문성수, 박태준, 김수연, 나유경,

Vol. 34, No. 6, pp. 1501-1516, 12월. 2024
10.13089/JKIISC.2024.34.6.1501, Full Text:
Keywords: Network Security, IDS/IPS, Random Forest, Anomaly Detection
Abstract

As network usage has increased, the frequency and sophistication of cyber-attacks through networks have also increased. Although security devices exist to mitigate these attacks, they generally rely on static rules, which limits their flexibility in responding to attacks beyond predefined rules. To address this issue, this paper proposes an intelligent intrusion detection and prevention system by complementing signature-based and flow-based models. Leveraging Random Forest and TranAD, our system analyzes individual packets and their sequential flows to identify both known and unknown attacks. Additionally, the system automatically updates packet filtering rules based on the detection results, allowing for immediate responses to attacks. Experiments with the system prototype demonstrates high accuracy ,showing its potential as a next-generation security solution.

Statistics
Cite this article