SDI Convergence - Nederlandse Commissie voor Geodesie - KNAW

5. RESULTS
Through the development of the Compartimos reference model, we have identified the
essential components for sharing address data on a data grid in an SDI environment.
We presented the computational viewpoint of Compartimos by describing the purpose
of each object and relating it to the OGSA data architecture. We analysed technology
choices for individual Compartimos objects and related these to current research. From
this analysis it is evident that there is a need for collaboration between grid and geospatial
communities to ensure harmonisation between respective standards and tools.
The service-oriented approached followed in both OGSA, ISO 19100 and OGC will
prove an advantage for collaboration and integration of respective technologies. The
results of our analysis contribute towards the mutual understanding between these two
communities. The analysis described in this article was a first investigation into the viability
of the data grid approach to national address databases in an SDI and has led to
further research questions that are discussed below.
The ISO 19100 series of standards together with OGC implementation specifications
have been implemented in a number of SDIs (Aalders, 2005). To grid-enable these
SDIs, requires grid-enabling these ISO standards and OGC implementation specifications.
Aloisio et al. (2005a) and Di et al. (2008) write about such efforts, but more implementations
are required to better understand the challenges under different circumstances.
Such implementations would also promote the development of tools to
streamline the grid-enablement. OGSA-DAI already provides uniform access to different
relational databases, similar to OGC web service for heterogeneous geographic
information. Future studies should investigate uniform access to spatial data through
OGSA-DAI, with or without making use of OGC web services. Also, interesting would
be a spatially enabled distributed query processing (DQP) of OGSA-DAI.
We based our initial research on the assumption that address data providers are
mostly local authorities in an SDI that can be trusted. In a Web 2.0 world, where the
citizens become the sources for data, this assumption does not hold anymore. Citizens,
living at an address, are the best available source to verify an address, but the
question is whether they can be trusted to provide accurate data. Goodchild (2008) and
Craglia et al. (2008) also raise this questions and future work should investigate how
such a ‘wikification’ of address data can be integrated into Compartimos.
Compartimos was developed for address data in an SDI and future research should
expand Compartimos for other types of spatial data. Incorporating recent research findings
on ontologies for interoperability would be relevant (Brodeur 2004, Shadbolt et al.,
2006). A reference model for data grids that caters for all kinds of geographic information
could be seen as the first step along the path of standardising geospatial data
grids. Also, research is required to better understand the requirements for grid-enabling
SDIs in terms of non-technical aspects, such as policies, legislation, agreements, human
and economic resources, and organisational aspects.
Finally, this research project started in 2005, before the current hype of ‘cloud computing’.
However, clouds, such as those by Amazon, IBM, Microsoft and the like, also
stand in line as the enabling platform for data sharing in an SDI. Instead of investing
servers and bandwidth, local authorities could buy scalable computing power and data
storage in a cloud without having to support an IT infrastructure. Thus, the viability of
data sharing in an SDI ‘in the clouds’ should be investigated.
117