Cloud Suitability Assessment Method for Application Software

Mesfin Workineh

, Nuno M. Garcia

2,3,4

and Dida Midekso

IT-doctorial program, Addis Ababa University, Addis Ababa, Ethiopia

Universidade da Beira Interior, Faculty of Engineering, Computer Science Department, Covilhã, Portugal

Instituto de Telecomunicações, ALLab-Assisted Living Computing and Telecommunications Laboratory, Covilhã, Portugal

Universidade Lusófona de Humanidades e Tecnologias, Lisbon, Portugal

Department of Computer Science, Addis Ababa University, Addis Ababa, Ethiopia

Keywords: Cloud Computing, Migration, Application Software, Suitability, Decision-making, Method.

Abstract: The advantages and initial adoption success stories of the Cloud computing inspire enterprises to migrate

their existing applications to the Cloud computing technology. As a result, the trend of migrating existing

application software to the Cloud grows steadily. However, not all applications are ideal candidates to be

ported. Moreover, very often client organizations do not have the appropriate methods to determine which

of their IT services are appropriate for migration. In this respect, a method is required to assess the

suitability of the existing applications before embarking on migration. This study designs a method to assess

Cloud suitability of exiting application software following the design science approach. The method is a

multi-step approach composed of seven activities, devised with the goal of reducing the risk of making

wrong migration decisions. Further research will be used to validate and refine the proposed method.

1 INTRODUCTION

The trend for adoption of Cloud computing has been

increasing from time to time but the migration of the

existing systems to Cloud solution is still in its

infancy (Banerjee, 2012; Loebbecke et al., 2012).

However, to benefit from Cloud solution, the

number of organizations migrating existing software

systems to Cloud computing environments have

been growing steadily (Juan-Verdejo et al., 2014;

Binz et al., 2011). Beside the potential advantages,

the initial success stories of Cloud computing

adoption inspire enterprises to migrate their

existing applications to a Cloud-based

architecture (Andrikopoulos et al., 2014, 2013; Juan-

Verdejo, 2012).

Migration of an application on to Cloud also

looks as an attractive investment for enterprises

(Banerjee and Mohapatra, 2013). However, not all

applications are ideal candidates to be ported to a

Cloud platform, or hosted on a Cloud infrastructure

(Chantry, 2009; Böhm et al., 2010; Abduljalil et al.,

2012; Jamshidi et al., 2013). In this respect, the

Cloud suitability of application software must be

assessed before embarking on migration. But, client

organizations lack appropriate method to assess

Cloud suitability of IT services (Loebbecke et al.,

2012; Banerjee, 2012).

Cloud suitability assessment is an initial activity

of the migration method (Khajeh-Hosseini et al.,

2012). The outcome of this assessment phase

determines whether or not to proceed with further

analysis.

Relatively less emphasis has been given in the

literature to Cloud suitability assessment method.

Literature also revealed that most of the Cloud

computing adoption decisions are made in

qualitative manner (Kaisler et al., 2012). Even the

existing quantitative methods aggregate the value for

all criteria as a single value to make decision (Deb,

2010; Beserra et al., 2012; Menzel and Ranjan,

2012; Menzel et al., 2013). But in real cases, the

value of certain criteria must achieve a minimum

benchmarked value and it might not be also

compensated by positive value of other criteria for

the migration to be effective. Hence, a new

systematic method which handles such kind of non-

compensated criteria independently is required.

The method we have proposed considers the

technological, the target Cloud, the risk willingness,

the application nature, and the organization and

business as decision area to assess Cloud suitability

598

Workineh, M., Garcia, N. and Midekso, D.

Cloud Suitability Assessment Method for Application Software.

DOI: 10.5220/0006365206260631

In Proceedings of the 7th International Conference on Cloud Computing and Services Science (CLOSER 2017), pages 598-603

ISBN: 978-989-758-243-1

of the existing application software. This method

guides the decision maker to make an informed

migration decision. As a result, the risk of making a

decision contrary to the organizational objectives

can be reduced.

The remainder of this paper is organized as

follows. Section 2 gives a brief description of the

related work. Section 3 presents the approach used

to develop the method. Section 4 introduces the

activities of Cloud suitability assessment method.

