Information Technology Reference
In-Depth Information
2.1 Data Acquisition with WSN
Sensor nodes, a technology appeared in the late 90s, are composed of a micro-
controller, memory, different sensors, battery and a radio module. Sensor nodes
can be interconnected into special networks called WSNs and interact among
them. Such WSN networks are used to study the environment and for acquir-
ing different variables related to weather (temperature, humidity, pressure, and
others).
Within a WSN, data are acquired by source nodes and sent via radio frequency
to a special node (known as sink) connected to the base station. The base station
coordinates all operations of the WSN and can be a personal computer (PC) or
embedded system. Furthermore, the base station can store or transmit via the
Internet all the information collected by sensor nodes.
WSNs nodes must meet requirements such as autonomy, low power consump-
tion, low cost, robustness and reliability. Unlike traditional wireless networks,
WSNs nodes use communications protocols specifically designed for working with
scarce energy sources and hardware resources. In addition, these protocols are
not compatible with TCP/IP networks.
2.2 Data Management
Data collected by AMS through WSNs, can be used to provide a solution to
many scientific and commercial problems (e.g., frost prevention, fire detection,
etc.). The data management process starts when WSN data are sent to re-
mote machines. Next, data are stored and processed in order to extract useful
information. Next subsections detail different technologies used to WSN data
management.
Traditional Technologies. Generally, the use of isolated machines such as
computers and mainframes is adequate to process low volumes of non-critical
WSN data. A typical use case of isolated machines is when low volumes of data
(in the order of Kbytes) are sent from the base station deployed in-field to a
remote server. The external server stores the data and then proceeds to run the
processing application.
Although this technology is easy to use, it presents some problems for (i) pro-
cessing large volume of data, (ii) scaling to a large number of WSN nodes and
(iii) ensuring availability 24 hours a day - 365 days a year. A possible solution to
solve these issues is by using powerful servers, mainframes and clusters in appro-
priate datacenter infrastructures. However, this solution generates prohibitive
economic costs, at least for agro-meteorological applications.
As the use of traditional technologies is not always suitable, different au-
thors proposed the use of Cloud Computing infrastructures for processing WSN
data [3,4,5,6,7].
Cloud Computing. Cloud Computing is a computing paradigm for applica-
tion development and the use of computing and storage resources [8]. Through
