摘要
In order to assure short time-to-market and high quality of new services, effective and efficient pre-deployment service interaction handling is necessary. Service creators and service providers need mechanisms for assessing the impact of newly introduced services on the existing ones. In particular, efficient methods for detecting undesired interactions are required. To satisfy these needs, the EURESCOM1 Project P509, Handling Service Interactions in the Service Life Cycle created the concept of interaction filtering which is a step up towards interaction detection. Filtering identifies service and feature combinations which are interaction-prone and to eliminate those which are unlikely to cause any interactions, which simplifies the actual interaction detection. Filtering uses a number of complementary methods which analyse different aspects of services. The paper presents and evaluates one of such methods called goal-based filtering. The method identifies interaction-prone combinations by analysing the relations between service properties and user's goals.