Analysis of ssd internal cache problem in a key-value store system

Won Seob Jeong, Yongseok Won, Won Woo Ro

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, we analyze how the journaling mechanism on a keyvalue store has an impact of the performance of a Solid State Drive (SSD). Journaling is a widely used technique in a modern database management system to provide high reliability and fast recovery from system failures. However, journaling mechanism writes the same data twice to both the journal area and data area and this write amplification causes performance degradation by increasing the storage traffic. Eliminating the impact of the write amplification is very challenging. This observation motivated us to change the approach for solving the performance degradation. Instead of handling the problem in system level, we focused on internal SSD level. For the purpose, we analyze the performance bottleneck of an SSD when journaling mechanism is used for a key-value store. A key-value store system model is proposed for the analysis. This paper reveals that frequent write operation to journal area can cause serious SSD performance degradation by increasing the overhead of cache line eviction.

Original languageEnglish
Title of host publicationProceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019
PublisherAssociation for Computing Machinery
Pages59-62
Number of pages4
ISBN (Electronic)9781450366427
DOIs
Publication statusPublished - 2019 Jan 10
Event2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - and its Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019 - Bali, Indonesia
Duration: 2019 Jan 102019 Jan 13

Publication series

NameACM International Conference Proceeding Series

Conference

Conference2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - and its Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019
CountryIndonesia
CityBali
Period19/1/1019/1/13

Fingerprint

Degradation
Amplification
Recovery

All Science Journal Classification (ASJC) codes

  • Software
  • Human-Computer Interaction
  • Computer Vision and Pattern Recognition
  • Computer Networks and Communications

Cite this

Jeong, W. S., Won, Y., & Ro, W. W. (2019). Analysis of ssd internal cache problem in a key-value store system. In Proceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019 (pp. 59-62). (ACM International Conference Proceeding Series). Association for Computing Machinery. https://doi.org/10.1145/3305160.3305213
Jeong, Won Seob ; Won, Yongseok ; Ro, Won Woo. / Analysis of ssd internal cache problem in a key-value store system. Proceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019. Association for Computing Machinery, 2019. pp. 59-62 (ACM International Conference Proceeding Series).
@inproceedings{95ba2acf46a743638572a457e6c4771a,
title = "Analysis of ssd internal cache problem in a key-value store system",
abstract = "In this paper, we analyze how the journaling mechanism on a keyvalue store has an impact of the performance of a Solid State Drive (SSD). Journaling is a widely used technique in a modern database management system to provide high reliability and fast recovery from system failures. However, journaling mechanism writes the same data twice to both the journal area and data area and this write amplification causes performance degradation by increasing the storage traffic. Eliminating the impact of the write amplification is very challenging. This observation motivated us to change the approach for solving the performance degradation. Instead of handling the problem in system level, we focused on internal SSD level. For the purpose, we analyze the performance bottleneck of an SSD when journaling mechanism is used for a key-value store. A key-value store system model is proposed for the analysis. This paper reveals that frequent write operation to journal area can cause serious SSD performance degradation by increasing the overhead of cache line eviction.",
author = "Jeong, {Won Seob} and Yongseok Won and Ro, {Won Woo}",
year = "2019",
month = "1",
day = "10",
doi = "10.1145/3305160.3305213",
language = "English",
series = "ACM International Conference Proceeding Series",
publisher = "Association for Computing Machinery",
pages = "59--62",
booktitle = "Proceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019",

}

Jeong, WS, Won, Y & Ro, WW 2019, Analysis of ssd internal cache problem in a key-value store system. in Proceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019. ACM International Conference Proceeding Series, Association for Computing Machinery, pp. 59-62, 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - and its Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019, Bali, Indonesia, 19/1/10. https://doi.org/10.1145/3305160.3305213

Analysis of ssd internal cache problem in a key-value store system. / Jeong, Won Seob; Won, Yongseok; Ro, Won Woo.

Proceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019. Association for Computing Machinery, 2019. p. 59-62 (ACM International Conference Proceeding Series).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Analysis of ssd internal cache problem in a key-value store system

AU - Jeong, Won Seob

AU - Won, Yongseok

AU - Ro, Won Woo

PY - 2019/1/10

Y1 - 2019/1/10

N2 - In this paper, we analyze how the journaling mechanism on a keyvalue store has an impact of the performance of a Solid State Drive (SSD). Journaling is a widely used technique in a modern database management system to provide high reliability and fast recovery from system failures. However, journaling mechanism writes the same data twice to both the journal area and data area and this write amplification causes performance degradation by increasing the storage traffic. Eliminating the impact of the write amplification is very challenging. This observation motivated us to change the approach for solving the performance degradation. Instead of handling the problem in system level, we focused on internal SSD level. For the purpose, we analyze the performance bottleneck of an SSD when journaling mechanism is used for a key-value store. A key-value store system model is proposed for the analysis. This paper reveals that frequent write operation to journal area can cause serious SSD performance degradation by increasing the overhead of cache line eviction.

AB - In this paper, we analyze how the journaling mechanism on a keyvalue store has an impact of the performance of a Solid State Drive (SSD). Journaling is a widely used technique in a modern database management system to provide high reliability and fast recovery from system failures. However, journaling mechanism writes the same data twice to both the journal area and data area and this write amplification causes performance degradation by increasing the storage traffic. Eliminating the impact of the write amplification is very challenging. This observation motivated us to change the approach for solving the performance degradation. Instead of handling the problem in system level, we focused on internal SSD level. For the purpose, we analyze the performance bottleneck of an SSD when journaling mechanism is used for a key-value store. A key-value store system model is proposed for the analysis. This paper reveals that frequent write operation to journal area can cause serious SSD performance degradation by increasing the overhead of cache line eviction.

UR - http://www.scopus.com/inward/record.url?scp=85063582971&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063582971&partnerID=8YFLogxK

U2 - 10.1145/3305160.3305213

DO - 10.1145/3305160.3305213

M3 - Conference contribution

AN - SCOPUS:85063582971

T3 - ACM International Conference Proceeding Series

SP - 59

EP - 62

BT - Proceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019

PB - Association for Computing Machinery

ER -

Jeong WS, Won Y, Ro WW. Analysis of ssd internal cache problem in a key-value store system. In Proceedings of the 2019 2nd International Conference on Software Engineering and Information Management, ICSIM 2019 - Workshop 2019 2nd International Conference on Big Data and Smart Computing, ICBDSC 2019. Association for Computing Machinery. 2019. p. 59-62. (ACM International Conference Proceeding Series). https://doi.org/10.1145/3305160.3305213