论文部分内容阅读
提出了一种启发式的多层客户-服务器排队网络模型的分析方法,适用于分析和预测运行在多处理器上或网络上的分布式软件系统的性能。首先提出多层客户-服务器排队网络的基本模型的概念,通过把基本模型分解为可以直接使用MVA方法求解的客户子模型和服务器子模型,提出一个迭代求解方法。然后将解决基本模型的方法扩展到多层客户-服务器排队网络模型。该方法具有收敛快、与具体的子模型求解算法无关等特点。
This paper proposes a heuristic analysis method for multi-tier client-server queuing network model, which is suitable for analyzing and predicting the performance of distributed software systems running on multiprocessors or networks. Firstly, the concept of the basic model of multi-layer client-server queuing network is proposed. An iterative solution method is proposed by decomposing the basic model into customer sub-models and server sub-models which can be directly solved by MVA method. Then the method of solving the basic model is extended to a multi-layer client-server queuing network model. The proposed method has the characteristics of fast convergence, independent of the specific sub-model algorithm.