摘要
为了解决传统RAID5机制小写性能低下问题,设计了一种高性能的冗余管理机制RMR(RedundancyManagement for RAID5).RMR通过异步冗余计算方法APC(Asynchronous Parity Calculating),对传统RAID5写更新机制的关键技术进行了优化,将集中计算校验块的同步操作分解为同步的并行镜像写操作和异步冗余计算操作.该操作分解了数据与其校验数据更新的紧耦合特性,提高了写操作并发度;通过合并多个版本镜像块写操作减少异或计算操作的次数.测试表明,通过RMR机制,可将小写性能大幅提高,峰值可达4倍以上.
In response to solve the RAID5 performance problem for small writes by Asynchronous Parity Caculating. the mechanism RMR is proposed. RMR optimized the write-update mechanism of traditional RAID5 and decomposed the concentrated operations of calculating parity blocks into writing mirror and asynchronous calculating. It shows from experiments that RMR will substantially improve the performance of small writes, the peak value will be over four times as before under the out-of-band mode. Under the in-band mode the performance has also been increased by 9.0%-29.9%.
出处
《西北师范大学学报（自然科学版）》
CAS
北大核心
2012年第3期52-55,59,共5页
Journal of Northwest Normal University(Natural Science)
基金
西藏自治区自然科学基金资助项目("藏民族颅面数据库建立与量化研究")
关键词
磁盘冗余阵列
冗余管理
小写更新
异步冗余计算
性能
RAID
redundancy management
small write-update~ asynchronous parity calculating
performance
