Abstract

An important class of networked systems is emerging that involve very large numbers of small, low-power, wireless devices. These systems offer the ability to sense the environment densely, offering unprecedented opportunities for many scientific disciplines to obtain detailed datasets for analysis. In this paper, we argue that a data handling architecture for these devices should incorporate their extreme resource constraints - energy, storage and processing - and spatiotemporal interpretation of the physical world in the design, cost model, and metrics of evaluation. We describe DIMENSIONS, a system that provides a unified view of data handling in sensor networks, incorporating long-term storage, multiresolution data access and spatio-temporal pattern mining.

Citations