You are on page 1of 21

Computer Standards & Interfaces 62 (2019) 98–118

Contents lists available at ScienceDirect

Computer Standards & Interfaces


journal homepage: www.elsevier.com/locate/csi

Toward a unified framework for Cloud Computing governance: An approach T


for evaluating and integrating IT management and governance models
Yassine Bounaguia, , Abdellatif Mezriouia, Hatim Hafiddia,b

a
STRS Lab, INPT Rabat, Morocco
b
SIME Lab, ENSIAS Rabat, Morocco

ARTICLE INFO ABSTRACT

Keywords: Cloud Computing is currently one of the major trends in the computer industry. It offers a wide range of both
Cloud Computing opportunities and challenges. The lack of governance causes a slowdown in the adoption process, especially with
Cloud Computing governance an absence of a comprehensive approach supporting the Cloud Computing governance objectives. Regarding this
Unified framework issue, our goal is to develop a new management and governance framework based on endorsed IT models;
IT models evaluation
namely ITIL, COBIT, and ISO/IEC 27001/2. These models have been developed for different purposes and as-
IT models integration
pects of IT governance. The proposed framework is an attempt to a unified approach by taking into consideration
the models’ dissimilarities. Before we go further with the final unification stage, this study's main objective is to
develop both the evaluation and the integration approaches with regwards to the IT models. The results,
therefore, are to shed light on the relevant outcomes emanate from an elaboration within an entirely unified CC
governance framework.

1. Introduction to-end security solutions; and complying with the regulations and
standards of good practices. According to the Cloud Security Alliance
Cloud Computing (CC) is a model to deliver IT resources via the [6], the main three CC security risks are unsecure interfaces and APIs,
network as an on-demand service. CC can allow organizations to out- data loss or leakage, and hardware failures. These risks respectively
source their IT infrastructures and services to a pay-as-you-go model account for 29%, 25%, and 10% of all the security risks in CC. The
based on the duration of use, the required service quality, etc. CC is service level agreements management realm also faces limitations and
currently the latest revolution in the IT industry. It promises huge op- challenges such as [7]: Difficulty to measure Cloud service providers
portunities for its users; particularly for organizations which could compliance; absence of mechanisms for automatic SLA/Service nego-
generate financial and operational interests. In [1], the National In- tiation and renegotiation; lack of procedures for supporting multiple
stitute of Standards and Technology (NIST) defines CC as “a model for Cloud service providers; and absence of SLA translation when moving
enabling ubiquitous, convenient, on-demand network access to a shared from one Cloud service provider to another. The Cloud migration pro-
pool of configurable computing resources that can be rapidly provi- cess would also need several changes in the IT architecture. For in-
sioned and released with minimal management effort or service pro- stance, migrating complex applications into Software as a Service
vider interaction”. NIST also defines five essential characteristics of CC model may require detailed planning and testing prior to their im-
(i.e. On-demand self-service, broad network access, resources pooling, plementation. This can go up to fully change the code of some legacy
rapid elasticity, and measured service), three service models (i.e. Soft- applications [8]. As a consequence, new requirements have emerged
ware as a Service, Platform as a Service and Infrastructure as a Service), and IT governance aspects should be reviewed and empowered [9,10].
and four deployment models (i.e. Private Cloud, Community Cloud, Governance is the key discipline to address all these challenges and
Public Cloud, and Hybrid Cloud). maximize the organizations’ return on investment in CC [11]. CC
There are numerous benefits that CC can offer. This can make its governance is defined as a set of policies, processes, roles, responsi-
adoption highly desirable [2,3]. Although it has a lot of advantages, this bilities, and practices used to manage and control CC adoption and
model also encounters challenges and risks that can limit its credibility implementation in accordance with business goals [12]. For these
and pervasiveness such as [4,5]: Ensuring the confidentiality, integrity purposes, an IT governance framework is highly recommended to allow
and availability of the organization's data and services; providing end- organizations to do different tasks such as: Maximizing the CC value;


Corresponding author.
E-mail address: bounagui@inpt.ac.ma (Y. Bounagui).

