Runtime execution management of distributed simulations

Distributed Modeling and Simulation (M&S) provides benefit from the ability to bring together a large number of simulations, across a network, to fulfill a specific requirement. However, this capability comes with the costs and complexity of coordinating all of the computing platforms for the startup, execution, shutdown and artifact collection of the simulation execution. Typically, an exercise event also requires many iterations of the simulation execution, necessitating the ability to perform these tasks in an efficient and repeatable manner. This paper discusses an approach to handle the runtime execution of a simulation exercise as part of the Executable Architecture Systems Engineering (EASE) research project. We discuss the methodologies used to control the overall execution of a distributed simulation as well as control the individual applications involved. We further present some of the current use cases for this approach and lessons identified.