SDN动态停滞超时时间优化算法  

Dynamic timeout optimization algorithm in SDN

作　　者：刘霞 杨桂芹[1] 邵军花[1] 蒋占军[1] LIU Xia;YANG Guiqin;SHAO Junhua;JIANG Zhanjun(School of Electronics and Information Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学电子与信息工程学院

出　　处：《传感器与微系统》2019年第10期118-121,共4页Transducer and Microsystem Technologies

基　　金：甘肃省自然科学基金资助项目(1610RJZA049);甘肃省高原交通信息及控制重点实验室开放课题项目(20161106)

摘　　要：针对当软件定义网络(SDN)数据量高峰时,SDN交换机有限的流表项资源会对网络性能有很大限制这一问题,提出了一种动态停滞超时时间优化(DTO)算法。先用动态指数平滑(DES)算法估计出下一时间周期内网络中的新增流表项数量,而后依据SDN中的实际负载状况,采用动态参数对流表项的停滞超时时间进行动态优化调整。经Mininet仿真证明:优化算法能有效提高流表匹配率、流表资源利用率以及网络平均吞吐量,提升整体网络性能。Aiming at the problem that at the peak of software defined network(SDN) data,the limited flow entry resources of SDN switches will have a great limitation on network performance,a dynamic timeout time optimization(DTO)algorithm is proposed. The algorithm estimates the number of new flow entry in the network in the next time period by dynamic exponential smoothing(DES)algorithm,at first.And then,dynamically optimize and adjust the stagnation time of the flow entry by dynamic parameters,according to the actual load status in the SDN.The simulation in Mininet shows that the optimization algorithm can effectively improve matching ratio of flow entry,the utilization of flow entry resources and the average network throughput, and improve the performance of the whole network.

关 键 词：软件定义网络(SDN) 动态指数平滑 流表项 停滞超时时间 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]