Information Technology Reference
In-Depth Information
On the one hand, in-field data acquisition process can be performed by mea-
suring instruments, weather stations and Wireless Sensor Networks (WSN) [1,2].
Compared to traditional measurement instruments and weather stations, WSNs
have the advantage that they can cover extensive areas with low cost devices
called sensor nodes. Moreover, sensor nodes low-power consumption and long
lifetime (over 2 years) allow long-term monitoring with low maintenance.
On the other hand, the WSNs data management include data remote access,
storage and data processing. This management process can be reliably and eas-
ily performed using Cloud Computing technologies [3,4,5,6,7]. The use of Cloud
Computing for data management allow to incorporate the benefits of this tech-
nology (data replication, fault tolerance, resources scalability, etc.) to AMS.
There are two main reason for using public Clouds in order to process and
store WSN data. The first one is the large volume of data generated by WSNs.
As an example, in the region of Cuyo there are up to 170000 hectares of crops
which can be instrumented with one sensor node per hectare. For this reason,
there are 170000 potential sensors that generate data, which must be processed
and stored in a proper infrastructure. The second one is the trac bottle neck
from the WSNs to an isolated private data center. Several Cloud providers offer
different types of public infrastructure resources which can be used to store and
process data in a profitable way. Today one of the leading providers is Amazon.
The Elastic Compute Cloud (EC2) toolkit service provides different types of
virtual machines (instances) for both processing and data storage. In addition,
due to the wide range of instances offered by Amazon, arises the need to identify
which of them has better performance, in terms of execution time and economic
cost, for processing frost prediction applications.
In this paper we propose a set of models, constructed from empirical data, that
can be used to estimate the performance and economic costs of Amazon EC2
instances applied to frost prevention applications processing. Although there are
other costs associated with the use of Amazon EC2 instances (like the ones
for data transfer), the target of our study are the economic costs for WSN
data processing. These ones are more relevant compared with the ones for data
transfer.
This paper is structured as follows. Section 2 introduces Agricultural Moni-
toring Systems based in WSN. Next, Section 3 surveys relevant related works.
Section 4 describes the application developed for frost prediction. Then, Sec-
tion 5 presents our proposal of models for each Amazon EC2 instance and the
methodology used to construct them. Finally Section 6 concludes this paper and
discusses future prospective extensions.
2 Agricultural Monitoring Systems Based in WSNs
In this section we provide an introduction to the technologies used to perform
both the data acquisition and data management in WSN's based AMS.
