Despite the fact
that hypnosis testing can be applied to process aggregate functions MAX and
MIN directly, some aggregate functions, such as SUM and COUNT, cannot
use this technique.
This section introduces some aggregate techniques in sensor networks.
Pipelined aggregation can increase throughput of the network, and cope with
transmission failures. Taking the advantages of a shared channel can reduce
the number of transmitted messages and improve the accuracy of aggregate.
Using hypnosis testing can also reduce total tra?±c in the sensor networks.
While processing aggregate function AVERAGE with hypnosis testing, a
user defined error range is required. If the aggregate error does not exceed
the threshold after discarding some readings, those data do not need to be
transmitted.
5.3 Continuous Query Processing
Query processing in sensor networks is a kind of distributed query processing.
It involves one global query processor and some local query processors at
each sensor node. We begin with the local query processors. In this section, we
discuss continues query processing which is suitable for local query processors.
Infinite real-time data streams are produced in wireless sensor networks,
which are the query objects of users. Continuous query is commonly used by
users. After a continuous query is proposed, the global query processor decomposes
the query into a set of sub-queries and sends them to the corresponding
sensors to process.
Pages:
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508