SDN中基于超时机制改善的QoE研究  

Research on QoE based on Timeout Mechanism Improvement in SDN

作　　者：施达雅 余庚[1,2] SHI Da-ya;YU Geng(College of Engineering,Fuzhou Institute of Technology,Fuzhou 350506,China;School of Guomai Information,Fujian University of Technology,Fuzhou 350014,China)

机构地区：[1]福州理工学院工学院,福州350506 [2]福建工程学院国脉信息学院,福州350014

出　　处：《光通信研究》2018年第4期10-13,共4页Study on Optical Communications

基　　金：福建省教育厅中青年教师教育科研基金资助项目(JAT170802)

摘　　要：随着大数据在软件定义网络(SDN)上的融合应用,网络因无法实时更新流表而导致数据流失效等一系列用户体验(QoE)问题逐渐呈现。为此,文章构思一种在具有突发特性的负载环境下基于停滞超时优化处理的算法,用于提升OpenFlow交换机流表资源可用性。该算法通过动态均衡流表资源和控制资源以及合理设置停滞超时来使数据流与流表项顺利适配,进而提高数据流转发效率。同时,算法将那些与数据流成功适配的流表项封存在流表中,有效地提升了流表在时间上的生存性。实验表明,算法中的停滞超时改善机制在QoE方面较固定停滞超时机制有明显的优化效果。With the integration of big data in the Software Defined Network （SDN） network applications, a series of Quality of Equipment （QoE）problems are emerging gradually, such as data flow failure which is caused by failure to update the flow table in real time. In order to solve this issue, this paper proposes an algorithm based on stagnation timeoutoptimization to enhance the availability of OpenFlow switch table resource in aburstload environment. The algorithm can make the data stream and the flow item match smoothly, and then improve the efficiency of data stream forwarding through dynamic balance table flow re source, control resources, and reasonable settingtimeout. At the same time, the algorithm seals the flow table entries that are successfully matched with the data flow in the stream table, and effectively improve the survivability of the flow table in time. Experimental results show that the stopping timeout improvement mechanism in the algorithm is more effective than the fixed timeout mechanism in terms of QoE.

关 键 词：软件定义网络 用户体验 资源 超时 数据流 

分 类 号：TN915[电子电信—通信与信息系统]