Perf test on Sun Vs HP app ser. Is it ok, if we have tested only one type?
I have a unique situation ??, need ur advise folks. we had done the performance testing (Batch, user scalability & stress) for our application which was on Sun machines (Vorlon app server), but now the Sun machines will be replaced by HP app servers (Which are expected to perform much batter then Sun stuff Ė in fact thatís the reason we r switching). HP boxes will have almost the same configuration as SUN, like the number of CPUís, CPU speed, RAM and stuff like that.
All other parts of the system, the DB, the network connectivity, the application residing on the app server will remain the same.
So, now the question is, do we need to carry out a full blown performance testing, with the new stuff in place (NOTE: Just the HP-App server will be new, in place of SUN app server)? What options can we explore other then a full blown perf test? What are the potential risks if we do not performa another round of perf test?
thankx in advance.
Education is a progressive discovery of our own ignorance......Will Durant
Re: Perf test on Sun Vs HP app ser. Is it ok, if we have tested only one type?
If the HP boxes are expected to perform better then configuring them "almost the same" as the Sun boxes may not be the best config to take advantage of the expected performance gains.
If you have already benchmarked your app on Sun is it too much time/expense to re-benchmark on HP and compare.
Finally how do you verify the cost of the new hardware has brought increased performance. One way to do it is not to performance test just put it live, if expectations are true Great! if not "heads will roll"..... I know which I prefer.