現代 NVMe SSD 具有多隊列架構以提高整體性能，其中提交隊列可以為來自主機的流提供服務。 由於 NAND 閃存中讀寫延遲的不對稱性，現代 NVMe SSD 通常使用 DRAM 空間中的寫入緩衝區和寫入映射緩存來延長壽命並減少 NAND 閃存中的寫入。 在本文中，我們將提出一種固態驅動器 (SSD) 內的多流感知 DRAM 分配策略。 所提出的策略是為每個流的寫緩衝區和寫映射緩存預測合適的DRAM空間，主要減少固定DRAM大小或可變DRAM空間下的NAND閃存中的寫入。 根據實驗結果，我們可以證明，所提出的策略可以為每個流預測合適的 DRAM 空間，以有效減少多流 SSD 中 NAND 閃存的寫入。

The modern NVMe SSDs have a multi-queue architecture to increase overall performance, where a submission queue can serve a stream from the host. Because of the asymmetry of read/write latency in NAND flash memory, the modern NVMe SSDs usually use the write buffer and the write mapping cache in DRAM space to extend the lifetime and reduce writes in NAND flash memory. In the paper, we will propose a multi-stream-aware DRAM allocation strategy inside solid-state drives (SSDs). The proposed strategy is to predict the suitable DRAM space for the write buffer and the write mapping cache of each stream to mainly reduce writes in NAND flash memory under a fixed DRAM size or variable DRAM space. According to the experimental results, we can show that the proposed strategy can predict a suitable DRAM space for each stream to effectively reduce writes in NAND flash memory in a multi-stream SSD.

[1] C. Siau, K.-H. Kim, S. Lee, K. Isobe, N. Shibata, K. Verma, T. Ariki, J. Li,
J. Yuh, A. Amarnath, Q. Nguyen, O. Kwon, S. Jeong, H. Li, H.-L. Hsu, T.-y.
Tseng, S. Choi, S. Darne, P. Anantula, A. Yap, H. Chibvongodze, H. Miwa, M. Yamashita, M. Watanabe, K. Hayashi, Y. Kato, T. Miwa, J. Y. Kang, M. Okumura,
N. Ookuma, M. Balaga, V. Ramachandra, A. Matsuda, S. Kulkani, R. Rachineni,
P. K. Manjunath, M. Takehara, A. Pai, S. Rajendra, T. Hisada, R. Fukuda, N. Tokiwa,
K. Kawaguchi, M. Yamaoka, H. Komai, T. Minamoto, M. Unno, S. Ozawa, H. Nakamura, T. Hishida, Y. Kajitani, and L. Lin, “13.5 a 512gb 3-bit/cell 3d flash memory
on 128-wordline-layer with 132mb/s write performance featuring circuit-under-array
technology,” in 2019 IEEE International Solid- State Circuits Conference - (ISSCC),
pp. 218–220, 2019.
[2] S. Nie, Y. Zhang, W. Wu, and J. Yang, “Layer rber variation aware read performance
optimization for 3d flash memories,” in Proceedings of the 57th ACM/EDAC/IEEE
Design Automation Conference, DAC ’20, IEEE Press, 2020.
[3] N. Megiddo and D. S. Modha, “ARC: A Self-Tuning, low overhead replacement
cache,” in 2nd USENIX Conference on File and Storage Technologies (FAST 03),
(San Francisco, CA), USENIX Association, Mar. 2003.
[4] H. Kim and S. Ahn, “BPLRU: A buffer management scheme for improving random writes in flash storage,” in 6th USENIX Conference on File and Storage
Technologies (FAST 08), (San Jose, CA), USENIX Association, Feb. 2008.
[5] S.-y. Park, D. Jung, J.-u. Kang, J.-s. Kim, and J. Lee, “Cflru: A replacement algorithm for flash memory,” in Proceedings of the 2006 International Conference on
Compilers, Architecture and Synthesis for Embedded Systems, CASES ’06, (New
York, NY, USA), p. 234–241, Association for Computing Machinery, 2006.
[6] J. Cui, W. Wu, Y. Wang, and Z. Duan, “Pt-lru: a probabilistic page replacement algorithm for nand flash-based consumer electronics,” IEEE Transactions on Consumer
Electronics, vol. 60, no. 4, pp. 614–622, 2014.
[7] A. Gupta, Y. Kim, and B. Urgaonkar, “Dftl: A flash translation layer employing demand-based selective caching of page-level address mappings,” SIGARCH
Comput. Archit. News, vol. 37, p. 229–240, mar 2009.
[8] D. M. Bates and D. G. Watts, Nonlinear regression analysis and its applications,
vol. 2. Wiley New York, 1988.
[9] H. Shim, B.-K. Seo, J.-S. Kim, and S. Maeng, “An adaptive partitioning scheme for
dram-based cache in solid state drives,” in 2010 IEEE 26th Symposium on Mass
Storage Systems and Technologies (MSST), pp. 1–12, IEEE, 2010.
[10] Y. Hu, H. Jiang, D. Feng, H. Luo, and L. Tian, “Pass: A proactive and adaptive ssd
buffer scheme for data-intensive workloads,” in IEEE International Conference on
Networking, Architecture and Storage (NAS), pp. 54–63, Aug. 2015.
[11] Y. Yao, M. Yan, X. Kong, X. Xu, W. Feng, and X. Xu, “An adaptive read-write
partitioning flash translation layer algorithm,” IEEE Access, vol. 7, pp. 179063–
179073, 2019.
[12] A. Tavakkol, M. Sadrosadati, S. Ghose, J. Kim, Y. Luo, Y. Wang, N. Mansouri Ghiasi, L. Orosa, J. Gómez-Luna, and O. Mutlu, “Flin: Enabling fairness and enhancing
performance in modern nvme solid state drives,” in 2018 ACM/IEEE 45th Annual
International Symposium on Computer Architecture (ISCA), pp. 397–410, 2018.
[13] “Snia iotta trace repository,” 2020.
[14] S. P. Council, “Spc trace file format specification,” 2002.
[15] C. Lee, T. Kumano, T. Matsuki, H. Endo, N. Fukumoto, and M. Sugawara, “Understanding storage traffic characteristics on enterprise virtual desktop infrastructure,”
in Proceedings of the 10th ACM International Systems and Storage Conference,
pp. 1–11, 2017.
[16] D. Narayanan, A. Donnelly, and A. Rowstron, “Write off-loading: Practical power
management for enterprise storage,” ACM Transactions on Storage (TOS), vol. 4,
no. 3, pp. 1–23, 2008.