Improving the Sustainability of Solid-State Drives by Prolonging Lifetime

被引:0
作者
Ke, Zhaokang [1 ]
Kang, Dingyi [2 ]
Yuan, Bo [3 ]
Du, David [1 ]
Li, Bingzhe [2 ]
机构
[1] Univ Minnesota Twin Cities, Minneapolis, MN 55455 USA
[2] Univ Texas Dallas, Richardson, TX 75083 USA
[3] Rutgers State Univ, New Brunswick, NJ USA
来源
2024 IEEE COMPUTER SOCIETY ANNUAL SYMPOSIUM ON VLSI, ISVLSI | 2024年
关键词
Sustainability; Solid-State Drives (SSDs); NAND Flash Memory; Error Correction Codes (ECC); NAND FLASH MEMORY;
D O I
10.1109/ISVLSI61997.2024.00096
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
In the big data era, vast amounts of data are processed and stored in cloud data centers using Solid-State Drives (SSDs). While SSDs are crucial for efficient data handling as the fast-tier in tiered storage, they also contribute significantly to embodied carbon emissions. In this paper, we explore a method to enhance SSD sustainability by reducing the maximum threshold voltage of flash cells, extending lifespan and improving write performance. However, this leads to decreased read performance due to a high Raw Bit Error Rate (RBER). To improve the SSD performance while prolonging its lifespan, we partition the device into two regions. One region maintains standard voltage levels to efficiently manage read-intensive tasks. The other with a reduced threshold voltage absorbs write-intensive data, leveraging the benefits of extended lifespan and enhanced write performance provided by low-voltage operation. Additionally, we propose a dynamic caching policy that caches frequently accessed data to the DRAM in SSD. Finally, our experimental results demonstrate that the proposed design improves the SSD lifespan by 8.7% to 20.3% compared to some previously proposed schemes while achieving average latency reductions ranging from 19.7% to 69.8%. Our work demonstrates the potential for enhancing the sustainability of SSDs while ensuring high performance, contributing to the reduction of the carbon footprint associated with storage systems in the era of big data.
引用
收藏
页码:502 / 507
页数:6
相关论文
共 15 条
[1]  
Cai Y, 2018, Arxiv, DOI arXiv:1711.11427
[2]   Enabling NAND Flash Memory Use Soft-Decision Error Correction Codes at Minimal Read Latency Overhead [J].
Dong, Guiqiang ;
Xie, Ningde ;
Zhang, Tong .
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS, 2013, 60 (09) :2412-2421
[3]  
Du Y., 2017, 2017 33 INT C MASS S, P1
[4]   The real climate and transformative impact of ICT: A critique of estimates, trends, and regulations [J].
Freitag, Charlotte ;
Berners-Lee, Mike ;
Widdicks, Kelly ;
Knowles, Bran ;
Blair, Gordon S. ;
Friday, Adrian .
PATTERNS, 2021, 2 (09)
[5]   Exploring and Exploiting the Multilevel Parallelism Inside SSDs for Improved Performance and Endurance [J].
Hu, Yang ;
Jiang, Hong ;
Feng, Dan ;
Tian, Lei ;
Luo, Hao ;
Ren, Chao .
IEEE TRANSACTIONS ON COMPUTERS, 2013, 62 (06) :1141-1155
[6]  
Jaeyong Jeong, 2014, Proceedings of the 12th USENIX Conference on File and Storage Technologies. FAST '14, P61
[7]   HAML-SSD: A Hardware Accelerated Hotness-Aware Machine Learning based SSD Management [J].
Li, Bingzhe ;
Deng, Chunhua ;
Yang, Jinfeng ;
Lilja, David ;
Yuan, Bo ;
Du, David .
2019 IEEE/ACM INTERNATIONAL CONFERENCE ON COMPUTER-AIDED DESIGN (ICCAD), 2019,
[8]   EFM: Elastic Flash Management to Enhance Performance of Hybrid Flash Memory [J].
Li, Bingzhe ;
Yuan, Bo ;
Du, David .
2021 IEEE 39TH INTERNATIONAL CONFERENCE ON COMPUTER DESIGN (ICCD 2021), 2021, :162-169
[9]  
Liu WH, 2019, DES AUT TEST EUROPE, P312, DOI [10.23919/DATE.2019.8714941, 10.23919/date.2019.8714941]
[10]   Reliability of Solid-State Drives Based on NAND Flash Memory [J].
Mielke, Neal R. ;
Frickey, Roberte. ;
Kalastirsky, Ivan ;
Quan, Minyan ;
Ustinov, Dmitry ;
Vasudevan, Venkatesh J. .
PROCEEDINGS OF THE IEEE, 2017, 105 (09) :1725-1750