Research On Service Composition Technologies In Multi-Service Systems

With the rapid development of the Internet and people's lives, all services providenew services to meet the new needs of users. From the point of view of the survey data,streaming media service is increasing year after year, and is predicted to occupy 90%of the Internet traffic in 2013. Multi business collaborative system itself has very goodopenness, scalability and compatibility, which makes the existing streaming media ser-vice and the new streaming media service receive very good deployment. Researcheson multi business collaborative system are of high academic value and practical appli-cations.This paper studies the factors affecting the performance of the business cooper-ative system, and builds a mathematical model for analysis, focusing on access con-trol technology and resources deployment technology, and studies service compositiontechnology, proposes corresponding strategies of optimization. First it constructs athree-layer service composition model, and this model decomposes the service com-position into three layer : Abstract Service Layer, Service Middle Layer and ServiceOverlay Network Layer. Service overlay network layer handles the actual physical nodein services, service middle layer processes the service module of business, and the ser-vice abstraction layer deals with user oriented services, which optimizes the structureof the system even much clear and simple. The candidate service pathes are selectedaccording to user the basic QoS requirements of users, and then linear programming al-gorithm is applied to locate service capacity for each path. While we adopt a dynamicthreshold method, triggering algorithms for automatically update to meet the actual de-mand changes of the users'requests in a video system . On the basis of consideringconsumers'demands, we build a Markov decision process model on the service com-position of multi business cooperative system. The optimization model uses iterationalgorithm, and simulation tests(including comparison with other algorithms) are madeto verify the algorithms. In the end system is designed as two layers: Virtual Layer,Physical Layer. In Physics Layer gives a kind of resource deployment algorithm isgiven for multi-business collaborative system , and this algorithm redeploys resourcesto cope with the changes for service requests, and the system resource demand changesdue to node joining and node exit. On Virtual Layer we will divide the self-independentnetwork based on the type of the business, and three-layer service composition model is adopted in this paper, the establishment of Markov decision process model helps tooptimize the service composition of single business services. This algorithm can effec-tively reduce computation complexity of global optimization algorithm, which greatlyimproves the efficiency and is highly effective.All the algorithms referred in this paper are verified by simulation tests, and somealgorithms have been deployed in the project of"next generation business cooperativesupporting environment". We implement the cooperative supporting optimization sys-tem. The approach solves the problem of searching service paths in the users'demandchanges and resource deployment adjustment, the performance of the system is greatlyincreased compared with other optimization algorithms.