The future suitability of simulation as a viable tool to analyze large, complex, and elaborate systems relies on the concurrent execution of the simulation program in parallel or distributed multiprocessing environments. In the last 15 years, research in parallel (and distributed) discrete event simulation (PDES) [1–3] has focused on attainable simulation execution speedups. The main reason PDES is not embraced by the general simulation community is the lack of research contributions toward simplifying the development of simulation models for concurrent execution. The authors, after citing a profusion of papers “bemoaning the acceptance problem” of PDES, attempt to respond to the call for a discussion of the future of PDES.
The authors point out that the PDES research community has not placed any importance on the role of the conceptual framework within the model development process. The “logical process” modeling methodology that has become the de facto standard for the PDES paradigm is dictated by the need to partition the simulation program for minimal processor synchronization. The authors prescribe a model-centric view as opposed to the current program-centric view of the simulation process in PDES. As a starting point, they provide four recommendations to reduce the disparity between PDES and prevailing discrete event simulation methodology.
Overall, the paper is a worthy endeavor and was a pleasure to read. The authors provide a brief overview of the simulation process by presenting some fundamental questions and their answers. Throughout the paper, work by prominent researchers is cited.
Researchers and practitioners in PDES must read this paper.