Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems

Narges Shahidi, Mahmut T. Kandemir, Mohammad Arjomand, Chita R. Das, Myoungsoo Jung, Anand Sivasubramaniam

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

14 Citations (Scopus)

Abstract

In the last decade, NAND flash-based SSDs have been widely adopted for high-end enterprise systems in an attempt to provide a high-performance and reliable storage. However, inferior performance is frequently attained mainly due to the need for Garbage Collection (GC). GC in flash memory is the process of identifying and clearing the blocks of unneeded data to create space for the new data to be allocated. GC is a high-latency operation and once it is scheduled for service to a block of a plane in a flash chip (each flash chip consists of multiple planes), it can increase latency for later arriving I/O requests to the same plane. Apart from that, the consequent high latency also keep other planes of the same chip, that are not involved in this GC, idle for a long time. We show that for the baseline SSD with modern FTL, GC considerably reduces the plane-level parallelism, causing significant performance degradation. There are several circuit-level constraints that make it difficult to allow subsequent I/O operations and/or GCs to be served concurrently from the same chip, but different planes, during the long latency GC. This paper proposes a novel GC strategy, called Parallel GC (PaGC), whose goal is to proactively run GC on the remaining planes of a flash chip whenever any of its planes needs to execute on-demand GC. The resulting PaGC system boosts the response time of I/O requests by up to 45% (32% on average) for different GC settings and across a wide spectrum of enterprise I/O workloads.

Original languageEnglish
Title of host publicationProceedings of SC 2016
Subtitle of host publicationThe International Conference for High Performance Computing, Networking, Storage and Analysis
PublisherIEEE Computer Society
Pages561-572
Number of pages12
ISBN (Electronic)9781467388153
DOIs
Publication statusPublished - 2016 Jul 2
Event2016 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2016 - Salt Lake City, United States
Duration: 2016 Nov 132016 Nov 18

Publication series

NameInternational Conference for High Performance Computing, Networking, Storage and Analysis, SC
Volume0
ISSN (Print)2167-4329
ISSN (Electronic)2167-4337

Other

Other2016 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2016
CountryUnited States
CitySalt Lake City
Period16/11/1316/11/18

Fingerprint

Flash-based SSDs
Flash memory
Industry
Degradation
Networks (circuits)

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications
  • Computer Science Applications
  • Hardware and Architecture
  • Software

Cite this

Shahidi, N., Kandemir, M. T., Arjomand, M., Das, C. R., Jung, M., & Sivasubramaniam, A. (2016). Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems. In Proceedings of SC 2016: The International Conference for High Performance Computing, Networking, Storage and Analysis (pp. 561-572). [7877126] (International Conference for High Performance Computing, Networking, Storage and Analysis, SC; Vol. 0). IEEE Computer Society. https://doi.org/10.1109/SC.2016.47
Shahidi, Narges ; Kandemir, Mahmut T. ; Arjomand, Mohammad ; Das, Chita R. ; Jung, Myoungsoo ; Sivasubramaniam, Anand. / Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems. Proceedings of SC 2016: The International Conference for High Performance Computing, Networking, Storage and Analysis. IEEE Computer Society, 2016. pp. 561-572 (International Conference for High Performance Computing, Networking, Storage and Analysis, SC).
@inproceedings{38963add135245eb8c2dd7d6fd90d7fd,
title = "Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems",
abstract = "In the last decade, NAND flash-based SSDs have been widely adopted for high-end enterprise systems in an attempt to provide a high-performance and reliable storage. However, inferior performance is frequently attained mainly due to the need for Garbage Collection (GC). GC in flash memory is the process of identifying and clearing the blocks of unneeded data to create space for the new data to be allocated. GC is a high-latency operation and once it is scheduled for service to a block of a plane in a flash chip (each flash chip consists of multiple planes), it can increase latency for later arriving I/O requests to the same plane. Apart from that, the consequent high latency also keep other planes of the same chip, that are not involved in this GC, idle for a long time. We show that for the baseline SSD with modern FTL, GC considerably reduces the plane-level parallelism, causing significant performance degradation. There are several circuit-level constraints that make it difficult to allow subsequent I/O operations and/or GCs to be served concurrently from the same chip, but different planes, during the long latency GC. This paper proposes a novel GC strategy, called Parallel GC (PaGC), whose goal is to proactively run GC on the remaining planes of a flash chip whenever any of its planes needs to execute on-demand GC. The resulting PaGC system boosts the response time of I/O requests by up to 45{\%} (32{\%} on average) for different GC settings and across a wide spectrum of enterprise I/O workloads.",
author = "Narges Shahidi and Kandemir, {Mahmut T.} and Mohammad Arjomand and Das, {Chita R.} and Myoungsoo Jung and Anand Sivasubramaniam",
year = "2016",
month = "7",
day = "2",
doi = "10.1109/SC.2016.47",
language = "English",
series = "International Conference for High Performance Computing, Networking, Storage and Analysis, SC",
publisher = "IEEE Computer Society",
pages = "561--572",
booktitle = "Proceedings of SC 2016",
address = "United States",

