基于连续缓存和二级缓存的DFTL改进算法

发布时间：2018-05-15 02:01

本文选题：NAND闪存 + 固态硬盘　；参考：《计算机研究与发展》2014年09期

【摘要】：DFTL(demand-based FTL)是一种根据负载访问特点动态加载映射项到缓存中的知名FTL(flash translation layer)算法,但是它没有考虑到请求的空间局部性,而且缓存中的一个映射项剔除就可能会导致翻译页的更新,缓存中映射项的频繁剔除又会导致额外的擦除操作.在DFTL的基础上,提出了SDFTL(sequential/second cache DFTL)算法.SDFTL新设置连续缓存和二级缓存,连续缓存通过预取映射信息,利用请求的空间局部性,提高了FTL对连续负载的处理性能;二级缓存通过暂存从一级缓存中剔除的、发生更新的映射项,并采取批量更新策略回写到闪存,减少了闪存的翻译页写回次数和擦除次数.利用实际负载做的实验结果显示,SDFTL相比DFTL缓存命中率平均提高41.57%,擦除次数平均减少23.08%,响应时间平均减少17.74%.
[Abstract]:DFTL(demand-based DFTL(demand-based) is a well-known FTL(flash translation layer algorithm that dynamically loads mapping items into the cache according to load access characteristics. However, it does not take into account the spatial localization of requests, and the removal of a mapping item in the cache may result in the updating of translated pages. Frequent culling of mapped items in the cache results in additional erasure operations. On the basis of DFTL, SDFTL(sequential/second cache DFTL algorithm. SDFTL sets up continuous cache and two-level cache. Continuous cache can improve the processing performance of FTL to continuous load by prefetching mapping information and utilizing the spatial localization of request. The second level cache removes the updated mapping items from the primary cache by temporary storage, and adopts batch update strategy to write back to flash memory, which reduces the number of page write-back and erasure times of flash memory. The experimental results show that the hit rate of SDFTL is 41.57 higher than that of DFTL, the erasure times are 23.08, and the response time is 17.74.
【作者单位】：杭州电子科技大学通信工程学院;
【基金】：国家自然科学基金项目(61100044)
【分类号】：TP333

【相似文献】