A flexible and scalable framework for QoS-aware web services composition

Abstract

Daily increasing number of web services has introduced web service composition as a decision making problem. It is because; selecting web services to be composed needs to meet QoS requirements. Local selection approaches in meeting user's global requirements are impotent. In contrast, global optimization approaches are capable of supporting user's global requirements in a composite service. Increasing equivalent web services leads to a performance reduction of global techniques which is unacceptable particularly in dynamic and real time applications. In this paper, a flexible and scalable framework is introduced for composition. User request model is one of this framework's components. In this model, user is allowed to determine global criterions of composite web services and consider different quality factors for each task of the request. A new discovery service with ranking capability is proposed in this framework. The proposed discovery service allows the composer to control bandwidth usage, discovery time, composition time, number of candidate web services, required memory and quality of composite web service. Simulation results show the adaptability of our framework from time and quality point of view. Regulating these parameters by composer makes composer more flexible, scalable and adaptable.