A Model of Encrypted Network Traffic Classification that Trades Off Accuracy and Efficiency

A Model of Encrypted Network Traffic Classification that Trades Off Accuracy and Efficiency

연구 분야: Cryptography

학회: Journal of Network and Systems Management

As the Internet industry evolves, the need for effective encrypted traffic classification (ETC) becomes critical for network management and cybersecurity. Meanwhile, existing deep learning (DL)-based methods struggle to balance model complexity with accuracy. A major challenge is to deploy these models on dominant network devices in a way that ensures fast and accurate traffic classification. In this paper, we propose FasterTrafficNet, a novel DL-based lightweight ETC strategy designed for deployment on low-configuration network devices. We have designed the core component of FasterTrafficNet using a PConv-based approach that efficiently extracts spatial features from data while reducing unnecessary computation and memory access, further enhancing the operational efficiency of the model. In addition, we integrated Do-Conv in place of conventional non-dot convolution to significantly increase the performance of the model without increasing the computational overhead during inference. We conducted a comparative analysis of FasterTrafficNet against seven advanced ETC methods using four publicly available benchmark datasets. Experimental results show that FasterTrafficNet, with 1.46 million model parameters, provides superior classification performance compared to the other methods. As a result, FasterTrafficNet demonstrates a lightweight approach to ETC that can be applied to large networks of devices.