多指标综合评价的负载均衡路由策略研究

发布时间：2018-02-17 08:53

本文关键词： OpenFlow Fat-tree 前K条最短路径负载均衡多指标综合评价　出处：《小型微型计算机系统》2017年02期 　论文类型：期刊论文

【摘要】：如何在规模庞大的数据中心网络内部进行负载均衡,实现网络资源的优化利用,是个亟待解决的问题.OpenFlow技术提供了一种可编程操作的流量控制方法,已成为近期解决负载均衡问题的热门关键技术.针对数据中心常见的Fat-tree网络拓扑,就数据流在OpenFlow网络中的路径设计问题,利用OpenFlow网络控制器集中控制的优势,在控制器上设计并实现了一种动态负载均衡路由策略.该策略首先选出源节点与目的节点间的前K条最短路径,然后根据获取的网络链路实时状态信息,对选出的前K条路径进行多指标综合评价,并从中选出数据流的最优转发路径.最后,采用Floodlight控制器和Mininet仿真平台,对提出的负载均衡路由策略进行可行性验证和性能验证.结果表明,该策略有效缓解了网络阻塞,提高了链路利用率.
[Abstract]:How to balance the load inside the large data center network and realize the optimal utilization of network resources is a problem to be solved urgently. OpenFlow technology provides a programmable flow control method. It has become a hot key technology to solve the problem of load balancing in the near future. Aiming at the common Fat-tree network topology in data center, this paper makes use of the advantage of centralized control of OpenFlow network controller to design the path of data flow in OpenFlow network. A dynamic load balancing routing strategy is designed and implemented on the controller, which first selects the first K shortest path between the source node and the destination node, and then according to the obtained real-time state information of the network link, the first K path between the source node and the destination node is selected. The first K path is evaluated synthetically and the optimal forwarding path is selected. Finally, the Floodlight controller and Mininet simulation platform are used. The feasibility and performance of the proposed load-balancing routing strategy are verified. The results show that the proposed strategy can effectively alleviate the network congestion and improve the link utilization.
【作者单位】：南京邮电大学计算机学院软件学院;
【基金】：国家自然科学基金项目(61472192)资助
【分类号】：TP393.06

【相似文献】