5G core network anomaly detection based on deep learning

• 1、School of Cyberspace, Beijing University of Posts and Telecommunications, BeiJing 100876

Abstract：In recent years, with the development of deep learning technology, network traffic analysis technology has made great progress, and more and more researchers have applied it to network traffic anomaly detection, network traffic classification, etc. In addition, with the large-scale application of the fifth-generation mobile communication technology, its security risks have also received more and more attention, such as signaling storms and unauthorized access. At present, most of the existing network traffic detection technologies are used for anomaly detection of Internet traffic, lacking targeted detection for special scenarios such as 5G core networks. In addition, there is currently a lack of public 5G core network anomaly detection data sets, making it difficult to verify the algorithm. In response to the above problems, this paper generates a 5G core network anomaly detection data set based on a simulation environment, considers the HTTP protocol attack inside the core network and the N2 interface signaling attack, and proposes an anomaly detection method for 5G core network traffic based on GNN+LSTM. This method is superior to the existing general traffic detection methods in terms of detection accuracy and detection performance.

Keywords： Network security Abnormal traffic detection Deep learning