An Intrusion Detection System (Ids) Schemes for Cybersecurity in Software Defined Networks

Abstract

The process of analysing and improving network traffic is of tremendous relevance to network management and multimedia data mining techniques. Security in Software Defined Networks (SDNs), which rely on a programmable controller in the middle, has recently emerged as the most challenging aspect of SDNs. Network traffic monitoring is critical for detecting and exposing intrusion anomalies in an SDN context. Thus, this study offers a thorough assessment of the NSL-KDD dataset using five separate clustering algorithms: K-means, Farthest First, Canopy, Density-based method, and Exception-maximization (EM). The software used to conduct the comparisons is the Waikato Environment for Knowledge Analysis (WEKA). In addition, the article introduces a knowledge discovery in databases (KDD)–based deep learning (DL) model for intrusion detection that is SDN-based. Initially, the dataset that is being used is clustered into four main attack types and one normal category. We will next go over the steps necessary to build a deep learning intrusion detection system that is based on SDN. The results provide an objective assessment of the several attack types present in the KDD dataset. Just like other methods, the results demonstrate that the proposed deep learning strategy provides better intrusion detection performance. As an illustration, the tested dataset demonstrates a detection accuracy of 94.21% when using the suggested approach.