Advanced faults patterns for WSN dependability benchmarking
Proceedings of the 13th ACM international conference on Modeling, analysis, and simulation of wireless and mobile systems
| Hi-index | 0.01 |
Recent advances in wireless sensor networks show an expansion both in the size of the network and in the variety of the applications that are to be executed. In this context, wireless network simulators play a key role in the design, development and testing of new wireless sensor networks applications and protocols. Most of the currently available wireless sensor network simulators work well for small to medium sized networks yet fail to scale to really large networks, mainly due to their single desktop machine architecture and its limited resources. In this paper, we present XS-WSNet, a wireless sensor network simulator that is designed with extreme scalability as a prime objective. First, we distribute the simulated wireless sensor network nodes on a variety of actual machines that communicate through a wired network. A large-scale simulation is then emulated by several smaller scale simulations that run concurrently and collaboratively. As smaller instances of simulations running on different machines usually induce asynchrony and non-determinism in the whole system, we also propose a distributed synchronization protocol, responsible for simulation accuracy in some simulation models. Our implementation (both asynchronized and synchronized versions) of XS-WSNet is fully evaluated in various contexts using a scalable benchmark application. Against its single machine version, the distributed simulator provides sensible matching results, exhibits both scale-up and speed-up of the simulation, and performs with linear slowdown with up to ten million simulated nodes.