基于链路监控的SDN恶意流量检测与防御  

SDN Malicious Traffic Detection and Defense Based on Link Monitoring

作　　者：赵新辉 张文镔[1] 王清贤 武泽慧[1] ZHAO Xinhui;ZHANG Wenbin;WANG Qingxian;WU Zehui(Information Engineering University, Zhengzhou 450001,China;School of Sports, Zhengzhou University, Zhengzhou 450001, China)

机构地区：[1]信息工程大学,河南郑州450001 [2]郑州大学体育学院,河南郑州450001

出　　处：《信息工程大学学报》2020年第1期61-67,共7页Journal of Information Engineering University

基　　金：国家重点研发计划资助项目(2019QY0501);河南省软科学研发计划资助项目(182400410108,192102210128)。

摘　　要：软件定义网络(Software Defined Network,SDN)是一种新的计算机网络架构,能够适应网络规模不断增长的趋势,已得到广泛部署。但SDN独特的架构引入了新的攻击面,攻击者可以针对不同的脆弱点进行攻击。基于网络监控的思想,提出一种SDN网络恶意流量检测和防御方法,通过监控链路流量并使用改进的卷积神经网络模型进行恶意流量识别,提高了恶意流量检测准确性和效率,可以检测网络中的恶意流量并进行防御。实验测试结果表明,文章的方法对恶意流量的检测准确率达到95%以上,并能进行有效防御,可以有效提升SDN安全性。SDN(software defined network)is a new computer network architecture,which can adapt to the growing trend of network scale and has been widely deployed.At the same time,more and more attention has been paid to the security of SDN.The unique architecture of SDN introduces a new attack surface,and attackers can attack different vulnerable points.Based on the idea of network monitoring,this paper proposes a detection and defense method for malicious traffic in SDN network.By monitoring the link traffic and using the improved convolutional neural network model to identify malicious traffic,the accuracy and efficiency of malicious traffic detection are improved,and malicious traffic in the network can be detected and defended.The experimental results show that the detection accuracy of this method for malicious traffic is more than 95%,and it can effectively defend and improve the security of SDN.

关 键 词：链路监控 软件定义的网络(SDN) 恶意流量 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]