Both very interesting ideas. I had not thought about that. However, I am concerned by the fact that in the HTTP GET statements, the IP of the web server and not of the BIG-IP is referenced. So if I either hard code that reference or use a variable, for the duration of that test run I will always go to that particular web server and the BIG-IP will not load balance as programmed. SO I either have to set up a seperate script which pulls all of the web server IP's/names out of a datapool and assigns them to a persistent variable at which time I can then use them in subsequent tests, I need to create a script to randomly/logically select a server, or I need to write several different scripts.