Abstract: The ever increasing number of data-intensive scientific visualizations motivate the need for high-performance data management support. One such example is given from a simulation and visualization system that assists the physician in the planning and execution of a novel, non-invasive, and very effective treatment of brain aneurysm. The output dataset of this blood flow simulation is very large and information rich that demands an effective means to gaining insights into the nature of time-varying flow fields. For this purpose, we built a computing environment which exploits efficient data storage and management techniques to provide a sufficient performance for real-time visualization. Our approach has substantial advantages for achieving higher access efficiency and storage utilization, yielding frame rates suitable for interactive visualization on common desktop workstations. Practical experiments indicate that the techniques developed in this environment can provide the basic building blocks for future efficient implementation of large complex data visualization systems.