绿色主干网络中一种高效的节能路由算法

An Energy-Efficient Routing Algorithm in Green Networks

随着互联网规模不断扩大,当前的网络为了能够支持最大的用户访问量并且能够同时保证服务质量,出现了过度供给的情况,正是这种过度供给的现象导致能耗问题日益严重.因此,以降低能耗为目标的绿色互联网的概念被提出,ICT部门也积极投身于相关研究工作当中.该文基于流量整合的思想,提出了一种高效的节能路由算法,具有节能效率高、节能效果好且不会使优化目标陷入局部最优的特点.该文首先对绿色互联网国内外的研究现状进行了综述和分析,通过对当前节能算法的分析,该文发现目前的一些节能算法容易陷入局部最优解,另外一些算法在寻路的过程中花费了大量时间.针对以上两个问题,该文基于捆绑链路设计了网络模型、功耗参数和数学模型,设计并实现了绿色互联网中的一种高效的节能路由算法.最后该文使用了CERNET2的网络拓扑和实际流量对算法进行了仿真实验和性能评价,并且与最短单一路径优先算法、快速贪心启发式算法进行了对比.实验结果表明,该文设计的算法在真实流量需求下最多可以节省整个网络56.3%的能耗,在节能效果方面相比快速贪心启发式算法有较大幅度的提升,与最短单一路径优先算法相比也有一定程度的提升,在运行时间和路由请求满足率方面也均明显优于快速贪心启发式算法和最短单一路径优先算法.

In recent years,with the rapid development of internet technologies,the increasing demand of customers on the network causes the expanded scale of internet.Current networks are typically over-provisioned to support the maximum number of customers whilst meeting their Quality of Service.This redundant design,however,has led to greenhouse gas emissions,and the increasingly serious problem of energy consumption.Therefore,the concept of the Green Networks is proposed to save energy consumption.Currently,the study on Green Networks has been the focus of domestic and international attention.This paper firstly analyzes the research prospects of the Green Networks.We focus our problem on the routing algorithm in backbone based on bundle links.According to our scenarios,network model,parameters of power consumption and the mathematical model are designed.Through the analysis of the current algorithms,we find that some of current algorithms are easy to trap in local optimum which affects the energy conservation results while others cost lots of time on finding paths.Secondly,an energy-efficient routing algorithm is proposed.The algorithm considers the problem of shutting down a subset of bundled link during the off-peak period,and make part of the ports,line cards and chassis to sleep.The main idea of the algorithm is to route each requirement in order.In the process of finding path,we try to choose the occupied path to integrate the flow.Thirdly,the algorithm is improved by the routing strategy.The improved algorithm is fast while avoiding the shortcoming that the greedy algorithm is easy to trap in local optimum.After that,we analyze the time complexity of algorithm.It shows that the proposed algorithm has low time complexity,and can also be implemented easily.Finally,we evaluate our algorithm based on the CERNET2 network topology and the actual network traffic.The results show that the proposed algorithm can reduce the energy consumption of the whole network devices on CERNET2 topology by up to 56.3%.In order to better evaluate our algorithm,we choose the Shortest Single Path First algorithm and Fast Greedy Heuristic algorithm as comparison algorithms.Shortest Single Path First algorithm is a heuristic algorithm based on link bundling and the main idea is to close the cables in network iteratively until all the cable can not be closed.Fast Greedy Heuristic algorithm is a an excellent heuristic algorithm based on greedy algorithm.Compared with the Shortest Single Path First algorithm and Fast Greedy Heuristic algorithm,the proposed algorithm has better energy saving effect.Furthermore,it is better than the Shortest Single Path First algorithm in terms of the running time and the routing success rate.