Cryptography Reference
In-Depth Information
the sink node. However, metrics such as fault tolerance and reliability necessitate the
deployment of additional sensors, yielding additional energy consumption, so that the
network can recover swiftly and deliver accurate sensed data to the sink, despite some
sensor or link failures. Hence, routing and data-dissemination protocols should con-
sider the trade-offs between fault tolerance, reliability, energy, and delay. Recall that
energy is a constraint that should be considered by any routing and data-dissemination
protocol to guarantee an efficient usage of the amount of energy available at each sensor.
7.2.6 Network Dynamics
Requirements such as limited-energy use (discussed previously) and goals such as mobility
have had direct impact on the design decisions of WSN network topology. In theory,
a deterministic sensor deployment approach would provide even coverage of the area
that has to be sensed. In addition, this approach would require fewer numbers of sensor
nodes for accomplishing the required sensing task. However, in a real environment with
an uneven terrain, it can be extremely challenging to apply a deterministic sensor deploy-
ment strategy. As a result, we are only left with the option of distributing the nodes in a
random fashion. Consequently, not all areas of the sensing region are evenly covered by
the sensors, leading to a coverage hole. In addition, there is a possibility of not all sen-
sor nodes in the network being connected with each other or even with the sink node.
In such situations, mobility plays an important role and becomes the main source of
network dynamics that can be used to solve problems. In any sensor network, the aggre-
gated data will be transmitted over some established paths between the source sensors
and the cluster heads or sink node. And the establishment of optimal paths depends on
whether the sensors are static or mobile. Hence, routing and data-dissemination proto-
cols can be classified based on whether a network is static or dynamic.
In a static network, every node in the network is static—that is, both the sensors
and the sink node remain in their fixed positions during their collaborative opera-
tion of monitoring a physical environment. Therefore, there is not much overhead
required to maintain routes between the sensors and the sink and between the sen-
sors themselves. In particular, the positions of the sensors and the sink can be learned
before data exchange, by exchanging some control messages. In certain cases, if the
terrain is familiar, node positions can be preconfigured in nodes before deployment.
Furthermore, neighbors of a given sensor do not change unless a new sensor has joined
the network or an existing sensor has left the network, either by its own will or because
of exhaustion of its battery life.
In a mobile network, either the sensors are moving or the sinks or cluster heads
are moving. As a result, the routes between the sensors and the sink are changing fre-
quently in such a dynamic environment. Hence, a currently active route could at any
time become inactive. This route instability would result in additional overhead and
delay in discovering valid routes for data transmission and forwarding. To overcome this
drawback, routing algorithms have been proposed in which the ordinary sensors and
sinks are designed to be static, whereas certain relay nodes such as cluster heads could be
mobile. One such example is the MULES-based architecture (Shah et al. 2003).
Search WWH ::
Custom Search