Enhancing XML Data Warehouse Query Performance by Fragmentation

TL;DR

This paper proposes and experimentally validates two algorithms for initial horizontal fragmentation of XML data warehouses to improve query performance and manageability in heterogeneous decision-support applications.

Contribution

It adapts derived horizontal fragmentation techniques from relational databases to XML data warehouses, focusing on dimensions' XML documents.

Findings

Fragmentation improves query response times.

Two alternative algorithms are compared experimentally.

The proposed approach enhances XML warehouse performance.

Abstract

XML data warehouses form an interesting basis for decision-support applications that exploit heterogeneous data from multiple sources. However, XML-native database systems currently suffer from limited performances in terms of manageable data volume and response time for complex analytical queries. Fragmenting and distributing XML data warehouses (e.g., on data grids) allow to address both these issues. In this paper, we work on XML warehouse fragmentation. In relational data warehouses, several studies recommend the use of derived horizontal fragmentation. Hence, we propose to adapt it to the XML context. We particularly focus on the initial horizontal fragmentation of dimensions' XML documents and exploit two alternative algorithms. We experimentally validate our proposal and compare these alternatives with respect to a unified XML warehouse model we advocate for.