Earth observation data collections from multiple missions, sensors and processing levels are very heterogeneous in structure, size and format. Managing digital data products in such collections requires a flexible product data model. The object-oriented product model used in the Data Information Management System (DIMS) is highly configurable, allows the definition of arbitrary metadata structures and is extensible during runtime. There is not one common product identification scheme, so each mission and product type can define its individual complex product identifier, thus using comprehensive and adapted data structures. The same applies to descriptive metadata which can be mission-specific. At the collection level, the ISO standards for geo-metadata are supported.
The actual products are stored in one or more file system archives, while the metadata is stored in an inventory that maps the object model to relational structures and ensures fast access even for very large amounts of products. Depending on the relational database vendor chosen, sophisticated geographic search capabilities can be exploited to allow fast searches for data product footprints, in addition to the usual query capabilities.
The archive file systems are managed by the digital library to ensure data integrity and safety and hide archive internals such as directory and file names. This allows optimal organization of the physical file storage without implications for external entities. Upon retrieval, the original product structure is restituted. All kinds of file systems can be used: simple local file systems, remote network storage locations and hierarchical storage management systems.