Machine Learning Based Anomaly and Intrusion Detection to mitigate DoS and DDoS attacks in Private Campus Networks

Inhalt:
Private Campus Networks (PCNs) are critical in providing customized connectivity for industries such as smart factories and autonomous driving. PCNs are customizable and provide essential revenue opportunities. Artificial intelligence and machine learning are key enablers in modern networks, optimizing and securing them. However, technology growth brings increased security risks. A Network Intrusion and Anomaly Detection System (NIADS) helps secure cellular networks and protect businesses from suspicious activities. With NIADS as the focus, this paper presents machine learning-based methods to predict network anomaly and intrusion. The uniqueness of this study lies in the dataset, the data used in training and testing were collected from live 5G SA PCNs available at RPTU Kaiserslautern. Subsequently, the paper presents details about generating and extracting normal, and anomalous datasets with various DoS and DDoS attacks such as TCP flood, UDP flood, Golden Eye, Hulk, Slowloris, and Slowloris HTTP. Machine learning methods like long short-term memory (LSTM) and Autoencoders are used with these datasets to predict anomalies and intrusion. The results show fair performance and demonstrate the applicability of such methods in private campus networks (PCNs).