I agree with both Marc and Siggi ;-) It depends on what you are trying to achieve with your load test.
When the goal is to load-test the *technical infrastructure* (i.e. regardless of the deployed application) the Siggi's way is the way to go. When the goal is to load-test the *application* then I would follow Marc's path. Let's face it: there is not much value in knowing that your appserver can handle n requests per second if half the responses only report an exception because of concurrency issues or resource locks in your app. cheers Dierk
