> From: Alex Jalali [mailto:[EMAIL PROTECTED] > Subject: RE: TomCat + mod_jk performance > > Maybe the problem is when I was testing static pages, it > wasn't tomcat as stand-alone vs. apache it was apache vs. > tomcat + connector + apache
That was hardly a fair test, was it? Adding processing to each request will never improve performance. > also I run the test locally on a windows machine. so perhaps java runs > slower on windows. Shouldn't matter much, if (big if) the JVMs are tuned comparably. The HotSpot JVM defaults for things like heap size and local allocation buffers are typically smaller on Windows, so out of the box, Java applications will appear to run better on Linux/Solaris. > Although i do have this question about non-static pages. Which do you > think is faster? let say you have 1GB ram and 2 CPUs. running > a) apache + 1 tomcat or > b) apache + 2 tomcat in cluster via mod_jk? I'd pick c: 1 Tomcat, no httpd. > would having 2 JVM on the same server that > has two CPU's run faster or it wouldn't make > any diffrence? The number of CPUs has no correlation with the number of JVMs. Tomcat is multi-threaded, and one instance will happily use all the CPUs you have if there's that much work to be done. > do you know of any tests done for clusters No, but clustering is bound to slow things down due to extra traffic among the members. - Chuck THIS COMMUNICATION MAY CONTAIN CONFIDENTIAL AND/OR OTHERWISE PROPRIETARY MATERIAL and is thus for use only by the intended recipient. If you received this in error, please contact the sender and delete the e-mail and its attachments from all computers. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
