Image Processing Reference
In-Depth Information
of multichannel MAC protocols. While this scheme traditionally falls into the category
of frequency-based scheduling, a new interesting constraint is added to the problem: the
usability of a given channel depends on the amount of power used in the adjacent com-
munication bands. While the transmission power of the devices is usually kept constant,
the distance between devices is not constant (and might also vary with time)—thus the
topology of the network at a given moment is an important factor in deciding a particular
channel allocation scheme [].
•
Network layer
: his is responsible for routing of the packets inside the sensor network. It is
one of the most studied topics in the area of WSN and a tremendous amount of literature
is available on this topic. he main design constraint for this layer is, as in all the previous
cases, the energy efficiency.
The goal of WSNs is to deliver sensed data (or data aggregates) to the base stations
requesting it. The concept of data-centric routing has been used to address this prob-
lem in an energy-efficient manner, minimizing the amount of traffic in the network. In
data-centric routing, each node is assigned a specific task based on the interests of the
base stations. In the second phase of the algorithm, the collected data is sent back to the
requesting nodes. Interest dissemination can be done in two different ways, depending
on the expected amount of traffic and level of events in the sensor network: the base sta-
tions can broadcast the interest to the whole network or the sensor nodes themselves can
advertise their capabilities and the base stations will subscribe to that.
Based on the previous considerations, the network layer needs to be optimized mainly
for two operations: spreading the user queries, generated at one or more base sta-
tions, around the whole network and then retrieving the sensed data to the requesting
node. Individual addressing of each sensor node is not important in the majority of the
applications.
Due to the high density of the sensor nodes, a lot of redundant information is available
inside the sensor network. Retrieving all this information to a certain base station might
easily exceed the available bandwidth, making the sensor network unusable. The solu-
tion to this problem is the data aggregation technique which requires each sensor node
to inspect the content of the packets it has to route and aggregate the contained infor-
mation, reducing the high redundancy of the multiple sensed data. This technique was
provedtosubstantiallyreducetheoveralltraicandmakethesensornetworkbehave
as an instrument for analyzing data rather than just a transport infrastructure for raw
data [].
•
Transport layer
: This layer appears from the need to connect the WSN to an external
network such as the Internet in order to disseminate its data readings to a larger com-
munity []. Usually the protocols needed for such interconnections require significant
resources and they will not be present in all the sensor nodes. The envisioned scenario
is to allow a small subset of nodes to behave as gateways between the sensor network
and some external networks. hese nodes will be equipped with superior resources and
computation capabilities and will be able to run the needed protocols to interconnect the
networks.
As devices become more and more powerful and available bandwidth increases due to
technological advances and improvements in the protocols, the transport layer attracts
moreattention.Whileprobablynotneededduringtheregularusagemodeofasensor
network, there are cases (e.g., in the case of abnormal behavior detected by a device in
the network) in which large amounts of data need to travel between the user and a spe-
cific node in the network. Examples also include the need of configuring or updating a
software image on a given device.
