부산대학교 도서관

The use of Internet of Things (IoT) devices is on the rise as they enter our homes and businesses, from smart appliances to increased use in agriculture, health care, and manufacturing. These IoT devices rely on their network connection to send and act on data they are acquiring from the environment around them. Because of this dependency, IoT devices are especially susceptible to denial of service (DoS) and distributed denial of service (DDoS) network attacks. In the following research, a simulated DoS attack will be performed the new Arduino Oplà device. The results are then fed into several machine learning algorithms to review predictability for detection based on packet loss and packet transit time. Results indicated the Decision Tree J48 algorithm with 5-fold cross-validation resulting in an f1-score of 0.848 best predicts the IoT network attack. The results show the methods in which these attacks can be performed, how susceptible the device is to DoS attacks, and the adverse effects that the attacks have on response time and packet loss.