NAND Flash Memory具有體積小、非揮發性、抗震性高、功耗較低以及存取速度快等優點，隨著技術的進步，NAND Flash Memory已從SLC (Single-Level Cell)發展至容量更大的MLC (Multi-Level Cell)及TLC (Triple-Level Cell)，讓單位體積的資訊儲存量大幅增加，因此近年來NAND Flash Memory廣泛使用在各類電腦、行動裝置、嵌入式裝置或大型儲存系統上。然而NAND Flash Memory仍然存在硬體的限制，NAND Flash Memory中的區塊 (block) 有抹除次數 (Erase Count) 的上限，當抹除次數到達上限後，會有壞掉的風險，且NAND Flash Memory並不支援覆寫(Overwrite)的功能，要抹除一個Block之前，必須將Block中所有有效的(Valid)頁面 (Page) 都複製並且搬移。因此許多關於NAND Flash Memory的快閃記憶體轉換層 (Flash Translation Layer, FTL) 被提出以管理資料位置、減少資料搬移， FTL在存放資料時，若能分類冷、熱資料，在抹除Block 時，挑選的熱資料多的Block會因熱資料易改動為無效(Invalid)的狀態，而達到減少資料搬移。本文運用Machine Learning 演算法改進現有的FTL，並藉由區分冷、熱資料來減少資料搬移、降低抹除次數並提升NAND Flash Memory壽命。

NAND Flash Memory has the advantages of small size, non-volatile, high shock resistance, low power consumption and fast access speed. With the advancement of technology, NAND Flash Memory has evolved from SLC (Single-Level Cell) to larger capacity. MLC (Multi-Level Cell) and TLC (Triple-Level Cell) increase the storage density. Therefore, NAND Flash Memory has been widely used in various computers, mobile devices, embedded devices or large storage systems in recent years. However, NAND Flash Memory still has hardware limitations. Blocks in NAND Flash Memory have a limit of Erase Count. When the number of erases reaches the limit, there is a risk of Data loss.Also, NAND Flash Memory does not support the overwrite function. Before you can erase a block, you must copy and move all valid pages in the block. Therefore, many flash translation layers (FTL) for NAND Flash Memory are proposed to manage data location and reduce data movement. If FTL can separate cold and hot data, when garbage collection started, choose the hot block can reduce live page copy. This paper uses the Machine Learning algorithm to improve the existing FTL, and to reduce live page copy, and improve NAND Flash Memory life by separating cold and hot data.

[1] Aayush Gupta, Youngjae Kim, and Bhuvan Urgaonkar, "DFTL: A Flash Translation Layer Employing Demand-based Selective Caching of Page-level Address Mappings," Proceedings of the 14th international conference on Architectural support for programming languages and operating systems, pp. 229-240, 07 - 11 March 2009.
[2] Z. Qin, Y. Wang, D. Liu, Z. Shao, and Y. Guan, "MNFTL: An efficient flash translation layer for MLC NAND flash memory storage systems," in 2011 48th ACM/EDAC/IEEE Design Automation Conference (DAC), 2011.
[3] Dawoon Jung, Jeong-UK Kang, Heeseung Jo, Jin-Soo Kim, and Joonwon Lee, "Superblock FTL: A superblock-based flash translation layer with a hybrid address translation scheme," ACM Transactions on Embedded Computing Systems (TECS), March 2010.
[4] Jesung Kim,Jong Min Kim, and S.H. Noh, "A Space-Efficient Flash Translation Layer For Compactflash Systems," IEEE Transactions on Consumer Electronics, pp. 366 - 375, 07 August 2002.
[5] Lee, Sang-Won, Park, Dong-Joo, Chung, Tae-Sun, Lee, Dong-Ho, Park, Sangwon, and Song, Ha-Joo, "A log buffer-based flash translation layer using fully-associative sector translation," ACM Transactions on Embedded Computing Systems (TECS), July 2007.
[6] Hyunjin Cho,Dongkun Shin, and Young Ik Eom, "KAST: K-Associative Sector Translation for NAND flash memory in real-time systems," in 2009 Design, Automation & Test in Europe Conference & Exhibition, 2009.
[7] Sungjin Lee,Dongkun Shin,Young-Jin Kim, and Jihong Kim, "LAST: locality-aware sector translation for NAND flash memory-based storage systems," in ACM SIGOPS Operating Systems Review, 2008.