Section 5 gives conclusions and future research.

2 RELATED WORK

Most of the migration approaches proposed in the

literature uses different approach to check the

alignment of legacy applications with Cloud

(Beserra et al., 2012; Khajeh-Hosseini et al., 2012;

Menzel and Ranjan, 2012; Menzel et al., 2013;

Andrikopoulos et al., 2014). But there are very few

exceptional works which tried to assess Cloud

suitability of enterprise application software (Deb,

2010; Kishore et al., 2011; Misra and Mondal, 2011;

Juan-Verdejo and Baars, 2013; Frey et al., 2013).

Deb (2010) proposed an approach to determine

suitability of enterprise applications for the Cloud

based on the Analytic Hierarchy Process (AHP)

approach. The method evaluates Cloud suitability of

applications in three dimensions: business,

technology and risk appetite of an enterprise.

Kishore et al. (2011) assessed Cloud suitability

of a particular web service using a Turing machine

approach and classified the web service as suitable

or unsuitable to be deployed over Cloud. The

authors consider properties of web services and

Cloud services as evaluation criteria.

Misra and Modal (2011) identify a company’s

suitability for migrating to the Cloud environment

and model Return on Investment from using Cloud

computing. Companies’ business key characteristics

as well as pre-existing IT resources were used to

identify suitability of companies for Cloud. They

used mathematical modeling approach to compute a

suitability index based on credit they assigned to

different factors. The model set an upper and a lower

cut-off point to assess organization as suitable, may

or may not be suitable or unsuitable.

Frey and Hasselbring(2011) proposed CloudMIG

approach to migrate software system to IaaS or PaaS

Cloud environments. The approach classifies

existing software systems regarding their Cloud

suitability into five classes as: Cloud incompatible,

Cloud compatible, Cloud ready, Cloud aligned, and

Cloud optimized.

Juan-Verdejo and Baars (2013) proposed a

framework to assess suitability of software

components migrating to a hybrid Cloud

deployment model. The framework was modeling

the interdependencies between the software

components taking into account many parameters.

The method we propose assesses the Cloud

suitability of targeted application based on multiple

criteria decision making approach and assign to

preference-ordered predefined classes. It considers

Cloud suitability assessment as classification or

sorting problem. Unlike other methods proposed in

the literature, our method doesn’t aggregate the

whole criteria as a single value to compute

suitability index. Rather, it classifies criteria as the

one that can be compensated or not compensated by

merit of other criteria to conduct sorting into two

stages. But none of the proposed methods in

literature performs like that.

3 RESEARCH METHODOLOGY

As the output of this research is an artifact that is a

method, design science approach is used to design

this method. Design science research is well suited

for a research that needs to create practical IT

artifacts such as construct, model, method, or an

instantiation (Hevner et al., 2004). Design science

research project must contain three clearly

identifiable and closely related cycles (relevance,

design and rigor) of activities (Hevner, 2007).

Hence, our research method was structured based on

these three-cycles of activities as shown in figure 1.

Figure1: Methodology adopted from (Hevner et al., 2004).

To ensure the rigorous of the research, proposed

method was grounding on theory and involves

expert evaluation. In the design cycle activity the

method is going to built based on multi-criteria

decision making approach (MCDMA) and evaluated

by the domain expert iteratively. In the following

subsection, MADMA steps used to designing the

cloud suitability assessment method are detailed.

Cloud Suitability Assessment Method for Application Software

599

4 CLOUD SUITABILITY

ASSESSMENT METHOD

Multi-criteria decision making approach is used to

develop Cloud suitability assessment method. The

method we propose involves a number of activities

as shown in the Figure 2. These activities are

generic, later to be tailored to a specific organization

and to specific multi-criteria decision making

techniques and tools.

Figure 2: Activities for Cloud suitability assessment.

4.1 Determine Hierarchical Structure

To assess the Cloud suitability of application

software, different decision areas or components

must be considered. Lantana Consulting Group

states that suitability is difficult to define and

measure precisely, and therefore it is easiest to look

at suitability in terms of components (Lantana

Consulting Group, 2011).

