Service Evolution Management Based on Usage Profile

Services have been increasingly used as the building blocks for decoupled and flexible applications. Service evolution is a critical issue because even small changes, if not compatible, can potentially affect a huge number of client applications. However, particularly in the context of large scale usage of a service, changes cause different impact on client applications according to its use. This paper proposes to focus on compatibility from the point of view of usage patterns in order to deal with service evolution issues in more flexible and less costly way. The idea is to summarize the behavior of client applications into usage profiles, from which metrics that represent the impact of changes can be derived. This valuable information may support service providers on decisions about service lifecycle. The paper discusses the adoption of usage profiles and presents a framework for the automatic evaluation of service changes impact during its lifecycle.