Lifetime improvement of NAND flash-based storage systems using dynamic program and erase scaling
Lifetime improvement of NAND flash-based storage systems using dynamic program and erase scaling
复制标题
DOI：
--
复制
发表时间：
2014-02
期刊：
影响因子：
--
通讯作者：
Jaeyong Jeong;Sangwook Shane Hahn;Sungjin Lee-;Jihong Kim
中科院分区：
文献类型：
--
作者：
Jaeyong Jeong;Sangwook Shane Hahn;Sungjin Lee-;Jihong Kim
NAND闪存的每比特成本已经通过半导体工艺缩放和多级技术（例如，10 nm节点TLC器件）。然而，作为最近先进技术的副作用的NAND闪存的寿命减少被认为是广泛采用基于NAND闪存的存储系统的主要障碍。在本文中，我们提出了一种新的系统级的方法，称为动态编程和擦除缩放（DPES），提高寿命（特别是，耐力）的NAND闪存。DPES方法基于我们的关键观察，即改变擦除电压以及擦除时间会显著影响NAND耐久性。通过以较低的擦除电压缓慢擦除NAND块，可以非常有效地提高NAND耐久性。通过修改NAND芯片以支持具有不同操作电压和时间的多种写入和擦除模式，DPES使闪存软件能够在动态编程和擦除缩放下利用NAND耐久性与擦除电压/速度之间的新的折衷关系。我们已经实现了第一个DPES感知的FTL，称为autoFTL，它提高了NAND的耐久性，而整体写入吞吐量的下降可以忽略不计。我们使用各种I/O跟踪的实验结果表明，autoFTL可以将P/E周期的最大数量提高61.2%，而现有的DPES不知道FTL的整体写入吞吐量下降不到2.2%。
The cost-per-bit of NAND flash memory has been continuously improved by semiconductor process scaling and multi-leveling technologies (e.g., a 10 nm-node TLC device). However, the decreasing lifetime of NAND flash memory as a side effect of recent advanced technologies is regarded as a main barrier for a wide adoption of NAND flash-based storage systems. In this paper, we propose a new system-level approach, called dynamic program and erase scaling (DPES), for improving the lifetime (particularly, endurance) of NAND flash memory. The DPES approach is based on our key observation that changing the erase voltage as well as the erase time significantly affects the NAND endurance. By slowly erasing a NAND blockwith a lower erase voltage, we can improve the NAND endurance very effectively. By modifying NAND chips to support multiple write and erase modes with different operation voltages and times, DPES enables a flash software to exploit the new tradeoff relationships between the NAND endurance and erase voltage/ speed under dynamic program and erase scaling. We have implemented the first DPES-aware FTL, called autoFTL, which improves the NAND endurance with a negligible degradation in the overall write throughput. Our experimental results using various I/O traces show that autoFTL can improve the maximum number of P/E cycles by 61.2% over an existing DPES-unaware FTL with less than 2.2% decrease in the overall write throughput.