Lightweight-Heavyweight Hybrid Approach for Application Traffic Classification

Traffic usage is increasing due to the continuous evolution of networks. As various applications for user convenience appear, various application traffic is also generated. Classifying application traffic is essential to manage high-capacity traffic according to various application traffic and usage. Until now, various studies have been conducted to classify application traffic, but most of them are studies that increase classification accuracy according to the development of deep learning. As environments with large amounts of traffic increase, research on the fast classification speed of application traffic is also needed in terms of efficient network management and network security. In this paper, we propose a system that applies thresholds and improves classification speed using two CNN algorithms. As a result of applying the proposed method, the accuracy increased by 0.06% and the processing speed increased by 0.11 seconds.