Reproducibility model for wireless sensor networks parallel simulations.

Abstract

Several wireless sensor networks (WSNs) simulations run in parallel computer architectures to improve their scalability. The main problem with this strategy is guaranteeing the reproducibility transparently to simulation users. We present a reproducible model for WSNs parallel simulations. The model uses a chunk partition strategy, in which all simulation elements are wrapped into chunks and simulated sequentially inside each chunk. We can simulate multiple chunks in parallel as often as possible by adding seed and a pseudo-random number generator in each of them, thus ensuring the same results. This model was integrated into the JSensor simulator and validated accordingly. An important aspect is that the new reproducibility model does not alter the simulation’s semantics, and it is transparent to the developer. The results demonstrated that our model could guarantee the reproducibility of stochastic WSNs parallel simulations performed in diferent computer systems with diferent threads.