360NorVic Article Swipe
YOU?
·
· 2021
· Open Access
·
· DOI: https://doi.org/10.1145/3458306.3460998
· OA: W3160440512
Streaming 360{\\deg} video demands high bandwidth and low latency, and poses\nsignificant challenges to Internet Service Providers (ISPs) and Mobile Network\nOperators (MNOs). The identification of 360{\\deg} video traffic can therefore\nbenefits fixed and mobile carriers to optimize their network and provide better\nQuality of Experience (QoE) to the user. However, end-to-end encryption of\nnetwork traffic has obstructed identifying those 360{\\deg} videos from regular\nvideos. As a solution this paper presents 360NorVic, a near-realtime and\noffline Machine Learning (ML) classification engine to distinguish 360{\\deg}\nvideos from regular videos when streamed from mobile devices. We collect packet\nand flow level data for over 800 video traces from YouTube & Facebook\naccounting for 200 unique videos under varying streaming conditions. Our\nresults show that for near-realtime and offline classification at packet level,\naverage accuracy exceeds 95%, and that for flow level, 360NorVic achieves more\nthan 92% average accuracy. Finally, we pilot our solution in the commercial\nnetwork of a large MNO showing the feasibility and effectiveness of 360NorVic\nin production settings.\n
