2014 47th Hawaii International Conference on System Science

Greening the Service Selection in Cloud Computing: the Case of Federated

ERP solutions
Holger Schrdl
Otto-von-Guericke University of Magdeburg
Chair of Business Informatics I
Universitaetsplatz 2
39106 Magdeburg / Germany
[email protected]

Abstract
The increasing industrial acceptance of cloud-based
IT services has led to a paradigm shift in the
development and operation of complex business
applications. IT managers are incorporating cloudbased IT services as replacements and/or as
enhancements for existing on-site solutions. This
strategy leads to the concept of a federated business
application, which consists of a variety of on-site and
cloud-based subparts, dynamically orchestrated to a
single solution. In this setting, the selection of
appropriate IT services is critical. Following the
discussions of environmental thinking in IT aspects
like Green IT and Green IS, the issue of appropriate
IT service depends on not only functionality and
costs, but additionally also on the environmental
impact of the service selection. To address this issue,
this paper presents a service selection model for
cloud-based services with a focus on the
environmental aspects of the selection process. This
is done by modeling a multi-criteria decision model
based on the rational choice theory. The proposed
model provides a selection process which leads to
maximal service functionality coverage with minimal
environmental impact for a given service provider
setting. The application of the model is illustrated in
a typical industry case.

1. Introduction
The market for cloud-based IT services is increasing
steadily [32]. This gives companies the opportunity
to select suitable services to restructure and improve
their IT infrastructures. The selection of an
appropriate service is challenging [6]. First, it is not
easy for companies to formulate their requirements
for a particular complex IT solution [10]. From the
service provider perspective, a similar problem lies in
the service description. In combining an unspecified
978-1-4799-2504-9/14 $31.00 2014 IEEE
DOI 10.1109/HICSS.2014.519

Paulina Simkin
University of Augsburg
Chair of Business Informatics and Systems
Engineering
Universitaetsstr. 13
86179 Augsburg / Germany
[email protected]

service request and a uncertain service description, it

is highly unlikely to find an optimal solution at first.
Moreover, service identification is a critical issue.
Several other aspects involve the service pricing, the
formulation of appropriate service level agreements
and security issues [24]. In the end, the so-called
provider lock-in gains more importance since a cloud
market lives from its dynamic structure. Provider
lock-in means in general that the replacement of
similar services from one service provider to another
faces relevant obstacles.
For isolated services, various solutions exist in
research and industrial practise. But with respect to
the composition of different isolated services to a
service bundle, the complexity rises and the
challenges increase. In addition to the previous
mentioned issues, the coordination between the
particular services brings up new questions.
Moreover, possible scaling effects may be achieved
from bundling services. Lastly, the lock-in problem
gains more importance since the change of a provider
may affect other elements of the service bundle.
Besides all these technical and economic issues,
ecological issues are becoming the focus of
discussions in cloud-based services [3]. There are
several works on the linkage between cloud
computing and the positive impact of environmental
issues like carbon emission and energy reduction [8].
Despite the fact that these discussions are at a very
early stage, the general scholarly consensus is that the
consideration of environmental issues in information
systems is a highly relevant issue for the future.
Therefore, all research that sheds light on this linkage
generates a value on the future scientific grounding.
The present work takes care of this linkage by
investigating the cloud-based service selection under
an environmental perspective. The key question is
how service selection for companies can be done by
minimizing the environmental impact of the decision.
To investigate this, a case of a federated ERP system
4200

that should be restructured by cloud-based IT

services is considered. The research question is: How
can the selection of cloud-based services for a
federated ERP system be modeled to obtain
maximum eco-efficiency in the whole system?
The article is structured as follows: in Section 2, the
foundations of cloud computing, federated ERP,
Green IS, and the research approach is displayed. In
Section 3, the service selection process is discussed
in detail, and the service selection model is
developed. This model is a multi-criteria decision
model with an emphasis on environmental aspects. In
Section 4, this model is applied to a case of federated
ERP to demonstrate the application of the model.
Implications of the model and limitations are
discussed in Section 5. The paper closes with a
summary and future research in Section 6.

billing for services. Other aspects are also the better

use of their own IT staff, a reduction of manual
processes, and an improved transparency of
processes. Emphasis on the consideration of cloudbased systems lies in the areas of security and trust
[47].

2.2 GreenIS and Green Cloud Computing

The research field Green IS includes measures and
solutions to understand the issues of the efficient use
of energy and the development of innovative and
environmentally friendly strategies and information
systems. These information systems can contribute
significantly to energy and resource efficiency in
companies [38]. A key aspect is the realization of
potential savings through the efficient design of
business processes using Green IS [7].
In differentiating the development of Green IT to
Green IS, the technical analysis refers to the design
and implementation of IT systems that help support
sustainable business processes. Thus, Green IS
addresses broader and deeper problems than Green IT
and thereby derives from a greater potential to
achieve energy reduction. Therefore, Green IT can be
considered as a branch of Green IS [46].
The realization of Green IS measures is typically
associated with capital expenditures for companies.
However, Green IS and the sustainable development
of organizations should not be seen as a cost factor,
but as an opportunity to improve company
productivity, reduce costs, and increase profitability
[46].
There are initial attempts to investigate the issues of
Green IS in the context of Cloud Computing. The
majority of this research is concentrated on the issue
of energy consumption by the cloud service providers
(see for example [37, 8, 48, 4, 13]). An extension of
this viewpoint is done by Baliga et al. when they
investigated energy consumption not only in data
centers, but also switching and data transportation
[3]. Still, however, this work considers only the
technology part of cloud service provisioning.
Environmental factors beyond this scope are not
considered here.

