System design for on-line distributed computational visualization and steering

We propose a distributed computing framework for network-optimized visualization and steering of real-time scientific simulations and computations executed on a remote host, such as workstation, cluster or super-computer. Unlike the conventional "batch" simulations, this system enables: (i) monitoring of an on-going remote computation using visualization tools, and (ii) on-line specification of simulation parameters to interactively steer remote computations. Using performance models for transport channels and visualization modules, we develop a dynamic programming method to optimize the realization of the visualization pipeline over a wide-area network to maximize the frame rate. We present experimental results to illustrate the effectiveness of this system.