Search

Several geo-information communities (e.g. oceanography, atmospheric science, earth observation, etc.) have developed tailored metadata specifications for data discovery, evaluation and use. They conceived these models either profiling standard models (e.g. ISO 19115 metadata specification) or enriching existing and well- accepted data models (e.g. THREDDS/OPeNDAP/netCDF data model) in order to capture and describe more semantics. These metadata profiles have generated a set of related catalog services that characterize the different Communities, initiatives and projects (e.g. INSPIRE, MERSEA, LEAD, etc.). In addition, specific catalog services have been generated by profiling standard catalog services which were designed to accomplish the general requirements coming from the geo-information community (e.g. OGC CS-W). Indeed, to implement catalog services interoperability is a near-term challenge in support of fully functional and useful discovery and sharing infrastructures for spatial data. To implement catalog services interoperability requires metadata profiles harmonization and discovery protocols adaptation and mediation. In an over- simplified way, these solutions may be considered catalogue of catalogues or catalogue broker components. We conceived a solution for making several well-accepted catalogue services interoperable (e.g. OGC services, THREDDS, ESA EOLI, MERSEA CDI, etc.). This solution has been experimented as a stand-alone application tool, called GI-go. More recently, we re-engineered this approach as a service-oriented framework of modular components. We implemented a caching brokering catalog service which acts as a broker towards heterogeneous catalogues services dealing with IGCD (Imagery Gridded and Coverage Data). This service is called GI-cat; it implements metadata harmonization and discovery protocols adaptation. GI-cat supports query distribution allowing its clients to discover and evaluate the datasets, managed by the federated heterogeneous catalogs, via a unique interface. Presently, it is possible to finalize selections by any combination of the following criteria: spatial extent, temporal extent, keywords (free text) and data source (i.e. Where, When, What, Who). We are experimenting another component, called GI-profile, which implements mediation services to profile the generated unified output according to a well-known metadata models (e.g. OGC CS-W). Connecting these two components and a specific GUI it is possible to work out catalog tools, tailored for different Communities, which work as real brokers.