In this step, all criteria and sub criteria in each

component must be identified iteratively to construct

hierarchical structure. The literature review and

stakeholders interviews are used to identify these

components and criteria in the hierarchical structure.

There have been a few attempts to model Cloud

suitability of enterprise application software, but

none of them incorporate all the necessary

components. Some of them consider the suitability

of applications for Cloud only from the technical

point of view (Frey and Hasselbring, 2011; Menzel

and Ranjan, 2012). The other studies consider it

from economical point of view (Khajeh-Hosseini et

al., 2012; Misra and Mondal, 2011). But in the

Cloud environment, the decisions taken at business

level will raise constraints to the technology and

vice-versa (Orue-Echevarria et al., 2012). This study

incorporate all together and identified five different

decisions areas from literature such as: technology,

business and organization, nature of application, risk

willingness and targeted Cloud (Banerjee, 2012;

Beserra et al., 2012; Orue-Echevarria et al., 2012).

4.1.1 Technology

A technology component refers to technology up on

which an application relies to give services. As the

Cloud computing technology gives its service based

on the Internet, the decision of migration of an

application to Cloud must consider the availability

of network infrastructure and the network

bandwidth. Network bandwidth is a critical factor to

be considered because higher bandwidth usually

means higher costs (Banerjee, 2012) and low

bandwidth may seriously hamper the availability of

the application in view of candidate workloads.

4.1.2 Risk

Risk refers to chance of dangers that are associated

with “living in the Cloud”. Migrating application is

expected to respect constraints imposed by the

Cloud provider and to provide expected quality of

service. These constraints may affect the enterprise

policies related to privacy and/or security, for

instance, to share virtual machine with other

customers of the provider. This condition may result

in security or privacy breaches. Therefore Cloud

suitability decision must consider the risk appetite of

the organization (Deb, 2010; Beserra et al., 2012).

4.1.3 Nature of Application Software

Nature of application software assesses how the

migrating application software characteristics fit

with the Cloud computing environment. To fit to

Cloud environment an application may need to be

adapted, therefore the migration complexity and cost

depends on the way that application was previously

designed. For instance an application implemented

as service oriented architecture can be migrated to

the Cloud platform in an easier manner when

compared to a composite application that is

implemented in a multi-tier architecture. Therefore,

different factors related to nature of application must

be taken into consideration to determine how well an

application is suited for Cloud.

4.1.4 Business and Organization

The criteria on Business and Organization refer to

the economical aspect of the migration and traits of

organization owning the application. The migration

of application software is not simply lifting and

putting the application to different platforms. It

requires assessing the application with respect to the

organization’s portfolio to determine how well it is

suited for Cloud environments. For instance, legal

CLOSER 2017 - 7th International Conference on Cloud Computing and Services Science

600

and regulatory constraints like enterprise-specific

policies, industry-specific laws and regulations, and

national privacy legislation that have to be respected

after the application and enterprise data have been

migrated (Juan-Verdejo, 2012). Beside socio

technical suitability, the migration of an application

must be economically feasible for the organization

to be able to reap benefits from the migration.

4.1.5 Targeted Cloud Environment

The targeted Cloud environment refers to different

characteristics of the targeted Cloud and constraints

that are imposed by Cloud providers. It is important

to consider characteristics of the Cloud as it is the

destination for the application to be migrated. The

characteristics of the Cloud like scalability,

availability and reliability (Kishore et al., 2011)

affects the suitability for an application running on

the Cloud. A migrated application is also expected

to satisfy the constraints imposed by the Cloud

provider, for instance, the access to the file system,

the number of files, or number of calls to specific

methods (Frey et al., 2013). Likewise, the Cloud

environment has to comply with security, privacy,

performance, availability and regulatory

requirements of the targeted application (Banerjee

and Mohapatra, 2013).

4.2 Define Suitability Classes and Profile

Garg et al.(2011) defined suitability of a Cloud

provider for customer requirements and quantified

as values between 0 and 1(as ratio scale). Kishore et

al. (2011) measure suitability using nominal

scale(yes/no). Other scholars measure suitability

using ordinal scale (Frey and Hasselbring, 2011;

Misra and Mondal, 2011).

