Employing aspect composition in adaptive software systems: a case study

Adaptive software is a closed-loop system which aims at adjusting itself at runtime in different situations. Such a system needs a set of sensors to monitor attributes of itself and its operating environment. Furthermore, it requires a set of effectors in order to make changes in its entities. These changes are essential for fulfilling system's non-functional and functional requirements. Aspect-Oriented Programming (AOP) is a promising way to develop these sensors and effectors through static and dynamic composition of advices. This paper presents the experience of employing aspect composition in engineering a sample adaptive software. The main objectives are exploring the difficulties of utilizing this approach, and investigating the effectiveness of aspect-based adaptation actions. A J2EE bookstore application, TPC-W, was selected as the case study, to instrument sensors by the aid of static aspects, and effectors using dynamic aspects. The findings are promising, and encourage us to continue this line of research for more complex systems.