Technology advancement in wireless sensor networks (WSN) is enabling a wide range of diverse applications such as homeland security, healthcare, environment monitoring, and mission-critical military operations. WSN offers a remarkable potential to bridge the gap between the physical world of sensors and a virtual world of information and services, thereby enabling us to assimilate a deep and broad understanding and control of the environment. However, before WSN is truly established as the environment-aware ubiquitous sensing, networking, and computing infrastructure, it is critical that these smart sensors deliver acceptable reliability. Due to severe resource constraints (for example, energy constraints) and usually harsh operational environments, sensor nodes are prone to failures. As a result, it is imperative to design fault tolerance into WSN so that the overall sensor network functionalities can be sustained without interruption despite sensor node failures. In this paper, a set of fault-tolerant models are presented, and reliability performance of these models is evaluated and compared through dynamic fault tree analysis. Reliability formulae and comparison results for different fault-tolerant designs generated from this study will provide useful insights and guidelines for designers in achieving or improving fault tolerance for WSN.
Received on November 3, 2007, revised on June 6, 2008