Specific situations in a production system, are instances of the jPDM model. They contain all their properties, characteristics, and specific bottlenecks.

Any change in one of the subsystems would result in a different situation with different properties and characteristics, thus in a different instance of the model. For example, changing the load on the system could result in a totally different bottleneck.

This is also the reason why performance tests in environments other than production will result in potentially different bottlenecks. The different environment has at least a different OS and hardware situation, not necessarily in the hardware and configuration being used, but in the whole condition of OS processes. Simulated scenarios such as performance tests therefore don't allow conclusions about bottlenecks or performance optimizations. They represent a different jPDM instance.

Since we use the model to ultimately analyze performance issues, the following approaches only make sense when there are actual performance issues. If there is no issue, that is, the defined SLAs are met, there is nothing to investigate or act upon.