EarthServer, funded by the European Commission, is establishing open access and ad-hoc analytics on extreme-size Earth Science data, based on and extending the rasdaman Array DBMS.
Database query languages serve as c/s interface for “mix & match” access to multi-source, any-size, multi-dimensional spatio-temporal data and their metadata, across all Earth sciences – in short: Big Earth Data Analytics. These interfaces are strictly based on OGC (for geospatial information) and W3C XQuery (for metadata) standards.
The underlying platform consists of rasdaman („raster data manager“), a multi-dimensional Array DBMS [Baumann, VLDBJ 1994] adding n-D arrays as a new attribute type to relations. The rasdaman query engine extends SQL with n-D array operators, based on Array Algebra which also guides server-side optimizations. Arrays of unlimited size are partitioned („tiled“) and stored in a standard relational DBMS.
Based on this, six services are being established as Lighthouse Applications, together covering all Earth sciences, each a mix of at least 100 TB of in-situ and imported data. Experiences made are being fed back into OGC and ISO standardization.