2. Research Background
2.1 Cloud Computing
Cloud computing is a topic that is given increasing
attention from IT managers in companies. The term
"Cloud Computing" goes back to a collaboration
announcement between Google and IBM [36]. Before
this time, various other technologies were discussed
in the market which may be considered as
predecessors of the term Cloud Computing like "Grid
Computing" [14], "Computer in the Cloud" [35] or
"Dreaming in the Cloud" [43]. A current definition of
"Cloud Computing" is given by NIST: Cloud
computing is a model for enabling ubiquitous,
convenient, on-demand network access to a shared
pool of configurable computing resources (e.g.,
networks, servers, storage, applications, and services)
that can be rapidly provisioned and released with
minimal management effort or service provider
interaction.[33]. Breaking this rather complex
definition down we can describe Cloud Computing as
the delivering of infrastructure, platform, and
software in a service model based on a pay-per-use
model provided to the customer [10]. The market for
Cloud Computing is segmented into three service
models: Infrastructure as a Service (IaaS), Platform
as a Service (Paas), and Software as a Service (SaaS)
[33]. Following a recent study by Sterling
Commerce, a software provider from Duesseldorf,
Germany, 87 percent of senior IT managers in
Germany plan the move to cloud-based information
systems in the B2B sector [23]. Main driver for this
development according to the survey is cost pressure:
most companies promise a cost reduction through the
use of cloud-based IT-structures due to a usage-based

2.3 Federated ERP Solutions

Enterprise Resource Planning (ERP) systems are
information systems to support all business processes
within one company [20]. In the classical view, ERP
systems incorporate different types of business
applications into one holistic system. These systems
comprise several functions and requirements which
are needed in the company. Depending on the

4201

specific requirements of a company, the installation

and configuration setting is determined by the
customer itself or by a service provider for the ERP
system. In contrast to this monolithic view, a
federated ERP system is an ERP system which
consists of system components that are distributed
within a computer network [5]. These system
components are provided by network nodes, which
are represented by different service providers. The
overall functionality of the ERP system is the
aggregation of the individual functionality of the
single nodes. The components, which are combined
to the entire ERP solution, may be developed and
provided by different service providers [1]. From a
consumer perspective, the collection of components
is perceived by the user as single ERP solution. From
a cloud perspective, the area of federated ERP
solutions is also discussed in the scientific
community of inter-cloud research [9].

this research, experimental evaluation based on an

artificial business case was selected to demonstrate
the usability of the proposed model. The business
case was developed with leading industry experts
from the area of cloud service providers and business
application vendors.

3. Service Selection for Federated, Cloudbased ERP Solutions

3.1 Service Selection Process
Throughout the course of this paper, the focus is on
the consideration of the supplier-related tasks, which
include essentially methods for the identification,
selection, and qualification of potential suppliers
[45]. Supplier identification is used to identify
service providers on a market who provide the
required procurement item, a cloud-based service.
Supplier selection covers different activities; firstly,
the potential service providers have to be analyzed
and rated. Sometimes it is necessary to add additional
information to the first results of the search because
the potential service providers do not provide enough
self-information to get a rating. Secondly, for rating
purposes, it would be beneficial to use existing
ratings for the service providers to integrate them into
a new ranking or to compare the existing ranking
with the new ranking. Supplier qualification follows
as the third step and leads to the long-term aspects of
the supplier-buyer relationship. After the selection of
relevant service providers, there is a need to maintain
and develop the supplier-buyer relationship for
forthcoming issues. In literature, these activities can
be found under the term supplier relationship
management [28]. Regarding the development of the
relationship between a service provider and a
customer, the selection process can be regarded as the
first step in their collaboration. The importance of
this first step is shown by the fact that customers
often buy services in a certain order, starting with
relatively simple services to more complex and
expensive ones [11].

