Designing a Geo-scientific Request Language - A Database Approach

A large part of sensor, image, and statistics data in the Earth Sciences naturally come as data points aligned to regular grids of some dimension, including 1-D time series, 2-D imagery, 3-D image time series and volumetric data, and 4-D spatio-temporal data. Frequently repositories have to host objects of multi-Terabyte size, in the future multi-Petabytes. Serving this information category, large multi-dimensional arrays, is a task of increasing importance among the Earth Sciences. Still, however, ad-hoc implementations with focused functionality prevail which lack the flexibility of SQL-enabled databases. The Web Coverage Processing Service (WCPS) geo raster model and language allows navigation, extraction, and ad-hoc analysis on multi-dimensional geoscientific raster data which can be geo-referenced or not. The request language has been designed to offer some key properties considered advantageous by the database community: it is formalized, declarative, data independent, optimizable, and safe in evaluation. WCPS has been adopted as an international standard by the Open GeoSpatial Consortium (OGC) in December 2008. The reference implementation is almost finished. Actually, the embedding into the modular framework of the OGC geo service standards has posed particular constraints which the design had to respect. We discuss conceptual model, query language, and the context using real-life use case scenarios.