Fingerprinting Technique for YouTube Videos Identification in Network Traffic

Recently, many video streaming services, such as YouTube, Twitch, and Facebook, have contributed to traffic, leading the possibility of unwanted inappropriate content minors or individuals at workplaces. Therefore, monitoring is necessary. Although traffic encrypted, several studies proposed techniques using data decipher users’ activity on web. Dynamic Adaptive Streaming over HTTP (DASH) uses Variable Bit-Rate (VBR) - most widely adopted technology, ensure smooth streaming. VBR causes inconsistencies in identification research. This research proposes a fingerprinting method accommodate for inconsistencies. First, bytes per second (BPS) are extracted from YouTube stream. Bytes Period (BPP) generated BPS, then fingerprints these BPPs. Furthermore, Convolutional Neural Network (CNN) optimized through experiments. The resulting CNN used detect streams VPN, Non-VPN, combination both VPN Non-VPN network traffic.