摘要
研究了代理服务器在流媒体传输和缓存方面对于网络带宽的消耗以及在缓存过程中内存分配和管理等方面存在的问题;在分析了现有的流媒体代理缓存技术的基础上,提出了基于代理服务器的流媒体动态共享缓存(DSB)算法。分析和实验结果表明,与现有的缓存技术相比,DSB算法能够有效地提高缓冲区的利用率,节省代理服务器的内存资源,节省网络带宽,同时能够为更多的客户端请求提供服务,并且可以有效地缩短请求延时。
With the increasing number of multimedia services, caching media objects in the memory of a proxy helps to reduce network bandwidth requirement and data delivery latency. The patching and caching approaches in the existing techniques have some limits. Data sharing among the multiple running buffers is not considered and user access patterns are not also insightfully considered in the memory management. Proposed dynamic shared buffers algorithms-DSB in the proxy to address these limits. This algorithms adaptively allocated memory buffers and fully utilized the caching data in the streaming sessions, aimed to reduce the memory scheduling and the network traffic. As compared with several existing caching algorithms, the proposed algorithm shown that it had achieved significant performce improvement by effectively using the shared buffers and reduced significantly the consumption of aggregate bandwidth.
出处
《计算机应用研究》
CSCD
北大核心
2007年第5期262-265,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(60573182
69883004)
关键词
流媒体
代理缓存
补丁预取
调度算法
streaming media
proxy caching
patch prefetching
scheduling algorithms