Inside the New Coolstreaming: Principles, Measurements and Performance Implications
Inside the New Coolstreaming: Principles, Measurements and Performance Implications
复制标题
DOI：
10.1109/infocom.2008.157
复制
发表时间：
2008-04
期刊：
影响因子：
--
通讯作者：
Bo Li;Susu Xie;Yang Qu;G. Y. Keung;Chuang Lin;Jiangchuan Liu;Xinyan Zhang
中科院分区：
文献类型：
--
作者：
Bo Li;Susu Xie;Yang Qu;G. Y. Keung;Chuang Lin;Jiangchuan Liu;Xinyan Zhang
基于P2P(Peer-to-Peer)的视频流已经成为一种很有前途的互联网视频分发解决方案。利用终端用户可用的资源，这种方法具有在互联网上进行扩展的巨大潜力。我们现在已经看到了商业P2P流媒体系统，它们比早期的学术系统大了几个数量级。我们相信，了解它的基本原理和局限性在未来系统的设计中很重要。CoolStreaming于2004年夏天首次发布，可以说是第一个成功的大规模P2P直播系统。从那时起，该系统已进行了重大修改并投入商业使用。本文通过揭示其设计选项和背后的基本原理，深入了解了新的CoolStream系统，并研究了它们对流媒体性能的影响。具体地说，通过利用从最近的现场事件广播和广泛的模拟获得的大量轨迹，我们研究了工作负载特征、系统动态以及各种系统参数的影响。我们证明了在这类系统中存在高度倾斜的资源分布，并且性能主要受系统动态特性的影响。此外，我们还证明了不同的系统参数之间存在内在的相关性和基本的权衡，这可以进一步研究以提高系统的性能。
The peer-to-peer (P2P) based video streaming has emerged as a promising solution for Internet video distribution. Leveraging the resource available at end users, this approach poses great potential to scale in the Internet. We have now seen the commercial P2P streaming systems that are orders of magnitude larger than the earlier academic systems. We believe understanding its basic principles and limitations are important in the design of future systems. The Coolstreaming, first released in summer 2004, arguably represented the first successful large-scale P2P live streaming system. Since then, the system has been significantly modified and commercially launched. This paper takes an inside look at the new Coolstreaming system by exposing its design options and rationale behind them, and examines their implications on streaming performance. Specifically, by leveraging a large set of traces obtained from recent live event broadcast and extensive simulations, we study the workload characteristics, system dynamics, and impact from a variety of system parameters. We demonstrate that there is a highly skewed resource distribution in such systems and the performance is mostly affected by the system dynamics. In addition, we show that there are inherent correlations and fundamental trade-off among different system parameters, which can be further explored to enhance the system performance.