基于多目标价值函数的流媒体代理缓存替换策略

A multi object value function based proxy cache replacement strategy for streaming media

在单代理结构上,分析了传统的缓存价值函数,综合多种目标设计了一种新的价值函数,提出了基于多目标价值函数的缓存替换算法MOC.理论分析和模拟实验表明,MOC在网络传输成本、命中率和字节命中率等重要的性能指标上大大优于传统的缓存替换算法.

The paper analyses the value function of traditional LRU and LFU algorithm and devises a new value function integrating multi objects in single proxy architecture, proposes a multi object value function based on proxy cache replacement strategy （MOC）. It has been proved that from the theoreti-cal analysis and experimental simulation, MOC is far better than the traditional caching replacement algorithms in important performance parameters such as transferring cost, hit ratio and byte hit ratio and can better meet network transferring needs for streaminz media.