为了提高web缓存性能,在已有缓存替换算法的基础上加入预测机制,提出了一种面向社交网站(SNS)用户访问行为特征的预测替换算法.通过研究SNS的用户行为模型,引入预测对象集,减小了替换风险,提高了缓存命中率.为了验证所提算法的性能,进行了大量仿真实验,结果表明,该算法在基于SNS使用行为的缓存方面,具有提高命中率的优越性.
A prediction algorithm based on user requests for prediction on user requests for social networking services(PUR-SNS) is proposed to improve the efficiency of web cache.The proposed algorithm is based on behavior pattern of user requests in SNS application and a forecasting mechanism is introduced.Simulations shows that PUR-SNS will improve the hit rate.