Candidate Talk: TCP and P2P: supporting Internet from layer 4 and layer 7

  • Shao Liu | Visiting Researcher - Microsoft Research

Two main trends of the evolution of the Internet are the sharply increasing bandwidth and the dominance of video traffics. These two trends bring new requirements to TCP and embrace the emerging peer-to-peer streaming applications. We have studied the layer 4 and layer 7 support of the Internet, in terms of designing a new multi-mode TCP, and studying the fundamental performance limit of P2P streaming.

For the TCP side, we introduce a notion of multi-mode TCP, which accepts configuration settings or signals from applications and detects network conditions to responsively or automatically switch among different working modes, including normal mode, high-speed mode, low-priority mode, real-time mode, multicast mode, etc. In this talk, I will focus on introducing a new high speed TCP congestion control algorithm, called TCP-Illinois. TCP-Illinois uses packet loss information to determine whether the window size should be increased or decreased, and uses queueing delay information to determine the amount of increment or decrement. TCP-Illinois achieves high throughput, allocates the network resource fairly, is incentive compatible with standard TCP, outperform most other TCP variants, and is suitable for large-scale incremental deployment.

For the P2P side, it has shown enormous potential in delivering (live) streaming for IPTV, Video-on-Demand and video conferencing. However, the fundamental capacity limit of the maximum supported streaming rate by P2P remains unclear, especially with topology and peering constraints. We develop the first unifying framework and approximation algorithms that can approach the streaming capacity of a general P2P network over arbitrary overlay topology, number of streaming sessions, and peer selection constraints. For some special IPTV cases, we also derive the exact capacity value and develop the optimal overlay multicast tree construction algorithm. The study of P2P streaming capacity not only establishes a benchmark of any P2P system performance, but also brings insight on designing a practically implementable, throughput and delay efficient protocol for ISP controlled peer-assisted streaming system.

Combining the TCP and P2P work, our research provides a unifying solution to support the future Internet that has a typical Mega-Giga bps throughput and is dominated by streaming traffic.

Speaker Details

Shao Liu received B.S. degree from Peking University, Beijing, and M.S. and Ph.D. degrees in electrical engineering from University of Illinois at Urbana-Champaign, where his advisors are Prof. R. Srikant and Prof. Tamer Basar. He is currently with Princeton University, where he is a Postdoctoral Researcher in the Department of Electrical Engineering, advised by Prof. Mung Chiang. He is also a visiting researcher in CCS group of Microsoft Research, advised by Dr. Jin Li and Dr. Phil Chou, and he once was a summer Intern in Microsoft Research Cambridge, working with Dr. Milan Vojnovic.His research interests include congestion control for communication networks, peer-to-peer streaming systems, with applications like IPTV, Video on Demand and Video Conferencing, service differentiation and quality of service, etc. His recent work includes the design of TCP-Illinois protocol and the analysis on the fundamental performance bounds for peer-to-peer living streaming systems. His email address is shaoliu@princeton.edu, and more information on his recent and past projects is available at http://www.princeton.edu/~shaoliu

Watch Next