摘要
为了提高P2P流媒体系统性能,深入分析了P2P流媒体系统直播和点播业务的异同,提出采用动态时间坐标系(OTCS)构建直播点播融合的流媒体系统,将系统节点的协作特性与观看内容解耦。此外,充分考虑到节点的异构性与不同频道内容的流行度,提出了融合系统的节点协作模型,通过分布式服务质量感知算法(DQAA)来进行优化分析,充分利用P2P流媒体系统中服务能力较强的节点的剩余带宽提高系统性能。从融合系统架构及节点协作策略两个层次描述了直播点播融合的系统特性。实验表明,融合系统能有效提高节点播放质量与系统整体性能。
To improve the performance of a P2P streaming media system, the paper deeply analyzes the similarities and differences between a P2P live system and a P2P VoD(Video-on-Demand) system, and based on this, presents a method to build a P'2P media streaming system integrated with live and VoD based on the Dynamic Time Coordinate System(DTCS). The integrated system decoulpes the collaboration feature with the particular content watched by users. In addition, taking consideration of the heterogeneity among peers and the popularity of different channels, the paper proposes a peer collaboration model for the integrated system. Based on this model, the paper addresses the Distributed QoS-Aware Algorithm(DQAA) to optimize the system, which makes full use of the peers' surplus bandwidth to improve system performance. The paper describes the features of the integrated system from two as- pects: overlay construction and cooperation strategy. The simulation results demonstrate that the integrated system can significantly improve the playback quality of peers and system performance.
出处
《高技术通讯》
CAS
CSCD
北大核心
2012年第3期263-270,共8页
Chinese High Technology Letters
基金
“十一五”国家科技支撑计划(2008BAH28804)和国家自然科学基金(60903218)资助项目.