Design and development of a simulation environment in OPNET using High Performance Computing

Most major simulation federations are designed and constructed based on the professional experience of one or a few experts with extensive experience in creating such federations. Organizations have a risk of losing their understanding of federation performance if one of their key people is no longer available. As very little objective data on federation performance exists, it is difficult for many facilities to create large federations in a manner that insures optimum performance. A detailed architecture and accompanying design guidelines or tools able to create viable interoperable implementations has not yet emerged which minimizes the amount of human intervention, tweaking, and custom software development. This paper proposes the creation of a simulation environment and the generation of initial simulations where data can be gathered that depicts some of the trade-offs needed to make intelligent decisions regarding the design and deployment of simulation federations. The project applies the expertise of University of Central Florida (UCF) in simulation, computer network modeling and high performance computing. The test-bed and data generation/analysis activity was performed in collaboration with Team Orlando, a partnership among the military services, industry, and academia working to leverage resources and contribute to the overall security of the United States. We captured objective data on the performance of a number of very complex federations and created detailed OPNET models of the computers, networks, and software applications that make up large federations. As the communication models were complex and a large number of different federation configurations were needed to be studied, the simulations were run on the new high performance computer (HPC) facility at the UCF. The results of this study will allow the simulation community to design more efficient federations.