Efficient simulation of wireless networks using lazy MAC state update

Scalable and efficient network simulation methods are the method of choice for evaluating and verifying wireless network protocols on a moderate to large scale. This need becomes obvious when simulating very large-scale wireless networks such as emerging ad hoc sensor networks in which the number of nodes can be the order of thousands or more, and with very high node density. Unfortunately, simulation of such large-scale wireless networks often requires excessively large amounts of computing resources and can be slow to complete. One approach to achieving higher performance in a large-scale network simulation is the use of parallel or distributed simulation techniques. However, the efficient distributed simulation of wireless ad hoc networks is still a daunting task. Therefore, we turn our attention to more traditional sequential simulation methods, and seek to reduce the overhead incurred in the medium access control (MAC) state update propagation between wireless nodes. We introduce a novel method called LAMP (lazy MAC state update), that substantially reduces this overhead, with no loss of accuracy. Using our wireless network simulation tool, we compare the efficiency of the LAMP approach to the more traditional approach, and show a performance improvement of up to a factor of eight, with no loss of accuracy.