论文部分内容阅读
该文给出了一个新的Web服务系统的模型，将服务器和通信访问网络考虑为一个整体，构成一个开放的马尔可夫排队网络模型，并将模型扩展到了多服务器系统；然后分析各模型参数对服务系统性能的影响，得到了一些有益的结论，提出了避免死锁的方法，证明了增加新服务器的速率原则，最后对比了几种提高服务器性能的方法在不同拥塞条件下的优劣。仿真结果证实了模型是有效的。该模型以及参数分析的一些结论有益于指导网络管理、规划设计和升级。
In this paper, a new model of Web service system is proposed, which considers the server and communication access network as a whole, forms an open Markov queuing network model and extends the model to a multi-server system. Then, the model parameters Some useful conclusions have been obtained on the performance of service system. Some methods to avoid deadlock have been put forward. The principle of increasing the rate of new server has been proved. Finally, several methods to improve server performance have been compared under different congestion conditions. The simulation results confirm that the model is effective. Some conclusions of this model and parametric analysis are useful in guiding network management, planning and design and upgrading.