Similar to Frey and Hasselbring (2011) our

Cloud suitability assessment method uses five

ordinal scales (extremely suitable, very suitable,

suitable, slightly suitable, and unsuitable) to measure

or describe the Cloud suitability of an application

software. These measurement scales are considered

as classes to sort applications software based on

their suitability for the Cloud migration. Sorting of

the application requires to compare aggregated value

of an application with some reference profiles that

distinguishes the classes.

Reference profile r

of a class k is defined as a

vector of local profiles (r

=(r

, r

, … , r

)) for

each criterion (C

; C

; . . . ; C

) where r

is the local

profile class k for C

criterion. The local profile

refers to the minimum performance on each positive

criterion or maximum performance on each negative

criterion that application software satisfies to be

belongs to a class. The local profile for each class is

determined by the organization’s decision analyst

based on the business needs of that organization.

Sometimes it is difficult to define the reference

profile of a class. In that case, it is required to define

the central profiles of a class (Ishizaka et al., 2013)

as an average value an application software system

must satisfy on each criterion to belong to a class. If

there is equal distance between two central profiles,

then the reference profile of a class is determined as

an average value of the central profile of a the class

and preceding class (r

=(r

+ r

k+1

)/2).

4.3 Categorize Criteria

Most of existing Cloud computing adoption decision

approaches are based on a qualitative approaches

(Kaisler et al., 2012). Even the existing quantitative

approaches evaluate the decision in such a manner

that the risk or cost of one criterion is compensated

by the merit of other criteria, but this is not always

true. There is a case where the deficiency in one

criterion cannot be compensated by merit of one or

more other criteria, for example, an application may

be evaluated to have high suitability scores in the

application nature, and business value, but it may

not be a good candidate for migration if the risk

exposure is higher than the level of risk an enterprise

is willing to accept.

Taking this into consideration the criteria used to

assess Cloud suitability of application software

system is categorized as screening criteria and

evaluation criteria. Screening criteria are a set of

criteria whose limitation should not be offset by the

strength of the other criteria. In such a case for an

application to be suitable for migration the minimum

limit of profit criteria must satisfied or must not

exceed the maximum limit of cost criteria.

Evaluation criteria are a set of attributes used to

assess Cloud suitability of a software system whose

value may be offset by the strength of the other

criteria. An application can assume any value in

these criteria from their domain.

4.4 Determine Relative Priority and Scale

Each dimensions and respective criteria have

decisive effect whether to move or not move legacy

software system to Cloud, but their relative

importance or weight is different (Menzel et al.,

2013). Therefore, such weight has to be set based on

Cloud Suitability Assessment Method for Application Software

601

the business needs of an organization and type of

migrating application.

There are different mechanisms to set such a

weight for criteria from decision maker preference

score. AHP is one of the mostly used methods to

determine relative weights of each attributes from

pair wise comparisons of each attributes in multi

criteria decision making approach.

Some criteria may not have standards of

measurements for instance security. For such kind of

criteria it required to set measurement scale and

possible value (domain) for criteria.

4.5 Select Sorting Model

Cloud suitability assessment method evaluates

application software using a set of criteria and

categorizes them into different classes. If these

classes are ordered such kind of a problem is said to

be sorting problem, otherwise, it is said to be

classification problem (Zopounidis and Doumpos,

2002). Sorting or classification model can be

developed by different disciplines such as: statistics,

artificial intelligence, and operation research. This

study considers operation research approach to

develop the method. The MCDMA has the

following advantages (Zopounidis and Doumpos,

2002): allows to incorporate decision maker’s

opinion; and needs not be data intensive to generate

a classification or sorting model.

4.6 Compute Suitability Index

The method we have proposed first assign a value

for application against each screening criteria and

then multiply with respective global weight to

classify the application using none compensatory

multi criteria classification method as that can be

adapted for migration, stay on primes, or

redeveloped for migration. Then if the application

can be adapted for migration, go to second phase of

analysis to show extent of its suitability.

In the second phase a weighted value of

evaluation criteria aggregated to give suitability

index using selected compensatory MCDMA. The

suitability index could then be used to assign

application into different suitability classes.

