A Flexible Approach for Assessing Service Compatibility at Element Level

Service evolution requires sound strategies to appropriately manage versions resulting from changes during service lifecycle. Compatibility addresses the graceful evolution of services by considering the effects of changes on client applications. Traditionally, compatibility assessment is restricted to the worst-case scenario, because the analysis is based on the documents that describe service properties (e.g. WSDL and XSD files). However, such compatibility may not be representative of the real impact on clients, because a non-compatible change may affect a not used element, such as the type of an optional parameter or an unused operation. This paper addresses service evolution on a finer grain fashion. We propose a versioning model and a compatibility assessment at element level in order to identify the impact points, propagation effects and categorization of changes on service elements during evolution. This approach provides important information for supporting service evolution by either maximizing version reusability (e.g redirecting of requests, overload balancing, etc.) and/or pinpointing the change impact points that basis approaches such as usage based analysis.