Live user-generated video streaming platforms like Twitch. tv generate a large portion of the Internet traffic. Millions of viewers daily watch user channels, although roughly 85% of all channels have less than 200 views during one session. Due to latency, Twitch. tv provides one or more servers for each of Twitch. tv’s supported countries. An alternative approach could enable peer-to-peer communication in order to utilize the capacities of the user devices. Solutions up to now, mainly offer only best effort delay guarantees on the distribution speed from initial seeders to all peers. In this paper, we present Chunked-Swarm, a swarm-based approach, which aims to offer predictable streaming delays, independently of the number of peers. Evaluation shows the various impact of the number of peers, number of video parts and chunks on the streaming delay. Being able to hold specific deadlines for up to 200 peers, predestines our solution to be suitable for the majority of Twitch. tv’s channels.
CITATION STYLE
Probst, C., Disterhöft, A., & Graffi, K. (2015). Chunked-swarm: Divide and conquer for real-time bounds in video streaming. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9247, pp. 198–210). Springer Verlag. https://doi.org/10.1007/978-3-319-23126-6_18
Mendeley helps you to discover research relevant for your work.