Reinforcement learning based background segment cleaning for log-structured file system on mobile devices

Chao Wu; Cheng Ji; Chun Jason Xue

doi:10.1109/ICESS.2019.8782508

Reinforcement learning based background segment cleaning for log-structured file system on mobile devices

Department of Computer Science

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

11 Citations (Scopus)

Abstract

With the adoption of Log-structured file system in mobile devices, the impact of background segment cleaning on system performance and storage lifetime becomes notable. Aggressive background segment cleaning solution generates excessive block migrations and impairs the endurance of NAND storage device, while a lazy solution cannot reclaim enough segments for subsequent I/O requests thus leading to the occurrence of foreground segment cleaning and prolonging I/O latency. In this paper, a reinforcement learning based approach is proposed to balance the trade-off. Through learning the behaviors of I/O workloads and the statuses of logical address space, the proposed approach can adaptively reduce the frequency of foreground segment cleaning by 68.57% on average, and decrease the number of block migrations by 71.10% over existing approaches.

Original language	English
Title of host publication	2019 IEEE International Conference on Embedded Software and Systems (ICESS)
Publisher	IEEE
ISBN (Electronic)	9781728124377
DOIs	https://doi.org/10.1109/ICESS.2019.8782508
Publication status	Published - Jun 2019
Event	15th IEEE International Conference on Embedded Software and Systems (ICESS 2019) - Las Vegas Convention Center, Nevada, United States Duration: 2 Jun 2019 → 3 Jun 2019 http://lcs.ios.ac.cn/icess2019/

Publication series

Name	IEEE International Conference on Embedded Software and Systems, ICESS

Conference

Conference	15th IEEE International Conference on Embedded Software and Systems (ICESS 2019)
Abbreviated title	ICESS 2019
Country/Territory	United States
City	Nevada
Period	2/06/19 → 3/06/19
Internet address	http://lcs.ios.ac.cn/icess2019/

Research Keywords

Endurance
Log-structured file system
Mobile device
Performance
Reinforcement learning
Segment cleaning

Access Output

10.1109/ICESS.2019.8782508

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{3867a120fac24eaab65828b9d7c27edb,

title = "Reinforcement learning based background segment cleaning for log-structured file system on mobile devices",

abstract = "With the adoption of Log-structured file system in mobile devices, the impact of background segment cleaning on system performance and storage lifetime becomes notable. Aggressive background segment cleaning solution generates excessive block migrations and impairs the endurance of NAND storage device, while a lazy solution cannot reclaim enough segments for subsequent I/O requests thus leading to the occurrence of foreground segment cleaning and prolonging I/O latency. In this paper, a reinforcement learning based approach is proposed to balance the trade-off. Through learning the behaviors of I/O workloads and the statuses of logical address space, the proposed approach can adaptively reduce the frequency of foreground segment cleaning by 68.57% on average, and decrease the number of block migrations by 71.10% over existing approaches.",

keywords = "Endurance, Log-structured file system, Mobile device, Performance, Reinforcement learning, Segment cleaning",

author = "Chao Wu and Cheng Ji and Xue, {Chun Jason}",

year = "2019",

month = jun,

doi = "10.1109/ICESS.2019.8782508",

language = "English",

series = "IEEE International Conference on Embedded Software and Systems, ICESS",

publisher = "IEEE",

booktitle = "2019 IEEE International Conference on Embedded Software and Systems (ICESS)",

address = "United States",

note = "15th IEEE International Conference on Embedded Software and Systems (ICESS 2019), ICESS 2019 ; Conference date: 02-06-2019 Through 03-06-2019",

url = "http://lcs.ios.ac.cn/icess2019/",

}

Wu, C , Ji, C & Xue, CJ 2019, Reinforcement learning based background segment cleaning for log-structured file system on mobile devices. in 2019 IEEE International Conference on Embedded Software and Systems (ICESS)., 8782508, IEEE International Conference on Embedded Software and Systems, ICESS, IEEE, 15th IEEE International Conference on Embedded Software and Systems (ICESS 2019), Nevada, United States, 2/06/19. https://doi.org/10.1109/ICESS.2019.8782508

Reinforcement learning based background segment cleaning for log-structured file system on mobile devices. / Wu, Chao ; Ji, Cheng ; Xue, Chun Jason.
2019 IEEE International Conference on Embedded Software and Systems (ICESS). IEEE, 2019. 8782508 (IEEE International Conference on Embedded Software and Systems, ICESS).

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - Reinforcement learning based background segment cleaning for log-structured file system on mobile devices

AU - Wu, Chao

AU - Ji, Cheng

AU - Xue, Chun Jason

PY - 2019/6

Y1 - 2019/6

N2 - With the adoption of Log-structured file system in mobile devices, the impact of background segment cleaning on system performance and storage lifetime becomes notable. Aggressive background segment cleaning solution generates excessive block migrations and impairs the endurance of NAND storage device, while a lazy solution cannot reclaim enough segments for subsequent I/O requests thus leading to the occurrence of foreground segment cleaning and prolonging I/O latency. In this paper, a reinforcement learning based approach is proposed to balance the trade-off. Through learning the behaviors of I/O workloads and the statuses of logical address space, the proposed approach can adaptively reduce the frequency of foreground segment cleaning by 68.57% on average, and decrease the number of block migrations by 71.10% over existing approaches.

AB - With the adoption of Log-structured file system in mobile devices, the impact of background segment cleaning on system performance and storage lifetime becomes notable. Aggressive background segment cleaning solution generates excessive block migrations and impairs the endurance of NAND storage device, while a lazy solution cannot reclaim enough segments for subsequent I/O requests thus leading to the occurrence of foreground segment cleaning and prolonging I/O latency. In this paper, a reinforcement learning based approach is proposed to balance the trade-off. Through learning the behaviors of I/O workloads and the statuses of logical address space, the proposed approach can adaptively reduce the frequency of foreground segment cleaning by 68.57% on average, and decrease the number of block migrations by 71.10% over existing approaches.

KW - Endurance

KW - Log-structured file system

KW - Mobile device

KW - Performance

KW - Reinforcement learning

KW - Segment cleaning

UR - http://www.scopus.com/inward/record.url?scp=85070907311&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85070907311&origin=recordpage

U2 - 10.1109/ICESS.2019.8782508

DO - 10.1109/ICESS.2019.8782508

M3 - RGC 32 - Refereed conference paper (with host publication)

T3 - IEEE International Conference on Embedded Software and Systems, ICESS

BT - 2019 IEEE International Conference on Embedded Software and Systems (ICESS)

PB - IEEE

T2 - 15th IEEE International Conference on Embedded Software and Systems (ICESS 2019)

Y2 - 2 June 2019 through 3 June 2019

ER -

Reinforcement learning based background segment cleaning for log-structured file system on mobile devices

Abstract

Publication series

Conference

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this