A Distributed Service Invocation Approach for Cross-Organizational Data-Intensive Web Services

The paper presents a distributed service invocation approach for cross-organizational data-intensive web services. Our approach applies one service proxy in each organizational domain, which is responsible for service scheduling and invocation. Service Invocation across different domain is realized by the coordination of service proxies. Three performance improvement methods, which use compression, cache with consistency and two-level cache technologies, are discussed with focus. The most suitable one of them will be chosen according to different service types and invocation metadata. The structure design and implementation of service proxy is also presented succinctly. Experimental evaluations demonstrate that our approach combined with these three methods gives significant performance improvements for service invocation.