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

Abstract

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