[en] The database system for the storage and retrieval of experimental and technical data at TEXTOR-94 has to be revised. A new database has to be developed, which complies with future performance and multiplatform requirements. The concept, to be presented here, is based on the commercial object database Objectivity. Objectivity allows a flexible object oriented data design and is able to cope with the large amount of data, which is expected to be about 1 TByte per year. Furthermore, it offers the possibility of data distribution over several hosts. Thus, parallel data storage from the frontend to the database is possible and can be used to achieve the required storage performance of 200 MByte per min. In order to store configurational and experimental data, an object model is under design. It is aimed at describing the device specific information and the acquired data in a common way such that different aproaches for data access may be applied. There are several methods forseen for remote access. In addition to the C++ and Java interfaces already included in Objectivity/DB, CORBA and socket based C interfaces are currently under development. This could also allow an access by non-supported platforms and enable existing legacy applications an integration of the database for storage and retrieval of data by a minimum of code changes