A Methodology to Predict the Performance of Distributed Simulations

Predicting the time-performance of a Distributed Simulation (DS) system may be of interest to evaluate system alternatives during the development cycle, before the system is implemented. In this paper, we introduce a methodology to predict the execution time of a DS system during its design phase. The methodology is based on a model-building approach that, basing on the design documents of the DS system, first produces its performance model and then evaluates it. The model includes components such as middleware to use (e.g., the HLA RTI), the set of DS execution hosts and the set of host interconnection networks. The methodology is applied to determine whether or not producing the distributed simulator of a given system may be advantageous in terms of execution time with respect to a conventional local simulator. An example use of the methodology is presented and validated by a comparison of the time-prediction with the actual execution time of the implemented DS system.