M.V. Ramakrishna, K. Kandaswamy, and J. Patel (Australia)
Sensor data management, dynamic optimisation
We envisage a future where a number of sensor networks exist, each with thousands of sensors. One sensor network may be monitoring and controlling traffic, while another is used for environmental prediction. Users would be able to pose queries to find if there is heavy traffic on a particu lar road, if the shipments have reached their retail outlets etc. In this paper we address the issue of efficiently pro cessing a large number of dynamic queries in such systems with streaming sensor data. Unlike the traditional database systems where the query optimiser generates query plans a priori, we need dynamic optimisation capability to process sensor data as the pattern data may change. We present the simple architecture of our system, and techniques to pro cess such queries dynamically. The experimental simula tion results presented show the tradeoffs of our techniques and their efficiency.
Important Links:
Go Back
