Research And Implementation Of A Highly Available Web Services Platform

The acceptance of web services in some Web based mission-critical systems posesan important problem on how to achieve their essential prerequisites such as highavailability. Current Web service technology still has serious lacks for providing thelevels of availability required by these mission-critical systems. Although there arespecifications dealing with reliable message exchange, availability has been largelyoverlooked.In this paper, we address this problem by proposing a novel platform named tosupport high performance and availability for Web services. The design of platformarchitecture is inspired by the infrastructure of prevalent cloud computing: it consists asingle service registry and multiple service nodes and is accessed by multiple clients.By dynamic service registration and supervision, the system can get the information ofService location and status in a timely and effective way to manage the Service nodeand service in the cluster. The platform provides high availability by replication strategybased on the service type, and state-restoration of Web services while running oninexpensive commodity hardware, and it delivers high aggregate performance to a largenumber of clients. Some empirical experiments have been conducted to prove theeffectiveness and efficiency of our platform.