Abstract—5G is the new generation of mobile networks intended to improve the performance of earlier generations while integrating a variety of use cases with varying requirements into a single network. With such a wide coverage and its pervasiveness into society, it is crucial to assess the cyber-risks inherent to its implementations. In particular, we question the suitability of network monitoring solutions such as intrusion detection systems with regards to 5G requirements.

This survey presents an analysis of intrusion detection systems in contexts related to 5G networks and propose a taxonomy to determine a set of suitable features. Among these features, it motivates the need for collaboration in order to overcome some of the challenges imposed by 5G networks such as heterogeneity or low-latency. We study Federated Learning (FL) as a candidate to enable collaboration in intrusion detection for 5G networks and discuss future research directions.

Index Terms—5G, intrusion detection system, machine learning, network security, federated learning.