Abstract: Cooperation of streaming media cache servers sited between the Internet backbone and an access network can increase the hit rate and keep load balancing. This paper presents a close coupling cooperation scheme, that the cache servers shared their individual caches, puts forward a cooperative cache replacement algorithm and a dynamic load balancing method. The experiments with NS2 simulation indicate that these mechanisms can make the cache system maintain better performance.

Key words: Streaming media, Cooperative cache, Cache replace algorithm, Load balancing

摘要： 位于因特网骨干网和同一接入网之间的流媒体缓存代理服务器相互协作，可以提高缓存命中率，保持负载平衡。该文提出了一种共享缓存空间的紧耦合的多代理服务器协作机制，给出了多代理协作的缓存替换策略和负载平衡算法。通过NS2模拟验证，该机制可以使系统保持更好的性能。

关键词: 流媒体, 缓存协作机制, 缓存替换算法, 负载平衡

CLC Number: 

• TP37