https://doi.org/10.1016/j.csi.2018.09.001
Received 11 July 2017; Received in revised form 13 September 2018; Accepted 21 September 2018
Available online 24 September 2018
0920-5489/ © 2018 Elsevier B.V. All rights reserved.
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

minimizing its inset challenges; handling all the CC governance aspects; governance could be implemented [9,12].
providing a widely accepted CC management and governance frame- Over the past decade, several IT management and governance
work; understanding the available capabilities; and defining the po- models have been developed. They are designed to provide good
licies that stand for the organizations’ needs. Furthermore, this frame- practices for IT leaders to develop, implement, monitor, and continually
work will provide effective means to check whether CC infrastructures improve IT control and governance [16]. These models can help orga-
and services meet the business expectations. The elaboration of such a nizations to solve the CC governance issue. According to [4,22], com-
framework requires meeting the following key objectives: monly used IT models are ITIL [23], COBIT [24], and ISO/IEC 27001/2
[25,26]. Although these models are not basically designed to respond to
• Ensure the effective management of CC security risks: The CC CC new risks and challenges, the use of a particular model or their
distributed nature and the removal of the security boundary separ- combination as a unified framework for CC has been recommended by
ating IT traditional data centers from external environments have several studies [9,27]. However, evaluating the appropriateness of
introduced several security risks [13]. Risk management is essential these models is still required. The evaluation process should provide the
to understand organizations exposure to CC risks as well as to de- ability to quantitatively measure the adequacy of the models’ process
velop suitable strategies for analyzing, assessing, mitigating, mon- elements; and thus, determine their corresponding coverage rate with
itoring, and reviewing the risks [14,15]. CC Risk management also regard to CC governance requirements. We would like to mention that a
helps organizations understand Cloud service providers responsi- process element is a fundamental and atomic unit for process definition;
bilities for ensuring continuous provision of CC services, review for example, activities, tasks, controls, clauses, etc. The term process
their IT continuity plans, and provide alternative processing and element is used in this study to provide a standardized nomenclature for
recovery capabilities [13]. different models.
• Align and communicate objectives: The business and IT objec- Based on the evaluation results, the unification of these models into
tives of CC adoption should be effectively aligned to achieve the a single comprehensive CC governance framework is required. The
desired outcomes [13]. They must be communicated both internally unification should focus on integrating the models as well as homo-
and with Cloud service providers. This should be a part of an on- genizing them by resolving their terminological and structural conflicts.
going aligning and communication program [16]. Thereby, the focus will be on how the models would be integrated to
• Integrate CC governance with existing IT governance ap- better cover the new CC governance requirements. The models’
proaches: As CC governance is an extension of IT governance, en- homogenization will be the subject of our future research.
suring the approaches concordance is a requirement [17]. Organi- Based on the following points, our main motivation is to quantita-
zations can therefore benefit from effective coexistence and tively demonstrate that the unification of ITIL, COBIT, and ISO/IEC
interaction between CC governance and existing IT governance ap- 27001/2 into a comprehensive framework is an effective means to
proaches. cover the insufficiencies in these models when they are applied to CC
• Ensure the conversion of IT rules and decisions into policies: governance.
The implementation of effective CC governance depends on how IT
rules and decisions are converted into policies. The conversion • The adoption of cloud computing presents challenges within IT
process should be a part of a policy management program that al- governance from which new CC governance requirements are de-
lows the elaboration, communication, tracking, and enforcement of rived.
policies with regard to the CC utilization [15]. Additionally, the • A CC governance framework that helps organizations address these
program should check both the correctness of new and updated challenges is required.
policies and the potential inter-policy relations that may affect their • The existing approaches for CC governance have shortcomings.
effectiveness (e.g. sub-sumptions and contradictions resulting from • The commonly used IT management and governance models are
change in policies) [18]. ITIL, COBIT, and ISO/IEC 27001/2.
• Adapt easily: CC governance can allow organizations to easily • Except for COBIT, these models do not take into account the new CC
adapt flexibility, scalability, and services in the Cloud as business governance requirements.
requirements evolve [13]. It can also provide the ability to handle a • These models deal with several aspects relevant to CC governance.
broad and complex ecosystem of services delivered by different • The unification of the three models may be an effective means to
Cloud service providers. Consequently, organizations can overcome overcome their insufficiencies.
IT complexity when changing their provisioning architecture from a
traditional to an on-demand model [19]. This paper attempts to answer the following questions:
• Ensure regulatory and contractual compliance: One of the major
concerns that would be addressed by CC governance is to ensure • What are the existing studies on CC governance?
that the use of CC infrastructures, platforms and services comply • What is the impact of CC on the IT models: ITIL, COBIT, and ISO/IEC
with existing laws and regulations as well as with the agreed con- 27001/2?
tractual requirements. It would particularly ensure the compliance • Why should these three models be applied to CC governance?
with the law and regulations in case the user and the provider are • What are the new CC governance requirements?
from different countries [20]. Regulatory and contractual yielding • How can we assess the suitability of these models with respect to the
should be a part of a compliance management system that allows the identified CC governance requirements?
organization to track and assess the impact of new regulations as • How can these models be integrated to better meet the new CC
well as to report breaches [15]. governance requirements?

