부산대학교 도서관

To prevent a website, network, or device from operating, a Distributed Denial of Service (DDoS) attacks transmits a large amount of data to it. This attack makes use of a “botnet,” which is an enormous collection of pilfered devices that simultaneously transmit a massive amount of requests and data to the target system. In a smart factory management, where a lot of devices are linked to each other via the Internet of Things (IoT), DoS attacks could be very risky. IoT devices are essential to smart factories, but these hacks have the ability to make them useless, which might have a lot of unfavorable effects. Downtime is a serious problem because it prevents Internet of Things (IoT) devices from working, which slows down production and raises costs. DDoS attacks may be employed as a diversion from riskier behaviors that might compromise security, such as unauthorized access and data breaches. Additionally, data corruption or loss might occur, harming the business's reputation and long-term operations. In the proposed model ML trained chip systems are capable of real-time data analysis. They identify patterns of typical activity and immediately identify anomalies that may indicate Distributed Denial of Service (DDoS) attacks. These anomalies not only immediately trigger alerts, but they also assist in identifying compromised IoT devices, enabling prompt and efficient action and safety measures. The model can manage new threats and continually adapting and learning new things. The building's managers and security personnel may see real-time data on a basic screen. In this research study, four distinct methodologies were used. Each provided a unique method for approaching challenges related to machine learning and categorization. XGBoost, K-Nearest Neighbors (KNN), Logistic Regression, and Gaussian Naive Bayes were among the techniques used. The investigation's conclusions indicate that XGBoost stood out as the top performer as it continuously produced the best results and showed exceptional performance throughout the range of tasks assessed.