> -----Original Message----- > From: Kaushal Shriyan [mailto:kaushalshri...@gmail.com] > Sent: Monday, November 14, 2016 8:46 AM > To: Tomcat Users List > Subject: Re: Apache Tomcat Version 7.0.59 on CentOS 6.8 to handle 2k > requests/second > > On Mon, Nov 14, 2016 at 8:06 PM, Mark Thomas <ma...@apache.org> wrote: > > > On 14/11/2016 14:28, Kaushal Shriyan wrote: > > > Hi, > > > > > > is there a way to configure tomcat to handle 2k requests/second (2k > > meaning > > > 2000 requests per second). How many cpu cores do i need to setup > > > tomcat > > to > > > handle 2k requests/second? How much physical memory the server > > > should > > have? > > > > That will depend on your application. > > > > My laptop will easily handle 10 times that for a simple application. I > > have seen much larger servers unable to handle more than a handful of > > requests a second for a large, complex, poorly performing application. > > > > > Any guidelines or rule of thumb to be followed? Any help will be > > > highly appreciable. > > > > Profile your application. > > > > Thanks Mark Thomas for the quick reply. > > Please suggest me any open source profiling application which i can start > with. I > will appreciate if you can help me understand the relation between 2k > requests/second vs cpu cores of the server. > > Thanks again in advance. > > Regards, > > Kaushal
There is no fixed relationship between requests per second and number of cores. It all depends on your application. Think of it this way: if each request takes 1 second, you'll get fewer requests per second than if each request takes 1 millisecond. Whether a request takes 1 second or 1 millisecond has more to do with your application than hardware. As for an open source profiler, try asking your favorite search engine. John
