Maurice, We might want to take this into another thread, as to not steal the original, but we should keep it on the list so that others can read it. If anyone objects, please re-port and reference the original.
> [I]t appears that you find apache fronting tomcat via mod_jk to be a > very workable setup. I do. > Do you find apache -> mod_jk -> tomcat to be "Fast" ? I find it "fast enough". Local network access of our test machines leaves us nothing to complain about, although we have never benchmarked the application with other front servers (or with no front server at all). We also do not have any clustering or anything like that at this time, so Apache/mod_jk doesn't have any complex decisions to make. > Do you find it as fast as tomcat alone ? (Well, by definition, > putting something in between two other things can only make a > process slower, but "apparency" is important on the web.) For us, fronting Tomcat is a necessity: we have several webapps running under different VMs, so we can't connect directly to Tomcat using the HTTP connector. We also have some static content that I personally feel is better served by Apache, but that's just because I'm a native bigot ;) > So the question really is: Is apache fronting tomcat via mod_jk > apparently fast ? (to your taste and needs, of course). I have no complaints. We're working with a single box, so Apache-Tomcat connections probably use shared memory behind the scenes, so that might be a factor in why we are so happy with performance ;) Don't forget that pound was designed to be a proxy and lb, whereas Apache was designed to be an entire web server with tons of features. Pound is likely faster due to sheer simplicity, since it doesn't need all kinds of hooks that Apache exposes for mod_* and all the rest. -chris
signature.asc
Description: OpenPGP digital signature
