SDN环境下基于CNN-BiLSTM的入侵检测研究  

Research on Intrusion Detection Based on CNN-BiLSTM in SDN Environment

作　　者：韩炎龙 翟亚红[1] HAN Yanlong;ZHAI Yahong(School of Electrical and Information Engineering,Hubei Institute of Automotive Industry,Shiyan Hubei 442002,China)

机构地区：[1]湖北汽车工业学院电气与信息工程学院,湖北十堰442002

出　　处：《佳木斯大学学报（自然科学版）》2024年第3期16-20,52,共6页Journal of Jiamusi University：Natural Science Edition

基　　金：湖北省教育厅科研计划重点项目(D20211802);湖北省科技厅重点研发计划项目(2022BEC008)。

摘　　要：软件定义网络(SDN)是一种将控制层和数据层分离的新型网络架构,在实现网络集中管理和可编程性的同时也面临易受到入侵攻击的问题。针对此问题设计了检测防御机制。利用深度学习算法,对数据集进行处理后,融合卷积神经网络(CNN)和双向长短期记忆网络(BiLSTM),设计了CNN-BiLSTM模型检测攻击,利用SDN可编程性设计了防御机制,搭建基于SDN的网络平台进行仿真实验。实验结果表明,所设计方法相较传统检测方法可更准确检测出入侵流量,并在检测出后有效实现了防御功能。Software Defined Network(SDN)is a new network architecture that separates the control layer from the data layer.While realizing centralized management and programmability of the network,it also faces the problem of vulnerability to intrusion attacks.A detection and defense mechanism is designed for this problem.After using the deep learning algorithm to process the data set,the CNN-BiLSTM model is designed to detect attacks by integrating the convolutional neural network(CNN)and the bidirectional long-term and short-term memory network(BiLSTM),and the defense mechanism is designed by using SDN programmability.A network platform based on SDN is built for simulation experiments.Experimental results show that the designed method can detect intrusion traffic more accurately than traditional detection methods and can effectively implement defense functions after detection.

关 键 词：软件定义网络 深度学习 卷积神经网络 双向长短期记忆网络 入侵检测 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]