近年來，非揮發性隨機存取記憶體(Non-Volatile Random Access Memory, NVRAM)，例如相變化隨機存取記憶體(Phase-change random access memory, PCM)等技術蓬勃發展，與傳統主記憶體所使用的DRAM相比，這些NVRAM不會因斷電即造成資料流失。除了有潛力取代DRAM或擴充DRAM之外，NVRAM的高密度與非揮發性之優勢也使其可以用來作為儲存裝置以儲存檔案系統資料。其出現讓傳統儲存架構之主記憶體與儲存裝置間的階層界線趨於模糊，使得現有之檔案系統在設計上需重新針對非揮發性記憶體之特性做更進一步的考量。從而促使了非揮發性記憶體檔案系統的誕生。然而，與DRAM相比，NVRAM的寫入延遲仍有一大段差距，非揮發性記憶體檔案系統之效能也受限於NVRAM的寫入延遲而造成效能低落。因此，本論文提出了利用減緩短期檔案的資料保留時間來改善非揮發性記憶體之寫入速度的方法。方法中，我們利用Multilevel Cell (MLC) PCM可以依據資料保留時間而選擇速度不同的寫入方式此一特性。例如，資料保留時間較長則寫入延遲時間也較長。反之，資料保留時間較短則寫入延遲時間較短。並找出生命週期相對較短的檔案，對其使用短期的資料保留時間之寫入方式，以期望改善系統整體存取記憶體之效能。
實驗中，我們的方法與對照組相比，有效的達成了平均10.61% 的記憶體存取速度改善，並且達成了平均79.83%的記憶體壽命改善．

Byte-addressable nonvolatile memory (NVM), which has a reading speed equivalent to dynamic random access memory (DRAM) and enables persistent data storage, has become a promising technology. Meanwhile, file systems for byte-addressable NVM (NVMFSs) have been proposed in recent years. However, the write latency of NVM is higher than that of DRAM. This high latency degrades the overall file access speed when write-intensive programs are involved. Previous studies have shown that the write latency of NVMs can be minimized by conducting retention relaxation of multilevel-cell (MLC) NVM. For example, in phase-change memory, a write operation that has a high number of set iterations offers high retention capability and high write latency. Moreover, a write operation with few set iterations offers low retention capability and low write latency. Hence, the average memory access latency of an NVMFS can be improved by relaxing the retention capability of file data. However, the use of a short-term retention write operation for all data causes frequent refreshing of the data, which degrades the bandwidth performance and lifetime of NVM memory. Therefore, a balance should be provided between the retention capability and write speed of MLC NVM.
In this paper, we observed that programs often generate short-lived files. These files only require a short-term retention capability in an NVMFS. According to this observation, a short-lived file manager (SFM) is proposed in this study to pick up short-lived files and perform short-term retention writing operations for these files. The results of this study reveal that the SFM improves the average memory access latency of each workload in the system by up to 19.8% (10.61% on an average) and increases the memory lifetime by up to 127.42% (79.83% on an average) compared to the baseline.

[1] E. Chen et al., “Advances and future prospects of spin-transfer torque random access memory,” IEEE Trans. Magn., vol. 46, no. 6, pp. 1873-1878, Jun. 2010.
[2] Z. Diao et al., “Spin-transfer torque switching in magnetic tunnel junctions and spin-transfer torque random access memory,” J. Phys. Condens. Matter, vol. 19, no. 16, article 165209, Apr. 2007.
[3] A. Raychowdhury, D. Somasekhar, T. Karnik, and V. De, “Design space and scalability exploration of 1T-1STT MTJ memory arrays in the presence of variability and disturbances,” in Proc. IEEE Trans. Electron Devices, USA, 2009, pp. 7-9.
[4] H. Akinaga, and H. Shima, “Resistive random access memory (ReRAM) based on metal oxides,” in Proc. IEEE, vol. 98, no. 12, pp. 2237-2251, Dec. 2010.
[5] A. Kawahara et al., “An 8 Mb multi-layered cross-point ReRAM macro with 443 MB/s write throughput,” IEEE J. Solid-State Circuits, vol. 48, no. 1, Jan. 2013.
[6] B. C. Lee, E. Ipek, O. Mutlu, and D. Burger, “Architecting phase change memory as a scalable DRAM alternative,” ACM SIGARCH Comp. Arch. News, vol. 37, no. 3, pp. 2-13, Jun. 2009.
[7] Micro and Intel, “Intel and Micron produce breakthrough memory technology,” 2015. [Online].Available: http://newsroom.intel.com/community/intel_newsroom/blog/2015/07/28/intel-and-micron-produce-breakthrough-memory-technology.
[8] F. T. Hady, A. Foong, B. Veal, and D. Williams, “Platform storage performance with 3D Xpoint technology,” in Proc. IEEE, vol. 105, no. 9, pp. 1822-1833, Sept. 2017.
[9] P. Sehgal, S. Basu, K. Srinivasan, and K. Voruganti. An empirical study of file systems on nvm. Presented at IEEE Sym. Mass Stor. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/7208283
[10] S. R. Dulloor et al., “System software for persistent memory,” in Proc. Eur. Conf. Comp. Sys., Amsterdam, Netherlands, 2014, pp. 15.
