Detecting Performance Antipatterns in Component Based Enterprise Systems

ions. For example one method may check the balance of an account while another may approve an order, and yet another may apply a payment to an invoice. Such sessions can create memory and latency issues since their creation and management can come with much overhead due to their bulky nature. The PAD tool identified two potential bloated session beans in the IBM Workplace application. The first potential instance of this antipattern was a session bean which had 8 entity bean relationships, 6 session bean relationships and was invoked in 11 of the 25 user actions executed. It also contained a high number of business methods (47). The second instance of this antipattern was a session bean which had 7 entity relationships 1 session relationship and was invoked in 6 of the 25 user actions. This session had 14 business methods. In general the rule of thumb is that there should be a 1 to 1 relationship between session and entity beans [5]. From the information (above), presented by the PAD tool upon identification of these potential antipattern instances, it seemed that these session beans were in fact bloated session beans. Unfortunately we were unable to contact the developers originally responsible for this code. However we did discover that this code was removed from later releases of the application which indicated that it was indeed problematic. No false positives were identified when we applied the PAD tool to the IBM workplace application and in fact all antipatterns identified were addressed in the 82 JOURNAL OF OBJECT TECHNOLOGY VOL 7, NO. 3