You are here


The architecture for large-scale data sharing is based on the integration of Grid, P2P and data-integration technologies. It is service-oriented, and abstracted into four layers:

Data Sharing Architecture
Grid-enabled architecture for large-scale data sharing
  • Data layer: this layer consists of a set of autonomous, distributed data sources. Each data source makes its own decision about the system and data, and great heterogeneities may exist among different sources.
  • Grid layer: this layer hides the heterogeneities exposed in the data layer, and presents an uniform view (i.e. services) of all resources to the upper layer. All resources are exposed as Grid services except data resources which are exposed as Data Access Services.
  • P2P layer: this layer organizes services using P2P models for the support of decentralized service discovery and schema mediation. Note that the distinction between this layer and the Grid layer may not be obvious, and sometimes these two layers may be mixed together (e.g. a Grid service is implemented in a P2P mode).
  • Application layer: this layer performs some data intensive operations, e.g., data analysis possibly spanning over multiple data sources.

For more details of the architecture, please consult the publications and the prototypes page.

SemanticWildNet Architecture

The Semantic WildNet prototype uses semantic web technologies to represent and integrate species sighting data, taxonomic databases, climate sensor data, vegetation data and spatial data enabling environmental scientists to reason across the integrated datasets. The system provides a semantically-unified view of: wildlife sighting data from the Environmental Protection Agency; species data from the Australian Museum and the National Herbarium; climate sensor data from the Bureau of Meteorology and topographic maps from Geosciences Australia. The data is represented in RDF (Resource Description Framework) and merged via a common biodiversity ontology (represented in OWL). Finally, SPARQL (a query language for RDF) is combined with Google Maps to provide an intuitive mapping interface to query the integrated datasets.