RDP - Measuring Application Response Times

[tofes]

Hi all,

I'm after a little advice of running AutoIT scripts on a remote Terminal Server over RDP.

Just to give a bit of background I have already created some AutoIT scripts in our environment to run load testing in our terminal services environment.

This has worked extremely well with the times of certain user actions to complete logged to a .csv file. I have then been able to graph these figures and determine the number of users our environment can handle before action times begin to increase and delay is noticed by the users.

We have been running these scripts via logon on the Terminal Server's themselves and recorded the time stamp difference between an action and the WinWaitActive command for example.

This has worked extremely well over the LAN, however, a major drawback that I have now noticed is that we are not taking in to account any delay in RDP refresh rates when running over a saturated WAN link.

My initial plan to overcome this was to re-run the scripts locally, that is, from a PC on the other side of the WAN log on to a Terminal Server and re-run the same scripts and again monitor the time between the user actions and the WinWaitActive functions. Of course, this is not working for me as the WinWaitActive command is now listening / running on the local PC rather than the remote Terminal Server.

Does anyone have any experience or ideas in how we could measure the effect of application response times through RDP / Terminal Server over a WAN link?

As mentioned I already have completed tests running locally on the Terminal Server throguh RDP logon scripts with 200-300 users, however, as mentioned this doesn't take in to account for any delay between Server and Client.

Any advice would be much appreciated, and if anyone has questions about how we have completed load testing up until now, our our results I'd be more than happy to share my experience.

cheers,

tofe.