A Trust Type Based Model for Managing QoS in Web Services Composition

Given the distributed, heterogeneous nature of Web services, the availability and reliability of Web services is uncertain. Moreover, service providers may have varying levels of quality. A way to understand and manage the quality attributes is needed. In this paper, we present an approach for the dynamic composition of Web services with quality of services (QoS) attributes. We adopt a concept of mediator to select credible services for service requestors. In our model, mediators evaluate services and assign trust ratings. A mediator includes a multi-dimensional quality model based on a set of non-functional properties that are transversal to all Web services, for example, their cost and reliability. In particular, we use the trust type inference rule to rate the trustworthy for selecting services optimally. Based on the QoS model, we present dynamic service selection approach to select and aggregate services to satisfy user's QoS requirements. This approach can be used to prune services that would be composed and thus avoid service composition execution plans that would be unacceptable to service requestors.