This was discussed in length on this mailing list back when the article was posted. I vaguely remember the discussion, other than it happened.
Checking the archives should give you what you are looking for. --Mark On Thu, Jun 18, 2009 at 3:04 PM, David Rosenstrauch<[email protected]> wrote: > Just wondering what thoughts the MINA/NIO brain trust here might have about > a blog post I came upon a while back. The author (a pretty well known > hard-core developer who's worked for Google, as well as his own startups) > makes the case that NIO is actually significantly slower than traditional > blocking Java I/O when used in conjunction with NPTL threading. > > Apparently, NPTL allows the traditional blocking I/O to scale beyond just a > few hundred or thousand threads, thereby making it a viable option again for > large-scale multi-threaded servers. > > Curious if anyone's got any thoughts/experiences that either confirm or > refute what he's saying. > > http://mailinator.blogspot.com/2008/02/kill-myth-please-nio-is-not-faster-than.html > > Thanks, > > DR >
