可重构路由器采用构件化的体系结构,使以构件为对象的能耗细粒度调整成为可能.论文面向路由器低能耗问题,基于部件能耗随业务负载变化的思想,在对构件化路由器低能耗模型分析的基础上,提出了基于构件重构的路由器能耗细粒度调整方法.该方法使用布尔二次指数平滑法预测网络负载变化,能够有效避免重构结果滞后于负载变化;使用遗传算法并行搜索构件配置空间,能够快速搜索出最优构件配置方案.使用真实网络流量进行实验,结果表明,该方法能够根据负载的动态变化进行构件重构,能够使路由器能耗跟随负载变化,与传统算法相比可以节省60%的算法运行时间,降低构件组25%能耗.
Reconfiguration routers have an architecture made up of components, which make the energy meticulous-grained scaling to be real. In this paper, we propose an energy meticulous- grained scaling algorithm to deal with the routers' energy problem, which is based on the idea that components energy can change with application load variety. It first predicts the future workloads of the applications with Brown's quadratic exponential smoothing method to make reconfiguration catch up with loads. Next, it adopts a genetic algorithm to parallel find the opti- mal reconfiguration policy. The real network traffic is used to check the algorithm. Experimental results demonstrate the approach can adapt the router's energy according to the change of net- work traffic, reduce the algorithm computing time by 60% and the energy consumption by 25%.