On the water frontMassive sensor network to monitor Hudson River

Now, here is something that sounds counterintuitive: Solar-powered submarine. Come again? Yes: The Beacon, New York-based Beacon Institute for Rivers and Estuaries, a nonprofit scientific research organization, has teamed up with IBM and several other research groups to develop a high-tech environmental-monitoring system for the state’s Hudson River which would transform the river’s 315 miles into an interconnected network of sensors. The sensors would collect data on the river’s biology and chemistry and transmit them to a central location for analysis by IBM-developed system. The system will take the information and create a virtual model of the river to simulate its ecosystem in real time. Comprising both distributed-processing hardware and analytical software, the system is designed to take heterogeneous data from different sources and make sense of it in real time. The software learns to recognize data patterns and trends and prioritizes useful data. If some data stream begins to exhibit even minor variations, the system automatically redirects resources toward it. The system will also be equipped with IBM’s visualization technologies; fed with mapping data, they can create a virtual model of the river and simulate its ecosystem in real time. The IBM system “enables us to do a great deal of work in the area of data integration and data management for very large volumes and different types of data,” says Harry Kolar, Global Alliance executive at IBM. “Another reason we are working in this sensor area is that we can actually build end-to-end solutions, meaning from the smallest device to a large back-end system.”

Sensor networks to monitor everything from sewage systems to battlefields have been under development for many years at companies such as Intel, Sun, and Siemens and at academic institutions such as the University of California, Los Angeles. What the “research community has not had is the making-meaning part,” says David Culler, a professor of computer science at the University of California, Berkeley. This is what the IBM system is intended to provide. “A lot of what the research community has been focused on is getting sensors and delivering them through reliable, energy-efficient networks to the computing infrastructure,” says Culler. “But once you have the data, what do you do with it, and how do you sort it?”

Back to the Hudson River. “In terms of having an integrated network of sensors, and given the magnitude of it for the Hudson River,