We describe a new benchmark approach which is needed for two reasons. First, today's computer systems almost exclusively use relational database management systems (DBMS), and, besides (purely) relational DB technology, only the new trend of object-relational technology is expected to obtain practical relevance. Second, today's (complex) software systems are developed in an object-oriented way. Thus, a benchmark approach is needed assessing the contributions of (object-)relational DBMS to object-oriented software development. Consequently, such a benchmark does not only have to 'measure' the efficiency of the DBMS itself but does also have to take the overhead to be spent for mapping object-oriented data structures to the DBMS interface into account. We describe how our approach achieves this goal and report on first measurement results.