Emerging distributed information systems combine local-area network information systems and wide-area network management to deliver multimedia information. Typically, these distributed information systems are based on multiple sites and multiple servers. In this worldwide network dimension, query processing balancing and data placement management have a key role to play, due to the large number of possible processing sites and due to the information types. We propose the design of an experimental distributed DBMS called TOSDHIM (TOtal System for Distributed and Heterogeneous Information Management) to provide high performance in distributed and heterogeneous information systems. In this paper, we present the architecture and implementation of TOSDHIM and discuss its operating characteristics. In addition, TOSDHIM provides a general, customisable platform for the development of new algorithms for distributed query optimization, storage management and scalable query processing of multimedia information. This flexibility is primarily due to a customisable architecture, integrating dynamic resource management and load balancing supervision, that permits autonomous, local dynamic query optimization to be made regarding data placement, query execution location and storage management.