[8] Chanik Park, Wonmoon Cheon, Jeonguk Kang, Kangho Roh, Wonhee Cho, and Jin-Soo Kim, "A reconfigurable FTL (flash translation layer) architecture for NAND flash-based applications," ACM Transactions on Embedded Computing Systems (TECS), July 2008.
[9] Bongjae Kim, Minkyu Park, Cheol Jeon, Chang Oan Sung, Yookun Cho, and Jiman Hong, "AAGC: an efficient associativity-aware garbage collection scheme for hybrid FTL," SAC '12 Proceedings of the 27th Annual ACM Symposium on Applied Computing, pp. 1785-1790 , March 2012 .
[10] M.-L. Chiang, and R.-C. Chang, "Cleaning Policies in Mobile Computers Using Flash Memory," Journal of Systems and Software, pp. 213-231, Nov. 1999.
[11] Mendel Rosenblum, and John K. Ousterhout, "The Design and Implementation of A Log structured," ACM Transactions on Computer Systems (TOCS), pp. 26-52, Feb. 1992.
[12] L.-P. Chang, "On Efficient Wear Leveling for Large-Scale Flash-Memory Storage Systems," SAC '07 Proceedings of the 2007 ACM symposium on Applied computing, pp. 1126-1130, March 2007.
[13] Ohhoon Kwon, Kern Koh, Jaewoo Lee, and Hyokyung Bahn, "FeGC: An efficient garbage collection scheme for flash memory based storage systems," Journal of Systems and Software, pp. 1507-1523, September 2011.
[14] M. Wu, and W. Zwaenepoel, "eNVy: a non-volatile, main memory storage system," in Proceedings of IEEE 4th Workshop on Workstation Operating Systems. WWOS-III, 1993.
[15] H. J. Kim and S. G. Lee, "An Effective Flash Memory Manager for Reliable Flash Memory Space Management," in IEICE Transactions on Information and System, 2002.
[16] Dongchul Park, and David H.C. Du, "Hot data identification for flash-based storage systems using multiple bloom filters," in 2011 IEEE 27th Symposium on Mass Storage Systems and Technologies (MSST), 2011.
[17] C.-H. H. a. C. Z. Yen-Chin Liao, "Data Analysis and Prediction for NAND Flash Decoding Status," in 2017 IEEE International Memory Workshop (IMW), 2017.
[18] S. N. a. H. Y. Tomoko Ogura Iwasaki, "Machine Learning Prediction for 13X Endurance Enhancement in ReRAM SSD System," in 2015 IEEE International Memory Workshop (IMW), 2015.
[19] A. Ng. [Online]. Available: https://zh-tw.coursera.org/learn/machine-learning#.
[20] "Clustering," [Online]. Available: http://www.csie.ntnu.edu.tw/~u91029/Classification.html.
[21] S. RAY, "Understanding Support Vector Machine algorithm from examples (along with code)," 2017. [Online]. Available: https://www.analyticsvidhya.com/blog/2017/09/understaing-support-vector-machine-example-code/.
[22] "分類工具(1) - 支持向量機," 2016. [Online]. Available: http://yourgene.pixnet.net/blog/post/115169530-%E5%88%86%E9%A1%9E%E5%B7%A5%E5%85%B7(1)---%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%A9%9F(support-vector-machine).
[23] Lynn, "機器學習的衰頹興盛：從類神經網路到淺層學習," 2016. [Online]. Available: https://www.stockfeel.com.tw/%E6%A9%9F%E5%99%A8%E5%AD%B8%E7%BF%92%E7%9A%84%E8%A1%B0%E9%A0%B9%E8%88%88%E7%9B%9B%EF%BC%9A%E5%BE%9E%E9%A1%9E%E7%A5%9E%E7%B6%93%E7%B6%B2%E8%B7%AF%E5%88%B0%E6%B7%BA%E5%B1%A4%E5%AD%B8%E7%BF%92/.
[24] J.-Y. K. a. D.-H. L. Bo-Kyeong Kim, "A new hash algorithm exploiting triple-state bucket directory for flash storage devices," IEEE Transactions on Consumer Electronics , November 2016.
[25] Chiang, Hung-Yi ;Hsieh, Sheau-Ling;and Lin, Cheng-Chung, "A hybrid framework Based on SVM and K-means for seismic," 2008.
[26] Yucheng Kao,and Szu-Yuan Lee, "Combining K-means and particle swarm optimization for dynamic data clustering problems," in 2009 IEEE International Conference on Intelligent Computing and Intelligent Systems, 2009.