一句话：提出 LeaFTL，用学习替代人工启发式来优化 SSD 的闪存页地址映射与 DRAM 使用。
现代固态硬盘（SSD）的存储控制器需要对闪存页进行索引，这一地址映射不仅影响数据访问性能，也决定设备内宝贵 DRAM 资源的利用效率。已有工作提出了多种地址映射方案与优化，但大多依赖人工设计的启发式规则。本文提出 LeaFTL，一种基于学习的闪存转换层（FTL），尝试用数据驱动的学习方法替代人工启发式来完成 SSD 的地址映射与相关优化。摘要未报告具体实验结果，因此其性能收益与 DRAM 开销改善需由正文评估确认。
In modern solid-state drives (SSDs), the indexing of flash pages is a critical component in their storage controllers. It not only affects the data access performance, but also determines the efficiency of the precious in-device DRAM resource. A variety of address mapping schemes and optimizations have been proposed. However, most of them were developed with human-driven heuristics.
@inproceedings{paperbot10046,
title = {LeaFTL: A Learning-Based Flash Translation Layer for Solid-State Drives},
author = {Jinghan Sun and Shaobo Li and Yunxin Sun and Chao Sun and Dejan Vucinic and Jian Huang},
booktitle = {ASPLOS},
year = {2023},
doi = {10.1145/3575693.3575744}
}