}

Shahidi, N, Kandemir, MT, Arjomand, M, Das, CR, Jung, M & Sivasubramaniam, A 2016, Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems. in Proceedings of SC 2016: The International Conference for High Performance Computing, Networking, Storage and Analysis., 7877126, International Conference for High Performance Computing, Networking, Storage and Analysis, SC, vol. 0, IEEE Computer Society, pp. 561-572, 2016 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2016, Salt Lake City, United States, 16/11/13. https://doi.org/10.1109/SC.2016.47

Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems. / Shahidi, Narges; Kandemir, Mahmut T.; Arjomand, Mohammad; Das, Chita R.; Jung, Myoungsoo; Sivasubramaniam, Anand.

Proceedings of SC 2016: The International Conference for High Performance Computing, Networking, Storage and Analysis. IEEE Computer Society, 2016. p. 561-572 7877126 (International Conference for High Performance Computing, Networking, Storage and Analysis, SC; Vol. 0).

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

TY - GEN

T1 - Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems

AU - Shahidi, Narges

AU - Kandemir, Mahmut T.

AU - Arjomand, Mohammad

AU - Das, Chita R.

AU - Jung, Myoungsoo

AU - Sivasubramaniam, Anand

PY - 2016/7/2

Y1 - 2016/7/2

N2 - In the last decade, NAND flash-based SSDs have been widely adopted for high-end enterprise systems in an attempt to provide a high-performance and reliable storage. However, inferior performance is frequently attained mainly due to the need for Garbage Collection (GC). GC in flash memory is the process of identifying and clearing the blocks of unneeded data to create space for the new data to be allocated. GC is a high-latency operation and once it is scheduled for service to a block of a plane in a flash chip (each flash chip consists of multiple planes), it can increase latency for later arriving I/O requests to the same plane. Apart from that, the consequent high latency also keep other planes of the same chip, that are not involved in this GC, idle for a long time. We show that for the baseline SSD with modern FTL, GC considerably reduces the plane-level parallelism, causing significant performance degradation. There are several circuit-level constraints that make it difficult to allow subsequent I/O operations and/or GCs to be served concurrently from the same chip, but different planes, during the long latency GC. This paper proposes a novel GC strategy, called Parallel GC (PaGC), whose goal is to proactively run GC on the remaining planes of a flash chip whenever any of its planes needs to execute on-demand GC. The resulting PaGC system boosts the response time of I/O requests by up to 45% (32% on average) for different GC settings and across a wide spectrum of enterprise I/O workloads.

AB - In the last decade, NAND flash-based SSDs have been widely adopted for high-end enterprise systems in an attempt to provide a high-performance and reliable storage. However, inferior performance is frequently attained mainly due to the need for Garbage Collection (GC). GC in flash memory is the process of identifying and clearing the blocks of unneeded data to create space for the new data to be allocated. GC is a high-latency operation and once it is scheduled for service to a block of a plane in a flash chip (each flash chip consists of multiple planes), it can increase latency for later arriving I/O requests to the same plane. Apart from that, the consequent high latency also keep other planes of the same chip, that are not involved in this GC, idle for a long time. We show that for the baseline SSD with modern FTL, GC considerably reduces the plane-level parallelism, causing significant performance degradation. There are several circuit-level constraints that make it difficult to allow subsequent I/O operations and/or GCs to be served concurrently from the same chip, but different planes, during the long latency GC. This paper proposes a novel GC strategy, called Parallel GC (PaGC), whose goal is to proactively run GC on the remaining planes of a flash chip whenever any of its planes needs to execute on-demand GC. The resulting PaGC system boosts the response time of I/O requests by up to 45% (32% on average) for different GC settings and across a wide spectrum of enterprise I/O workloads.

UR - http://www.scopus.com/inward/record.url?scp=85017265462&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85017265462&partnerID=8YFLogxK

U2 - 10.1109/SC.2016.47

DO - 10.1109/SC.2016.47

M3 - Conference contribution

AN - SCOPUS:85017265462

T3 - International Conference for High Performance Computing, Networking, Storage and Analysis, SC

SP - 561

EP - 572

BT - Proceedings of SC 2016

PB - IEEE Computer Society

ER -

Shahidi N, Kandemir MT, Arjomand M, Das CR, Jung M, Sivasubramaniam A. Exploring the Potentials of Parallel Garbage Collection in SSDs for Enterprise Storage Systems. In Proceedings of SC 2016: The International Conference for High Performance Computing, Networking, Storage and Analysis. IEEE Computer Society. 2016. p. 561-572. 7877126. (International Conference for High Performance Computing, Networking, Storage and Analysis, SC). https://doi.org/10.1109/SC.2016.47