Steering simulations on high performance computing resources

Computational Steering (CS) of numerical simulations has been developed over the last three decades. While it has succeeded in some of its chief aims, the uptake and impact of CS has not been as great as anticipated. This paper aims to investigate the reasons for this, and from this analysis to provide an enhanced CS framework, taking into account both modern developments in end-user devices and changes in the architecture of very large High Performance Computing (HPC) systems (supercomputers). We also consider the impact on CS of the recent interest in Dynamic Data Driven Application Systems (DDDAS) and Cyber-Physical System (CPS). As the beginning phase of our research, we present a general-purpose framework that provides CS and HPC as Web services to widen their uptake and use. Key advantages of this approach, include the re-usability, modularity, real-time 3D web visualization, and capacitating users to access CS and HPC service on portable devices.