Towards the Robustness of High-performance Execution of Multiscale Numerical Simulation Codes Hosted by the Cyberinfrastructure of CAVS @ MSU

Ciorba, F. M., Carino, R.L., & Banicescu, I. (2009). Towards the Robustness of High-performance Execution of Multiscale Numerical Simulation Codes Hosted by the Cyberinfrastructure of CAVS @ MSU. MSU.CAVS.CMD.2009-R0008. Mississippi State University: CAVS.

Abstract

The research work in this report addresses the ‘self-optimization’ (flexible workflow scheduling methods capable of adapting to changes in the resources’ availability and capability) and ‘self-healing’ (reliable workflow scheduling methods capable of recovering from failures) properties of the cyberinfrastructure autonomic computing system (cyber-ACS). The goal of this research work is to employ state-of-the-art workflow scheduling methods, to enable the autonomous execution of multiple complex computational workflows, while meeting the system-wide set of desired goals. Towards achieving this research goal, a theory of robustness for such methods is needed, including definitions and analyses of robustness, diversity, redundancy, and optimality and their relationship to one another. The research work in this report is the first step towards a robustness theory for autonomic complex workflow scheduling methods for high-performance multiscale numerical methods.