Intelligent Detection System for SMF-AMF Stream Multiplexing Attacks in 5G-Enabled Smart Cities

In this paper, we propose a detector for stream multiplexing attacks in 5G networks for Smart Cities, which exploit vulnerabilities in the communication between the Session Management Function (SMF) and the Access and Mobility Management Function (AMF). We introduce two stream multiplexing attack variants and apply the proposed detector to a large IEEE 5G flow-based dataset. The detector, which is machine learning-based, achieved promising results. The Gradient Boosting model performed best, with an accuracy of 72.67%, AUC of 0.760959, recall of 88.84%, precision of 75.84%, and an F1 score of 81.83%. These results highlight the most relevant user traffic attributes in distinguishing attacks from legitimate traffic. Furthermore, the research contributes to the security and resiliency of 5G networks for Smart Cities, which is critical for the uninterrupted operation of public safety and critical infrastructure. Our detector demonstrates an improvement over prior work and provides a scalable, effective solution for securing urban 5G networks.