A Heuristic Solution For Multilayer Client-Server Queuing Network Model

摘要：

A new heuristic solution for multilayer client-server queuing network that is called NLQN is proposed to predict performance of distributed software system based on existed queuing network analysis technique. The main concept of the method is called the basic model that can be decomposed into two sub-models: client sub-model and server sub-model. These two submodels can be solved using traditional queuing network analysis method such as MVA algorithm. The client sub-model calculates think time for server sub-model, and the server sub-model calculates waiting time for client sub-model. The overall model can be solved by iteration from one submodel to the next sub-model. Then the basic model is extended to multiplayer client-server queuing network. This solution has advantages of fast convergence, being dependent of specific MVA and eliminating interlocking effect existed in multilayer clientserver queuing network. The error range of the method is from 5％ to 15％ according many tests.

关键词： Client-server network layered queuing network software performance model

作者: Dengshi Li Bin Yan Xi Guo

作者单位: School of Mathematics and Computer Science, Jianghan University. Wuhan, China 430056 School of Computer, Wuhan University. Wuhan, China 430072

会议类型: 国际会议

会议名称: 2010 International Conference on Measurement and Control Engineering(2010年IEEE测量与控制工程国际会议 ICMCE2010)

会议地点: 成都

会议语种:英文

页码: 688-692

在线出版日期: 2010-11-16（万方平台首次上网日期，不代表论文的发表时间）