Finally, notice
a two-dimensional range query ???list all the events in [0.3??’0.5, 0.4??’0.8]??™. The
range of the query [0.3 ??’ 0.5, 0.4 ??’ 0.8] intersects with a set of zones and the
309
Jinbao Li, Zhipeng Cai, and Jianzhong Li
query can be resolved by routing the request to the nodes that own those
zones.
5 Query Processing in Sensor Networks
We can observe some physical phenomena in the natural environment by
querying and analyzing data from wireless sensor networks, e.g., monitoring
the level of rainfall in a certain region, obtaining the status of a disaster area,
supervising the vehicles in a big city, etc. Most queries about the data in
sensor networks can be classified into three types [2, 3]:
??? Historical query: it queries the historical data gathered from a wireless
sensor network, e.g., ???list the average level of rainfall in region A for 1999.??™
??? Snapshot query: it retrieves the status of a sensor network at a given
moment, e.g., ???return the current temperature of all sensors.??™
??? Continuous query: it concerns the status of a sensor network for a given
time interval, e.g., ???For the next 36 hours, list the average rainfall every
30 minutes in region A.??™
Queries in sensor networks can be expressed by a SQL-like languagewhich
has been introduced in the above section. Here, we will discuss several up-todate
query processing techniques in sensor networks.
5.1 Centralized and Distributed Query Processing
There are two kinds of techniques to process queries in sensor networks, centralized
query processing and distributed query processing.
Pages:
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498