Browser Identification Based on Encrypted Traffic

Network traffic encryption brings security to network communication, however it also brings challenges to network monitoring. As more and more major websites use encryption protocol to protect imformation of visitors, it is a burning issue to identify client when session is encrypted. In this paper, we present browser identification of HTTPS client based on packet length. The sequence of request packet length is different among browsers and our experiment shows that it is possible to recognize what kind of browser the traffic comes from according to length sequence. We theoretically analyze the possibility of using the request packet length to identify browsers and show the method of using length sequence to establish dictionary. The dictionaries are used to distinguish unknown traffic flow. Our experiment results show that we can get accurate results of browser identification in HTTPS communication through packet length analysis.