Encrypted 5G over-the-top voice traffic classification using deep learning

With the commercialization of fifth-generation (5G), rapid popularity mobile Over-The-Top (OTT) voice applications brings huge impacts on traditional telecommunication call services. Tunnel encryption technology such as Virtual Private Networks (VPNs) allow OTT users to escape supervision network operators easily, which may cause potential security risks cyberspace. To monitor harmful in context 5G, it is critical identify encrypted traffic. However, there no comprehensive study typical traffic identification. This paper mainly focuses analyzing 5G specifically. We propose employing Long Short-Term Memory (LSTM) and Convolutional Neural (CNNs) traffic, identification performance used deep learning methods three different scenarios. verify proposed approach, we collect 28 types non-OTT from experimental network. Experimental results prove effectiveness robustness approach