2.4 Research Approach

The main objective of this research is to create a
decision model for cloud-based IT services with a
focus on aspects of environmental impact. The
research uses a design science approach. Design
science is a research method to solve organizational
problems by creating and evaluating IT artifacts [21].
These IT artifacts are denoted as constructs, models,
methods,
and
instantiations
[31].
Several
methodological approaches for the conduction of
design science research are proposed in the literature
(see for example [31, 39, 27]). For this research, we
have adapted the design science research
methodology from Kuechler & Vaishnavi. The first
process step, problem recognition, has been
addressed in the introduction. It could be shown that
the inclusion of environmental aspect in the decision
making for service selection is relevant, but missing
in research and industrial practice. This leads to the
main research question: How to develop a decision
model with emphasis on the environmental aspects of
the IT service selection? In the second process step,
we suggest the adaption of a multi-criteria decision
model. This suggestion is based on a broad literature
study, which identified relevant measures and
existing artefacts for assessing these measures. The
new model is developed in the third step by
integrating environmental aspects as criteria groups
into the decision model. This development extends
existing decision models and leads to model which is
varying in the solution approach based on the type of
criteria integration. For the evaluation of the
proposed model, we follow the guidelines for design
science research according to Hevner et al. [21]. In

3.2 Selection Criteria

In IT service management, the selection of services
and their corresponding service providers is a
challenging issue [40]. Mainly, this is because
services are generally characterized by intangibility,
inseparability of product and consumption, difficulty
of standardization, and perishability [22]. It is vital
for the service provider to know the specific
requirements of the service consumer when they try

4202

to make their decisions for a particular service

offering [18].
The selection of IT services is based on various
criteria. These criteria can be categorized in three
different areas: First, functional criteria have to be
taken into account [17]. In the functional area, it is
necessary to ensure that the functional requirements
of the service consumer are met by the cloud service.
Second, consumers decide for a service based on
non-functional criteria [42]. Criteria in this area are,
for example: price, quality, and image of the service
provider. Third, emerging ecologic criteria come into
play [3, 30, 16]. The discussion of the ecological
impact of cloud computing is gaining more
momentum. It is assumed that substituting on-site IT
services with cloud-based services has a positive
impact on ecological issues (see for instance an
industry study by Accenture and WSP Environment
& Energy [2]). Several criteria have to be considered
for an ecological perspective. In the forefront, there
are rather obvious factors like carbon emissions,
waste, and equipment recycling ratios. Additionally,
factors like noise, heat, and carbon footprints of
tangible aspects (for example, if consultants have to
implement parts of the IT services on location and
need to travel to the company) are taken into
considering. Research on these and further aspects of
the ecological description of cloud services is only in
their beginning [15].

which are typically defined as explicit or implicit. In

the first case, each decision alternative is represented
by its behavior in each criterion. The challenge is to
find the best alternative for a decision maker or
sorting the alternatives in relation to a specific goal.
In the second case, the alternatives are not explicitly
known, and the challenge to find an alternative is by
solving a mathematical model. These are called
multiple-criteria design problems and are discussed
in the area of mathematical programming.
Formulating a multi-criteria decision problem consist
of three key elements: First, the criterion space. The
criterion space consists typically of a set of criterion
functions that represent the relevant criteria of the
items to be selected. Second is the decision space,
which is the set of possible decisions for the problem.
Third, constraints are used to describe the
relationship between different criteria with respect to
the decision space.
There is a large list of methods which can be used to
solve multi-criteria decision models. The most
prominent methods are the Weighted Sum Model
(WSM), Value Analysis (VA) and Data Envelop
Analysis (DEA). It is necessary to take care of
applying the appropriate solution method to the
characteristics of the model.
The discussion of using multi-criteria selection
methods for optimal service selection in Cloud
Computing has already started (see for example
Rehman et al. [40]). This work provides a general
approach to use multi-criteria decision models for the
cloud service selection based on very common
attributes like cost and performance. The work is
extended to an investigation of multi-criteria
selection on the IaaS service level for Cloud
Computing services [41]. On one hand,
environmental aspects are not considered as factors
for decision making. On the other hand, the question
remains whether these investigations will hold true
when considering the SaaS service level, which is
necessary for selecting services for business
applications.

3.3 Service Selection as a Decision Problem

Making a decision for an item which is described by
various criteria can be a challenging task. It is not
only about making the right decision for a single
attribute, but rather for the whole combination, which
may contain contradictions, making it difficult to find
an optimal solution. To address this, operation
research provides methods to address this problem. In
particular, the described case can be considered as a
multi-criteria decision making problem. In daily life,
we can see the problem of conflicting criteria. For
example, in buying decisions, the criterion of cost
often contradicts the criterion of quality. It is unusual
that the cheapest offer provides the highest quality.
By looking at the IT service industry, we see a
conflict between the cost of providing a particular
service to the ratio of customer satisfaction. These
small examples show that making the right decision
in multi-criteria decision environments needs a
systematic approach to determine the optimal
solution.
Research in the field of multiple criteria decision
making started in the early 1960s [26]. There are
different types of multi-criteria decision problems,

3.4 Problem Formulation

We denote   as the service with index ,
selected from the set of services X. The service  is
provided from a service provider  and is able to
fulfill the requirements of a certain service to the
degree of
 . These requirements may not only be
functional requirement, but also non-functional
requirements (e.g. quality, performance), risk
constraints and regulatory issues. Services might be
grouped together in service classes. These service
classes act as an umbrella for services providing

