In this paper, we focus on load-balancing across
clouds containing replicas of services. And since
cloud services are payable, we try to minimize, the
most possible, the cost to pay by a Cloud customer.
The literature shows that MCDA techniques are
indeed effective and can be used for cloud service
selection. Also, several works do reveal that TOPSIS
and both outranking methods (ELECTRE and
PROMETHEE) are more suitable for this purpose. If
the number of available services is very large, then
TOPSIS is appropriate because of its computational
simplicity, so we will use this later in order to rank
service alternatives and in order to select the best one.
And, since our system changes dynamically (Cloud
states change dynamically according to service
requests), we will use Discrete Time Markov Chain
(DTMC) that is well established analytical tool for
understanding dynamic systems behavior (Kemeny
and Snell, 1976), and it has been applied to a variety
of practical problems in real-world domains.
The remainder of the paper is structured as
follows: In Section 2 some related works are
presented. In Section 3, first the overall model of the
cloud service selection, in multiple cloud
environments, is presented, and then how this model
can minimize the cost and in the same time how it
balances the load between the different clouds, is
discussed. The paper is concluded and an outlook on
future work is given in Section 4.
2 LITERATURE REVIEW
Making a decision involves that there are alternative
choices to be taken into account and in a such
situation we need to choose the one that best fits with
our goals, objectives, desires, values and so on.
Several researches show that MCDA techniques are
effective and can be used for cloud service selection.
In this regard, several approaches which are based
on MCDM techniques, that assist a user in making a
service selection decision in the Cloud environment,
exist. For instance TOPSIS (Qu and Chen, 2009),
AHP (Zuo et al, 2008), and PROMETHEE (Karim et
al, 2011) were applied for service selection. Chen et
al. In (Qu and Chen, 2009), the authors developed a
general QoS-based service selection method. By
importing the proposed QoS ontology into OWL-S
standards, the proposed method can express Web
service's nonfunctional attributes in a semantic and
extensible way. Web service QoS based selection is
formulated as a multi-criteria decision making
(MCDM) which can be solved by using different
MCDM models to evaluate QoS criteria of the
candidate Web services. The values of quality
parameters of a Web service are normalized to a non-
negative real-valued number where higher normalized
values represent higher levels of service performance.
(Zuo et al., 2008) focused on the problem that how
to select the optimal service among many Web
services which all meet the functional needs,
establishes an index system for Web services products
selection from four aspects, namely the supply side,
the user, product and environment. Based on this, the
authors collected the views of 30 experts by Analytic
Hierarchy Process (AHP) method and calculated the
weight of each index at all levels based on the data
collected from questionnaire survey. In the overall
sample data analysis, the authors put two types of
sample data namely business operation experts and
academics for comparative analysis. The Web
services selection model proposed could provide the
reference to Web services managers when they
selecting Web services, and also contributed to in-
depth research on the adoption of Web services based
information system.
In (Karim et al., 2011), the authors proposed to
use an enhanced PROMETHEE model for QoS-based
web service selection. The first enhancement of
authors was to take into account the QoS
interdependency by using the Analytical Network
Process (ANP) to calculate the weight/priority
associated with each criterion. User's QoS
requirement is not considered in the original
PROMETHEE model. As a consequence, the second
enhancement of authors was to check the outranking
flows of each service with respect to the request in the
ranking step, so that they knew how well a service
satisfies the user requirement.
(Chen et al., 2012) proposed a system that enables
automatic conflict detection between the user's
criteria and enterprise policies in cloud service
selection for enterprises. Their system checks various
conflicts which result from the violation of enterprise
policies and inconsistency in a cloud service user's
requirements. Then, it selects an appropriate service
that satisfies the user's requirements and also
complies with enterprise policies, using constraint
programming. Zia et al. (Rehman et al., 2011).
presented the cloud service selection problem as a
multi-criteria decision making problem by proposing
a mathematical framework for multi-criteria cloud
service selection.
A few numbers of existing approaches, however,
simultaneously consider user requirements (as well as
cost) and load balancing between Clouds. In contrast
to the above research works and discussion, we
approach cloud service selection in a multiple cloud