论文部分内容阅读
预送作为主动cache，是cache机制由时间局部性向空间局部性的拓展．文章提出服务器主动预送的两种模式．基于单个URL的模式利用客户请求的马尔可夫链特征获取文档的时序相关模型，可进行多级预送．基于会话的模式包括基于文档属性和会话整体语义的途径，重点研究基于文档属性的途径，给出基本的聚集算法，探讨了文档兴趣的定量表达，提出反映访问时序的属性向量距离算法．对于预送性能的度量，给出请求命中率、会话命中率、预送效率和预送代价等度量方法，同时，完成大量实验，对客户行为分析的这两种模式进行比较．文章提出的由服务器访问记录提取客户行为模式的方法，不但适用于文档预送，对于服务器站点设计和ISP（internetserviceprovider）的服务规划也有重要价值．
Pre-delivery as an active cache, is the cache mechanism from the time to local space to expand. The article proposes two modes of server proactive delivery. Based on a single URL model using the Markov chain characteristics of customer requests to obtain the document timing model, multi-level pre-delivery. Session-based patterns include approaches based on document attributes and the overall semantics of the conversation, focusing on approaches based on document attributes, giving basic aggregation algorithms, discussing quantitative expressions of document interest, and proposing an attribute vector distance algorithm that reflects access timing. For the measurement of the pre-delivery performance, this paper gives the measurement methods of request hit rate, session hit rate, pre-delivery efficiency and pre-delivery cost. At the same time, a large number of experiments are completed and the two models of customer behavior analysis are compared. The method proposed by the article to extract customer behavior patterns from server access records not only applies to document pre-delivery, but also has important value to server site design and ISP (internetserviceprovider) service planning.