How to Choose the Right Cloud - Advanced Design Approaches to Emerging Software Systems - page 223

over the long term. This means that one has
to estimate the utilization over the cloud for a
significant period of time. In these estimations
the “pay as you go” billing system offered by
cloud computing providers is evaluated in terms
of elasticity measured in resource utilization. An
IT manager will predict daily average and peak
demand measured for example as the number of
servers required and then he will be able to com-
pare utility computing versus privately owned
infrastructure. Also hardware expenditures should
be taken into consideration into economic models.
Hardware resource costs decline at variable rates
a fact that may lead to unjustifiable expenditures
compared to actual resource usage. Cloud com-
puting can track changes to hardware costs and
pass them through the client more cost effectively.
Klems et al. (2009) propose a framework for
determining the benefits of cloud computing as an
alternative to privately owned IT infrastructure.
The model presented is based on the business
scenario and the comparison of costs between
the two alternatives. The business scenario is
defined by the business domain and objectives,
the demand behavior and the technical require-
ments. For example, the business domain defines
whether an application will be used at a Business
to Business level or Business to Client level, or
for internal use. The business goals will point out
particular benefits coming from web hosting in
the cloud such as short time to market, reduced
costs, and software licenses violations. Demand
behavior also is an important factor that affects
the performance of services and applications in
the Web according to Kleims (2009). Demand
behavior can be seasonal, temporary or caused
by batch processing jobs.
Related studies that discuss the cost of familiar
to cloud computing models like grid computing
are (Kondol et. al., 2009) and (Optitz et al., 2008).
Performance trade- offs and monetary costs of
cloud computing compared to desktop grids are

analyzed in (Kondol et. al., 2009). The above
comparison involves two relevant architectural
platforms, cloud computing and volunteer comput-
ing, that present similar principles. Performance
comparison is quantified in terms of execution,
platform construction, application deployment and
completion times. Cost comparison is performed
in terms of technical requirements such as project
resource usage. The costs of relevant aspects of
cloud computing such as grid computing is ad-
dressed also in (Optitz et al., 2008). The study
analyzes different types of costs and determines
the total costs of a resource provider. Relevant
cost for resource providers include hardware,
business premises, software, personnel and data
communication expenses.
Practitioners on the other hand seem to be bigot
supporters of utility computing. Miller (2009)
states that cloud computing is a type of web-based
computing that allows easy and constant access
to applications and data from all over the world
through an internet connection and facilitates
group collaboration. Though he mentions that
cloud computing is not suitable for any case, stress-
ing the advantages and disadvantages of cloud
computing. Regarding costs he refers that cloud
computing reduces hardware and software costs
and increases the productivity of the employees
as they have access to their files and applications
from home as well. Among the disadvantages of
cloud computing related to costs Miller (2009)
mentions that cloud computing requires fast and
instant internet connections. Also data confiden-
tiality in the cloud is a subject under examination
that may cause economic loss (McGowan, 2009).
Knight (2009) argues that the dilemma between
cloud computing and on-premise development
is wrong and should be substituted by the ques-
tion of which is the right combination of cloud
and premise based assets. The combination of
the two approaches can indeed exploit the best
of both worlds.