For the sake of this question, let me assume that I have an application running on one instance of Tomcat that takes 10 seconds to complete a client request. If I cluster 10 machines together does it help with the through-put, i.e. meaning it might take 1 second to complete the request now, or does it just make the session redundant in case of failure. Just wondering about which setup would serve me the best and be the fastest for our clients.
Thanks for any input on this.
