Deep Learning Approach for SDN-based Intrusion Detection and Prevention System

• Main Authors: LIN, JHIH-REN, 林志仁
• Other Authors: LEE, TSUNG-HAN
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/td4888

碩士 === 國立臺中教育大學 === 資訊工程學系 === 106 === In recent years, Software Defined Network (SDN) has been widely used in cloud computing and will be adopted in 5G. In the past, when the traditional network needs to change settings, it is necessary to modify the network deviece individually and quite time-consuming. SDN separate the control plane and the data plane in the network and it uses a centralized management method to mange network devices. Through the SDN Controller, network administrator can easily managent network and settings network devices. In the SDN, OpenFlow Controller focuses on the operation of the network. OpenFlow Controller used the OpenFlow Protocol to define flow table, and decision packets how to forward. The detection of network attacks is less of a concern. In this paper, an intrusion detection and prevention system based on software-defined network architecture and deep learning have been proposed. By modifying the SDN mechanism and the OpenFlow Protocol, made OpenFlow Controller and OpenFlow switch both can detect metwork attacks. When the suspected network traffic detected, OpenFlow Controller extract the feature and deep learning model uses it to identify network traffic. When the detection result is an attack, the OpenFlow Controller will send command to prevent the attack.