4203

Minimizing costs:

similar functionality and helps the model to narrow

the selection of an appropriate service within a
service class to ensure no double-selection of services
for the same requirements. Each service  is related
to its specific environmental impact  . The costs of
the service are subdivided in two parts: the
integration costs 
 and the compensation costs

 .
The integration costs are the costs related to the effort
of integrating the single services into the basic
business application. The compensation costs are
those costs which arise because of a lack of
functional coverage of the service. A single service
 can be written as tupel:

min  



with the constraints:

1)  , ,  from
for   

service

classes

3)
 1
The constraints are interpreted as follows: First, the
selected services have to cover all the required
service functionalities. Therefore, different services
have to be from a different service class to avoid that
one service functionality will be provided by two
different services which provide the same
functionality. Second, all impact and cost measures
should be positive. Third, the service coverage of a
single service should be in the range of 0 (no
coverage) to 1 (full coverage).
The solution of the problem depends on the modeling
of the service orchestration. In the simple case of
connecting single services together with no or limited
interaction, the problem is a linear programming
problem and can typically be solved easily by using
the Simplex method. For a higher integration ratio
between the selected services, the model may
develop to a dynamic programming model. It can be
assumed that in this case, uncertainty will be an
essential part of the service orchestration. Therefore,
the problem may develop to a probabilistic dynamic
programming model. As soon as we leave the level of
dyadic relationships between the suppliers and
consider a supply network, the problem will no
longer remain linear. The study of these types of
problems will lead to new insights in the
requirements for service markets in the future. The
inclusion of uncertainty and the consideration of
supply networks will be part of the future work.

The aim of the selection process is to find a set of

services which cover the required functionality
maximum while concurrently minimizing the
environmental impact and costs. As a decision
problem, it can be formulated as follows:

max  


with  as the functional orchestration of all

selected services. The term functional orchestration
describes to integration of the selected services with
the existing system environment on a system level,
data level, and process level. The orchestration
covers all tasks to integrate and run the selected
services in combination with the existing platform as
a single solution. In a simple setting, the services
might by orchestrated by just selecting the single
services without any integration. In this case, the
orchestration is just an adding of services and would
lead to a sum in the model formulation. In a complex
setting, service might be highly depending on each
other. Therefore, orchestration would eventually lead
to high individual programming effort, which would
lead to a multiply formulation and to a non-linear
model. This type of general orchestration concept is
necessary to reflect the specific requirements for
federated ERP systems. In this type of market, based
on the component structure of the ERP system, the
orchestration might be different beginning with the
mentioned simple setting up to the complex situation
of individual programming effort for component and
service integration.

4. Model Evaluation
According to the selected design science approach,
several methods can be used to evaluate the proposed
model. In this paper, the decision was made to use an
experimental evaluation for the proposed model. The
decision for an experimental evaluation was based on
the fact that the business case of service selection in

Minimizing the environmental impact:

min  


different

2)
 ,  , 
 ,

 0

 = ( ,
 ,  , 
 ,

 )

4204

federated information systems is just at the beginning

of practical implementation, and established industry
cases are not available yet. Therefore, to evaluate the
proposed model and to demonstrate the application of
the proposed model in industrial practice, an artificial
business case has been developed to serve as an
evaluation foundation. This artificial approach may
be classified as a simulation, which is an accepted
means to evaluate a proposed artifact [21].
This business case has been developed based on
intensive round-table discussions with industry
experts in the area of data center provisioning and
cloud services providers. There was a first discussion
with eight industry experts to gain a common
understanding of the business case and to collect
existing experiences in current industrial practice. In
addition, three experts from a leading ERP solution
provider were involved in the discussion of the
scenario. The round table took one day of discussion
in a moderated group setting. All experts agreed on
the fact that providing this kind of service is one of
the future challenges in information system
provisioning and that an enhanced understanding of
these selection processes would have a significant
impact on future business models for IT service
providers. The discussion was voice recorded and
analyzed immediately after the session to extract the
relevant issues. Based in these first findings, a
business case was developed as a foundation for the
second round table. Once again, the proposed
business case was structured discussed with the same
11 industry experts and refined to its current state.
This refinement took again one day of round table
discussion in a grouped environment and was voice
recorded again. Based on the experiences and
knowledge of future user requirements, the following
business case has been developed to represent the
core of the future challenge. Therefore, this case may
be considered representative of the addressed
problem from an industry experts view.
The business case may be described as follows:
Assume an existing ERP solution which is
implemented on-site by an industry company. This
ERP solution should be enhanced in three functional
areas: First, for the CRM (Customer Relationship
Management) model, an address validation service
should be integrated to enhance the quality of the
customer addresses. Second, the company wants to
eliminate all physical fax machines by introducing a
cloud-based fax service. Third, the IT department
wants to switch to a backup service level with valueadded functionality for the company users. To
achieve these enhancements, the company decides to
follow a federated ERP system approach rather than
implementing the enhancements by themselves.

