This paper presents a cross validation of network system models in two simulation tools, namely the ONE Simulator and Scenargie, for the simulation of Delay or Disruption Tolerant Networks (DTNs). The study compares the simulation results from three network system models provided by the twosimulators as well as their efficiency for a commonly used DTN scenario. The results show a fundamental problem inherent to the time stepped approach that may introduce an unacceptable level of inaccuracies in the predicted DTN performance.
Also, limiting such inaccuracies gives the time stepped approach no runtime performance superiority over the event driven approach. Further, the study shows that the DTN performance is highly sensitive to the link establishment overheads, which may change the predicted end-to-end message delivery latencies by an order of magnitude. Moreover, other details of the network system are important when the target network is expected to be highly loaded with communication traffic.