kernel articles, which sometimes go over my head (altho I'm sure <br> some find them as incredibly useful as I and apparently you found this, <br> and if that someone is the guy/gal writing the driver for that peripheral <br> I'm going to be buying...), this is exactly the type of thing that I can <br> make use of, therefore, exactly the type of thing that I have in mind when <br> it comes time to fill in that subscription reauthorization. <br> <br> Duncan <br> http://lwn.net/Articles/128979/rss 2005-03-24T08:09:24+00:00 geertj Compliments for the article. The background and history given here is very useful, and to me makes the difference between any Linux news site and the quality articles at LWN.<br> http://lwn.net/Articles/128971/rss 2005-03-24T05:16:09+00:00 corbet Yes, Stephen's patch include all of those algorithms - most of them are already in the kernel, he just repackaged them into modules. http://lwn.net/Articles/128963/rss 2005-03-24T03:35:03+00:00 jwb Are all these algorithms actually implemented in the proposed patch? It could be interesting to <br> play with, because it makes TCP a game. Individual sysops may choose an algorithm that gives <br> them a slight advantage while ruining aggregate performance. That sounds pretty fun.<br>