Impact Analysis improved data caching

In the February 2008 PTFs for WebSphere Studio Asset Analysis V5.1, there is an Impact Analysis (IA) implementation improvement that substantially reduces the amount of CPU usage and JVM heap turn over . It also improves recovery of JVM memory in Java garbage collection. Improvements are more apparent for large impact analyses.

Without this improvement:

• Memory leaks were causing memory usage to increase by 20-40% with each run of a large IA.
• Garbage collection ran more and more frequently with each run of a large IA.
• A large IA took longer each time it ran.

With this improvement:

• Memory usage remains consistently reduced for each run of a large IA.
• The recurrence of Java garbage collection is greatly reduced.
• Each consecutive run of a large IA takes the same amount of time. (The first run may still take more time than subsequent runs, as it still requires some database query setup that is then reused by the following runs).