Abstracts - 2006
CarTelDB: Continuous Query Processing in an Intermittently Connected World
Yang Zhang, Bret Hull, Vladimir Bychkovsky, Hari Balakrishnan & Samuel Madden
Over the past few years, the "first generation" of wireless sensor computing systems have taken root [13, 14], and the idea of thinking of a sensor network as a streaming data repository over which one can run continuous queries [7, 9], with optimizations such as "in-network" aggregation , is now well-established. This approach works well for a class of applications that are characterized by static sensor nodes with relatively low data rates, where the primary function of the sensor network ("sensornet") is to periodically monitor a remote environment or to track some event of interest.
We believe that the next generation of sensornets will display much higher degrees of mobility and significantly higher data rates. For example, media-rich sensors, such as cameras to capture images and video, chemical sensors to monitor pollution, vibration (acceleration) sensors to monitor car and road conditions, and cellular and 802.11 (Wi-Fi) radio sensors to map wireless network conditions, connected to the millions of automobiles in urban and suburban areas of the world can dramatically improve the scale and fidelity of spatio-temporal sensing of a wide range of important phenomena.
There are many issues in the successful design and implementation of such mobile, high-data-rate sensor systems, of which this project focuses on one: query processing. Motivated by the success of first-generation systems that have viewed the "sensornet as a streaming database," we adopt a similar programming model. The goal is to enable users to connect to a central server (which we call the portal), declaratively specify (primarily via continuous queries) what data they are most interested in collecting, and deliver responses to the portal. The portal takes care of distributing queries to the mobile nodes, each of which has a local query processor.
The combination of mobility and high data rates, however, introduces two crucial differences from previous systems [2, 3, 7, 8] that adopt a similar philosophy:
The combination of these two properties -- unaddressed in previous work on query processing -- motivates a new framework for specifying and processing continuous queries. In CarTelDB we focus on two main ideas:
We have implemented CarTelDB under Linux in the context of the CarTel project (http://cartel.csail.mit.edu/). A small number of cars equipped with CarTel boxes are currently in daily use, collecting data from GPS receivers, Wi-Fi interfaces, cameras, and the cars' standard on-board diagnostics (OBD) interfaces. We use the data collected from this real (albeit lightly used) system to conduct a series of trace-driven simulations of different prioritization policies expressed using our query language extensions. Our results demonstrate the usefulness of our language features as well as the clear need for data prioritization in bandwidth constrained settings.
This work is funded in part by the National Science Foundation under Award Number CNS-0205445 and by Quanta Corporation.
 M. Balazinska, H. Balakrishnan, S. Madden, and M. Stonebraker. Fault-tolerance in the borealis distributed stream processing system. In SIGMOD, 2005.
 D. Carney, U. Centiemel, M. Cherniak, C. Convey, S. Lee, G. Seidman, M. Stonebraker, N. Tatbul, and S. Zdonik. Monitoring streams - a new class of data management applications. In VLDB, 2002.
 S. Chandrasekaran, O. Cooper, A. Deshpande, M. J. Franklin, J. M. Hellerstein, W. Hong, S. Krishnamurthy, S. R. Madden, V. Raman, F. Reiss, and M. A. Shah. TelegraphCQ: Continuous dataflow processing for an uncertain world. In CIDR, Jan. 2003.
 FON home page. http://en.fon.com/.
 J. Hwang, M. Balazinska, A. Rasin, U. Cetintemel, M. Stonebraker, and S. Zdonik. High-availability algorithms for distributed stream processing. In ICDE, 2005.
 S. Madden, M. Franklin, J. Hellerstein, and W. Hong. Tag: A tiny aggregation service for ad-hoc sensor networks. In OSDI, 2002.
 S. Madden, M. Franklin, J. Hellerstein, and W. Hong. The design of an acquisitional query processor for sensor networks. In SIGMOD, 2003.
 R. Motwani, J. Widom, A. Arasu, B. Babcock, S.Babu, M. Data, C. Olston, J. Rosenstein, and R. Varma. Query processing, approximation and resource management in a data stream management system. In CIDR, 2003.
 P. Bonnet, J. Gehrke, and P. Seshadri. Towards sensor database systems. In Conf. on Mobile Data Management, 2001.
 M. Richtel. San Francisco gets proposals for free citywide wi-fi net. New York Times, Feb. 2006.
 M. Shah, J. Hellerstein, and E. Brewer. Highly-available, fault-tolerant parallel dataflows. In SIGMOD, 2004.
 B. Tedeschi. Big wi-fi project for philadelphia. New York Times, 2004.
 G. Tolle, et al. A macroscope in the redwoods. In SenSys, 2005.
 N. Xu, et al. A wireless sensor network for structural monitoring. In SenSys, 2004.