面向云计算的服务性能模型研究

为了衡量云计算的服务性能。分析了云计算服务的处理过程，提出了一个以排队论为基础的云计算服务性能模型。以Amazon SimpleDB的更新延迟为例，仿真验证了该模型，并以此为基础，分析比较了模型中的关键因子。实验表明，提升服务装置的性能可以有效提升云计算服务的性能，且在云计算部署后期，虚拟机实例的性能是服务性能提升的瓶颈。

Research on a service performance model orienting cloud computing

In order to evaluate the performance of cloud service, this paper analyzed the process of cloud service and presented a mathematical model of service performance base on queuing theory. Take the update latency of Amazon SimpleDB as an example, simulation results demonstrated the validity of this method. On this basis, the key factors were analyzed and compared in this paper. Experiment results showed that improving service processing unit could improve efficiently the performance of cloud service, and the performance of virtual machine could be the bottleneck in the later period of cloud computing deployment.