Lectures: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23

Overview

Papers:

• Gehrke, Johannes, and Samuel Madden. "Query Processing in Sensor Networks." In Pervasive Computing, Jan-March 2004, pp. 46-55.
• Deshpande, Amol, Carlos Guestrin, Samuel Madden, Joseph Hellerstein, and Wei Hong. "Model-Driven Data Acquisition in Sensor Networks." In Proc. of International Conference on Very Large Databases, 2004.

These papers discuss issues related to query processing on a new class of hardware -- wireless sensor networks (sensornets). Sensornets consist of many tiny, battery-powered, sensor-equipped computers that are usually deployed for purposes of monitoring some remote environment. As such, they have relatively high loss and failure rates and strict energy limitations, requiring some novel query processing techniques.

The first paper summarizes two well known query processing systems for sensor networks, TinyDB and Cougar. They both provide a declarative language that can be used to extract data from large networks and have special features designed to deal in particular with the energy constraints of these networks.

The second paper discusses a system called BBQ that uses statistical models to reduce energy consumption and improve query accuracy in the face of network losses.

As you read the papers, consider the following questions:

1. In what ways is a sensor network database different from a traditional database? In what way is it like a traditional database?
2. What special features do TinyDB/Cougar have for reducing energy consumption?
3. In the 'Model Driven' paper, how can a statistical model be used to reduce energy consumption?
4. How can a statistical model be used to improve answer quality in the face of network loss?