On the usage of container and container orchestrators as a computational infrastructure for simulation experiments

The execution of simulation experiments becomes increasingly resourceintensive, either due to the increasing scale and complexity of the simulation model or the nature of the experiment itself. Popular approaches, like simulation-based optimisation and data farming, require extensive computational resources. Therefore, an appropriate methodology is needed to distribute simulation workloads onto complex computing infrastructures efficiently. Containerisation and container orchestration are promising approaches toward this goal. This paper discusses the requirements needed to utilise containerisation for simulation, gives an example of a container-based computational infrastructure for experimentation, and shows how containerisation and container orchestration benefit simulation execution.