摘要
在流媒体CDN中,采用"推拉"结合的流媒体内容分发方式可进一步减少客户的平均启动时间,降低网络资源的消耗,根据流媒体文件的部分访问特性给出了一种流媒体内容的部分推送策略,并利用源服务器统计的历史访问信息,提出了目的代理缓存服务器的一种随机选择算法。实验表明该策略及算法可提高系统性能。
In streaming media CDN,pushing copies of the new content to caching proxies ahead of time,while allow others to be pulled from the origin server,is an effective way to reduce the client start time and save network resource.Based on the characteristics of the intra-video skewness,this paper has proposed a method that partially pushing streaming media objects to cache proxies,and put forward an stochastic algorithm of destination selection using the history access information the origin server recorded.The experiments indicate that these mechanisms can make the system maintain better performance.
出处
《计算机工程与应用》
CSCD
北大核心
2007年第25期162-164,185,共4页
Computer Engineering and Applications
关键词
流媒体
CDN
内容分发
缓存代理
streaming media
CDN
content distribution
cache proxy
