RasDaMan is a database management system (RasDaMan DBMS) that provides doctors with fast and efficient access to networked archives of radiological images. The raster RasDaMan DBMS supports storage and retrieval of raster data of any size, dimension and base type. The raster RasDaMan DBMS offers a conceptual model with a C/C++ API and a declarative data description language (DDL) and DML to define and manipulate arbitrary arrays across heterogeneous platforms (PC/UNIX), and a physical data organization supporting heterogeneous storage media; transparent compression decompression; specialised raster indexes; algebraic query optimization based on index evaluation; and access statistics. From the users' point of view, the two key features of RasDaMan DBMS are: high-level semantic modelling of raster data as arrays of any size and dimension over complex base types; flexible declarative query language, based on a powerful formal imaging model. The enabling raster server technology for these high-level services is: intelligent query optimization, based on schema information, access patterns; data independence: the application selects its preferred data format, this also includes transparent compression/decompression; leading-edge storage architecture; transparent integration of secondary and tertiary storage media; all classical DMBS services available: multi-user support, recovery, data independence, etc. As a complement to the DBMS, one interactive healthcare application, which makes use of the raster DBMS functionality, has been implemented.