Current medical practice needs images as a major diagnostical tool. However, as clearly demonstrated in some cases, the storage and retrieval of medical images becomes a growing problem. The volume of images and the problems to make extra copies contribute to this problem. Rapid developments in technology indicate the feasibility of an Image Information System (IMAGIS), coupled to the existing Dutch Hospital Information System ("ZIS"), by which functions as archiving, communication, processing and representation of medical image data can be realized. Such a system, however, incorporates such a complexity, that its realization can only be achieved through several stages of development. One of the very first stages can be the set-up of a digital mass data storage part. Hereby the evolution of optical storage media (digital optical disks and jukeboxes containing them) seems much promising, taking into account data lifetime and price/performance characteristics of storage media. The division of the storage part of an IMAGIS prototype into multiple layers seems inevitable, because of performance characteristics of the various storage devices available. A computer simulation model can be of great help to determine the composition of a storage configuration and of the algorithms for assignment of images to a specific layer. In order to meet future trends in technology a simulation model has been constructed with adjustable parameters. In the model each component, which is simulated, is assigned a process description giving in detail the actions a component has to perform, related to the actions of other components. An attractive configuration may be obtained by a systematic choice of different sets of parameter values. The input of the simulation model is derived from technical data and data representing the current use of medical images. This implies the need for knowledge of when and how often a medical image is requested, of the time a medical specialist examines the image etc. Part of this knowledge can be obtained by using the information available in the existing Hospital Information System, developed by BAZIS, and being applied in 20 hospitals in the Netherlands (11,500 beds together). Preliminary simulation results presented illustrate the usefulness of our problem approach, bringing the realization of an IMAGIS prototype one step further.