B+-tree index optimization by exploiting internal parallelism of flash-based solid state drives

Hongchan Roh, Sang Hyun Park, Sungho Kim, Mincheol Shin, Sang Won Lee

Research output: Contribution to journalArticle

44 Citations (Scopus)

Abstract

Previous research addressed the potential problems of the harddisk oriented design of DBMSs of flashSSDs. In this paper, we focus on exploiting potential benefits of flashSSDs. First, we examine the internal parallelism issues of flashSSDs by conducting benchmarks to various flashSSDs. Then, we suggest algorithm-design principles in order to best benefit from the internal parallelism. We present a new I/O request concept, called psync I/O that can exploit the internal parallelism of flashSSDs in a single process. Based on these ideas, we introduce B+-tree optimization methods in order to utilize internal parallelism. By integrating the results of these methods, we present a B+-tree variant, PIO B-tree. We confirmed that each optimization method substantially enhances the index performance. Consequently, PIO B-tree enhanced B+-tree's insert performance by a factor of up to 16.3, while improving point-search performance by a factor of 1.2. The range search of PIO B-tree was up to 5 times faster than that of the B+-tree. Moreover, PIO B-tree outperformed other flashaware indexes in various synthetic workloads. We also confirmed that PIO B-tree outperforms B+-tree in index traces collected inside the Postgresql DBMS with TPC-C benchmark.

Original languageEnglish
Pages (from-to)286-297
Number of pages12
JournalProceedings of the VLDB Endowment
Volume5
Issue number4
DOIs
Publication statusPublished - 2011 Jan 1

Fingerprint

Flash-based SSDs

All Science Journal Classification (ASJC) codes

  • Computer Science (miscellaneous)
  • Computer Science(all)

Cite this

Roh, Hongchan ; Park, Sang Hyun ; Kim, Sungho ; Shin, Mincheol ; Lee, Sang Won. / B+-tree index optimization by exploiting internal parallelism of flash-based solid state drives. In: Proceedings of the VLDB Endowment. 2011 ; Vol. 5, No. 4. pp. 286-297.
@article{31684fcb8bdd4e5d8f795f461e9915d7,
title = "B+-tree index optimization by exploiting internal parallelism of flash-based solid state drives",
abstract = "Previous research addressed the potential problems of the harddisk oriented design of DBMSs of flashSSDs. In this paper, we focus on exploiting potential benefits of flashSSDs. First, we examine the internal parallelism issues of flashSSDs by conducting benchmarks to various flashSSDs. Then, we suggest algorithm-design principles in order to best benefit from the internal parallelism. We present a new I/O request concept, called psync I/O that can exploit the internal parallelism of flashSSDs in a single process. Based on these ideas, we introduce B+-tree optimization methods in order to utilize internal parallelism. By integrating the results of these methods, we present a B+-tree variant, PIO B-tree. We confirmed that each optimization method substantially enhances the index performance. Consequently, PIO B-tree enhanced B+-tree's insert performance by a factor of up to 16.3, while improving point-search performance by a factor of 1.2. The range search of PIO B-tree was up to 5 times faster than that of the B+-tree. Moreover, PIO B-tree outperformed other flashaware indexes in various synthetic workloads. We also confirmed that PIO B-tree outperforms B+-tree in index traces collected inside the Postgresql DBMS with TPC-C benchmark.",
author = "Hongchan Roh and Park, {Sang Hyun} and Sungho Kim and Mincheol Shin and Lee, {Sang Won}",
year = "2011",
month = "1",
day = "1",
doi = "10.14778/2095686.2095688",
language = "English",
volume = "5",
pages = "286--297",
journal = "Proceedings of the VLDB Endowment",
issn = "2150-8097",
publisher = "Very Large Data Base Endowment Inc.",
number = "4",

}

B+-tree index optimization by exploiting internal parallelism of flash-based solid state drives. / Roh, Hongchan; Park, Sang Hyun; Kim, Sungho; Shin, Mincheol; Lee, Sang Won.

In: Proceedings of the VLDB Endowment, Vol. 5, No. 4, 01.01.2011, p. 286-297.

Research output: Contribution to journalArticle

TY - JOUR

T1 - B+-tree index optimization by exploiting internal parallelism of flash-based solid state drives

AU - Roh, Hongchan

AU - Park, Sang Hyun

AU - Kim, Sungho

AU - Shin, Mincheol

AU - Lee, Sang Won

PY - 2011/1/1

Y1 - 2011/1/1

N2 - Previous research addressed the potential problems of the harddisk oriented design of DBMSs of flashSSDs. In this paper, we focus on exploiting potential benefits of flashSSDs. First, we examine the internal parallelism issues of flashSSDs by conducting benchmarks to various flashSSDs. Then, we suggest algorithm-design principles in order to best benefit from the internal parallelism. We present a new I/O request concept, called psync I/O that can exploit the internal parallelism of flashSSDs in a single process. Based on these ideas, we introduce B+-tree optimization methods in order to utilize internal parallelism. By integrating the results of these methods, we present a B+-tree variant, PIO B-tree. We confirmed that each optimization method substantially enhances the index performance. Consequently, PIO B-tree enhanced B+-tree's insert performance by a factor of up to 16.3, while improving point-search performance by a factor of 1.2. The range search of PIO B-tree was up to 5 times faster than that of the B+-tree. Moreover, PIO B-tree outperformed other flashaware indexes in various synthetic workloads. We also confirmed that PIO B-tree outperforms B+-tree in index traces collected inside the Postgresql DBMS with TPC-C benchmark.

AB - Previous research addressed the potential problems of the harddisk oriented design of DBMSs of flashSSDs. In this paper, we focus on exploiting potential benefits of flashSSDs. First, we examine the internal parallelism issues of flashSSDs by conducting benchmarks to various flashSSDs. Then, we suggest algorithm-design principles in order to best benefit from the internal parallelism. We present a new I/O request concept, called psync I/O that can exploit the internal parallelism of flashSSDs in a single process. Based on these ideas, we introduce B+-tree optimization methods in order to utilize internal parallelism. By integrating the results of these methods, we present a B+-tree variant, PIO B-tree. We confirmed that each optimization method substantially enhances the index performance. Consequently, PIO B-tree enhanced B+-tree's insert performance by a factor of up to 16.3, while improving point-search performance by a factor of 1.2. The range search of PIO B-tree was up to 5 times faster than that of the B+-tree. Moreover, PIO B-tree outperformed other flashaware indexes in various synthetic workloads. We also confirmed that PIO B-tree outperforms B+-tree in index traces collected inside the Postgresql DBMS with TPC-C benchmark.

UR - http://www.scopus.com/inward/record.url?scp=84863758568&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84863758568&partnerID=8YFLogxK

U2 - 10.14778/2095686.2095688

DO - 10.14778/2095686.2095688

M3 - Article

VL - 5

SP - 286

EP - 297

JO - Proceedings of the VLDB Endowment

JF - Proceedings of the VLDB Endowment

SN - 2150-8097

IS - 4

ER -