Apprehending Mirai Botnet Philosophy and Smart Learning Models for IoT-DDoS Detection

Abstract

This paper aims at imparting acquaintance to the researchers an insight into the IoT metamorphosis from a security point of view. This paper presents a state-of-the-art apprehension of the IoT botnet landscape with a close analysis of Mirai. We have elucidated the characterization of the IoT-specific network behaviors such as limited endpoints, sleep time between packets, packet size, etc. that have turned out to be of substantial efficacy to contemporary learning algorithms, including neural networks. The learning algorithms have been reliable to be efficient enough for distributed denial of service (DDoS) attacks detection. We have evaluated the existing learning models and have proposed an efficient IoT-DDoS defense solution. Finally, we have concluded the research with prospective extensions.