CAPS: A contribution-aware P2P streaming system

This paper focuses on incentive, free rider, and fairness issues of p2p streaming systems. P2P streaming systems, applying p2p model over streaming applications, significantly improve the system scalability. However, instability and loose control of p2p model also impose uncertainty factors to p2p streaming systems. This paper proposed a contribution-aware p2p model, denoted as CAPS- Contribution Aware P2P Steaming, to fairly managing system resources for p2p streaming systems. Through fairly resource allocation, stream-quality proportional to peers' contribution is supported in CAPS and thus stimulates peers' desires to stay with system and improve stability and content availability. In the proposed CAPS system, peers form simple multi-chain architecture and a distributed scheme for peers to manage resources in CAPS is proposed. In addition, CAPS provides a strategy to pick out uncooperative peers. Simulation results demonstrate that CAPS not only supports fair resource allocation but also provides high system utilization.