4.7 Assign to Specific Class

A degree of alignment during the reengineering

process may be different for applications in each

class and this degree of alignment is considered as

the weight for the class, the weight can be generated

from the preference score of the decision maker.

The local reference profile, which is set by the

decision analyst, is multiplied by the weight of

respective class to determine the global reference

profile or global central profile of a class. Then the

global reference profiles are aggregated as a single

value using similar approach used to aggregate the

suitability index. Finally the suitability index of an

application compared against this aggregated

reference profiles of a class to assign a candidate

application software to a specific class.

5 CONCLUSION AND FUTURE

WORK

This research proposed a method to assess Cloud

suitability of application software in five different

decision areas. The three closely related cycles of

activity of design science was identified and used to

structure this research. In design cycle, based on a

MCDMA seven steps proposed to assess Cloud

suitability. The proposed method guides the decision

maker to make an informed migration decision.

Unlike other existing method this method

proposed two stage evaluation approaches to assess

Cloud suitability of legacy application based on two

different groups of criteria. It is unique approach in

considering suitability assessment as multi-criteria

sorting problem.

This research is a work-in-progress to assess the

Cloud suitability index of an application. In future

research, list of criteria in each decision area and

their respective measurement scale will be identified

and validated using expert evaluation. Finally, the

method will be validated empirically taking specific

multi-criteria decision making approach.

ACKNOWLEDGMENTS

The authors acknowledge the contribution of COST

Action IC1303 AAPELE Algorithms, Architectures

and Platforms for Enhanced Living Environments.

Authors affiliated with the Instituto de

Telecomunicações also acknowledge the funding for

the research by means of the program FCT project

UID/EEA/50008/2013. (Este trabalho foi suportado

pelo projecto FCT UID/EEA/50008/2013).

CLOSER 2017 - 7th International Conference on Cloud Computing and Services Science

602

REFERENCES

Abduljalil, S.M., Hegazy, O., Ammar, H.H., 2012.

Evaluation of Architecture Styles for Cloud

Computing Applications. Int. J. Rev. Compt. 11, 1–11.

Andrikopoulos, V., Darsow, A., Karastoyanova, D.,

Leymann, F., 2014. CloudDSF – The Cloud Decision

Support Framework for Application Migration, in:

International Federation for Information Processing.

Presented at the ESOCC 2014, pp. 1–16.

Andrikopoulos, V., Strauch, S., Leymann, F., 2013.

Decision Support for Application Migration to the

Cloud: Challenges and Vision, in: CLOSER 2013.

Presented at the 3rd International Conference on

Cloud Computing and Services Science, pp. 149–155.

Banerjee, J., 2012. Moving to the Cloud: Workload

Migration Techniques and Approaches, in: In High

Performance Computing (HiPC). Presented at the 19

th International conference on, IEEE, pp. 1–6.

Banerjee, J., Mohapatra, B., 2013. An Approach to Cost-

Effective Transformation of Workloads towards Cloud

Delivery Models. Int. J. Comput. Appl. 82, 47–53.

Beserra, P.V., Camara, A., Ximenes, R., Albuquerque,

A.B., Mendonça, N.C., 2012. Cloudstep: A step-by-

step decision process to support legacy application

migration to the cloud, in: Maintenance and Evolution

of Service-Oriented and Cloud-Based Systems.

Presented at the 2012 IEEE 6th International

Workshop, IEEE, pp. 7–16.

Binz, T., Leymann, F., Schumm, D., 2011. CMotion: A

Framework for Migration of Applications into and

between Clouds, in: 2011 IEEE International

Conference on Service-Oriented Computing and

Applications (SOCA).IEEE Computer Society, pp.1–4.

Böhm, M., Leimeister, S., Riedl, C., Krcmar, H., 2010.

Cloud computing and computing evolution.

Chantry, D., 2009. Mapping Applications to the Cloud.

TechEd Spec. Ed. 19, 1–9.

Deb, B., 2010. Assess enterprise applications for cloud

migration Using the Analytic Hierarchy Process to

evaluate apps for the cloud.

Frey, S., Hasselbring, W., 2011. The CloudMIG

Approach: Model-Based Migration of Software

Systems to Cloud-Optimized Applications. Int. J. Adv.