Therefore, services have to be identified and selected,

which provide the necessary functionality and can be
integrated into the existing solution. For the address
validation, an online service has to be selected which
is able to receive address data from the ERP system,
validate them with a reliable result, and send the
validation result back to the ERP system. For the fax
solution, an online communication service has to be
selected which is able to transform different message
formats to a fax format in order to send it to the
corresponding recipients. The systems backup should
change from an internal solution to a cloud-based
backup service to ensure a better service level for the
backup functionality. For these three services, a
selection has to be made. The case situation can be
obtained from Fig. 1:

Figure 1. Federated ERP business case

In the IT service market, several service providers

exist who are able to provide these services. Some of
the IT service providers may be able to provide only
one particular service, whereas others may be able to
provide multiple services. For structuring purposes,
the services from the proposed business case will be
categorized by service classes. These service classes
consolidate all services from the IT service market
into one specific requirement.
This means that the selection problem for the
proposed business case is to select three services in
three different service classes. To apply the model, a
market structure is necessary to be able to conduct
the selection process based on the proposed decision
model. The market structure to reflect the major
market mechanism (service providers for unique
services and service providers for multiple services)
is defined as follows: for the first service of address
validation, three services are available. Two of them
come from service provider 1, and the third one
comes from provider 2. For the second service,
online fax, four services are available. One comes
again from service provider 1, one comes from

4205

service provider 3, and two come from service

provider 4. For the last requirement, the online
backup solution, two services are available, one of
which from provider 2 and the other from provider 3.
For every service, certain assumptions for the
characteristic have been made. First, for every
parameter (SC, EI, IC, CC), a value range was
defined which is aligned with a typical decision
situation in practice. For Service Coverage (SC), it
was assumed that the identified services are covering
the requirements quite well. Therefore, the range was
defined from 80% to 100%. For Environmental
Impact (EI), the assumption is that the impact will
not be higher than 50% of the possible impact. For
the Integration Costs (IC) and the Compensation
Costs (CC), no limitation was chosen because costs
may vary significantly and there is no reason which
leads to a limitation on the lower or upper limit. After
these restrictions, the parameter value of the specific
parameter was selected randomized by using the
Random function implemented in Excel. The
overview of the complete setting can be obtained
from Table 1 to Table 3.

The settings for the environmental impact are

assumed for every single service. It is assumed that
the environmental impact may decrease when two or
more services are provided by a single service
provider due to bundling effects. This assumption is
not yet part of the proposed model but will be part of
the future work on this topic.

Service
Service Class 2 Service Class
Class 3

Table 3. Settings for Service Costs (IC+CC)

Service
Service Class 2 Service Class
Class 3

SP1

SP2

Sp3

Sp4

x_AV-1
x_AV-2
x_AV-3
x_OF-1
x_OF-2
x_OF-3
x_OF-4
x_OB-1
x_OB-2

100
80
0
100
0
0
0
0
0

0
0
90
0
0
0
0
90
0

0
0
0
0
80
0
0
0
90

0
0
0
0
0
70
90
0
0

Service
Service Class 2 Service Class
Class 3

Table 2. Settings for Environmental Impact (EI)

SP1

SP2

Sp3

Sp4

x_AV-1
x_AV-2
x_AV-3
x_OF-1
x_OF-2
x_OF-3
x_OF-4
x_OB-1
x_OB-2

30
30
0
30
0
0
0
0
0

0
0
40
0
0
0
0
40
0

0
0
0
0
20
0
0
0
30

0
0
0
0
0
40
30
0
0

SP2

Sp3

Sp4

x_AV-1
x_AV-2
x_AV-3
x_OF-1
x_OF-2
x_OF-3
x_OF-4
x_OB-1
x_OB-2

40
140
0
40
0
0
0
0
0

0
0
100
0
0
0
0
30
0

0
0
0
0
140
0
0
0
60

0
0
0
0
0
120
90
0
0

The proposed decision problem has been solved by

using Solver for Excel 2010. The Simplex-LP
module was chosen since the problem can be
regarded as a linear optimization problem with
integer constraints. The solution to the selection
decision can be obtained from Table 1: the servicesprovider combinations which are marked in grey with
a 1, have to be selected to solve the formulated
decision problem. Therefore, the solution for the
proposed setting is to select the services

Services which are not provided by a particular

service provider are marked with 0.

Service

SP1

At the models current development stage, service

costs and compensation cost are combined into one
cost factor. Again, the bundling of two or more
services from one service provider may have a
positive impact on the cost factor, especially on the
decreasing of integration costs. This assumption is
not yet part of the model and has to be investigated in
future research.

Table 1. Settings for Service Coverage (SC)

Service

Service

x_AV 1 and x_OF 1 and x_OB 2

with x_AV-1 and x_OF-1 from service provider 1
(SP1) and x_OB-2 from service provider 3 (SP3).

