Duration: 1:24:10

1,520 views

14

With P2P live video streaming, peers viewing videos also assist the server in streaming the videos, thereby significantly reducing server infrastructure cost. In recent years, there have been several large-scale deployments of P2P live video systems; for example, PPstream has 350 million downloads, about 12 million active users every day, and thousands of channels. These P2P streaming systems have several fundamental performance problems including large channel switching delays, long playback lags, and poor performance for less popular channels. We propose a new cross-channel P2P streaming framework, called View-Upload Decoupling (VUD). VUD strictly decouples peer downloading from uploading, bringing stability to multichannel systems and enabling cross-channel resource sharing. We propose a set of peer assignment and bandwidth allocation algorithms to properly provision bandwidth among channels, and introduce sub-stream swarming to reduce the bandwidth overhead. We evaluate the performance of VUD using simulations and a PlanetLab implementation. In the second part of the talk, we show how multiple channel P2P systems can be modeled by product-form stochastic models. We apply the product-form theory to both traditional isolated channel systems and VUD designs. Using the classical theory of product-form networks, we show how to develop efficient algorithms to calculate performance measures, how to derive rules-of-thumb via asymptotic analysis, how to obtain the optimal size of the VUD groups in an asymptotic setting, and how to optimize the VUD groups for equitable support for small and large channels. This second part of the talk is a summary of the Infocom 2009 best paper award. This work is joint work with Yong Liu and Di Wu of Polytechnic Institute of NYU.