摘要
软件定义网络(SDN)控制和数据平面的分离以及全局可编程控制器的实现有助于网络入侵检测系统(NIDS)监控网络的整体安全性,在基于SDN的NIDS中已经实现了机器学习方法,然而机器学习方法可能带来高误报率。一种替代解决方案是使用深度学习。深度学习不仅能够自动发现数据中的相关性,还能有效地检测零日攻击。研究了基于SDN的网络入侵检测系统、用于网络入侵检测的数据集以及SDN入侵检测系统的深度学习方法,并比较了SDN中利用深度学习方法进行网络入侵检测的最新研究成果。
The separation of Software Defined Network(SDN) control and data planes and the implementation of global programmable controllers help Network Intrusion Detection Systems(NIDS)to monitor the overall security of the network.Machine learning methods have been implemented in SDN-based NIDS, however,machine learning methods may bring high false positive rate. An alternative solution is to use deep learning. To automatically discover correlations in data and effectively detect zero-day attacks, the SDN-based network intrusion detection system, the data set for network intrusion detection and the deep learning method of SDN intrusion detection system were investigated. The latest research results of network intrusion detection using deep learning method in SDN were compared.
作者
张阳玉
吕光宏
李鹏飞
ZHANG Yangyu;LYU Guanghong;LI Pengfei(College of Computer Science(College of Software Engineering),Sichuan University,Chengdu Sichuan 610065,China)
出处
《计算机应用》
CSCD
北大核心
2019年第S02期147-151,共5页
journal of Computer Applications
关键词
深度学习
入侵检测
软件定义网络
deep learning
intrusion detection
Software Defined Network(SDN)
