There is a need to be able to run multiple, container level, simulations, in a (near) simultaneous manner. Each of the simulations will run using a different version of the same deployment unit (i.e. identical GAVs except for the version id), with the intent being to produce information that can be used to perform the following kinds of analysis:

1. A/B Comparison
2. Champion/Challenger

A/B Comparison

In this case, the results of the two simulations should be used to determine the differences resulting from the execution of rules and processes found in each version of the deployment unit. In particular, the analysis should look for differences in:

• Output - This includes differences in data produced by the rules/processes
• Working Memory/Process Variables - The internal data elements that exist during and after the execution of rules/processes
• Session state - The state information associated with the KIE session(s) and/or processes contained within session(s)

Champion/Challenger

For this post-simulation analysis, the relevant data is found in the performance timings that can be gathered during the simulation runs.