Currently, various approaches have been put forward to establish It is important to note that our approach is not specifically designed
methodologies for CC governance. Most of them aim at improving or to support the unification of ITIL, COBIT, and ISO/IEC 27001/2 but
extending traditional IT governance and making it Cloud-aware rather to provide a general and independent approach from the type of
[11,21]. However, there is no framework that can accomplish all CC the models being used. Thereby, the main proposition of this work is
governance objectives. The existing approaches have limitations. For the actual approach that is being put forward and not the application of
instance, either they are limited to one specific governance domain or this approach to particular IT models.
they only provide guidance instead of advancing a holistic approach. In Thus, the paper objectives are: First, to conduct a state-of-the-art
addition, none of the available studies can provide details on how CC review of the main works and research studies on CC governance;

99
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

second, to review and analyze the impact of CC on IT governance (what is moving to the Cloud); the delivery dial (SaaS, PaaS, and IaaS);
models; third, to identify the CC governance requirements and then the deployment dial (Public, Private, and Hybrid); the Cloud formation
evaluate ITIL, COBIT, and ISO/IEC 27001/2 with regard to these re- dial (Internal/External, Proprietary/Open, Parameterized/De-para-
quirements; and finally, to set up the stage for the unification through meterized, and In-sourced/Outsourced); the enterprise risk manage-
integrating the evaluated models. ment dial; and the control dial (the necessary control modifications for
The paper is organized as follows: The second section presents the a specific Cloud solution). The study, however, remains limited as it
state-of-the-art review. The review and unification of IT models are only provides high-level description of the requirements without ex-
presented and discussed in the third section. In the fourth section the actly specifying how these requirements would be met. This limitation
evaluation of IT models is proposed. The latter's integration is detailed has been justified by the authors as it should not be one-size fits-all or
in the fifth section. The final section, then, focuses on the paper's unique proposition for integrating and extending IT models. Conse-
overview and future work. quently, it is the organization role to put into effect the necessary
methods and procedures to accomplish the integration and extension.
2. State-of-the-art review Moreover, relevant aspects of CC governance, such as process lifecycle
management and service level agreement management, are totally
As mentioned above, CC governance is the key discipline that allows missing.
organizations to ensure the effective control over their CC infra- He [17] has proposed a lifecycle process model for CC that helps
structures and services. Currently, several approaches have been de- organizations to govern their Cloud-based IT services. The model pro-
veloped to deal with the CC governance issue. The relevant approaches vides guidelines to manage CC services and assets as well as to align CC
will be discussed and analyzed in this section. governance with business goals. The model serves as an input for Cloud
Karkošková and Feuerlicht [4] have proposed a CC governance service providers to analyze their existing capabilities and enhance
lifecycle that provides guidance on the implementation and the con- them to better cover Cloud service users’ requirements. The author has
tinuous improvement of CC governance activities. The authors have divided CC governance into five main domains: strategic planning, or-
based their model on SOA governance through adapting its methodo- ganizational alignment, service lifecycle management, policy manage-
logical components. They have also extended them to include CC ment, and service level agreements management. Each domain high-
governance. The model has also been based on literature reviews on lights the processes and activities required to ensure effective CC
SOA and IT governance. The proposed Cloud computing governance governance. Nevertheless, the study still has got some limitations in
lifecycle consists in four main phases: Planning, definition, im- terms of the model's processes which are defined with high abstraction
plementation, and monitoring. Each phase includes the required ac- level and do not provide sufficient details; the linkage between the
tivities to implement and continually improve CC governance. To de- different roles and the model processes is absent; moreover, the in-
