Step 3 Perform your RAS To get a Baseline or to Test a Change

Step 3. Perform your RAS - To get a Baseline or to Test a Change

• Start the VLF application in your PC browser and move around in it. Use the STATS=Y option. Do not use developer=Y. It is important in applications that are used for long periods of time to “warm them up” before commencing a test. Delays that are experienced the first time that something is used when it is going to be used many times anyway, will just confuse your results and make them atypical of what a real end-users experience most of the time. This may lead you to wrong conclusions about what is the most significant improvement you can make.  
• Locate the L4WEB job on the server that is servicing your browser.
• Get positioned at step 1 of your RAS.
• Record the current CPU time for your L4Web job on the server.
• Perform the RAS exactly
• Record the perceived elapsed response time (PERT) for each step.
• Save the STATS=Y information in the VLF window.
• Check the CPU time of your L4Web on the server and record the delta (ie: how much CPU time was used on the server to complete execution of your RAS).
• Divide the CPU time used by the number of interactions in your script to get an average CPU time per server interaction (this is why not performing the script exactly each time will ruin your results).