摘要
数据收集是指在断连前把用户将来可能访问的数据预先存储到本地缓存的过程,在研究基于失效报告的缓存策略的基础上,提出了一种超图模型的自动数据收集技术。理论分析和实验结果表明,该技术开销较低,可以有效提高Cache命中率和降低平均访问时间,尤其适用于缓存较大和注意力频繁转移的情况。
Data hoarding is the process of pre-fetching important data into local cache for future operation before disconnection. Having researched many invalidation report-based cache strategies, a data hoarding technique of hyper graph model is proposed. The theoretical analysis and experiment results all indicate that the technique is low-cost and can effectively improve cache hit rate and shorten average response time, especially fits big caches and the environment of frequent attention shift.
出处
《计算机应用与软件》
CSCD
2009年第6期118-120,200,共4页
Computer Applications and Software
基金
湖南省自然科学基金项目(04JJ3045)
湖南省教育厅一般项目(07C215)