Composite Web Services Selection, Deployment And Implementation Of Key Technologies

Service-Oriented Architecture (SOA), as a key architecture to realize the model of Service-Oriented Computing (SOC), provides technical support for distributed and heterogeneous resources sharing and interoperation. Many research works on Web Services have been done in both industry and academy, which is one of the most essential standards to implement SOA, and consequently lead to the rapid increase of Web services over the Internet. On the bright contrast to the limitation of single Web services, diversity of user requirements emerges, which eventually results in the high development of Web service composition technology and wide use of composite services.Recent research on composite services rarely focuses on service behavior modeling, composite service dynamic selection and deployment optimization. Considering the issues above, this paper conduct the following research works:Issues of dynamic selection on composite service are studied, as well as analysis of relations between time and space features and quality of service. Composition distributed invoking framework is designed and quality information history database (HDB) is created to store its time and space features. A prediction algorithm is presented to predict quality data at specific time and space by using HDB. With this method, users can invoke atomic services of composite service at different time and space and eventually obtain better results.Composition deployment optimization is explored, for which a management framework is proposed. Algorithms are designed for initialization, optimization, and backtracking of user's service deployment requirements under this framework, and deployment results can be verified in advance. This framework can reduce time and space resource cost during Web service deployment, and guarantee its usability and safety.To study Web service behavior features, a set of NORM rules are defined to describe Web service behavior. Web service model is proposed to describe Web service and constrain the complex interactions among service behaviors within services, which can help reduce the difficulty of service dynamic discovery and composition later. The transformation from composition plan to BPEL is studied, and an optimization approach to this plan is presented. With the BPELEngine parallel and nested processing on optimized composition plan, the execution efficiency of composite Web service can be improved.Current status of campus energy consumption is analyzed and Web application features in campus energy efficiency supervision are elaborated. Combining practical examples, Fudan Campus Energy Efficiency Supervision Platform is developed by applying the above research results. The data and services collected by this platform provide experiment and verification environment for composition applied research.