Thanks to mobility and large coverage, 6G mobile networks introduce satellites and unmanned aerial vehicles as aerial base stations (ABS) in the 6G era. Instead of using a wired backhaul in 5G and its predecessor, an ABS leverages a wireless channel to a core network (CN). However, such a wireless channel design introduces new security challenges. In this paper, we present that passive attackers could sniff the ABS-CN wireless channel and identify what users are doing based on deep learning methods. We collect GTP protocol data on our testbed and use convolutional neural networks to classify 5 types of encrypted App traffic, like IG and TikTok. Experiment results proved the effectiveness of the proposed method, revealing the confidential data leakage problem on the 6G wireless ABS-CN channel.
|Title of host publication||ACM MobiCom '23|
|Subtitle of host publication||Proceedings of the 29th Annual International Conference on Mobile Computing and Networking|
|Publisher||Association for Computing Machinery (ACM)|
|Publication status||Published - 2 Oct 2023|