Global scale sensor networks: opportunities and challenges

While research in sensorwebs often focuses on extremely low-power, inexpensive sensors coupled with ad hoc, occasionally-connected networks, the recent tragic tsunami associated with the Sumatra earthquake leading to the loss of more than 280,000 lives in the Indian Ocean demands a fresh look at networks and grids on a global scale often involving highly sophisticated sensors. The extraordinary growth in network bandwidth coupled with initiatives such as the not-for-profit National Lambda Rail has enhanced grid computing and grid storage technologies making a truly global sensorweb not only possible, but also imperative. The same system lowers most barriers to scientific research on vast amounts of new data and lowers the costs of operations and maintenance. The Earth and planetary sciences are moving toward an interactive global observation capability necessitating a new generation of cyberinfrastructure. At UCSD and Scripps NSF ITR projects including ROADNet, LOOKING, and the OptIPuter are prototyping real time control of remote instruments, visualization of large data objects, metadata searching of federated data repositories, and collaborative analysis of complex simulations and observations. A single-purpose national or global warning system, for tsunamis for example, is nearly impossible to maintain because of the time scales of major events -- decades to centuries. However, a broad, scientifically based, multipurpose system can be maintained and will grow as new technologies become available and new approaches in cyberinfrastructure replace the old.