Can anyone point me to any resources on applying a methodology/process for determining hardware capacity requirements(disk/ram/cpu/network) for high-volume, web site application?
I'm looking for a way to derive metrics to figure out how many of what type of servers to buy. The application architecture is undecided (MS COM vs. EJB), so a generalized process would be helpful. Planning tools would also be helpful.
Yours is the first posting to the distributed computing area. And your question would have to be a hard one. Planning for a high volume web site application requires judicious decision making. I don't know of a documented architecture process that would allow you to create these estimates. There are two sources I will explore for you in the next few days. One is friends of mine at Netscape who have to do this regularly. The other are system simulation tools. We used a simulation tool on a recent project to simulate volume and throughput. Try doing a web search on 'system simulation' and see what happens.