Study On Requirement Group-based Web Service Scheduling Model

With the rapid development of Internet technologies, many enterprises publish their businesses on the Internet to provide more convenient services for their customers, and to get better user interactions. Web Service, which is self-described, platform independent, cross-language and loose coupling, is applied by more and more companies for its good characteristics. As the technology of Web Service is developing rapidly, the amount of applications based on Web Service is increasing, resulting in a fast growing of the amount of Web Service requests on the Internet. Too many concurrent requests pose a big challenge for the load capability and performance of Web Service. Therefore, how to decrease the load, and how to improve the performance of Web Service in the environment of massive Web Service requests is a problem that must be solved as soon as possible.In this thesis, we propose a requirement group-based Web Service scheduling model (RGBWSSM) on the basis of the fact that many concurrent Web Service requests are similar with each other. The model constructs a requirement group for concurrent similar requests, and searches Web Services that can provide service for this requirement group based on requirement constraints and Web Service capabilities. By the method that solving many concurrent similar requests in a united way, the load of Web Service is decreased, and the performance is improved. And at the same time, the model constructs a fault-tolerant service group for the function-replaceable Web Services to enhance the availability of Web Service in the environment of massive Web Service requests.The main research and innovative work mainly includes following points:First, according to the specific condition in the environment of massive Web Service requests, we design a framework of a requirement group-based Web Service scheduling model. The framework provides middleware supports for service requesters and service providers. It implements the transparent interaction between service requesters and service providers, and improves the performance of Web Services in the environment of massive Web Service requests.Second, we propose a requirement group model and a detail method of how to construct a requirement group. We use ontology to build a semantic-described requirement model, and construct a requirement group, which can meet all the needs of its member requirements, through the comparison of ontology concept of concurrent requirements'elements based on a precondition that the operation type is the same and the constraints are compatible. An algorithm about how to construct a requirement group is brought forward.Third, after building a service ontology model, we propose a service matching algorithm based on requirement constraints to find Web Services, and construct a service group for function-replaceable Web Services. When a service is failed during invocation, the service group may replace a valid Web Service with this failed Web Service. The good fault-tolerant capability of service group can satisfy the need of high availability of Web Service in the environment of massive Web Service requests.Forth, in order to validate the feasibility and effectiveness of the model, a prototype system is developed. By executing a typical example in economic area and analyzing the results, we mainly validate that the model is useful in decreasing the load and average response time of Web Service. The results prove that this system is of considerable referential importance in improving performance of Web Service in the environment of massive Web Service requests.