![](bg2.png)
(Manjunatha et al., 2010) describe an approach
based on a domain specific language (Deursen,
Kling, Visser, 2000) that allows the development of
a so called cloud-mobile-hybrid application.
Basically, in this kind of applications, the core
functionality is provided in a Cloud Computing
scenario, whereas a small and tiny client application
makes use of this Cloud Service to allow a mobile
consumer to use the Cloud Service from a mobile
device.
In (Mishra, Elespuru, Shakaya, 2009) the authors
describe a mobile MapReduce system that allows to
solve portions of a problem on mobile devices.
Therefore, this approach could be seen as one of the
first Software-as-a-Service implementations that is
heavily based on mobile devices.
Furthermore, some work has already been
published with respect to mobilde devices as Web
Service providers, e.g. (Li, Chou, 2011) describe an
approach based on a modified HTTP protocol that
allows to provide Web Services on mobile devices.
Furthermore, in (Jansen, 2012) another approach
for providing Web Services on mobile devices based
on standardized protocols is described. Additionally,
this approach provides the ability to overcome some
of the usual problems by providing services on
mobile devices, such as frequent network changes
and so on.
3 EXAMPLE SCENARIOS
Just to show how reasonable it might be to have a
certain service running on a mobile device, this
section describes two scenarios that can be
implemented with the help of the describe approach.
The first example is sort of comparable to a
location based service: one of the most important
facts about mobile devices is, that these kind of
devices are more like a pack of different sensors,
than a single device. Usually, mobile devices
nowadays are equipped with a GPS sensor that
allows to track the position of a device, an
Accelerometer that allows to track the acceleration
of the device, a compass to track the heading of the
device and many other sensors as well. Therefore, it
makes perfect sense either to use the informations
provided by these sensor in order to provide
contextualized information to the owner of the
device while using a specific software, or to make
use of these kind of information in order to share
informations with others. Here, the first example
scenario is a fairly easy one related to the current
position of the device. Imagine Person A wants to
know the current temperature at a certain location. In
order to get this question answered, Person A can
just raise the question for the current temperature
along with the geo-coordinates of the location he/she
is interested in, to a Cloud Service. Then a mobile
device that runs the approach described in this paper
will retrieve the question raised to the Cloud and
can, if the device is currently located within the area
of the location in question, answer the question
about the temperature.
The second scenario is a completely different
one: Another major advantage of mobile devices is
the number of devices available. As said before
(IDC, 2011), already in 2011 the number of
smartphones increased 300 million units. Therefore,
if an approach similar to the one described in this
paper would be deployed at least to a subset of all
available smartphones, this would lead to a
tremendous amount of computational power.
Furthermore, another positive aspect of smartphones
is the fact that these kinds of devices are connected
permanently to the internet usually. Beside using the
tremendous computational power of these devices,
also other scenarios might be reasonable, e.g.
making a survey among customers might lead to a
question send to the Cloud and answered by a
tremendous number of mobile users in a very narrow
time range. Here, of course, the feedback of the
owner of the mobile device is important, what
provides a new scalability dimension for mobile
Cloud Computing based services.
4 IMPLEMENTATION
In order to describe the example implementation of
the presented approach, this section first provides a
classification of the approach. The second
subsection describes the approach more clearly and
provides a presentation of the example
implementation.
4.1 Classification of the Described
Approach
According to the NIST definition for Cloud
Computing (Mell, Grand, 2011) Cloud Computing
consists basically of three different service models.
Within this definition the most low level service
model is the Infrastructure-as-a-Service model, in
which infrastructural resources are provided on a
flexible basis. The most top level service model is
the Software-as-a-Service model in which a
CLOSER2012-2ndInternationalConferenceonCloudComputingandServicesScience
148