velop their model, the authors have first adapted and redefined the SOA tegration with other IT governance approaches is missing.
governance vitality method and the SOA governance reference model; Ahmad and Janczewski [20] have proposed a governance lifecycle
second, they have specified a maturity level for CC governance. The framework for public CC. Their work results in an approach to mitigate
proposed model, however, remains limited. It only provides guidance CC security risks and improve CC governance and insurance. The fra-
instead of suggesting an integrated approach for CC governance. Fur- mework has been designed based on legal principles and on the uni-
thermore, CC governance aspects have not all been taken into account; fication of relevant clauses of international standards, such as COBIT or
for example, the service level agreement management or the Cloud ISO/IEC 27001. The unified clauses have been categorized into three
service provider auditing. distinct control types: The technical and operational controls; the
Brandis et al. [21] have proposed a model that serves as a holistic management controls; and the legal controls. The intersection between
framework addressing CC governance. The authors have defined an the unified clauses, the CC layers, and the statutory laws constitutes the
ontology that supports a semantic-based formalization of the model. CC governance framework. The proposed framework still suffers from a
The framework provides an approach that combines enterprise archi- number of drawbacks; for instance, the study has only focused on in-
tecture management, IT governance, and CC governance. The frame- formation security instead of suggesting a holistic approach for CC
work has been developed around the three following paradigms: The governance, the authors have not addressed how IT governance models
business–IT alignment paradigm; the governance paradigm; and the are unified and integrated, and the study only provides high-level re-
cloud paradigm. The authors have distinguished two types of levels quirements rather than providing particularities about these require-
corresponding to the two first paradigms (i.e. horizontal levels and ments.
vertical levels). The horizontal levels provide consistency between Guo and Song [15] have proposed a CC governance model that
business objectives, operations, and IT infrastructures. The vertical le- stresses the organizations needs in terms of policy schemes, service
vels offer compliance and governance and consider data and informa- profiles, services management, and governance processes. The model
tion, governance requirements as well as roles and responsibilities. The also emphasizes the need for service lifecycle management, visibility,
cloud paradigm meanwhile focuses on assessing Cloud services to allow and contextualization. To develop the model, the authors have made a
a more feasible application of the model in architectural settings be- clear distinction between the policy model, the management model,
lieved to be in complex. The proposed framework, however, only ad- and the operational model. This is to allow adjustment to requirements
dresses the semantic consideration and its modeling through ontologies. change and to different kinds of external events. The proposed model,
Besides, the authors did not provide details on the framework archi- though, remains limited; for example, the authors have not mentioned
tecture, processes, and activities. It is important to note that the study how the continuity of critical business processes is performed or how
does not consider or deal with all the CC governance aspects; for in- their model would integrate with existing IT governance models. Fur-
stance, the service level agreement management or the CC regulatory thermore, the study only provides high-level requirements instead of
compliance. coming up with details on how these requirements would be achieved
Bailey and Becker [9] have proposed a framework that defines the by the authors model.
necessary steps to implement CC governance. The authors have based The Cloud Security Alliance (CSA) has published the “Security
their work on existing IT models as COSO, ISO/IEC 27000, COBIT, or Guidance for Critical Areas of Focus in Cloud Computing” [28]. The
ITIL. The framework allows organizations to align CC with business report is intended to supply guidance for organizations in supporting
value and deliver that value via an optimal resource allocation and business objectives while managing and mitigating CC security risks.
performance management. To build the framework, the authors have The CSA has also released the Cloud Controls Matrix (CCM) [29] to
deconstructed CC governance into six concentric dials: The process dial provide Cloud service users with fundamental controls for CC security.

