A Novel Rate Control Strategy for Adaptive Video Streaming in P2P Overlays

This work proposes a novel P2P-DASH architecture that jointly leverages the intrinsic cooperation property of P2P and the adaptivity of Dynamic Adaptive Streaming over HTTP (DASH) to deliver videos to a large population of users with heterogeneous streaming requests. The system is made of several overlays, each distributing a different DASH representation; every peer within the system autonomously enforces a newly devised rate control strategy that steers its movements from one overlay to another, in order to maintain a good viewing experience and to simultaneously guarantee the successful operation of the P2P overlays. The effectiveness of the solution is demonstrated through simulation and it indicates that the P2P-DASH platform is able to guarantee its users a very good performance, much more satisfying than in a conventional P2P environment where DASH is not employed. Moreover, through the comparison with a reference system modeled via the Integer Linear Programming (ILP) approach,it is shown that our system outperforms such reference architecture in terms of viewing quality; at the same time, the number of users streaming the video at the desired bit rate is extremely close to the optimum of the ILP setting in the most favorable scenarios, and only moderately suboptimal in the worst.