Peer-to-Peer Live Video Distribution under Heterogeneous Bandwidth Constraints

This paper presents a new collaborative peer-to-peer (P2P) streaming framework for heterogeneous bandwidth capacity of clients. The proposed architecture aims for two primary goals: 1) providing higher streaming quality by making best use of the extra available bandwidth that might exist among heterogeneous clients and 2) providing robustness and resilience to high churn rate of peers by introducing redundancy, both in network paths (multisender overlay) and data (multilayered video content). For achieving these goals, we employ a gossip-based data-driven scheme for partnership formation and layered video coding for bandwidth adaptivity. To solve multilayer bitstream allocation problem, we proposed two algorithms, namely optimized transmission policy and graceful degradation scheme algorithms. The proposed system has a complete self-regulation and in a decentralized fashion. Extensive simulations show that the proposed system achieves higher quality of service through peer-assisted streaming and layered video coding. The results also show that the system outperforms some previous schemes in system overhead and reliability for dynamic node behavior.