Botnet-based IoT network traffic analysis using deep learning

IoT networks are increasingly being connected to a wide range of devices, and the number of devices connected has significantly increased in recent years. As a consequence, the number of vulnerabilities to IoT networks has also been increasing tremendously. In IoT networks, botnet-based Distributed Denial of Service attack is challenging due to its dynamic behavior. The sensors and actuators connected to IoT networks are low-powered and have less memory. Because of their inherent vulnerability, IoT devices can always be compromised by an attacker and be used to form a large botnet. A detailed analysis of IoT botnet attacks is presented in this article, along with statistics and the architectures of the botnet. We also survey the existing literature on IoT botnet traffic analysis and present a taxonomy of attack detection methods. We particularly focus on deep learning-based methods and conduct a comparative study to evaluate their performance on IoT traffic analysis. We identify the current issues and research challenges in this field, and we conclude by highlighting some future research directions.