Fortifying the 5G Core: Analysing and Detecting DoS Attack from Compromised NFs using HTM

The rapid increase in mobile data traffic, the growing number of connected devices, and the variety of applications are creating significant demands on the core networks of 5G and beyond (5GB). While 5G and 5GB technologies are effectively meeting these demands through strategies like network slicing, they also introduce vulnerabilities to specific attacks, such as session mirroring and resource overallocation, which arise when core functions like the SMF (Session Management Function) are compromised. These attacks can deplete network resources, resulting in Denial of Service (DoS). In this work, we address these attacks by analyzing their impact using various traffic metrics. We propose an anomaly detection approach based on Hierarchical Temporal Memory (HTM), which can effectively manage real-time traffic anomalies by leveraging previous learning states. Compared to other neural network models (Recurrent Neural Network, Long Short Term Memory, Multi-Layer Perception), our results show that the HTM-based detection method delivers superior performance in real-time traffic, particularly in terms of anomaly score, accuracy, precision, recall, and F1 score metrics.