3. If wakeup events are rare, as is
often the case in sensor networks, the more e?±cient sleep state of senderand
receiver-based solutions result in an overall lower power design than the
212
Chapter 8 Wakeup Strategies in Wireless Sensor Networks
rendezvous-based one. Note that once packets arrive more frequently, the network
is basically no longer in the quasi-dormant state, but moves into a more
awake state where it needs medium access, as was explained in Section 1.2.
Section 5.2 will further detail the interaction between MAC and asynchronous
wakeup.
The power e?±ciency of sender- versus receiver-based alternatives depends
on the fraction of time a sleeping node has to be awake either to listen for
or send information. Both essentially correspond pretty closely to the basic
scheme that was shown in Figure 4 of Section 1.3 and are well-analyzed by
the analysis there. In principle, a sender-based solution with a wakeup tone
results in the shortest active time, and therefore the lowest power consumption.
However, in the case of a wakeup event, all neighboring nodes to the
sender will wake up. Whether the overall average power consumption is still
better depends on the frequency of such wakeup events and the density of the
network. If the active fraction is the same for a sender- and receiver-based solution,
their main di?®erence lies in the probability of beacon/announcement
collisions. In the case of receiver-based wakeup, where sleeping nodes periodically
send out the announcements, this probability is a function of the
network density and the period.
Pages:
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351