摘要
针对P2P直播存在瞬间拥塞，导致用户启动延迟增大、系统服务拒绝率升高，从而降低用户体验的问题，提出了一种基于能力感知的用户访问控制算法。该算法可通过合理控制用户节点的加入速率，避免过多用户同时竞争带宽资源；优先允许高带宽用户节点接入，提高系统的服务能力；兼顾低带宽用户的等待时间，防止其因过长等待而离开。对该算法进行的数学建模分析、模型分析和相关实验表明，该算法能使P2P直播系统有效应对瞬间拥塞问题，改善用户服务质量。与两种控制方案的性能对比显示，该算法的平均延迟可分别降低17％和29％，拒绝率可分别减少60％和75％。
Flash crowds often make users suffer from long startup delays, high reject rate and bad service. A novel user access control algorithm based on capacity-aware is proposed to tackle the problems flash crowds bring. The algo- rithm controls users to enter at a proper rate to prevent the high user arrival rate to slow down the increase of system scale. Also, it gives high-bandwidth users the priority to enter the system to increase the system service capacity as quickly as possible. Moreover, it considers the waiting time of the low-bandwidth users and prevent these users from leaving the system for waiting too long. A model for evaluation of the new user access control algorithm is also given. The model analysis and related experiments reveal that the new algorithm can effectively deal with the flash crowd problems and improve the quality of service for users by reducin~ the startuD delav and service reiect rata.
出处
《高技术通讯》
CAS
CSCD
北大核心
2012年第12期1250-1256,共7页
Chinese High Technology Letters
基金
国家自然科学基金（61003266）,973计划（2012CB315802）和内蒙古自治区自然科学基金（2011MS0902）资助项目.
