Optimal data-space partitioning of spatial data for parallel I/O

Hakan Ferhatosmanoǧlu, Divyakant Agrawal, Ömer Eǧecioǧlu, Amr El Abbadi

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

It is desirable to design partitioning methods that minimize the I/O time incurred during query execution in spatial databases. This paper explores optimal partitioning for two-dimensional data for a class of queries and develops multi-disk allocation techniques that maximize the degree of I/O parallelism obtained in each case. We show that hexagonal partitioning has optimal I/O performance for circular queries among all partitioning methods that use convex non-overlapping regions. An analysis and extension of this result to all possible partitioning techniques is also given. For rectangular queries, we show that hexagonal partitioning has overall better I/O performance for a general class of range queries, except for rectilinear queries, in which case rectangular grid partitioning is superior. By using current algorithms for rectangular grid partitioning, parallel storage and retrieval algorithms for hexagonal partitioning can be constructed. Some of these results carry over to circular partitioning of the data-which is an example of a non-convex region.

Original languageEnglish
Pages (from-to)75-101
Number of pages27
JournalDistributed and Parallel Databases
Volume17
Issue number1
DOIs
Publication statusPublished - 1 Jan 2005
Externally publishedYes

Fingerprint

Parallel I/O
Spatial Data
Partitioning
Query
Hexagon
Grid
Spatial Database
Range Query
Parallelism
Retrieval
Maximise

Keywords

  • Data-space partitioning
  • Disk and page allocation
  • Parallel I/O
  • Range query
  • Two-dimensional data

ASJC Scopus subject areas

  • Information Systems
  • Theoretical Computer Science
  • Computational Theory and Mathematics

Cite this

Optimal data-space partitioning of spatial data for parallel I/O. / Ferhatosmanoǧlu, Hakan; Agrawal, Divyakant; Eǧecioǧlu, Ömer; El Abbadi, Amr.

In: Distributed and Parallel Databases, Vol. 17, No. 1, 01.01.2005, p. 75-101.

Research output: Contribution to journalArticle

Ferhatosmanoǧlu, H, Agrawal, D, Eǧecioǧlu, Ö & El Abbadi, A 2005, 'Optimal data-space partitioning of spatial data for parallel I/O', Distributed and Parallel Databases, vol. 17, no. 1, pp. 75-101. https://doi.org/10.1023/B:DAPD.0000045550.56749.75
Ferhatosmanoǧlu, Hakan ; Agrawal, Divyakant ; Eǧecioǧlu, Ömer ; El Abbadi, Amr. / Optimal data-space partitioning of spatial data for parallel I/O. In: Distributed and Parallel Databases. 2005 ; Vol. 17, No. 1. pp. 75-101.
@article{026fb6010b3f45aeabe9ad35e79d7df9,
title = "Optimal data-space partitioning of spatial data for parallel I/O",
abstract = "It is desirable to design partitioning methods that minimize the I/O time incurred during query execution in spatial databases. This paper explores optimal partitioning for two-dimensional data for a class of queries and develops multi-disk allocation techniques that maximize the degree of I/O parallelism obtained in each case. We show that hexagonal partitioning has optimal I/O performance for circular queries among all partitioning methods that use convex non-overlapping regions. An analysis and extension of this result to all possible partitioning techniques is also given. For rectangular queries, we show that hexagonal partitioning has overall better I/O performance for a general class of range queries, except for rectilinear queries, in which case rectangular grid partitioning is superior. By using current algorithms for rectangular grid partitioning, parallel storage and retrieval algorithms for hexagonal partitioning can be constructed. Some of these results carry over to circular partitioning of the data-which is an example of a non-convex region.",
keywords = "Data-space partitioning, Disk and page allocation, Parallel I/O, Range query, Two-dimensional data",
author = "Hakan Ferhatosmanoǧlu and Divyakant Agrawal and {\"O}mer Eǧecioǧlu and {El Abbadi}, Amr",
year = "2005",
month = "1",
day = "1",
doi = "10.1023/B:DAPD.0000045550.56749.75",
language = "English",
volume = "17",
pages = "75--101",
journal = "Distributed and Parallel Databases",
issn = "0926-8782",
publisher = "Springer Netherlands",
number = "1",

}

TY - JOUR

T1 - Optimal data-space partitioning of spatial data for parallel I/O

AU - Ferhatosmanoǧlu, Hakan

AU - Agrawal, Divyakant

AU - Eǧecioǧlu, Ömer

AU - El Abbadi, Amr

PY - 2005/1/1

Y1 - 2005/1/1

N2 - It is desirable to design partitioning methods that minimize the I/O time incurred during query execution in spatial databases. This paper explores optimal partitioning for two-dimensional data for a class of queries and develops multi-disk allocation techniques that maximize the degree of I/O parallelism obtained in each case. We show that hexagonal partitioning has optimal I/O performance for circular queries among all partitioning methods that use convex non-overlapping regions. An analysis and extension of this result to all possible partitioning techniques is also given. For rectangular queries, we show that hexagonal partitioning has overall better I/O performance for a general class of range queries, except for rectilinear queries, in which case rectangular grid partitioning is superior. By using current algorithms for rectangular grid partitioning, parallel storage and retrieval algorithms for hexagonal partitioning can be constructed. Some of these results carry over to circular partitioning of the data-which is an example of a non-convex region.

AB - It is desirable to design partitioning methods that minimize the I/O time incurred during query execution in spatial databases. This paper explores optimal partitioning for two-dimensional data for a class of queries and develops multi-disk allocation techniques that maximize the degree of I/O parallelism obtained in each case. We show that hexagonal partitioning has optimal I/O performance for circular queries among all partitioning methods that use convex non-overlapping regions. An analysis and extension of this result to all possible partitioning techniques is also given. For rectangular queries, we show that hexagonal partitioning has overall better I/O performance for a general class of range queries, except for rectilinear queries, in which case rectangular grid partitioning is superior. By using current algorithms for rectangular grid partitioning, parallel storage and retrieval algorithms for hexagonal partitioning can be constructed. Some of these results carry over to circular partitioning of the data-which is an example of a non-convex region.

KW - Data-space partitioning

KW - Disk and page allocation

KW - Parallel I/O

KW - Range query

KW - Two-dimensional data

UR - http://www.scopus.com/inward/record.url?scp=7544222418&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=7544222418&partnerID=8YFLogxK

U2 - 10.1023/B:DAPD.0000045550.56749.75

DO - 10.1023/B:DAPD.0000045550.56749.75

M3 - Article

VL - 17

SP - 75

EP - 101

JO - Distributed and Parallel Databases

JF - Distributed and Parallel Databases

SN - 0926-8782

IS - 1

ER -