Building a Sensor Grid for Real Time Global Positioning System Data Galip Aydin 1,2 , Zhigang Qi 1,2 , Marlon E. Pierce 1 , Yehuda Bock 3 , and Geoffrey C. Fox 1,2 1 Community Grids Lab 2 Department of Computer Science Indiana University 501 North Morton Street Suite 224 Bloomington, IN 47404 3 Cecil H. and Ida M. Green Institute of Geophysics and Planetary Physics Scripps Institution of Oceanography La Jolla, CA 92093 {gaydin, zqi ,mpierce, gcf}@indiana.edu, ybock@ucsd.edu Abstract. We describe the architecture of our streaming sensor grid system. Using a topic-based publish/subscribe methodology, we are able to build a scalable system for managing real-time data streams produced by the California Real Time GPS Network. The architecture is based on atomic, extensible elements called filters that receive, modify, and republish 1 Hz GPS data streams in our deployment. Our filter approach can be extended to include sophisticated data analysis and event detection applications. Keywords: Real time data streams, global positioning system, sensor webs, publish/subscribe middleware, message-oriented middleware. 1. Introduction Recent advancements in sensor technologies such as micro-circuitry, nano-technology and low-power electronics have allowed sensors to be deployed in a wide variety of environments [1-6]. The trend in this field shows that in the near future thousands of sensor nodes will be deployed either individually or as part of sensor networks in a large variety of application domains. Environmental monitoring, air pollution and water quality measurements, detection of the seismic events, and understanding the long-term motions of the Earth crust are example areas where the extent of the deployment of sensor networks can easily be seen. Extensive use of sensing devices