DraLCD:一种新的数据中心流量工程方法

发布时间：2018-06-08 21:45

本文选题：流量均衡 + 软件定义网络(SDN)　；参考：《电子学报》2017年05期

【摘要】：流量均衡是为了避免网络拥塞而作为流量工程中的路由优化目标提出来的,由于数据中心网络的流量特性,使得传统IP网络的流量工程方法不一定适合.为此,本文在SDN(Software Defined Network)的框架下,提出了一种基于链路关键度的自适应负载均衡流量工程方法:DraLCD(Dynamic Routing Algorithm based on Link Critical Degree).该方法通过对全局视图的网络管控,并充分利用了网络中存在的冗余路径,在完成细粒度流量均衡的同时,能够降低控制器的计算开销以及与交换机之间的通信开销,最终完成路由优化的目标.最后,基于DraLCD设计的原型系统,通过在Mininet仿真平台中部署并进行仿真实验,与现有的等开销多路径路由算法ECMP(Equal-Cost MultiPath)以及GFF(Global First Fit)路由算法相比较,能够明显地提升网络性能.
[Abstract]:Traffic balancing is proposed as a route optimization goal in traffic engineering in order to avoid network congestion. Because of the traffic characteristics of data center network, the traditional IP network traffic engineering methods are not always suitable. In this paper, under the framework of SDN Software defined Network, an adaptive load-balancing traffic engineering method based on link criticality is proposed. By controlling the global view of the network and making full use of the redundant paths in the network, the method can reduce the computational overhead of the controller and the communication overhead between the controller and the switch while the fine-grained traffic equalization is completed. Finally, the goal of routing optimization is completed. Finally, the prototype system based on DraLCD is deployed and simulated on Mininet simulation platform. Compared with the existing equal-cost multipath routing algorithm ECMPI Equal-Cost multipath and GFFFN Global first Fit, the network performance can be improved obviously.
【作者单位】：清华大学网络科学与网络空间研究院;西安通信学院;广西民族大学信息科学与工程学院;
【基金】：国家973重点基础研究规划(No.2012CB315806) 国家863高技术研究发展计划(No.2015AA015601) 国家自然科学基金重点项目(No.61432009,No.61462009) 教育部博士学科点专项基金(No.20130002110058) 广西自然科学基金(No.2014GXNSFAA118358)
【分类号】：TP393.06

【相似文献】