DOI QR코드

DOI QR Code

Performance analysis and prediction through various over-provision on NAND flash memory based storage

낸드 플래시 메모리기반 저장 장치에서 다양한 초과 제공을 통한 성능 분석 및 예측

  • Lee, Hyun-Seob (Division of Computer Engineering, Baekseok University)
  • 이현섭 (백석대학교 컴퓨터공학부)
  • Received : 2022.02.26
  • Accepted : 2022.03.20
  • Published : 2022.03.28

Abstract

Recently, With the recent rapid development of technology, the amount of data generated by various systems is increasing, and enterprise servers and data centers that have to handle large amounts of big data need to apply high-stability and high-performance storage devices even if costs increase. In such systems, SSD(solid state disk) that provide high performance of read/write are often used as storage devices. However, due to the characteristics of reading and writing on a page-by-page basis, erasing operations on a block basis, and erassing-before-writing, there is a problem that performance is degraded when duplicate writes occur. Therefore, in order to delay this performance degradation problem, over-provision technology of SSD has been applied internally. However, since over-provided technologies have the disadvantage of consuming a lot of storage space instead of performance, the application of inefficient technologies above the right performance has a problem of over-costing. In this paper, we proposed a method of measuring the performance and cost incurred when various over-provisions are applied in an SSD and predicting the system-optimized over-provided ratio based on this. Through this research, we expect to find a trade-off with costs to meet the performance requirements in systems that process big data.

최근 급격한 기술의 발달로 다양한 시스템에서 발생하는 데이터양이 증가하고 있으며, 많은 양의 빅데이터(big data)를 처리해야 하는 엔터프라이즈 서버(enterprise server)와 데이터 센터(data center)의 경우 비용이 증가하더라도 높은 안정성과 고성능의 저장 장치를 적용하는 것이 필요하다. 이러한 시스템에서는 고성능의 읽기/쓰기 성능을 제공하는 SSD(solid state disk)를 저장 장치로 사용하는 경우가 많다. 그러나, 페이지 단위로 읽기 쓰기를 하고 블록단위로 지우기 연산을 해야하고 쓰기 전 지우기 연산을 수행해야 하는 특징 때문에 중복 쓰기가 다발할 경우 성능이 저하되는 문제가 있다. 따라서 이러한 성능 저하 문제를 지연시키기 위해 SSD의 내부적으로 초과 제공(over-provision) 기술을 적용하고 있다. 그러나 초과 제공 기술은 성능 대신 많은 저장공간의 비용을 소모하는 단점이 있기 때문에 적정 성능 이상의 비효율적인 기술의 적용은 과대한 비용을 지불하게 만드는 문제가 있다. 본 논문에서는 SSD에서 다양한 초과 제공을 적용하였을 때 발생하는 성능과 비용을 측정하고, 이를 기반으로 시스템에 최적화된 초과 제공 비율을 예측하는 방법을 제안했다. 본 연구를 통해 빅데이터를 처리하는 시스템에서 성능의 요구사항을 만족하기 위한 비용과의 절충점(trade-off)를 찾을 수 있을 것으로 기대한다.

Keywords

Acknowledgement

This paper was supported by 2022 Baekseok University Research Fund

References

  1. S. S. Chae, R. Mativenga, J. Y. Paik, M. Attique & T. S. Chung. (2020). DSFTL: An efficient FTL for flash memory based storage systems. Electronics 9(1), 145, https://doi.org/10.3390/electronics9010145
  2. W. Xie, Y. Chen & P. C. Poth. (2017). ASA-FTL: An adaptive separation aware flash translation layer for solid state drives. Parallel Computing, 61. 3-17. https://doi.org/10.1016/j.parco.2016.10.006
  3. I. B. Zion. (2020). Key-value FTL over open channel SSD, 12th ACM International Conference on Systems and Storage. 192-192.
  4. H. S. Lee, H. S. Yun & D. H. Lee (2011). HFTL: hybrid flash translation layer based on hot data identification for flash memory. IEEE Transactions on Consumer Electronics, (4), 2005-2011. https://doi.org/10.1109/TCE.2009.5373762
  5. H. S. Lee & D. H. Lee. (2010). An efficient index buffer management scheme for implementing a B-tree on NAND flash memory. Data & Knowledge Engineering. 69(9), 901-916. https://doi.org/10.1016/j.datak.2010.03.004
  6. H. S. Lee, S. W. Park & D. H. Lee (2013). RMSS: an efficient recovery management scheme on NAND flash memory based solid state disk. IEEE Transactions on Consumer Electronics, 59(1), 107-112. https://doi.org/10.1109/TCE.2013.6490248
  7. S. Kim & Y. Son. (2021). Optimizing Key-Value Stores for Flash-Based SSDs via Key Reshaping. IEEE Access 9, 115135~115144. https://doi.org/10.1109/ACCESS.2021.3105428
  8. J. H. Park, D. J. Park, T. S. Chung & S. W. Lee. (2021). A Crash Recovery Scheme for a Hybrid Mapping FTL in NAND Flash Storage Devices. Electronics, 10(3), 327. https://doi.org/10.3390/electronics10030327
  9. M. Sacks. (2021). Incentives for the over-provision of public goods. Journal of Economic Behavior & Organization, 191, 197-213. https://doi.org/10.1016/j.jebo.2021.08.033
  10. R. Liu, Z. Tan, L. Long, Y. Wu, Y. Tan & D. Liu. (2022) Improving Fairness for SSD Devices through DRAM Over-Provisioning Cache Management. IEEE Transactions on Parallel and Distributed Systems, 1-1.
  11. Q. Li et al. (2021). RAMBO: Resource Allocation for Microservices Using Bayesian Optimization. IEEE Computer Architecture Letters, 20(1), 46-49. https://doi.org/10.1109/LCA.2021.3066142
  12. S. Elashri & A. Azim. (2021). Work-in-Progress: An Energy-Aware Optimization Model for Real-Time Systems Analysis and Design. International Conference on Embedded Software (EMSOFT), 45-46.
  13. H. Shen & L. Chen. (2018). Resource Demand Misalignment: An Important Factor to Consider for Reducing Resource Over-Provisioning in Cloud Datacenters. IEEE/ACM Transactions on Networking, 26(3), 1207-1221. https://doi.org/10.1109/tnet.2018.2823642
  14. L. Chen & H. Shen. (2017). Considering resource demand misalignments to reduce resource over-provisioning in cloud datacenters. IEEE Conference on Computer Communications. 1-9.
  15. D. Bega, M. Gramaglia, M. Fiore, A. Banchs & X. Costa-Perez. (2020). DeepCog: Optimizing Resource Provisioning in Network Slicing With AI-Based Capacity Forecasting. IEEE Journal on Selected Areas in Communications, 38(2), 361-376. https://doi.org/10.1109/jsac.2019.2959245