100
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

The CCM is an amalgam of compliance, governance, and technical explain their components, without specifying how these components
controls that have been aligned with various industry-accepted stan- are elaborated. The low level describes the approaches that provide a
dards and control frameworks; such as ISO/IEC 27001/2 or COBIT. The detailed description of their components. We would like also to mention
CCM provides a domain-based approach consisting of 133 controls and that the values assumed by the solution applicability column are either
is developed around 16 security domains. The CSA works are, however, partially applicable or totally applicable. The partially applicable stu-
restricted to cloud security and risks management instead of proposing dies are those that have been published with missing components which
a comprehensive CC governance approach. Additionally, CC govern- would be developed in the future by the authors. The totally applicable
ance aspects have not all been taken into consideration; for instance, studies, on the contrary, refer to the studies that do not present any
the alignment of CC with business goals or the continuous improvement missing components.
of CC security domains. Despite the quality of the studies analyzed in this section, limita-
Saidah and Abdelbaki [12] have proposed a CC governance fra- tions have been identified. However, a framework that meets the needs
mework that intends to enhance the Guo's governance model [15] by of every type of Cloud service user and also accomplishes all CC gov-
adding the security controls of the CCM [29]. To develop the frame- ernance objectives has not yet been developed. This has been agreed
work, the authors have first identified the gaps in the Guo model. upon by several studies; for instance, [9,13,15]. The state-of-the-art
Second, they have identified the controls of the CCM that relate to each review achieved in this section has emphasized the need of a new CC
of the model processes. Finally, they have added, modified and updated governance framework. This framework should provide an integrated
the missing components. The framework has been developed around approach that takes into account all CC governance objectives. It should
the five following stages to ensure the continuous improvement of CC particularly furnish solutions to the identified limitations.
governance processes: Strategic trigger; define and align; build and
implement; deliver and measure; and operate and feedback. Although 3. IT models review and unification
the study has proposed an enhanced version of the Guo's model, it does
present the same limitations we have already identified. Currently, there are several IT management and governance models.
The European Network and Information Security Agency (ENISA) They can help organizations achieve the CC governance objectives. The
has published a report entitled “Cloud Computing Benefits, Risks and role of these models in CC governance is one of the major trends in IT
Recommendations for Information Security” [30]. The report's aim is to [9]. Even though some studies consider CC as a new paradigm that
assist Cloud service users in assessing the security risks and benefits of requires to define a new set of models, others consider it as a tech-
CC. The report proposes a detailed review of the political, organiza- nology based on other existing and available technologies’ models [36].
tional, technical and legal risks. The ENISA has also released the “Cloud Understanding the existing IT governance and management models is,
Computing Information Assurance Framework” [31] to help organiza- then, important to any discussion of CC governance. In this section, the
tions assess the CC security risks, compare different Cloud service following IT models will be discussed and analyzed: ITIL v3, COBIT 5,
provider offers, and obtain assurance from the selected Cloud service and ISO/IEC 27001/2:2013.
providers. In addition, the ENISA has released the report “Security and We would like to point out that the process of selecting these models
Resilience in Governmental Clouds” [32]. The report supplies govern- is first based on their ability to meet the CC governance objectives
mental organizations with a decision-making model to determine which outlined in Section one. Second, their relevance and level of use by the
Cloud architectural solution best accommodates their security re- scientific community and the industry are also taken into account.
quirements [33]. Similarly, The Committee of Sponsoring Organiza- Though these three models may not gather the varied needs of all or-
tions of the Treadway Commission (COSO) has released the “Enterprise ganizations, studies (e.g. [22,37–39]) still agree that these models are
Risk Management for Cloud Computing” [34] based on the COSO En- actually the widely adopted frameworks/standards in IT management
terprise Risk Management framework [35]. The CC framework targets and governance. Accordingly, many organizations endure from the
to both identify the CC security risks and their impact and to provide models insufficiency with regard to CC challenges and requirements.
strategies for monitoring and mitigating those security risks. Never- Third, COBIT and ISO/IEC 27001/2 provide guidance on what should
theless, the works published by ENISA and COSO are limited to CC risk be done by IT processes through defining the overarching goals of
management. Consequently, they should be combined with other IT business “what”. The organization is hence responsible for choosing the
models in order to provide a holistic CC governance approach. way to achieve those goals “how”. ITIL however provides guidance on
The review and analysis of CC governance studies have shown the “how” IT processes would be planned, designed, and implemented.
existence of various approaches with different backgrounds, archi- Thus, the use of these three models can help organizations cover both
tectures, and characteristics. In Table 1 we summarize these studies and the “what” of CC governance and the “how” of CC management
provide other criteria for comparison. We would like to mention that [9,37–39]. Finally, although these three models are not specific to CC,
the values assumed by the abstraction level column are either high or studies (e.g. [13,40,41]) still agree that they are comprehensive enough
low. The high level is used to designate the approaches that simply and deal with many relevant aspects to support CC governance. We

Table 1
Comparative analysis of the existing studies on CC governance.
Authors, References Years Scope of CC governance Abstraction level Process oriented Lifecycle approach Solution applicability

CSA, [29] 2016 Focuses on security governance Low Yes No Totally applicable
Karkošková and Feuerlicht [4] 2016 All CC governance aspects High Yes Yes Totally applicable
Brandis et al. [21] 2014 All CC governance aspects High No No Partially applicable
Bailey and Becker [9] 2014 All CC governance aspects High Yes No Partially applicable
Saidah and Abdelbaki [12] 2014 All CC governance aspects High Yes Yes Partially applicable
COSO [34] 2012 Focuses on risks governance Low No No Totally applicable
ENISA [30–32] 2012 Focuses on risks governance Low No No Totally applicable
2009
2011
He [17] 2011 All CC governance aspects High Yes Yes Totally applicable
Ahmad and Janczewski [20] 2011 Focuses on security governance High Yes Yes Totally applicable
Guo and Song, [15] 2010 All CC governance aspects High Yes Yes Totally applicable

101
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Table 2.
Comparative analysis of the attributes of ITIL, COBIT, and ISO/IEC 27001/2.
Attribute ITIL COBIT ISO/IEC 27001/2

