The Approach To Dynamic Selection And Composition For Running Environment Sensitive Information Services

Modern information technology represented by the computer and the Internet develops with an amazing speed. Information grows with an unprecedented speed. Service computing can resolve the problem of the use of information resources in the distributed network environment, and provide a high quality solution for meeting the needs of users. In the current Internet environment, there is a large amount of information service resources. Thus, the effective service selection and composition have become an important research topic.At present, the running environment of information service resources is increasingly complex and diverse. Service selection and composition are facing many challenges. On one hand, QoS will be changed dynamically in the dynamic and unpredictable environment, which will lead to degrade the QoS of the selected service seriously or fail to select the service. Thus, the user's experience will be degraded seriously. On the other hand, with the development of green computing, many people pay more and more attention to the energy consumption of computing system. The running environment is closely related with the energy comsuption of service. The running environment may cause the high energy consumption of service, which will lead to seriously waste the energy resource. Current researches focus on how to optimize QoS in service selection and composition, seldom consider the impact of running environments on service selection and composition. This paper proposes the approach to QoS modeling with the running environment, and evaluates the services in terms of QoS history records in the dynamic environment, then conducts dynamic service selection and multipath service composition with QoS optimization and energy optimization respectively, which involves the following four aspects.Firstly, running environment sensitive QoS model is proposed. This paper describes the running environment of the service, models the status of running environment, establishes running environment sensitive QoS model. Secondly, QoS history record based information service evaluation is proposed, this paper uses QoS history records, filters out some services based on optimization model and evaluate the left services in terms of Integer Programming model. The evaluation value can show the service's capability of meeting user's need, provide the basis for the service selection and composition.Thirdly, QoS aware dynamic service selection is proposed. Based on the two former contributions, aiming at QoS optimization, this paper proposes the model of QoS aware dynamic service selection, which includes two phases:generating candidate queues in design phase; dynamically select service in running phase. Fourthly, energy aware multipath information service composition is proposed. This paper uses the energy consumption model in connection with the load in the running environment. For the continuous arrival of mutiple requests, this paper proposes the model of multipath service composition with the optimization of energy consumption, and then proposes the optimization algorithm, which includes two parts:finding feasible paths for service composition; allocating the flows among the paths.This paper mainly considers the running environment affecting the service selection and composition, which proposes the approach to dynamic selection and composition for running environment sensitive information services. Experimental results of simulation and prototype system show the proposed approach can adapt to the dynamic running environments and resolve the problem of service selection and composition in dynamic running environment. This work also provides important reference values for building the efficient application systems based on information services.