基于可信计算的We b服务调用关键路径研究

摘要

随着Web服务数目的增长,企业SAP(Systems Applications and Products in data processing)系统完成一次We b服务调用的响应时间开销越来越大,并且所调用的服务有部分是不可信的,严重束缚了服务请求的执行效率和调用质量.引入可信计算,即诚信度概念,结合工程学中的关键路径概念,创新性地提出一种基于诚信度算法的We b服务调用关键路径确定方法.针对服务注册中心提供的满足用户需求的We b服务,用诚信度算法对其进行计算,把诚信度值低于用户期望值的那部分We b服务排除在外;把可信We b服务调用过程抽象成用带权有向无环图表示的AOE(Activity On Edge)网图,使用关键路径确定算法,在图中找到一条服务调用关键路径.在关键路径上的所有服务都是关键服务,通过SAP系统客户端使用缓冲管理、业务代理等设计模式,减少关键服务的调用时间,来提高关键路径的工效,缩短服务调用响应周期,进而提高服务调用的整体执行效率和质量.通过实际系统开发,验证该算法的可行性和有效性.本系统已在某集团所属全部企业运行,降低了企业的运营成本,提高了工作效率,受到用户的满意度好评.%With the large growth in number of Web services and the demand for registries, there is a growing response time-spending on completing a Web service call of SAP(Systems Applications and Products in data processing)enterprise system, and some of these Web services are untrusted. It seriously hinders the efficiency and quality of service's request. This paper introduces the trusted computing, namely the concept of integrity. It combines critical path in engineering, pro-poses a determine algorithm of critical path of a Web service call based on integrity algorithm innovatively. For the Web services that meet the requester's demand registered in the UDDI(Universal Description, Discovery and Integration), it uses integrity algorithm to calculate them, excludes part of the Web services of which integrity values are below requester's expectations;it puts the process of a trusted Web service call abstracted into AOE(Activity On Edge)network diagram represented by weighted directed acyclic graph, finds a critical path of a Web service call in the graph by means of critical path determine algorithm. Services on the critical path are all critical services. By using the SAP system client buffer man-agement, business agents and other design patterns, it reduces the call time of critical services to improve the efficiency of the critical path, shortens service call response cycle, further improves the overall efficiency and quality of service calls. It develops the practical system to verify the feasibility and effectiveness of the algorithm. The system has been run in a group. It reduces enterprise's operating costs, improves work efficiency, and is praised by the user satisfaction.