互联网能耗和性能优化的权衡模型  

Trade-off model for Internet power consumption and performance optimization

作　　者：周三奇[1] 陈佳[1] 张宏科[1] 

机构地区：[1]北京交通大学电子信息工程学院下一代互联网互联设备国家工程实验室,北京100044

出　　处：《系统工程理论与实践》2014年第8期2106-2112,共7页Systems Engineering-Theory & Practice

基　　金：国家重点基础研究发展计划"973"(2013CB329100);国家自然科学基金(60833002)

摘　　要：互联网能耗和性能互为矛盾已经成为网络研究人员的广泛共识.目前的研究方法通常仅将其中一个方面作为主要问题,而将另外一方面作为次要问题来解决.因此,如何将互联网能耗和性能作为同等重要的指标进行协同优化,是本文所要研究的问题.本文提出了一种通用优化模型来实现减小互联网能耗的同时提升网络性能.为将其应用于实际,本文通过引入并定义代价函数和功率-带宽函数,将该通用模型具体化为一个线性规划模型.评估结果显示,该线性规划模型的最优解能够同时降低网络的最大链路利用率和能耗,并且使得这两者中的每一个指标均近似于对其进行单一优化的模型的最优解,从而实现了对网络能耗和性能的协同优化.与当前互联网相比,在最大链路利用率方面,本文所提出模型的最优解最多可将其减小60.6%,最少减小45.8%;在能耗方面,该模型的最优解最多可将其减小83.3%,最少减小56.7%.It has been widely accepted in network research community that there is a trade-off between Internet power consumption and performance. Some approaches were proposed to take only one of them as primary while considering the other one as secondary. Hence, it is the problem to be investigated in this paper which is how to take the Internet power consumption as same importance as performance to optimize in a collaborative way. In this paper, we propose a general optimization model to reduce the Internet power consumption while improving the network performance. To make it practical, we materialize it to a linear model by introducing and defining cost functions and power-bandwidth function. The evaluation results show that, the optimal solution of the linear programming model simultaneously reduces the network maximum link utilization and power consumption. Moreover, each of the two indices is respectively and approximately equal to the corresponding one-side optimal solution. Thus, the network power consumption and performance are collaboratively optimized. Comparing to the existing Internet, for maximum link utilization, the highest reduction of the optimal solution of the proposed model is 60.6% and the lowest is 45.8%; for power consumption, the highest reduction of the optimal solution of the model is 83.3% and the lowest is 56.7%.

关 键 词：能量效率 负载均衡 多物流 流量工程 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]