Intrusion Detection System for SDN-enabled IoT Networks using Machine Learning Techniques

Abstract

With the rapid increase in smart devices and lowering prices of sensing devices, adoption of the Internet of Things (IoT) is gaining impetus. These IoT devices come with a greater threat of being attacked or compromised that could lead to the Denial of Service (DoS) and Distributed Denial of Service (DDoS). The high volume of IoT devices with a high level of heterogeneity, amplify the possibility of security threats. So far, the security of IoT devices is a big research challenge. But to enable resilience, continuous monitoring is required along with an adaptive decision making. These challenges can be addressed with the help of Software Defined Networking (SDN) which can effectively handle the security threats to the IoT devices in a dynamic and adaptive manner without any burden on the IoT devices. In this paper, we propose an SDN-based IoT Anomaly detection system which detects abnormal behaviors and attacks as early as possible. Three machine learning (ML) techniques, that is support vector machines (SVM), k-nearest neighbour (KNN) and multilayer perceptron (MLP), are used to design an IDS which is proposed to be deployed at the SDN controller to monitor and learn the behavior of IoT devices over time and any deviation from the normal behaviour is labelled as an attack. We test our algorithms on the two benchmark datasets. We present comparison of the results of the three ML techniques which demonstrate comparable detection accuracy.