ABSTRACT
Internet of Things (IoT) is a combination of networks that have the ability to collect and share the information through the web. Although IoT is used in areas such as healthcare, smart cities, agriculture, industrial automation and home automation, it is also used in other fields. The challenges faced by the IoT devices are security, privacy, confidentiality, integrity, and technical complexity. In Distributed Denial of Service (DDoS) attacks, the attacker sends massive service requests to the IoT device which cannot handle such high traffic, resulting in the disruption of services to the end user. Modern types of DDoS attack are still more complex, and they are nearly impossible to identify or mitigate using classic intrusion detection systems and methods. In the past few years, machine learning plays a significant role in identifying DDoS attacks effectively. We conducted a literature review on different types of DDoS attacks based on ML techniques. The model’s performance is tested on several classifiers to choose the optimal classifier with high accuracy. We next go over each of the ML methods for IoT security in detail. The limitations and problems of current research are discussed.
