Sarang, I had to accomplish a similar issue a number of years ago at a different job. How I solved the problem was to put up a test environment with my test users on one hub, a single link to another hub which the server was attached, measured the traffic flow over a period of time then used some basic statistics to extrapolate out, added a fudge factor and hoped we were correct. We installed the required lines to the ISP which I believe were 2 t1s at the time, and hoped we were right. We were okay for about 2 years before we had to upgrade because of traffic load. The process was crude and with more money I could have setup a more exact testing environment which would have yielded better results. Measuring system and traffic performance is a art as well as a science, IMHO. jim On Fri, Nov 13, 2009 at 3:00 AM, lonetwin <lonetwins@xxxxxxxxx> wrote: > Hello everyone, > > we have developed a webapplication which will be put into production > soon. We expect close to a 1000 users to have access to the webapplication > and assume 300 ( 30%) users to be using it at a given point of time. We > have > tomcat taking care of the UI and jboss is the application server with mysql > as the db. > > Now we want to calculate the bandwidth that will be needed for this > application to be accessed by the 300 users. We are going to host it in our > premises, and want to know what is the bandwidth that we will have to get > from our ISP so that all users are able to access our application. > > Any help/idea on this ? > > thanks, > Sarang > -- > redhat-list mailing list > unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe > https://www.redhat.com/mailman/listinfo/redhat-list > -- James R. Jones System Manager UAF-BBC PO Box 1070 Dillingham, AK 99576 907-842-8312 -- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
