DDoS Attack Detection in Edge-IIoT using Ensemble Learning

Every Edge-IIoT device and network is susceptible to attacks because they are connected to the internet. The number of IoT devices grows daily due to the rapid advancement in technology. The server goes down as a result of a flood of requests in a DDoS attack, which is a common type of intrusion in the IoT. As a consequence, the business may experience upset clients, a decline in sales, and a decline in client confidence. Even if they do not steal anything or carry out a long-term offensive, DDoS attacks can cause significant harm to a business's productivity, uptime, and reputation. This study aims to identify normal or malicious DDoS attacks in an Edge-IoT network (DDOS traffic). The proposed study utilizes XGBoost and an ensemble of SVM, Decision Tree, and Naive Bayes through hard voting to predict normal and malicious traffic using the dataset Edge IIoT. In addition, our findings indicate that XGBoost outperformed the hard-voting ensemble classifier by 11%.