Successfully Mapping DASH Over a P2P Live Streaming Architecture

Abstract

This paper proposes to combine peer-to-peer (P2P) and dynamic adaptive streaming over HyperText Transfer Protocol (HTTP) (DASH), leveraging the scalability and self-organization properties of the former and the capability of the latter to adapt the rate of the delivered video to varying operating conditions. The devised P2P-DASH architecture is a multioverlay P2P platform, where every peer implements a decentralized rate control strategy that steers its transitions from one overlay to another, so as to achieve a good viewing quality and also preserve the good functioning of the entire system. At peer’s site, the new architecture displays the following salient features: 1) an augmented pool of neighbors, to provide every node with the knowledge of the streaming process within adjacent overlays; 2) a priority mechanism to speed up the delivery of video chunks to peers that newly enter the system or switch from the current to a new overlay, in order to quickly turn them into active peers; and 3) some alternative rate control algorithms to guarantee, with different degrees, more resources to overlays that deliver higher streaming rates, therefore reducing the misalignment among streaming processes which occur in different overlays, and ultimately improve the viewing quality of the single user. System behavior is extensively investigated through simulation, and some clear design guidelines are provided.