数据迁移算法研究

Study of Data Migration Algorithm

分级存储根据数据的访问频率、保留时间、容量等指标将数据存储在不同性能的存储介质上,是降低海量数据存储成本的有效解决方案,如何对存储的数据进行迁移是分级存储管理中的关键技术。该文研究了三种不同的数据迁移算法,基于存储空间的高低水位法、基于数据访问率的Cache替换迁移算法和信息生命周期管理中的价值评估法,对其优缺点进行了对比。综合应用三种迁移算法,可以实现数据在各级存储层次之间自动智能的迁移,最终达到提高系统性能的目的。

Hierarchical storage based on the data access frequency,retention time,capacity and so on indicators,store data on different per formance storage media,Which is an effective solution for reducing the cost of mass data storage.How to migrate storaged data is the key technology of hierarchical storage management.The paper studies three different data migration algorithms,high-low level method based on storage space,cache replacement algorithm based on the data access rate and data valuation method based on ILM,and compares their advantages and disadvantages.Integrated application three kinds of migration algorithm,it can achieve intelligent automatic migration of da ta among all levels of memory hierarchy.Finally,it can achieve the goal of improving storage system performance.