Research On QoS-aware Service Composition In Resource-constrained Environment

With the development of Web service, more and more services are published and used widely.There are a lot of functional similar services with but with different Quality of service (QoS). Inmost cases a single service is not enough to fulfil customers’ complex requirements, so servicecomposition are needed. Now the service consumers also focus on the QoS of the services. Thus asignificiant research problem in service compositions is how to select a compostion that the QoS ofthe each composite service is being maximized and the overall QoS of the composition is beingmaximized.Moreover, many services are deployed in resource-constrained environment. The composition isnot reliable and available by reason of the limited resource situation. By studying on Web servicetechnology and resource situation, a method that extends the WSDL with the information ofresource the service needs and the resource situation where the service deployed is proposed. Itmakes services deployed dynamically possible. The usability of service which depends on QoS andresource is proposed to evaluate the availability of service running in resource constrainedenvironment.The framework of QoS-aware Web service composition in resource constrained environment isproposed. Then this paper proposes a method to maximize the composition’s end to end QoS andrealizes it. The first step is filtering the services with low usability. And the second step is clusteringthe services based on QoS for each service task. Finally, the better candidate service of each clusteris to be selected and the improvd particle swarm optimization algorithm is introduced for the globaloptimization of composite service. Experiments show that the proposed algorithm can quicklyachieve the global optimal composite services with high quality of the composite service and withhigh success rate.