AngelCast: Cloud-based peer-assisted live streaming using optimized multi-tree construction

Increasingly, commercial content providers (CPs) offer streaming solutions using peer-to-peer (P2P) architectures, which promises significant scalability by leveraging clients' upstream capacity. A major limitation of P2P live streaming is that playout rates are constrained by clients' upstream capacities - typically much lower than downstream capacities - which limit the quality of the delivered stream. To leverage P2P architectures without sacrificing quality, CPs must commit additional resources to complement clients' resources. In this work, we propose a cloud-based service AngelCast that enables CPs to complement P2P streaming. By subscribing to AngelCast, a CP is able to deploy extra resources (angel), on-demand from the cloud, to maintain a desirable stream quality. Angels do not download the whole stream, nor are they in possession of it. Rather, angels only relay the minimal fraction of the stream necessary to achieve the desired quality. We provide a lower bound on the minimum angel capacity needed to maintain a desired client bit-rate, and develop a fluid model construction to achieve it. Realizing the limitations of the fluid model construction, we design a practical multi-tree construction that captures the spirit of the optimal construction, and avoids its limitations. We present a prototype implementation of AngelCast, along with experimental results confirming the feasibility of our service.