Name IT Infrastructure Library Control Objectives for Information and Information technology – Security techniques – Code of
related Technology practice for information security management
Organization Developer OGC, UK ISACA ISO/IEC
Version (Year) v3 (2007) V5 (2012) Second edition (2013)
Scope IT service management IT governance Information security
Focus How What How
Features Service delivery and support Control objectives Information security management system
Certification Certification of personnel N/A Certification of organization and personnel
Usage Guidelines Methodology Guidelines
Domain of application Service part of IT domains All IT domains All information security domains
Implementation guidance Guidance on processes implementation is Generic and need customization Generic and need customization
given
Architectural components 26 processes distributed across 5 lifecycle 37 processes grouped into 5 domains 14 clauses distributed across 35 categories and 114
stages and 4 functions controls

would like to mention that all the three models have been updated to 3.3. ISO 27001/2
their latest version to ensure that the new needs and requirements of IT
governance are taken into account. The ISO/IEC 27k series of standards have been published by the
International Organization for Standardization (ISO) and the
3.1. ITIL International Electro technical Commission (IEC). These standards
provide security controls and best practices to establish, implement,
ITIL is a widely accepted framework for IT service management and maintain, and continuously improve organizations’ Information
delivery. It has been developed and distributed by the Office of Security Management System (ISMS). The ISO/IEC 27001 and ISO/IEC
Government Commerce (OGC) in UK. ITIL is currently generating con- 27002 standards are the core components of the ISO/IEC 27k series
siderable interest in IT service management. It focuses on aligning IT responsible for implementing the ISMS. The ISO/IEC 27001 specifies
services with business, improving the quality of IT service delivery, and the cyclic management approach to continuously improve organiza-
reducing IT costs [23]. The framework describes how IT resources and tions’ ISMS. It also describes the concept of “Statement of Applicability”
services are organized to ensure value delivery and documents processes, which refers to the process of identifying the appropriate controls
functions, and roles. ITIL provides IT leaders with a multidimensional (addressed in the “Annex A” of ISO/IEC 27001) that should be applied
structure, a common semantic, and a lifecycle that describes the main to each lifecycle phase. Particularly, the annex addresses the informa-
stages, processes, and activities for IT service management. tion security controls and control objectives, from the ISO/IEC 27002,
Although ITIL is not designed to support Cloud service management, that should be implemented within each specific lifecycle phase. The
the effort required to extend this framework is reduced [42]. In fact, ISO/IEC has released the ISO/IEC 17788 [51] and ISO/IEC 17789 [52]
according to [43,44], nearly all ITIL processes remain applicable and standards to help organizations face the new security and privacy issues
valuable for the provisioning of IT services through CC. ITIL is, how- raised by CC adoption. The ISO/IEC 17788 specifies the CC overview
ever, limited to IT services management and delivery. Therefore, its and vocabulary, whereas the ISO/IEC 17789 describes the CC reference
adoption as a framework for CC governance will not provide a com- architecture. The ISO/IEC has also released the ISO/IEC 27017 [53]
prehensive approach in which all CC governance aspects are taken into and ISO/IEC 27018 [54] to, respectively, provide the code of practices
account. for information security in CC and ensure the protection of personally
identifiable information for CC services.
3.2. COBIT Although the ISO/IEC 27k Series of standards provide a risk-based
approach that covers the confidentiality, availability, integrity, and
The Information Systems Audit and Control Association (ISACA) has compliance aspects, these are still limited to the information security
published COBIT as an IT management and governance framework. domain. Thereby, their combination with other IT models is required.
COBIT helps organizations to manage IT risks, align IT strategy with We would like to mention that only ISO/IEC 27001 and ISO/IEC 27002
business goals, and ensure legal and regulatory compliance. COBIT is a (referred in this paper as 27001/2) are considered in this study as they
supporting tool-set that allows organizations to bridge the gap among are the core standards of the ISO/IEC 27k Series.
control requirements, technical issues, and business risks [45]. The
framework provides a set of generally accepted measures, indicators, 3.4. Unification of IT models
and processes to both maximize the IT value and develop the appro-
priate IT governance capabilities [46]. In addition to IT governance, Despite the diversity of the models involved (see Table 2), the
ISACA has developed COBIT to provide control and governance of CC, choice of a particular standard/framework or their combination de-
supply best practices for maximizing its value, and manage its asso- pends on the organization, the personnel, the maturity, and the effec-
ciated risks [47,48]. ISACA has released the “IT Control Objectives for tiveness of models’ process elements. Obviously, using a single model is
Cloud Computing” [13] to assist organizations encounter the CC gov- not enough. According to Sahibudin et al. [22], IT models are not
ernance challenges as well as develop a comprehensive governance comprehensive enough on their own to provide an efficient IT gov-
strategy around CC infrastructures and services. ernance and management system. Additionally, despite the duplications
Although COBIT provides a more comprehensive approach than and overlaps that exist between these models, their unification will help
ITIL and ISO/IEC 27001/2 [42,49,50], its adoption for CC governance organizations to cover the “what, how, and scope” of IT governance
presents limitations. One of its disadvantages is that it requires a deep [42]. This can provide organizations with the ability to integrate dif-
knowledge to implement its processes. Additionally, guidelines to im- ferent practices from different models, thus, allowing a more effective
plement the framework or research studies that investigate the frame- IT control and governance [39,55].
work's utilization are almost absent [49]. The unification of multiple models can help organizations better