[11] J. Xu, and S. Swanson, “NOVA: A log-structured file system for hybrid volatile/non-volatile main memories,” in Proc. USENIX Conf. File and Storage Technol. (FAST’16), USA, 2016, pp. 323-338
[12] S. Qiu and A. L. N. Reddy, “NVMFS: A hybrid file system for improving random write in nand-flash SSD,” in Proc. IEEE Sym. Mass Stor., USA, 2013, pp. 1-5.
[13] T. C. Huang, and D. W. Chang, “TridentFS: A hybrid file system for non‐volatile RAM, flash memory and magnetic disk,” Software: Practice and Experience, vol. 46, no. 3, pp. 291-318, Mar. 2016.
[14] J. Ou, J. Shu, and Y. Lu, “A high performance file system for non-volatile main memory,” in Proc. Eur. Conf. Comp. Sys., USA, 2016, article 12.
[15] Y. Zhang, and S. Swanson. A study of application performance with non-volatile main memory. Presented at IEEE Sym. Mass Stor. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/7208275
[16] E. H. M. Sha, X. Chen, Q. Zhuge, L. Shi, and W. Jiang, “A new design of in-memory file system based on file virtual address framework,” IEEE Trans. Comp., vol. 65, no. 10, pp. 2959-2972, Jan. 2016.
[17] M. Zhang, L. Zhang, L. Jiang, Z. Liu, and F. T. Chong, “Balancing performance and lifetime of MLC PCM by using a region retention monitor,” in Proc. IEEE HPCA., USA, 2017, pp. 385-396.
[18] M. Zhang, L. Zhang, L. Jiang, F. T. Chong, and Z. Liu. Quick-and-Dirty: Improving Performance of MLC PCM by Using Temporary Short Writes. Presented at IEEE Int. Conf. Comp. Des. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/8119274
[19] R. S. Liu, D. Y. Shen, C. L. Yang, S. C. Yu, and C. Y. M. Wang, “NVM duet: Unified working memory and persistent store architecture,” ACM SIGARCH Comp. Arch. News, vol. 42, no. 1, pp. 455-470, Mar. 2014.
[20] L. Sun, L. Youyou, and J. Shu, “DP2: Reducing transaction overhead with differential and dual persistency in persistent memory,” in Proc. ACM Int. Conf. Comp. Frontiers, Italy, 2015, article 24.
[21] J. Dean, and S. Ghemawat, “MapReduce: simplified data processing on large clusters,” Comm. ACM, vol. 51, no. 1, pp. 107-103, Jan 2008.
[22] J.C. Mogul, E. Argollo, M. Shah, P. Farabosch, “Operating System Support for NVM Hybrid Main Memory” In Proc. Conf. Hot topics in operating systems, USA, 2009, pp.14.
[23] C.W. Smullen, V. Mohan, A. Nigam, S. Gurumurthi, and M. R. Stan. Relaxing non-volatility for fast and energy-efficient STT-RAM caches. Presented at IEEE High Performance Comp. Arch. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/5749716
[24] Z. Sun et al. “Multi retention level STT-RAM cache designs with a dynamic refresh scheme,” in Proc. IEEE/ACM Int. Sym. Microarchitecture, Brazil, 2011, pp. 329-338.
[25] J. Coburn et al., “NV-Heaps: making persistent objects fast and safe with next-generation, non-volatile memories,” in Proc. Int. Conf. ASPLOS, USA, 2011,pp. 105-118.
[26] H. Volos, A. J. Tack, and M. M. Swift, “Mnemosyne: Lightweight persistent memory,” in Proc. Int. Conf. ASPLOS, USA, 2011, pp.91-104.
[27] R.S. Liu, C. L. Yang, and W. Wu, “Optimizing NAND flash-based SSDs via retention relaxation,” in Proc.Conf. File and Storage Technol., San Jose,USA,2012, pp.11.
[28] D.K.H. Yu, and J.W. Hsieh. Retention-Time Relaxation Scheme for MLC Flash-Memory Storage Systems. Presented at IEEE Int. Conf. Embedded and Real-Time Computing Systems and Applications. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/8607232.
[29] D. Ellard, J. Ledlie, P. Malkani, and M. Seltzer, “Passive NFS tracing of email and research workloads,” in Proc. Symp. File and Storage Technol., San Francisco, USA, pp. 203-216.
[30] R. Friedman, and D. Sainz, “File system usage in android mobile phones,” in Proc. ACM Int. Systems and Storage Conf., USA, 2016, article 16.
[31] N. Binkert et al., “The GEM5 Simulator,” SIGARCH Comp. Arch. News, vol. 39, no. 2, May 2011.
[32] J. Katcher, “Postmark: A new file system benchmark,” Network Appliance, Tech. Rep. TR3022, 1997.
[33] V. Tarasov, E. Zadok, and S. Shepler, “Filebench: A flexible framework for file system benchmarking,” USENIX; login, vol. 41, no. 1, pp. 6–12, Mar. 2016.
[34] A. Grosskurth, and M. W. Godfrey. A reference architecture for web browsers. Presented at 21st IEEE ICSM. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/1510168