ROADNet: A network of SensorNets

As sensor networks become denser and more widely deployed, the potential develops for interconnecting these networks to combine datasets, share technological solutions, and to conduct cross-disciplinary research and monitoring operations that rely on several signal domains simultaneously. To that end, the real-time observatories, applications and data management network (ROADNet) research project is connecting multiple sensor networks deployed by collaborating research projects into a single network in order to support a variety of research topics including coastal ocean observing, microclimatology and seismology. This paper gives a brief overview of the ROADNet project and discusses some of the implementation challenges we uncovered while building and maintaining the ROADNet system. We encountered challenges on several fronts including building effective programming abstractions for sensor networks, building tools for managing large-scale data in a scalable manner, and building efficient tools for deploying and managing hundreds of sensors. We discuss how these challenges were addressed and some of the lessons learned from collaborations with domain scientists using our network to conduct their research