4206

network then act as suppliers as well as customers

and the service selection face increased challenges.
One of these challenges is that the information
availability within a supply network over multiple
tiers decreases with the rising complexity of the
network [29]. Therefore, it is not obvious how to get
all information needed to apply the decision model. A
second obstacle is the rising complexity of cost
calculation, especially of integration costs, within the
supply network. This limitation arises because of the
decreased availability of information and as well
because of the problem of requirements specification
of services [12]. Answering these open questions
requires a further investigation of the situation.
Extending the experimental simulation will first give
insights as to how to approach these questions.
The proposed model is on a generic level highly
appropriate for a Green IS selection model.
Nevertheless, one serious challenge of the proposed
model is the quantification of the necessary
parameters, especially the environmental impact. In
current research, the environmental impact of IT
services can only be calculated for simple situations
[34, 25]. In a more complex one, the calculation of
environmental impact is not easy to achieve [44].
With an increasing understanding of linkage between
determining factors and more sophisticated methods
for environmental measurement, this situation will
become more transparent. This will lead to a better
application of the proposed model to increasingly
complex situations. The identification and
measurement of environmental impact factors for
cloud services will be part of future work of this
research.

Service
Class 3

Service Class 2

Service Class 1

Table 4. Solution for the service selection problem

Service

SP1

SP2

Sp3

Sp4

SC Sum
(constraint)

x_AV-1

x_AV-2

x_AV-3

x_OF-1

x_OF-2

x_OF-3

x_OF-4
x_OB-1
x_OB-2

0
0
0

0
0
0

0
0
1

0
0
0

Sum (Services)

This service selection has a service coverage rate of

97%, a mean environmental impact of 30, an average
implementation
cost
(integration
costs
+
compensation costs) of 40, and it provides the
optimal solution for service selection with
environmental focus.

5. Discussion
The experimental evaluation with an artificial
business case is on the one hand in alignment with
the generic rules for design science research [19], but
on the other hand can only be regarded as a first step
for a practical implementation of the model. The
business case was based on the experiences and
requirements knowledge of industry experts, but was
kept small to give a clear insight into the
functionality and the application of the model. To
understand the expected behavior of the proposed
model, the business has to be significantly enlarged
by considering more services, more service
providers, and more parameters for the service
description. This can be achieved by extending the
simulation data significantly and replacing the
predefined market setting used in this paper by a
randomized market setting based on a random link
between service providers and services. Taking this
randomized market setting as a foundation, the
simulation can be conducted with large numbers of
market participants to reflect a distributed and
fragmented market aligned to the practical situation.
The proposed model fits to the situation because
there are only dyadic relationships between the
company and the service providers. Complexity
increases significantly, if there will be a complete
supply network for IT service, where multiple tiers of
suppliers exist, and one supplier may handle the
sourcing request from the company to the supplier
one tier lower. This is a typical case in complex cloud
computing service markets. Suppliers in the supply

6. Summary and Outlook

The aim of this paper was to develop a decision
model for IT service selection that focuses on the
environmental impact of the service selection and the
resulting service combination. In the emerging IT
service industry there is an increasing acceptance of
cloud-based services as a substitution and/or
enhancement to one-site business applications, as
well as an increasing pressure for environmentalfriendly IT. Following this, we have developed a
multi-criteria decision model which incorporates the
environmental impact of an IT service with respect to
the corresponding service provider which has a
minimal environmental impact for a certain
maximum service coverage. The application of the
proposed model has been shown on a typical industry
business case.

4207

Environments for Scaling of Application Services", in

Algorithms and Architectures for Parallel Processing, D.
Hutchison, T. Kanade, J. Kittler, J.M. Kleinberg, F.
Mattern, J.C. Mitchell, M. Naor, O. Nierstrasz, C. Pandu
Rangan, B. Steffen, M. Sudan, D. Terzopoulos, D. Tygar,
M.Y. Vardi, G. Weikum, C.-H. Hsu, L.T. Yang, J.H. Park,
and S.-S. Yeo, Editors. 2010. Springer Berlin Heidelberg:
Berlin, Heidelberg.

This model can be considered as a first attempt to

include environmental thinking into the IT service
selection and, moreover, in the IT service industry
itself. In the next steps, the model will be extended
by the dynamic linkage between the different service
providers. From the current perspective, this model
works in classic small and midsized situations. The
models behavior in a larger setting is an open
question. It can be assumed that scaling effect by
distributing a single service functionality to multiple
service providers will gain more importance to ensure
service quality. As a third aspect, there is the
inclusion not only of one-time costs, but also of
recurring costs like monthly service fees for example.
This would lead to a more long-term coverage of the
selection process.

[10] Buyya, R., C.S. Yeo, S. Venugopal, J. Broberg, and I.

Brandic, "Cloud computing and emerging IT platforms:
Vision, hype, and reality for delivering computing as the
5th utility", Future Generation Computer Systems, 25(6),
2009, pp. 599616.
[11] Devlin, J.F., "An analysis of choice criteria in the
home loans market", International Journal of Bank
Marketing, 20(5), 2002, pp. 212226.
[12] Doukidis, G. and K. Pramatari, "Supply Chains of the
Future and Emerging Consumer-based Electronic
Services", LNCS, 3746, 2005, pp. 571581.