Softw. 4, 342–353.

Frey, S., Hasselbring, W., Schnoor, B., 2013. Automatic

conformance checking for migrating software systems

to cloud infrastructures and platforms. J. Softw.

Maint. Evol. Res. Pract. 25, 1089–1115.

Garg, S.K., Versteeg, S., Buyya, R., 2011. SMICloud: A

Framework for Comparing and Ranking Cloud

Services, in: Fourth IEEE International Conference.

Presented at the Utility and Cloud Computing, IEEE

Computer Society, pp. 201–218.

Hevner, A.R., 2007. A Three Cycle View of Design

Science Research. Scand. J. Inf. Syst. 19, 87–92.

Hevner, A.R., March, S.T., Park, J., Ram, S., 2004.

Design Science In Information Systems Research. MIS

Q. 28, 75–105.

Ishizaka, A., Pearman, C., Nemery, P., 2013. AHPSort: an

AHP based method for sorting problems. Int. J. Prod.

Res. doi:DOI: 10.1080/00207543.2012.657966.

Jamshidi, P., Ahmad, A., Pahl, C., 2013.

Cloud Migration

Research: A Systematic Review. IEEE Trans. Cloud

Comput. 1, 142–157.

Juan-Verdejo, A., 2012. Assisted migration of enterprise

applications to the Cloud–A hybrid Cloud approach,

in: Tagungsband Des 4. Workshops. Presented at the

Business Intelligence, pp. 14–27.

Juan-Verdejo, A., Baars, H., 2013. Decision support for

partially moving applications to the cloud: the

example of business intelligence, in: Proceedings of

the 2013 International Workshop on Hot Topics in

Cloud Services. ACM, pp. 35–42.

Juan-Verdejo, A., Zschaler, S., Surajbali, B., Baars, H.,

Kemper, H.-G., 2014. InCLOUDer: A formalised

decision support modelling approach to migrate

applications to cloud environments, in: 2014 40th

EUROMICRO Conference Software Engineering and

Advanced Applications. IEEE, pp. 467–474.

Kaisler, S., Money, W.H., Cohen, S.J., 2012. A Decision

Framework for Cloud Computing. Presented at the

45th Hawaii International Conference on System

Sciences.

Khajeh-Hosseini, A., Greenwood, D., W.Smith, J.,

Sommerville, I., 2012. The cloud adoption toolkit:

supporting cloud adoption decisions in the enterprise.

Softw. Pract. Exp. 42, 447–465.

Kishore, M.S.N., Jayakumar, S.K.V., Reddy, G.S.,

Dhavachelvan, P., Chandramohan, D., Reddy, N.P.S.,

2011. Web Service Suitability Assessment for Cloud

Computing. Presented at the LNCS-Springer,

Springer, Verlag Berlin Heidelberg, pp. 622–632.

Lantana Consulting Group, 2011. Quality Reporting Data

Architecture Suitability Analysis.

Loebbecke, C., Thomas, B., Ullrich, T., 2012. Assessing

Cloud Readiness at Continental AG. MIS Quarterly

Executive 11, 11–23.

Menzel, M., Ranjan, R., 2012. CloudGenius: Decision

Support for Web Server Cloud Migration, in: 21

International Conference on World Wide Web. ACM,

pp. 979–988.

Menzel, M., Schönherr, M., Tai, S., 2013. (MC2)2:

criteria, requirements and a software prototype for

Cloud infrastructure decisions. Softw. Pract. Exp. 43,

1283–1297. doi:10.1002/spe.1110.

Misra, S.C., Mondal, A., 2011. Identification of a

company’s suitability for the adoption of cloud

computing and modelling its corresponding Return on

Investment,. Math. Comput. Model. 53, 504–521.

Orue-Echevarria, L., Alonso, J., Escalante, M., Schuster,

S., 2012. Assessing the Readiness to Move into the

Cloud, in: International Conference on Cloud

Computing. Springer International publishing, pp. 12–

20.

Zopounidis, C., Doumpos, M., 2002. Multicriteria

classiﬁcation and sorting methods: A literature

review,. Eur. J. Oper. Res. 138, 229–246.

Cloud Suitability Assessment Method for Application Software

603