Machine Learning DDoS Detection for Generated Internet of Things Dataset (IoT Dat)

Network infrastructure faces a lot of attacks, including attacks on integrity and confidentiality of the network packets along with their destinations and sources as well as attacks on network availability. Distributed Denial of Service (DDoS) emanates from various attack sources and focuses on the network, services, and hosts' availability. DDoS attacks are difficult to trace back to actual attackers, can lead to catastrophic service loss, and are launched with ease, making them one of the most dangerous attacks. This research simulates an Internet of Things network in-home setting of 100 nodes using OMNeT++ simulation tool, including a DDoS attack. Regular and attack-injected traffic is generated to evaluate the accuracy of detecting DDoS attacks in IoT networks using machine learning technqiues. A new IoT Dataset called IoT Dat is generated with different scenarios of normal traffic and traffic with attacks of different intensities of 5, 10, and 20. The authors will make this dataset publicly available. Moreover, machine learning techniques are used to assess the efficiency of attack detection.