The costs for massive data storage and other related equipment are large enough and the network bandwidths available today are adequate to suggest sharing resources to support these needs for a variety of research groups on the Duke campus. We are therefore developing a model to provide partially centralized support for the computational and data storage needs of our efforts in bioinformatics and computational biology, neural analysis and neuroengineering, global climate change, geographic information systems, multi-scale modeling and in other areas as they arise.