A simulation system for modeling and analysis of large scale sensor networks

Research in wireless sensor networks often involves simulation systems as a large number of sensor nodes always are unavailable so far, and the operation of them is expensive as well. Unfortunately, the widely used simulators do not scale for sensor networks with 2000 nodes or less. In this paper, we propose a scalable simulation system VWSN (virtual wireless sensor network), to meet the needs of large scale sensor networks simulation. VWSN can diminish memory consumption when maintaining quite simulation precision based on different virtual node modes. It also can dramatically decrease the interference computation complexity from O(N²) to approximate linearity by using the filtering square to get the set of nodes affected by the transmission quickly. As a result, VWSN is able to scale much better for sensor networks under the same conditions. Analytical and simulation results show that VWSN decrease the memory usage and save the runtime of simulations significantly.