Exploring the optimal substream scheduling and distribution mechanism for data-driven P2P media streaming

• Published in: Computer communications 2014-05, Vol.44, p.14-25
• Main Authors: Hu, Chao, Chen, Ming, Xing, Changyou, Zhang, Guomin
• Format: Article
• Language: English
• Subjects: Applied sciences; Computer science; control theory; systems; Computer systems and distributed systems. User interface; Data-driven overlay; Exact sciences and technology; P2P live streaming; Scheduling and distribution; Software; Substream
• ISSN: 0140-3664; 1873-703X
• DOI: 10.1016/j.comcom.2014.02.018

Most P2P live streaming systems divide video stream into fine-grained blocks, and paralleled distribute these blocks in overlay network to utilize the bandwidth and storage resources of end users, which improves the scalability. Although transmitting video stream with these blocks greatly guarantees the system robustness, it also brings long distribution delay and high protocol overhead issues. Therefore, aggregating these blocks to substreams and delivering them in data-driven mode can promote the distribution efficiency under the network environment with peer churn. In this paper, we study the substream scheduling and distribution mechanism in data-driven overlay network, and formulate it as a minimax problem. Subsequently, we propose a global centralized approach to solve this problem, and derive the optimal substream scheduling and distribution scheme. According to the design idea of global centralized solution, we present a distributed substream scheduling and distribution mechanism, which is adaptive to dynamic P2P network. Finally, we compare the performance of global centralized solution and distributed substream scheduling and distribution mechanism with block-based live streaming. The experiment results show substream-based design achieves better efficiency in video stream dissemination, and distributed substream scheduling and distribution mechanism can preserve high performance when peer churn appears.