Title: Rate oscillation breaks in HTTP on-off distributions: a DASH framework

Authors: Koffka Khan; Wayne Goodridge

Addresses: Department of Computing and Information Technology, The University of the West Indies, Trinidad and Tobago, West Indies ' Department of Computing and Information Technology, The University of the West Indies, Trinidad and Tobago, West Indies

Abstract: The chances of multiple video players sharing a bottleneck link becomes increasingly likely as dynamic adaptive streaming over HTTP (DASH) remains popular. Multi-player DASH scenarios across bottleneck links pose many user-perceived quality of experience (QoE) challenges. Existing solutions still do not address the adequate sharing of network resources. We present a rate oscillation breaks in HTTP on-off distributions (ROBIN-HOOD) framework for DASH. A streaming approach implementing this framework allows player(s) receiving high bitrates to demand lower quality video segments, while the player receiving the lowest bitrate demands higher quality video segments. This mechanism allows the weaker players to improve their share of available network resources. The result is better user-QoE for players. Two approaches implementing the framework, B-DASH (heuristic) and S-MDP (stochastic), are compared in live internet experiments. It is shown that Robin-Hood improves user-QoE. In addition, the MDP stochastic S-MDP Robin-Hood approach outperforms the heuristic B-DASH Robin-Hood approach.

Keywords: multiple players; DASH; dynamic adaptive streaming over HTTP; bottleneck link; QoE; quality of experience; Robin-Hood; B-DASH heuristic; S-MDP stochastic.

DOI: 10.1504/IJAACS.2020.110742

International Journal of Autonomous and Adaptive Communications Systems, 2020 Vol.13 No.3, pp.273 - 296

Received: 07 Dec 2018
Accepted: 13 Feb 2020
Published online: 28 Oct 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article