A weakest link-driven global QoS adjustment approach for optimizing the execution of a composite web service

摘要

In service-oriented computing, service composition pertains to developing solutions quickly to satisfy new business requirements. Service composition provides flexible integration across distributed and heterogeneous web services to create a composite web service (CWS). However, global quality of service (QoS) of a CWS is dependent on the QoS of multiple component (or partner) web services that are integrated to form the CWS. In this context, an important research challenge is how to optimize the QoS of a CWS by detecting and adapting to runtime anomalies in component web services. A vital research issue is how to provide optimized CWS by effectively monitoring and identifying the weakest links (i.e., component web services with worst performance metrics). In this paper, a novel global QoS adjustment approach based on sensitivity analysis is proposed for detecting a component web service acting as a weakest link or a bottleneck in a CWS. This is demonstrated with an experiment, and an evaluation is performed for representing the effectiveness and performance of the proposed approach.