首页> 外文会议>International Conference on Cloud Computing and Security >Workload-Aware Page-Level Flash Translation Layer for NAND Flash-Based Storage Systems
【24h】

Workload-Aware Page-Level Flash Translation Layer for NAND Flash-Based Storage Systems

机译:用于基于NAND闪存的存储系统的工作负载感知页面级闪存转换层

获取原文

摘要

Demand-based flash translation layer is an efficient page-level flash translation layer, which can effectively reduce the RAM (Random Access Memory) footprint of NAND flash-based storage systems. However, this demand-based flash translation layer does not consider the spatial locality of workloads. In this paper, a new workload-aware page-level flash translation layer is proposed for NAND flash-based storage systems. The proposed flash translation layer maintains three caches in RAM to cache mapping entries, which are the on-demand mapping entry cache, frequent mapping entry cache, and dirty mapping entry cache. Considering both temporal locality and spatial locality of workloads, the on-demand mapping entry cache is designed to store the on-demand mapping entries and sequential mapping entries. Considering the access frequency of workloads, the frequent mapping entry cache is designed to cache the most frequently accessed mapping entries. To decrease the number of updates to translation pages, the dirty mapping entry cache is designed to cache the dirty mapping entries and flush the dirty mapping entries belonging to the same translation page to NAND flash memory in a batch mode. The experimental results show that the proposed flash translation layer performs better than existing page-level flash translation layers.
机译:基于需求的闪存翻译层是高效的页面级闪存翻译层,可以有效减少基于NAND闪存的存储系统的RAM(随机存取存储器)占用空间。但是,此基于需求的Flash翻译层未考虑工作负载的空间局部性。在本文中,针对基于NAND闪存的存储系统,提出了一个新的可感知工作负载的页面级闪存转换层。拟议的闪存转换层在RAM中维护三个高速缓存以高速缓存映射条目,它们是按需映射条目高速缓存,频繁映射条目高速缓存和脏映射条目高速缓存。考虑到工作负载的时间局部性和空间局部性,按需映射条目缓存被设计为存储按需映射条目和顺序映射条目。考虑到工作负载的访问频率,频繁映射条目缓存被设计为缓存最频繁访问的映射条目。为了减少对翻译页面的更新次数,脏映射条目缓存被设计为缓存脏映射条目,并以批处理方式将属于同一翻译页面的脏映射条目刷新到NAND闪存中。实验结果表明,所提出的Flash翻译层的性能要优于现有的页面级Flash翻译层。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号