DOOM: a Training-Free, Real-Time Video Flow Identification Method for Encrypted Traffic
Youting Liu, Shu Li, Chengwei Zhang, Chao Zheng, Yong Sun, Qingyun Liu · 2020
The internet traffic is mainly dominated by videos, limiting video transmission is thus required to manage the traffic when network congestion occurs. Besides, service providers want to guarantee video transmission for customers' satisfaction. However, traffic encryption is becoming increasingly common, making conventional payload-based methods unsuitable, while traffic analysis (TA) still works. In this paper, we implement and evaluate machine learning (ML)-based TA for video flow identification and propose a method named detecting ON-OFF mode (DOOM), which is a training-free, real-time and lightweight video flow identification method suitable for Gateway or Internet Backbone Provider (IBP) middle-boxes.