Portal Access to Parallel Visualization of Scientific Data on the Grid

Abstract

Visualizing extremely large and time-variant data presents a challenge in overcoming slow disk and network access. The Weather Research and Forecasting (WRF) model, used to model mesoscale weather phenomena, generates large amounts of storm system information that is visualized to better understand storm formation. WRF faces these caveats in disk and network bandwidth limitations. Using the open-source Visualization ToolKit (VTK) and the additional functionality of ParallelVTK, we devised an effective solution to view these data sets at a visually pleasing frame rate. In addition to viewing the data sets in real-time, it is also feasible to render off-screen and create a movie for reusable viewing. These remote jobs can be formed and launched from a remote portal environment using grid tools.