The Study Of Composite Service Distributed Execution Optimizing

Composite service is an important technology to implement inter-organizationagility application system integration in open environment. The current applicationbased composite service commonly uses centralization execution pattern. Because ofthe limitation of centralization execution pattern system in the aspect of messagetransmission efficiency, enforcement efficiency, throughput of system and flexibility,the distributed execution pattern of composite service become the focus of discussionin current research.This paper has studied the distributed implementation optimization technique ofcomposite service. It has also proposed distributed implementation optimization ofcomposite service including two steps: partition and distributed deployment forcentralization modeling. In the step of partition, this paper proposes the operation flowdependency graph modeling of composite service. It splits data stream and controlstream into the subassembly service of composite service system. For the compositeservice system after partition, this paper has build a weighted directed acyclic graphmodel for composite service system and build a network map of correspondingnetwork topology model for distributed execution environment, convert the problem ofdistributed deployment to graph mapping problem which corresponds to runtime ofminimized composite service system, and get a heuristic mapping algorithm based oncritical path finally.In conclusion, this paper has verified this optimization algorithm of distributedcomposite service system by experiment, and shown that the distributedimplementation system is better than centralization in performance.