7. References

[13] Duy, T.V.T., Y. Sato, and Y. Inoguchi, "Performance

evaluation of a Green Scheduling Algorithm for energy
savings in Cloud computing", in Distributed Processing,
Workshops and Phd Forum (IPDPSW), Atlanta, GA, USA.

[1] Abels, S., N. Brehm, A. Hahn, and J. Marx-Gmez,

"Change management issues in Federated ERP systems: an
approach for identifying requirements and possible
solutions", International Journal of Information Systems
and Change Management, 1(3), 2006, pp. 318335.

[14] Foster, I. and C. Kesselman, The grid: Blueprint for a

new computing infrastructure, 2nd edn., Morgan
Kaufmann, Amsterdam ;, Boston, 2004.

[2]
http://www.accenture.com/SiteCollectionDocuments/PDF/
Accenture_Sustainability_TheEnvironmentalBenefitsofMo
vingtotheCloud.pdf, accessed 6-10-2013.

[15] Frobse, K., Steigert das Product-Service System

Cloud-Computing die Oekoeffizienz?, Leuphana
University of Lueneburg, November 2012.

[3] Baliga, J., R.W.A. Ayre, K. Hinton, and R.S. Tucker,

"Green Cloud Computing: Balancing Energy in Processing,
Storage, and Transport", Proceedings of the IEEE, 99(1),
2011, pp. 149167.

[16] Garg, S.K., C.S. Yeo, and R. Buyya, "Green Cloud

Framework for Improving Carbon Efficiency of Clouds", in
Euro-Par 2011 Parallel Processing, D. Hutchison, T.
Kanade, J. Kittler, J.M. Kleinberg, F. Mattern, J.C.
Mitchell, M. Naor, O. Nierstrasz, C. Pandu Rangan, B.
Steffen, M. Sudan, D. Terzopoulos, D. Tygar, M.Y. Vardi,
G. Weikum, E. Jeannot, R. Namyst, and J. Roman, Editors.
2011. Springer Berlin Heidelberg: Berlin, Heidelberg.

[4] Berl, A., E. Gelenbe, M. Di Girolamo, G. Giuliani, H.

de Meer, M.Q. Dang, and K. Pentikousis, "Energy-Efficient
Cloud Computing", The Computer Journal, 53(7), 2010,
pp. 10451051.
[5] Brehm, N. and J. Marx-Gmez, "The Web ServiceBased Combination of Data and Logic Integration in
Federated ERP Systems", in: Proceedings of the 2007
IRMA2007, Vancouver, British Columbia, Canada, May
19-13, 2007. 2007. IGI Publ: Hershey, Pa.

[17] Goscinski, A. and M. Brock, "Toward dynamic and

attribute based publication, discovery and selection for
cloud computing", Future Generation Computer Systems,
26(7), 2010, pp. 947970.

[6] Briscoe, G. and A. Marinos, "Digital ecosystems in the

clouds: towards community cloud computing", DEST
2009, 2009.

[18] Grace, D. and A. OCass, "Child care services: An

exploratory study of choice, switching and search
behaviour", European J. of Marketing, 37(1/2), 2003,
pp. 107132.

[7] Buhl, H.U. and M. Jetter, "Die Verantwortung der

Wirtschaftsinformatik fr unseren Planeten",
WIRTSCHAFTSINFORMATIK, 51(4), 2009, pp. 317
321.

[19] Gregor, S., "Design Theory in Information Systems",

Australasian Journal of Information Systems, 10(1), 2007.
[20] Gronau, N., Enterprise resource planning und supply
chain management: Architektur und Funktionen,
Oldenbourg, Mnchen [u.a.], 2004.

[8] Buyya, R., A. Beloglazov, and J.H. Abawajy, "EnergyEfficient Management of Data Center Resources for Cloud
Computing: A Vision, Architectural Elements, and Open
Challenges", CoRR, abs/1006.0308, 2010.

[21] Hevner, A.R., S.T. March, J. Park, and S. Ram,

"Design Science in Information Systems Research", MIS
Quarterly, 28(1), 2004, pp. 75105.

[9] Buyya, R., R. Ranjan, and R.N. Calheiros, "InterCloud:

Utility-Oriented Federation of Cloud Computing

4208

[22] Hill, C.J., "Differences in the Consumer Decision

Process for Professional vs. Generic Services", Journal of
Services Marketing, 2(1), 1988, pp. 1723.

[37] Nunez, M., B. Miller, L. Liu, H. Wang, X. Liu, X. Jin,

W.B. He, Q.B. Wang, and Y. Chen, "GreenCloud: A New
Architecture for Green Data Center", in Proceedings of the
6th ICAC-INDST, Barcelona, Spain. 2009.

[23] http://www.itbusiness.de/marktforschung/articles/250098/, accessed 511-2013.

[38] Ortwerth, K. and F. Teuteberg, "Green IT/IS

Forschung Ein systematischer Literaturreview und
Elemente einer Forschungsagenda", in: Tagungsband der
MKWI 2012, D.C. Mattfeld and S. Robra-Bissantz,
Editors, MKWI 2012, Braunschweig, 29.02.-02.03.2012.
2012. Gito: Berlin.

