Parallel Compression of 3D Meshes for Efficient Distributed Visualization

In a distributed visualization environment transmission time is dominant because of the amount of data to be moved and the limitations of available bandwidth. In this paper we address the problem to speed up the compression operation of large Triangulated Irregular Networks (TIN) using commodity clusters. In our case TINs represent isosurfaces extracted from volumetric data sets. The proposed parallel compression algorithm is based on Edgebreaker, one of the most powerful connectivity compression algorithm, and it exploits mesh partitioning produced during the parallel isosurface extraction operation. In this way a high speed-up of the compression module and a considerable improvement of the visualization system are obtained.