Reliability of network simulators and simulation based research

Abstract

Network simulators have gained an increased importance in the research community to an extent that approximately 11% of the research papers that appeared in the proceedings of leading international conferences used network simulators to test their claims. It is anticipated that due to the several advantages network simulators will become more prevalent in the research community in future. However, with the growing popularity of the network simulators, challenges in the form of consistent revelations about their inaccuracies have started to seriously question the use of such simulators and simulation based research. This paper attempts to answer the questions (1) How can assumptions/simplifications in modeling affect realism of simulations? (2) How credibility of simulation based research be improved?, and (3) What necessary steps must be taken to make network simulators credible? The paper provides deliberated recommendations for improving the authenticity of simulation based experiments. The authors believe that the discussion in this paper will greatly facilitate the repeatability of simulation based research and will help researchers to select the most suitable (credible) simulator.