Scalable RTI-Based Parallel Simulation of Networks
Proceedings of the seventeenth workshop on Parallel and distributed simulation
Proceedings of the seventeenth workshop on Parallel and distributed simulation
| Hi-index | 0.00 |
Abstract: This paper explores the feasibility of using the United States Department of Defense's High Level Architecture (HLA) for federating an existing simulation with itself. Although the HLA was not originally designed for self-federation, it can be used as a mechanism for distributing a simulation while leaving most of its fundamental algorithms intact. This paper concentrates on the performance of such self-federations using a particular hardware/software platform. We address the feasibility question: is the performance of the system good enough to consider self-federating a large aviation simulation? Our approach involves building a meta-simulation, that is, a set of simple federates that resemble the aviation simulation in its characteristics relevant to distributed performance (computation, communication) without replicating the simulation in its entirety. Our main conclusion is that a large computational granularity-in excess of 200 milliseconds per time step, or 4.2 times as much computation as communication-must be reached to achieve acceptable results.