Youtube移动端加密视频传输模式快速识别

Youtube Streaming Communications Mode Quick Identification on the Encrypted Flow from Mobile

近年来,用户对视频业务的需求飞速增长.Youtube作为当今最受欢迎的互联网供应商之一,拥有超过一亿用户,人们每天在Youtube上观看超过数百万小时时长的视频,其中半数访问来源于移动终端.通过研究自适应传输模式的策略、中间人攻击进行明文密文比对,我们发现即使是在加密情况下,只需提取视频前两条流的少量数据包的四个特征,便能快速、准确识别出来自不同手机终端及手机浏览器访问Youtube的视频传输模式.本文使用的A-I-P-FP方法,能够快速识别来自移动端youtube加密流的视频传输模式,准确率超过90%.

In recent years,the demand for video services is grow ing rapidly. You Tube which is the most popular video sharing website at present has over a billion users – almost a third of all people on the Internet. Every day,people watch hundreds of millions of hours of You Tube videos and generate billions of views. More than half of You Tube view s come from mobile devices. In this paper,we present a fast and accurate method for identifying Youtube streaming communications mode on the encrypted flow from mobile. We found the video streaming communications techniques can be identified even they are encrypted based on four properties of the captured packets in first two flow ： total size of TLS fingerprint packets,SYN-ACK inter-arrival time,TLS protocol version and the acknumber of first tree TLS application data packets after we comparing the strategy of video streaming communications techniques and using man in the middle attack to find the distinct details among them. Our results show that A-I-P-FP method is sufficient to identify Youtube streaming communications mode with an accuracy over 90%.