面向网络系统性能优化的资源最优配置方法

Optimal resources allocation method for network performance optimization

如何根据网络系统的实际情况优化资源配置,充分利用有限的资源保证系统的服务质量(QoS),是网络系统性能优化研究中的一个核心问题。针对该问题,提出了一种基于PEPA nets的网络资源配置建模与分析方法。该方法首先根据网络资源配置方案的一般情形,建立了完全共享和部分共享两种资源配置方案的通用PEPA nets模型;然后,利用PEPA nets操作语义推导出了一个与模型相关的、带时间延迟标记的标记变迁系统;利用该标记变迁系统隐含对应的CTMC求解出了系统状态空间的稳态概率分布,进而由此分析得到系统性能指标,从而选出最优的资源配置方案。最后,通过试验分析验证了该方法的可行性和有效性。

The core of the problem of the network performance optimization is how to optimally allo- cate the resources of the network system and make full use of the limited resources to guarantee the QoS of the system. An optimal resources allocation method was proposed based on PEPA nets. Ac- cording to general schemes for resources allocation, the PEPA nets models were established for the two schemes of fully-shared and partially-shared resources allocation. A labelled transition system （LTS） with time delay could be generated according to operational semantic information for PEPA nets. CTMC corresponding to LTC was used to solve for the model＇s steady state probability distribution, acquire the performance index of the system and therefore choose the optimal resources allocation scheme. Finally, the results of experiments verify the feasibility and effectiveness of the method.