Computational fluid dynamics (CFD) linked with
virtual reality (VR) visualization techniques offer comfortable means to explore interaction of fluids and gases with complex surfaces in
the field of engineering or physics amongst others. Huge data sets, in the range of many gigabytes, require sophisticated
storage schemes to enable efficient access during post-processing.
In this paper we introduce approximate geometric ranking methods for CFD data using off-the-shelf RDBMS by significantly extending
the efficient indexing structure RI-tree. We further present preliminary, but very promising performance results of our ongoing
research.