On Using Weather Information for Efficient Remote Data Collection in WSN∗

Even though real-time data access to remote sensor data is desirable for many WSN applications, collecting remote data reliably from a WSN is a challenging task since many components of the WSN are subject to harsh elements of nature. A deployed system’s performance and lifetime can be impaired if this harsh environment of the deployment site is not considered. The deployments of LUSTER, a WSN for ecological monitoring, show that weather has significant impact on the connection availability between a WSN and a back-end. In this paper, we show how the weather of the deployment site can affect the performance of LUSTER, the connection to the back-end in particular. In addition, we propose a weather-aware delay-tolerant networking scheme in the context of LUSTER. Our trace-driven simulation result shows that despite the inaccuracy of the weather information and the simplicity of the algorithm, the sensor data transmission failure rate can be reduced significantly.