Title :
Aggregate Threshold Queries in Sensor Networks
Author :
Sharfman, Izchak ; Schuster, Assaf ; Keren, Daniel
Author_Institution :
Dept. of Comput. Sci., Technion, Haifa
Abstract :
An important class of queries over sensor networks are network-wide aggregation queries. In this work we study a class of aggregation queries which we refer to as aggregate threshold queries. The goal of an aggregate threshold query is to continuously monitor the network and give a notification every time an aggregated value crosses a predetermined threshold value. Aggregate threshold queries are of particular importance in a wireless sensor environment, since they allow network-wide events to be detected, with a minimum expenditure of energy. Such network-wide events might include, for example, the variance in sensor readings exceeding a certain threshold. We present an efficient algorithm for implementing arbitrary aggregate threshold queries over sensor networks. Our algorithm is based on a novel geometric approach by which an arbitrary aggregate threshold query can be split into a set of numerical constraints on the readings of the individual sensors. These constraints are used by the individual sensors to monitor their readings. The constraints are constructed so that as long as none of the constraints are violated, it is guaranteed that the aggregated value has not crossed the threshold. Experiments we performed on real-world data indicate that by employing these constraints, sensors are able to reduce the number of transmissions required for implementing the query by orders of magnitude, thus significantly reducing energy consumption.
Keywords :
distributed algorithms; wireless sensor networks; aggregate threshold queries; energy consumption; network-wide events; wireless sensor networks; Aggregates; Computer science; Control systems; Monitoring; Network servers; Operating systems; Sensor phenomena and characterization; Sensor systems; Temperature sensors; Wireless sensor networks;
Conference_Titel :
Parallel and Distributed Processing Symposium, 2007. IPDPS 2007. IEEE International
Conference_Location :
Long Beach, CA
Print_ISBN :
1-4244-0910-1
Electronic_ISBN :
1-4244-0910-1
DOI :
10.1109/IPDPS.2007.370297