Some previous works have been proposed to utilize the idle time of a workload to improve the system performance of solid-state drives (SSDs). This is because SSDs that adopt NAND flash memory will face two critical issues: (1) ”out-of place updates” and (2) ”garbage collection (GC)”. However, the GC activities will cause the long-tail latency and degrade the system performance. Therefore, how to correctly detect the idle time of a workload to perform the appropriate operations in SSDs (e.g., the GC activities) is a key point.
In the paper, we will propose an idle time detection strategy to predict the idle time of the real workloads. The detection strategy is to apply three proposed voting methods as investors to the revised Martingale and Anti-Martingale models that are used in economics for investment transactions. The experimental results also show that the prediction accuracy of the proposed strategy can be stably about 80% in average.

[1] Q. Wei, B. Gong, S. Pathak, B. Veeravalli, L. Zeng, and K. Okada“Waftl: A workload adaptive flash translation layer with data partition,” in 2011 IEEE 27th Symposium on Mass Storage Systems and Technologies (MSST), (Denver, CO, USA), pp. 1–12, IEEE, IEEE, May 2011.
[2] Y. Wang, Z. Qin, R. Chen, Z. Shao, Q. Wang, S. Li, and L. T. Yang, “A real-time flash translation layer for nand flash memory storage systems,” IEEE Transactions on Multi-Scale Computing Systems, vol. 2, no. 1, pp. 17–29, 2016.
[3] Q. Zhang, X. Li, L. Wang, T. Zhang, Y. Wang, and Z. Shao, “Lazy-rtgc: A real-time lazy garbage collection mechanism with jointly optimizing average and worst performance for nand flash memory storage systems,” ACM Transactions on Design Automation of Electronic Systems (TODAES), vol. 20, no. 3, pp. 1–32, 2015.
[4] W. Kang, D. Shin, and S. Yoo, “Reinforcement learning-assisted garbage collection to mitigate long-tail latency in ssd,” ACM Transactions on Embedded Computing Systems (TECS), vol. 16, no. 5s, pp. 1–20, 2017.
[5] W. Kang and S. Yoo, “Dynamic management of key states for reinforcement learning-assisted garbage collection to reduce long tail latency in ssd,” in Proceedings of the 55th Annual Design Automation Conference, DAC ’18, (New York, NY, USA), pp. 1–6, Association for Computing Machinery, 2018.
[6] W. Kang and S. Yoo, “Q-value prediction for reinforcement learning assisted garbage collection to reduce long tail latency in ssd,” IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems(Early Access), vol. 0, no. 0, pp. 1–1, 2019.
[7] J. Li, B. Huang, Z. Sha, Z. Cai, J. Liao, B. Gerofi, and Y. Ishikawa, “Mitigating negative impacts of read disturb in ssds,” ACM Transactions on Design Automation of Electronic Systems (TODAES), vol. 26, no. 1, pp. 1–24, 2020.
[8] C.-Y. Liu, Y.-M. Chang, and Y.-H. Chang, “Read leveling for flash storage systems,” in Proceedings of the 8th ACM International Systems and Storage Conference, pp. 1–10, 2015.
[9] N. J. Balsara, Money management strategies for futures traders, vol. 4. John Wiley & Sons, 1992.
[10] R. L. Weissman, Mechanical trading systems: Pairing trader psychology with technical analysis, vol. 220. John Wiley & Sons, 2005.
[11] K. T. Teoh, S. P. Y. Ng, H. R. Khoo, S. Y. Voon, and P. L. Yeep, Effectiveness of martingale strategy in gambling and investment. PhD thesis, UTAR, 2015.
[12] “Pyramid Trading Strategy-How To Turn Small Trades into Huge Trades.” http://www.learntotradethemarket.com/forex-tradingstrategies/ pyramid-money-management-strategy.
[13] “How to use Pyramid Trading to Build on Winners.” https://forexop.com/strategy/pyramid-trading/.
[14] W. Bux and I. Iliadis, “Performance of greedy garbage collection in flash-based solid-state drives,” Performance Evaluation, vol. 67, no. 11, pp. 1172–1186, 2010.
[15] R. S. Sutton and A. G. Barto, Reinforcement learning: An introduction. Adaptive Computation and Machine Learning series, Cambridge, MA, USA: MIT press, 2018.
[16] L. Duan, D. Zhan, and J. Hohnerlein, “Optimizing cloud data center energy efficiency via dynamic prediction of cpu idle intervals,” in 2015 IEEE 8th International Conference on Cloud Computing, pp. 985–988, IEEE, 2015.
[17] M. S. Hossain and I. Savidis, “Dynamic idle core management and leakage current reuse in mpsoc platforms,” in Proceedings of the ACM/IEEE International Symposium on Low Power Electronics and Design, pp. 49–54, 2020.
[18] C. Fu, Y. Zhao, M. Li, and C. J. Xue, “Maximizing common idle time on multicore processors with shared memory,” IEEE Transactions on Very Large Scale Integration (VLSI) Systems, vol. 25, no. 7, pp. 2095–2108, 2017.
[19] C. Kim, D.-H. Kim, W. Jeong, H.-J. Kim, I. H. Park, H.-W. Park, J. Lee, J. Park, Y.-L. Ahn, J. Y. Lee, et al., “A 512-gb 3-b/cell 64-stacked wl 3-d-nand flash memory,” IEEE Journal of Solid-State Circuits, vol. 53, pp. 124–133, Jan 2017.
[20] M. Kwon, J. Zhang, G. Park, W. Choi, D. Donofrio, J. Shalf, M. Kandemir, and M. Jung, “Tracetracker: Hardware/software co-evaluation for large-scale i/o workload reconstruction,” in 2017 IEEE International Symposium on Workload Characterization (IISWC), (Seattle, WA, USA), pp. 87–96, IEEE, IEEE,Oct2017
[21] S. P. Council, “Spc trace file format specification,” 2002.
[22] SINA, “Snia iotta trace repository,” 2020.