The Web Service-based Distributed Application Deployment Strategies

The availability of PC servers and the improvement of their performance provide a good hardware environment for the building of large-scale commercial applications. Computing environment composed of PC server cluster adapts to the change and development much more easily, especially for those rapidly developing enterprises. The technology of distributed component plays an important role in this computing environment, and the software system by distributed component technology affords the system's scalability guarantee. In the distributed system consisting of multi-servers, the deployment of the components becomes an important factor to use the computing resources reasonably and to reduce the whole using costs.This paper discusses the fundamental knowledge of distributed component; analyzes and compares the popular component standards including CORBA, COM / DCOM, J2EE&Web Service and then analyzes and compares the load balancing strategies used in these component standards. Using these load balancing strategies as references, this paper discusses far more deeply the technology of Web Service which is better fit for the internet computing environment.The chain commercial management businesses are always geological distributed extensively. For the e-commercial application system used in such businesses, this paper analyzes deeply the software deployment strategies and the load balancing strategies in the building of large-scale distributed application system using the technology of Web Service; discusses the different deployment methods of the business middleware using Web Service interface and database service between different middleware servers and database servers and their influences on the whole performance; design the performance observation methods.On this basis, for two kinds of typical applications: large amounts of data relevancy retrieval and complex business logic computing, this paper proposes several different deployment strategies, then tests their performance using real data, provides test results and find out the load bottleneck by analyzing and comparing the test results.