Achieving adaptivity through strategies in a distributed software architecture

Designing information systems which are able to modify their structure and behavior at runtime is a challenging task. This is due to various reasons mostly related to questions such as what should be changed, when should be changed, and how should be changed at runtime in order to maintain the functionalities of a system and, in the slime time, to personalize these functionalities to the current user, services requests and situations, as well as to improve its performances. The systems which manage to address properly these aspects are considered adaptive. Our approach to design adaptive systems exploits strategies to implement the decisional support and to ensure an efficient modularity, reusability and evolvability of the architectural model. In this paper we describe the main types of the strategies defined in our solution, as well as how these strategies are exploited at run-time in the context of an actual case study in the financial domain.