PIRES Karine1,2, SIMON Gwendal1,2
Communication dans une conférence avec acte
VideoNext 2014 : 1st Workshop on Design, Quality and Deployment of Adaptive Video Streaming, 02-02 december 2014, Sydney, Australia, 2014, pp. 13-18
Live game streaming platforms such as Twitch allow gamers to broadcast their gameplay over the Internet. The popularity of these platforms boosts the market of eSport but poses new delivery problems. In this paper, we focus on the implementation of adaptive bitrate streaming in massive live game streaming platforms. Based on three months of real data traces from Twitch, we motivate the need for an adoption of adaptive bitrate streaming in this platform to reduce the delivery bandwidth cost and to increase QoE of viewers. We show however that a naive implementation requires the reservation of a large amount of computing resources for transcoding purposes. To address the trade-off between benefits and costs, we formulate a management problem and we design two strategies for deciding which online channels should be delivered by adaptive bitrate streaming. Our evaluations based on real traces show that these strategies can reduce the overall infrastructure cost by 40% in comparison to an implementation without adaptive streaming.
1 : RSM - Dépt. Réseaux, Sécurité et Multimédia (Institut Mines-Télécom-Télécom Bretagne-UEB)
2 : IRISA - Institut de recherche en informatique et systèmes aléatoires (UMR CNRS 6074 - Université de Rennes 1 - INRIA - INSA de Rennes - ENS de Cachan - Télécom Bretagne - Université de Bretagne Sud)
Adaptive video, Cloud computing, Content delivery