Smart-IoT: A denial of service protection system utilizing machine learning
Network Security, IoT, DDoS, Machine Learning, SDN.
The popularization of devices known as the Internet of Things (IoT) has boosted the number of Distributed Denial of Service (DDoS) attacks in recent years. This threat takes advantage of the security limitations of these devices as well as their geographic location to maximize the impact of attacks. Developing mechanisms to detect and mitigate this new paradigm is a current challenge in the area of network security. This work proposes a defense mechanism integrated with the IoT network controller that uses Machine Learning (AM) techniques for attack detection and the flexibility of software-defined networks (SDN) to perform mitigation. The proposed system uses random samples to classify network traffic and the OpenFlow protocol to apply real-time mitigation measures. The preliminary results obtained by the system in an emulated environment using three datasets indicate that the proposed approach is promising to be embedded in specific purpose hardware.