CAREER: Next-Generation Peer-to-Peer Streaming: Theory and Design

New York University, New York NY

Investigators

Abstract

Peer-to-Peer (P2P) technology has enabled large-scale streaming services on the Internet. The success of the next-generation P2P streaming systems largely hinges on new designs delivering a higher level of efficiency, robustness and incentives. At the same time, the increasing popularity and high-streaming rate of P2P streaming systems have already generated significant traffic stress inside ISPs' networks. The operations of future P2P streaming systems will have significant impact on the stability and efficiency of the global network infrastructure. The PI and his PhD students are developing new theory and design for the next generation P2P streaming systems. They study the service differentiation in P2P streaming to address peer heterogeneity and provide incentives. They derive the minimum delay bounds for realtime P2P streaming and propose streaming algorithms to approach the bounds. They apply the low-delay streaming algorithms to live video streaming and 3D streaming. They develop ISP-friendly P2P streaming design to largely reduce the traffic stress imposed on ISP networks by P2P streaming applications. They also build prototypes for the proposed new designs. The research of this project will lead to open protocols, software and tools, allowing a multitude of developers, researchers, and companies to expand, improve, and evolve the next-generation P2P streaming systems. Important insights and key techniques developed in this project can also be generalized to study data streaming in general distributed systems, such as Content Delivery Networks and distributed data centers.