Abstract
In this paper, we study the important issues for execution of real-time queries in a wireless sensor system in which sensor nodes are distributed to monitor the events that have occurred in the environment. Three important objectives in processing the real-time queries are: (1) to minimize the number of missed deadlines, (2) to minimize the processing costs, especially in data communication; and (3) to provide temporally consistent sensor data values for query execution. To reduce the data transmission cost and delay time in gathering the right versions of data items for a query, we propose the Parallel Data Shipping with Priority Transmission (PAST) scheme to determine where and how to execute a real-time query. To meet the deadlines of the queries, a deadline-driven priority policy is adopted to schedule the transmission of sensor data versions to the coordinator node.
| Original language | English |
|---|---|
| Pages (from-to) | 188-199 |
| Number of pages | 12 |
| Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Volume | 3358 |
| DOIs | |
| State | Published - 2004 |
Bibliographical note
Funding Information:In this paper, we have studied how to process real-time queries in a wireless sensor system. Although many queries in a wireless sensor network processes real-time properties, the problem has been greatly in previous research although query processing in sensor networks has received a lot of research interests in recent years. To our best knowledge, this is the first study on the problem. Real-time queries are associated with a deadline on their completion times. In processing a real-time query, an important requirement is the temporal consistency. This issue has also been greatly ignored in the previous work in sensor data management. In this paper, we adopt the notion of relative consistency for processing real-time queries. We propose a synchronized parallel scheme called Parallel Data Shipping with Priority Transmission (PAST) to collect multiple versions of data items for execution of the query with smaller data transmission cost and transmission delay and at the same time to meet the processing constraints of the queries, i.e., deadline, data currency and result requirements. To meet the transmission deadline, a deadline-based priority scheduling algorithm is designed in PAST to transmit the data versions from a Acknowledgement. This work was supported, in part, by NSF grants IIS-0208758 and CCR-0329609.