-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Konstantin,
On 4/15/2010 7:17 PM, Konstantin Kolinko wrote: > The main competitors are APR connector vs. Nio connector, as both > provide multiplexing aka polling, allowing to serve more sockets than > the count of your worker threads. > > The APR connector sure has a bit more longer history, dating back to TC 5.5. > The Nio connector is more modern, available since TC 6.x only, but > that is several years already. My (as yet unpublished) performance tests show that APR and NIO offer virtually the same performance. Make sure you set sendFile="true" if you are doing any significant static-file serving through your Tomcat instance. >> less stable than the pure Java code > > Note, that that also depends on the JRE implementation that you would be > using. +1 The APR code is more likely to actually crash your JVM than the pure-Java implementation, but I haven't seen any recent complaints about APR crashing at all. Jeffrey is having a bear of a time getting APR working in his environment due to a number of issues that are all piling up. My advice: set up a testing environment where you can play and perform load testing. Compare the performance of the NIO versus APR connectors. - -chris -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkvI0V0ACgkQ9CaO5/Lv0PB1awCbBSPmpnyo1zPSn0pEmXwBBTVm 1kkAoJtNgbJHSRBCEbQpjhiFhg6ucAS1 =U+zf -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
