Temperature readings, for example,
are usually the same from each of a cluster of sensors. If every node transmits its
sensed data to the sink, these multiple packets will contain duplicate readings. Also,
for applications that ask only for averages, sums or total counts from sensor nodes,
331
332 Kai-Wei Fan, Sha Liu, and Prasun Sinha
transmissions form individual sensors in the cluster only consume more energy and
reduce the bandwidth. .
Data Aggregation is a mechanism to reduce the size of packet transmissions. Instead
of sending all data to the sink, data aggregation processes the data in-network,
and only sends processed data to the sink. For example when the sensing task is to
collect the average temperature, data aggregation allows nodes to combine multiple
readings into one report containing the average temperature T collected from N
sensor nodes. When two packets, each of them containing (T1, N1) and (T2, N2) respectively,
are aggregated, they can be combined into one packet with (T, N) where
T = (T1 ?— N1 + T2 ?— N2)/(N1 + N2) and N = N1 + N2. Aggregation also can
be applied when the sensing task is to collect the MAX or MIN value of the sensor
readings. This can effectively reduce the traffic load, and therefore reduce energy
consumption.
There are many aggregation approaches. One is opportunistic aggregation. Packets
are aggregated opportunistically in the network when multiple packets are met
at the same node while being forwarded to the sink.
Pages:
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531