All information systems process data and therefore, each system needs mechanisms to access data. In this paper we will first define a classification of information systems w.r.t. their requirements on data access and second, develop and discuss different approaches in order to fulfill these requirements. Emphasis will be given to data intensive systems that are reading and writing huge amounts of data. The overall application scenario is defined by engineering design frameworks and data modeling using the STEP standard (ISO 10303). Having classified arbitrary types of data sources, we will discuss multiple CORBA-based data access strategies. They will be evaluated and compared with Intra/Internet-based approaches using a modular prototype system. This system has been built according to the Java language mapping of the STEP data access interface (SDAI) which has been co-developed by us.