SDI Convergence - Nederlandse Commissie voor Geodesie - KNAW

An Analysis of Technology Choices for Data Grids in a Spatial
Data Infrastructure
Serena Coetzee and Judith Bishop
Department of Computer Science, University of Pretoria, Pretoria, South Africa
{scoetzee, jbishop}@cs.up.ac.za
Abstract
The concept of grid computing has permeated all areas of distributed computing,
changing the way in which distributed systems are designed, developed and implemented.
Data grids enable the sharing of data in a virtual organisation and are typically
implemented for data federation in data-intensive environments. So far, they have been
applied to traditional data (text, image, sound). We present a scenario that describes
for the first time how data grids can be applied to enable the sharing of address data in
a spatial data infrastructure (SDI). Consolidating spatial data from distributed heterogeneous
sources into a single centralised dataset requires, amongst others, a considerable
human coordination effort. A data grid consolidates data directly from the distributed
sources, thereby eliminating the effort. We present a reference model called
Compartimos (Spanish for ‘we share’), that is based on the Open Grid Services Architecture
(OGSA) but is customised for sharing address data in an SDI, and we analyse
existing technologies, such as the Globus Toolkit, ISO 19100 standards and Open
Geospatial Consortium (OGC) web service implementation specifications, for Compartimos.
This article advances the mutual understanding between data grids and SDIs
and sheds light on a future technological solution that could overcome some of the
data sharing impediments that are experienced in SDIs today. Finally, results from the
analysis and future directions for research are discussed.
Keywords: spatial data infrastructure (SDI), data grid, data sharing, service-orientation,
web service, grid computing, address data, geographic information, GIS.
1. INTRODUCTION
A grid is a system that is concerned with the integration, virtualisation, and management
of services and resources in a distributed, heterogeneous environment that supports
virtual organisations (collections of users and resources) across traditional administrative
and organisational domains (real organisations). A data grid is a special
kind of grid in which data resources are shared and coordinated (OGF, 2007b). How
virtual organisations collaborate and share resources in order to achieve a common
goal is described as the ‘grid architecture’ in The Anatomy of the Grid (Foster et al.,
2001) and The Physiology of the Grid (Foster et al., 2002). This has subsequently
evolved into the Open Grid Services Architecture (OGSA) published by the Open Grid
Forum (OGF, 2006), a vision of a broadly applicable and adopted framework for grids.
The OGSA data architecture (OGF, 2007a) describes the interfaces, behaviours and
bindings for manipulating data within the broader OGSA.
The work reported in this article is part of a research project on ‘Distributed Address
Management’, which has the objective of establishing whether the data grid approach
is an option for national address databases in an SDI and if so, what the design imperatives
for such an approach are. In earlier work, we used a novel evaluation framework
for national address databases to evaluate existing information federation models,
as well as the data grid approach, for the use in address databases for national
107