[24] Jensen, M., J. Schwenk, N. Gruschka, and L. Lo

Iacono, "On Technical Security Issues in Cloud
Computing", in IEEE International Conference on Cloud
Computing, Bangalore, India. 2009.

[39] Peffers, K.E., T. Tuunanen, M.A. Rothenberger, and S.

Chatterjee, "A Design Science Research Methodology for
Information Systems Research", Journal of Management
Information Systems, 24(3), 2008, pp. 4577.

[25] Kalange, P.R. and Mr. Prashant Vyankatrao Raut,

"Green Cloud Computing: Energy Efficiency and
Environmental Sustainability", Int. J. of Science,
Engineering and Technology Research, 1(6), 2012.

[40] Rehman, Z.u., F.K. Hussain, and O.K. Hussain,

"Towards Multi-criteria Cloud Service Selection", in Fifth
International Conference on Innovative Mobile and Internet
Services in Ubiquitous Computing, IMIS, Seoul, Korea
(South). 2011.

[26] Kksalan, M., J. Wallenius, and S. Zionts, Multiple

criteria decision making: From early history to the 21st
century, World Scientific, Singapore, 2011.
[27] Kuechler, B. and V. Vaishnavi, "On theory
development in design science research: anatomy of a
research project", European Journal of Information
Systems, 17(5), 2008, pp. 489504.

[41] Rehman, Z.u., O.K. Hussain, and F.K. Hussain, "Iaas

Cloud Selection using MCDM Methods", in 2012 IEEE 9th
International Conference on e-Business Engineering
(ICEBE), Hangzhou, China.

[28] Large, R., Strategisches Beschaffungsmanagement,

Gabler, Wiesbaden, 2006.
[29] Lazzarini, S.G., F.R. Chaddad, and M.L. Cook,
"Integrating supply chain and network analyses: The study
of netchains", Journal on Chain and Network Science, 1(1),
2001, pp. 722.

[42] Reiff-Marganiec, S., H.Q. Yu, and M. Tilly, "Service

Selection Based on Non-functional Properties", in ServiceOriented Computing - ICSOC 2007 Workshops, E. Nitto
and M. Ripeanu, Editors. 2009. Springer Berlin Heidelberg:
Berlin, Heidelberg.

[30] Liang, D.-H., D.-S. Liang, and C.-P. Chang, "Cloud

Computing and Green Management", in 2012 Second
International Conference on Intelligent System Design and
Engineering Application (ISDEA), Sanya, China.

[43]
http://arstechnica.com/business/news/2007/04/dreaming-inthe-cloud-with-the-xios-web-operating-system.ars,
accessed 8-30-2013.

[31] March, S.T. and G.F. Smith, "Design and natural

science research on information technology", Decision
Support Systems, 15(4), 1995, pp. 251266.

[44] Ren, C., D. Wang, B. Urgaonkar, and A.

Sivasubramaniam, "Carbon-Aware Energy Capacity
Planning for Datacenters", in 2012 IEEE 20th International
Symposium on Modelling, Analysis & Simulation of
Computer and Telecommunication Systems (MASCOTS),
Washington, DC, USA.

[32] Marston, S., Z. Li, S. Bandyopadhyay, J. Zhang, and

A. Ghalsasi, "Cloud computing The business
perspective", Decision Support Systems, 51(1), 2011,
pp. 176189.

[45] Schrdl, H., "Value enhancement of strategic supply

networks for value bundles through digital social
networks", in Proceedings of the 10th ICEB, Shanghai,
China, Dec 1-4, 2010. 2010: Shanghai, China.

[33] Mell, P. and T. Grance, The NIST Definition of Cloud

Computing: Recommendations of the National Institute of
Standards and Technology, NIST Special Publication(800145), September 2011.

[46] Watson, R., "Green IS: Building Sustainable Business

Practices", Information Systems: A Global Text, 2008,
pp. 117.

[34] Misra, V., P. Barford, M. Squillante, K. Le, O. Bilgir,

R. Bianchini, M. Martonosi, and T.D. Nguyen, "Managing
the cost, energy consumption, and carbon footprint of
internet services", in the ACM SIGMETRICS international
conference, New York, New York, USA.

[47] Weinhardt, C., A. Anandasivam, B. Blau, N. Borissov,

T. Meinl, W. Michalk, and J. Ster, "Cloud-Computing Eine Abgrenzung, Geschftsmodelle und
Forschungsgebiete", WIRTSCHAFTSINFORMATIK,
51(5), 2009, pp. 453462.

[35]
http://www.technologyreview.com/infotech/19397/?a=f,
accessed 9-2-2013.

[48] Younge, A.J., G. von Laszewski, L. Wang, S. LopezAlarcon, and W. Carithers, "Efficient resource management
for Cloud computing environments", in 2010 International
Conference on Green Computing (Green Comp), Chicago,
IL, USA.

[36]
http://www.nytimes.com/2007/10/08/technology/08cloud.h
tml, accessed 8-2-2013.

4209