Development of lightweight intrusion detection and prevention system for the internet of things

Abstract

In this thesis, a thorough and critical analysis of the existing intrusion detection systems in IoT with major focus on anomaly-based detection systems is performed. Then, a real-time framework for traffic generation and data collection for building an IoT dataset for IDS evaluation and testing is developed. Later, a lightweight machine learning based IDS has been proposed which detects multiple cross-layer attacks with high accuracy and precision. The experimental evaluation confirms that proposed IDS is capable of detecting cross-layer attacks effectively. Finally, a lightweight host intrusion prevention system is proposed to prevent critical IoT nodes from flooding attacks in the fourth phase. The proposed prevention system conserves the computation and communication power of the victim thereby ensuring maximum availability of the service provided by the victim node. Therefore, deploying the proposed IDS and IPS ensures robustness of IoT networks and devices. newline