What's the best way to test concurrency?

Hello Everybody,

Being new to database design, I have the following question: if I have two different database designs to solve one real world problem, which is the best way to test performance of both designs under the closest you could get to "real world" circumstances without putting the databases live Basically, I have a database with tables with short and narrow primary keys, some denormalisation and clustered indexes that do not coincide with the primary keys. On the other hand, I have a totally (3rd. form) normalised database with longer primary keys that also happen to be clustered indexes for the corresponding tables. I also have a bunch of stored procedures to perform the same tasks for both designs. I would like to know which design will work best (in terms of response time) when I put them live (no application involved here, by putting live, I just mean firing stored procedures to process data). I have been told there is a tool called Data Factory (from the Quest suite) to generate meaningful test data, but I am not sure if this tool also will help me simulate an environment with multiple users and heavy load. If it doesn't, can anybody suggest another tool for this particular purpose

Apologies if the question is pretty basic. I do appreciate your help on this.

Squig.