Accelerating YouTube with video correlation

In this paper, using long-term data traces, we present an in-depth measurement study on the characteristics of YouTube, the most successful site providing a new generation of short video sharing service. We find that YouTube videos have noticeable differences compared with traditional videos, making it difficult to use conventional strategies, such as peer-to-peer, to reduce the server workload. However, the video correlation presented in YouTube opens new opportunities. We design a novel peer-to-peer short video sharing system based on video correlation, in which peers are responsible for re-distributing the videos that they have cached. We address a series of key design issues to realize the system, including a novel architecture design, an efficient indexing scheme and a source rate allocation mechanism. We perform extensive simulations, which show that the system greatly reduces the server workload and improves the playback quality.