An Architecture for Low Overhead Grid, HPC, and Cloud-Based Simulation as a Service

This chapter describes the development of the MITRE Elastic Goal-Directed simulation framework (MEG), designed to provide modelers and analysts with access to (1) grid, cloud, and HPC computing support, (2) a wide range of Design of Experiments (DOE) Design of Experiments (DOE) methods, and (3) robust data processing and visualization. We review the motivation and use cases for MEG, the architecture and functionality of the current framework, discuss recent examples of its use, address its current challenges and future development, and finally consider its applicability and advantages in the context of cyber-physical systems.