Improving the speed and scalability of distributed simulations of sensor networks

Distributed simulation techniques are commonly used to improve the speed and scalability of wireless sensor network simulators. However, accurate simulations of dynamic interactions of sensor network applications incur large synchronization overheads and severely limit the performance of existing distributed simulators. In this paper, we present two novel techniques that significantly reduce such overheads by minimizing the number of sensor node synchronizations during simulations. These techniques work by exploiting radio and MAC specific characteristics without reducing simulation accuracy. In addition, we present a new probing mechanism that makes it possible to exploit any potential application specific characteristics for synchronization reductions. We implement and evaluate these techniques in a cycle accurate distributed simulation framework that we developed based on Avrora. In our experiments, the radio-level technique achieves a speedup of 2 to 3 times in simulating 1-hop networks with 32 to 256 nodes. With default backoffs, the MAC-level technique achieves a speedup of 1.1 to 1.3 times in the best case scenarios of simulating 32 and 64 nodes. In our multi-hop flooding tests, together they achieve a speedup of 1.5 to 1.8 times in simulating networks with 36 to 144 nodes. The experiments also demonstrate that the speedups can be significantly larger as the techniques scale with the number of processors and radio-off/MAC-backoff time.