层次型混合归档系统是数据存储领域的研究热点,在工业界也被广泛接受,小到个人存储大到数据中心都在使用混合存储系统.针对大数据中心的能耗问题,引入更为廉价低能耗的光介质存储,建立阵列、在线光盘库、离线光盘库构成的混合归档系统,采用低能耗磁光混合的存储架构,在牺牲陈旧文件读取速度的情况下,大幅降低了存储能耗;在研究分析数据的一致性保证机制以及多级存储系统的弹性设计的基础上,针对光介质的读写特性,重点研究了磁光混合归档系统的文件缓存和预取机制.采用基于整体访问频率的数据迁移策略解决了热度文件的访问效率不受影响的问题,采用基于I/O特征预测模型的预取算法,提升了多级存储结构的命中准确度.实验结果表明,所构建的系统能够有效节省能源并可维持数据检索查询的时效性.%Hierarchical hybrid archiving system is a research hotspot in the field of data storage and is also widely accepted in the industry.The hybrid storage systems have been used by not only personal storage but also the data center.In order to solve the problem of energy consumption in large data center,a hybrid archiving system composed of array,online optical disk library and off-line optical disk library with low cost and low energy consumption has been introduced and low energy consumption magneto-optic hybrid storage architecture has been adopted.Based on the study of data consistency guarantee mechanism and the elastic design of multi-level storage system,the optical read/write characteristics of optical media has been investigated as well as the characteristics of magneto-optical hybrid.The file cache and prefetch mechanism of the archiving system has been adopted and the data migration strategy based on the overall access frequency is adopted to solve the problem that the access efficiency of the thermal file is not affected.The prefetching algorithm based on the I/O characteristic prediction model has also been adopted and the multi-level storage structure has been improved.The experimental results show that the proposed system can effectively save energy and maintain the timeliness of data retrieval query.
展开▼
