Interactive visualization of large, computational flow grids is important to many disciplines in science and engineering. To address the problem of "big data", in which the size of modern data sets often exceeds the capacity of mainstream computing platforms, development of a new demand-paging system for efficient out-of-core visualization is proposed. The new approach applies innovations in spatial prediction and data partitioningto enable desktop applications to selectively read the contents of large data sets from servers connected by local area or wide area network. This will significantly reduce the platform resources required for interactive visualization, thereby relieving a major limitation to collaboration over geographically distributed design environments.Phase I will investigate technical feasibility by implementing a proof-of-concept system using a particle tracer developed with the Visualization Toolkit (VTK). Benchmark tests will be run using data sets provided by the Numerical Aerospace Simulation (NAS) facility. In Phase II, a large-scale prototype will be developed, integrated with commercial software such as FieldView, and tested with an advanced Internet connection. The resultingcapability will significantly increase the number of individuals and organizations that can visualize massive data sets produced by facilities such as NAS.
Potential Commercial Applications:The proposed innovation can realize immediate commercial potential as a collaborative option/enhancement to FieldView, an existing commercial software product with an established market. By greatly reducing the cost of desktop visualization, this technology will amplify the utility and benefits of large data sets for the aerospace community and other big data fields such as geophysics and meteorology. The technology can also bring the benefits of large-scale fluid simulation to the much broader manufacturing community, for use in the development of commodity products such as refrigerators, ink jet printers, golf clubs, etc. The commercialization strategy includes collaborative alliances with both Kitware and Intelligent Light, developers of the vtk and FieldView products respectively, in order to maximize the potential for commercial success.