102
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

overcome the CC governance challenges. Each model presents strengths iv Result analysis: The final phase of the evaluation method analyzes
and limitations variously when applied to CC governance. The uni- the findings. The models’ evaluation outcomes are then compared
fication will enable organizations to lessen the limitations and to better and synthetized.
fulfill the CC governance requirements. As each model determines its
own structure, terminology, definitions, and quality systems, etc. we More details about the evaluation method and its application can be
can then notice an increase in the complexity of models’ unification found in [59]. This section will first present the CC governance re-
[56]. We thus consider the unification process as a result of executing quirement identification. Second, it will try to provide details on the
the following unification methods. These methods have been designed evaluation approach; finally, it will discuss and analyze the findings.
to support the evaluation, integration, and homogenization of multiple
models to elaborate a unified framework for CC governance. 4.1. Requirements identification

• Evaluation method: It provides a quantitative approach to assess The first phase of the evaluation method concerns the identification
the accuracy with which a specific model validates a set of CC of the CC governance requirements. These requirements are the main
governance requirements. This method is performed by identifying component of the evaluation method. They have been identified
the CC governance requirements, mapping the models’ process ele- through a systematic research review of CC governance studies. To
ments to these requirements, and calculating the models’ coverage elaborate the systematic research review, we relied on the approach
rate. proposed by Brereton et al. [60]. As show in Fig. 2, the review process
• Integration method: It furnishes an approach to integrate the is divided into the three following steps: Plan the review; conduct the
evaluated IT models. This method focuses on identifying the pos- review; and analyze and document the review. These steps are de-
sible combinations that can exist between the models’ process ele- scribed as follows:
ments. The objective is to elaborate new integrated practices of (i.1) Plan the review:
these process elements which better address the CC governance re- The first step deals with the identification of the research questions
quirements. and the development of review protocol. First, we have defined the
• Homogenization method: It affords an approach to resolve the three following research questions that constitute the basis of the re-
structural heterogeneities that exist between the models in the de- search strategy for extracting the literature.
scription of their process elements. The purpose of the homo-
Q1: What are the existing approaches for Cloud Computing gov-
genization method is to provide a homogenized representation of
ernance?
the process elements that establish an integrated practice.
Q2: What are the types of approaches for cloud computing gov-
As we have already mentioned in the introduction, these three ernance?
methods are not specifically designed to support the unification of the
Q3: What are the CC governance requirements?
three IT models included in this study. They are rather designed to
support the unification of multiple models regardless of their type. The Second, based on the identified research questions, we have devel-
main contribution of this research is the approach that we have put oped the review protocol. The protocol supplies a detailed plan of the
forward to evaluate, integrate, and homogenize IT models and not the review process. This includes, for instance, the necessary steps to follow
application of this approach to particular IT models. The development when conducting the review or the conditions to apply when selecting
of the three methods has been based on the following research studies: the primary studies. Based on the developed protocol a search strategy
[22,37,39,56–58]. In the following sections, the evaluation and in- has been elaborated.
tegration methods will be described. The homogenization method will (i.2) Conduct the review:
be the subject of what follows. The second step of the review process focuses on: identifying the
relevant research studies; selecting the primary studies; extracting the
4. Evaluation method required data; and synthetizing the extracted data. First, to identify the
relevant studies, we have searched the following databases with the
This section is intended to provide details about the evaluation search string presented in Table 3: ACM Digital Library; IEEE Xplore;
method. The objective is to develop a method that provides the ability Science Direct; Springer Link; Google Scholar; and Wiley. We have
to measure the coverage rate of models’ process elements with regard to extracted a significant amount of peer-reviewed publications from 2010
CC governance requirements. For this purpose, an approach has been to 2017. Second, we have used the inclusion and exclusion criteria
elaborated. It is used to evaluate ITIL, COBIT, and ISO/IEC 27001/2. As presented in Table 4 to select the primary studies among the identified
shown in Fig. 1, the evaluation method embodies the following phases: relevant studies. Applying the inclusion and exclusion criteria has re-
sulted in the selection of 32 primary studies (see Table A1 of the ap-
i Requirements identification: The first phase of the evaluation pendix). Finally, the CC governance requirements are identified and
method focuses on identifying the CC governance requirements. then synthesized. We would like to point out that the requirements
These are identified through a systematic research review of the identification process is based on reviewing and synthesizing the CC
literature. The CC governance requirements constitute the basis governance requirements that have been addressed by the primary
upon which the models can be evaluated and subsequently in- studies. Thus, for each primary study, a list of requirements is elabo-
tegrated. rated by taking into account both the abstraction level as well as the
ii Models mapping: The second phase of the evaluation method context.
achieves the models mapping. This phase focuses on mapping the To ensure a better understanding, structuring, and consistency of
models’ process elements to the identified CC governance require- these requirements, they have been organized into three hierarchical
ments. The mapping is accomplished through a comparative analysis levels. The first level gives the high-level functional requirements. They
between the requirements of the process elements and CC governance. are called Required Domains (RDs). The next level is Required Controls
iii Models evaluation: The third phase of the evaluation method (RCs). They provide more details about the RDs. Finally, the last level is
measures the coverage rate of the mapped process elements to de- Required Activities (RAs). They highlight the low-level functional re-
termine their score. This score expresses the percentage of the CC quirements. The motivation behind the proposed structuring of CC
governance requirements that are validated by the mapped process governance requirements is, first, to allow their definition at different
elements. levels of abstraction and precision; second, to ensure efficiency and ease

