For the storage of large data stocks with spatial reference, the establishment of a geo-data storage system in the form of geo-databases is a good option. Spatial databases thus usually form the basis of geoinformation systems and spatial data infrastructures (GDI) with the associated geoweb services. Spatial databases use spatial data types and indices to store spatial information. Building on this, the integration of spatial operators and functions enables the processing of spatial data directly in a database.

In addition to the well-proven relational data structuring, data can also be stored graph-oriented or non-relational, for example in the so-called NoSQL databases. In most cases vector data are stored with a projection in the database and spatially indexed. In certain cases, however, the storage of binary raster data (aerial photographs, elevation models, maps, etc.) can also be useful. In our projects we use among others PostgreSQL/PostGIS, MySQL, SQLite/Spatialite, MS SQL Server. We use the very efficient binary HDF5/netCDF formats in combination with GDAL Virtual Raster (VRT) to store, process and dynamically evaluate very large data volumes via the map server.

The most important aspect when using databases in productive GIS is a thorough database design. This includes planning and creating a database. In relational databases, the structure is usually modeled by an ER diagram. Database design also includes considerations of storage space, backup, reliability, and cost.


  • Management and storage of laser scanner data
  • GDI for infrastructure data of power and network providers, transport companies, logistics, health care ...
  • Storage of pipeline networks or roads in the form of topological data as a basis for route planning (node-edge model)
  • individual OpenStreetMap (OSM) data management
  • Data management of GIS shells, e.g. access of a QGIS plugin to a geodatabase
  • Access from mapserver or geoserver for map rendering