Detecting Distributed Denial of Service Attacks in Internet of Things Networks Using Machine Learning in Fog Computing
Karrar Falih Hassan, Mehdi Ebady Manaa · 2022 5th International Conference on Engineering Technology and its Applications (IICETA) · 2022
The number of Internet of Things (IoT) devices and the data produced by these devices have increased dramatically in recent years. IoT networks have benefited from the arrival of 5G, which aims to link more devices quickly and reliably. However, many of these devices are unsecured, leaving the Internet open to assault. Since IoT devices are resource-restricted, attackers may simply mount a larger and more complex DDoS assault using IoT devices. Entropy-based detection is technique proposed by other researchers, and it has a high sensitivity and a good degree of reliability. In this paper, authors propose an entropy-based framework with the machine learning algorithms to optimize the detection result. The proposed system consists of three parts: the first part is to create a sliding time window to calculate the entropy, and the second is using the Euclidean distance to make early detection possible during the DDoS but not after the collapse. The third part is the use of machine learning algorithms (KNN and SVM), to improve the accuracy of the results in early detection. The results showed that the proposed system is able to increase the accuracy from 98.6% to 100%, and efficiently detect attacks in real-time for (ICMP, TCP, UDP, HTTP and DNS) DDoS attacks.