1 Introduction
Because of significant gains in computational power within the past decade, scientists are now able to perform simulations at unprecedented spatial and temporal resolutions. The sheer size of data generated from these simulations, however, becomes a great challenge for effective visualization and data analysis. Not only that, it is difficult to perform such data intensive tasks in machines far away from where the data are stored. In addition, the memory and storage space available to most workstations is hardly enough to host the entire dataset. As a result, it becomes necessary to shift the visualization computations as close as possible to where the data reside, in many cases to the supercomputers where the simulations are performed.