A Study Of Performance Optimization Strategies In P2P Live Streaming Services

P2P live streaming business has been widely used in recent years,with the expansion of business types and user scale,how to optimize P2P live streaming performance,reduce system consumption and improve user experience becomes to be important research topics.In this paper,we first present the unique requirements and challenges of P2P live streaming systems,comprehensively summarize current performance optimization strategies from overlay construction,data scheduling and dissemination aspects,then we introduce our previous work—how to improve user experience such as playback continuity,initial buffering delay by using network coding in P2P live streaming systems.For some novel live streaming businesses such as online stock system,online auction system,entertainment game and remote education,we study their real-time and interactive property and solve the following two optimization problems:1.we study the network fairness problem in real-time and interactive streaming businesses.Network fairness was originally proposed to treat the problem of fair resource/bandwidth allocation to different nodes in network environment with serious congestion,in this paper,we propose a novel definition of application-based network fairness,we illustrate the significant benefits of fairness scheduling for user profits in real-time an interactive live streaming services,and we propose a preliminary idea of server-side control to guarantee application-based network fairness in such businesses;2.we study the playback lag and initial offset placement problem in real-time and interactive streaming systems,we propose a proportional offset placement mechanism,through which users could achieve differentiated services and optimized average playback lag,we also construct a virtual tree+mesh dissemination manner, which increases the opportunity of data exchange among different peers and enables more economic server bandwidth consumption under tight resource environment.We make abundant simulation experiments based on the theories we proposed, and through careful analysis of experiment results,we testify the fundamental feasibility and effectiveness of these performance optimization strategies,and prove a great benefit to obtain better user experience.