Exploring sensor data management

The increasing availability of cheap, small, low-power sensor hardware and the ubiquity of wired and wireless networks has led to the prediction that `smart evironments' will emerge in the near future. The sensors in these environments collect detailed information about the situation people are in, which is used to enhance information-processing applications that are present on their mobile and `ambient' devices. Bridging the gap between sensor data and application information poses new requirements to data management. This report discusses what these requirements are and documents ongoing research that explores ways of thinking about data management suited to these new requirements: a more sophisticated control flow model, data models that incorporate time, and ways to deal with the uncertainty in sensor data.