| With the service computing and cloud computing technology developing, more and more products have been provided as services to serve the society through networks. In such a situation, how to effectively select the right services to build effective solution, and how to utilize service computing theory flexibly in practice to more efficiently operate and run the information flow, capital flow and logistics of traditional industries, are two serious problems. In the paper, we are focus on the following problems of service composition and application in business evironment.(1) Tranditional service composition model can not effectively reflect the group influence. For users in different group, the same QoS values and user preference may produce different results. But the group influence can not be reflected in tranditional QoS model. In tranditional QoS model, it’s resolved by only QoS value and user perference, group influence can not be reached.(2) Tranditional service composition model can not effectively reflect supplier strategy. From the point of view of the service supplier, they maybe add some strategy while providing services. These strategies will inevitably affect the selection result. Supplier strategies can be classified as constraint strategy and incentive strategy. They can not be described effectively through utility function and global constraints of tranditional service composition model. So tranditional service composition model should be extended to support supplier strategy.(3) The WSLA protocol is lack of coordination part to result low efficiency in the service composition. When a large number of services are composed and interacted, the cost and latency increase fast, In order to improve the service executive efficiency and reduce the service cost, service customers and service provders both needs to express coordination in service composition. But there is no coordination part in WSLA, so coordination can not be exactly described in service level agreement of service composition.For the above problems, this paper gives out some new ideas and methods. It introduces group influence utility model, service supplier strategy model and Co-WSLA collaborative strategy model. It expands the traditional system architecture of services computing, and then applies new models into healthy management, medical insurance and other traditional areas to improve the operational efficiency and service quality of traditional business models.(1) We propose a group influence based service composition model. Through group influence utility function, it solves the problem that traditional service compostion model can’t effectively reflect the group influence.(2) We propose a supplier strategy based service composition model. First, we analyse the service supplier constraint strategy and incentive strategy while providing services. Then we introduce a rule based method to extend traditional service composition model to describe strategies. And then we add a rule preprocess algorithm before traditional model to realize the support of supplier strategies.(3) We propose a Co-WSLA collaborative strategy based service composition model, In Co-WSLA protocol, we introduce a new Party type called Collaborative Party to extend WSLA protocol. This party is used to describe the collaborative agreement in protocol. Co-WSLA is suitable for the service composition environment which service do not ask for much of the real-time but the number of interactions are large. Through batch accumulated and lazy execution, the executing efficiency of composed services are much improved. At last, we apply the model to healthy insurance area. Through the new model, efficiency, quality and cost are much improved. |
PDF Full Text Request