Large-Scale Datasets Visualization on Networks Using Dynamic Pipeline Decomposition and Mapping

Datasets of tens of gigabytes are becoming common in computational and experimental science. Providing remote visualization of these large datasets with adequate levels of quality and interactivity is an extremely challenging task, particularly for scientists who collaborate in widely distributed locations and their primary access to visualization resources is a desktop computer. This paper describes a remote visualization system for large-scale terrain rendering, which adaptively decomposes and maps the terrain visualization pipeline such as filtering, refinement, rendering, and display onto a wide-area network. The prototype system was implemented on China next-generation Internet backbone. Scientists dispersed in three cities could visualize 3D flight simulation at about 30 frames per second; furthermore, the need for data replication to local desktops was eliminated.