基于访问量预测的教育资源网格副本放置策略  

Forecast of access traffic based education resource grid replication strategy

作　　者：邹立达[1] 刘方爱[1] 马艳[1] 

机构地区：[1]山东师范大学信息科学与工程学院,济南250014

出　　处：《计算机工程与应用》2009年第13期103-106,135,共5页Computer Engineering and Applications

基　　金：国家自然科学基金No.90612003;山东省科技攻关计划No.2006GGB01101~~

摘　　要：副本放置策略是数据网格副本机制的关键问题。针对教育资源网格环境,提出了基于访问量预测的副本放置策略(FATBRS),并介绍了基于历史数据的访问量预测(FATTHD)和基于人工神经网络的访问量预测(FATANN)两种访问量预测方法。FATBRS策略有效降低了用户的访问延迟,仿真实验证实了该策略性能的优越性。Grid replication strategy is the key problem of replica mechanism.The forecast of access traffic based education resource grid replication strategy （FATBRS） is presented for education resource grid.Two methods of forecasting accessing traffic are introduced,which are the forecast of access traffic based on historical data（FATTHD） and the forecast of access traffic based on artificial neural networks （FATANN）.FATBRS reduces the access delay efficiently and it is proved through the analysis of experiments.

关 键 词：副本放置 教育资源网格 访问量 人工神经网络 

分 类 号：TP393.01[自动化与计算机技术—计算机应用技术]