For example, this could mean that data is not transmitted
right away, but is bu?®ered instead and transmitted in larger, yet less frequent,
communication epochs. The benefit of this strategy is reducing the overhead
of radio startup [11].
The required communication rate Rdata of a node is not a constant, however.
In many applications, there are long periods with only very sporadic
tra?±c, interspersed by an occasional burst of high activity. Consider, for example,
a network to monitor forest fires that should remain active for years.
During most of its lifetime, the network will be quasi-dormant, with only infrequent
network management packets being sent, such as to check on the
status of the network every hour. Since the network is static, attributes such
as self-localization or routing tables likely do not need to be updated more
frequently either. However, once a fire is detected, this triggers nodes to send
data, possibly engaging in local collaboration, all of which result in a period
of high tra?±c activity. A possible scenario is depicted in Figure 3.
197
Curt Schurgers
Fig. 3. A typical activity profile.
In this chapter, the focus is on those quasi-dormant periods during which
there is only very sporadic network activity, which could constitute a vast
portion of the total system lifetime. In this case, the communication rate
Rdata is extremely small. Furthermore, the amount of information during a
communication epoch is lower bounded by Lmin, which is a function of the
minimum packet size.
Pages:
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328