Once the average is computed,
it is sent back to the neighboring nodes for update.
Table 3. Asynchronous Di?®usion Algorithm
Step Description
(1) For each node N with uniform probability
(2) Obtain clock readings from neighbor nodes
(3) Compute the average of the clock readings
(4) Send the average reading back to neighbors,
which then update the clock
This approach is considered as the timing technique of type (3) since no
precise time server is needed. Although the asynchronous di?®usion algorithm
will converge to an average value, the amount of overhead is large since each
node needs to perform the algorithm. Also, the convergence varies with the
size of the sensor networks and range of the transmission radius.
In summary, these timing techniques may be used for di?®erent types of
applications; each of them has its own benefits. All of these techniques try to
address the factors influencing time synchronization while designed according
to the challenges as described in Section 4. Depending on the types of services
required by the applications or the hardware limitation of the sensor nodes,
some of these timing techniques may be applied.
7 Conclusions
The factors influencing time synchronization and design challenges are described
in Sections 3 and 4, respectively. They are to provide guidelines for
231
Weilian Su
developing time synchronization protocols. The requirements of sensor networks
are di?®erent from traditional distributed computer systems.
Pages:
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380