103
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Fig. 1. Evaluation method.

Fig. 2. Steps to achieve the systematic research review.

of use; and finally, to maximize the potential of exploring a large set of


Table 3. CC governance requirements. More details about the approach followed
Search string. to identify these requirements can be found in [10].
Cloud Computing This approach has enabled us to identify the four following key
governance RDs: Cloud Migration (CM); Information Security (IS); Risk
[AND] Management (RM); and Service Level Agreements (SLA). The overall
Services 〈OR〉 Data 〈OR〉 Risk 〈OR〉 Performance 〈OR〉 Migration 〈OR〉 Information
objectives of these RDs are as follows: First, the CM required domain
security 〈OR〉 Confidentiality 〈OR〉 Service level agreement 〈OR〉 Compliance
〈OR〉 Service provider
aims to provide the required controls and activities for planning, ex-
[AND] ecuting, evaluating, and monitoring CC migration. Second, the IS re-
Governance 〈OR〉 Requirements quired domain intends to supply the required controls and activities
relating to the implementation of CC security measures. Third, the RM
required domain aims to furnish the required controls and activities

Table 4.
Inclusion and exclusion criteria.
Inclusion Studies that propose an approach for the governance of Cloud Computing or the governance of its subdomains (e.g. data governance, information security
governance, etc.).
Studies in English in the form of a peer-reviewed manuscript.
Studies in English published by recognized organizations (e.g. ISACA, CSA, etc.).
Exclusion Studies that do not explicitly propose an approach for the governance of Cloud Computing or the governance of its subdomains.
Studies developed for a very specific context (e.g., eHealth, national security, etc.).
Non-English and non-peer reviewed Studies, white papers, theses, books, and chapters.

104
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

dealing with the identification, analysis, mitigation, monitoring, and (ii.1) Knowledge acquisition:
review of CC security risks. Finally, the SLA required domain intends to The first step concerns the acquisition of knowledge about the
provide the required controls and activities concerned with the initia- models involved. This step is first performed by analyzing the archi-
tion, negotiation, establishment, monitoring, enforcement, and termi- tectural components of the models in order to identify how each model
nation CC service level agreements. describes its process elements (see the last row of Table 2). Second, the
It can be noted that each RD is composed of several RCs. Similarly, process elements of each model are analyzed in order to identify the
each RC comprises various RAs that describe the Cloud service user and Candidate Process Elements (CPEs). The analysis process focuses on
the Cloud service provider requirements. An example illustrating the performing a high-level examination of each process element descrip-
identified CC governance requirement of the SLA required domain is tion. Thus, if the description of a process element is adequate or par-
given in Table A2 of the appendix. Moreover, Fig. A1 of the appendix tially adequate to satisfy the overall objective of a RD, this process
shows all the identified RCs and RAs with regard to the four RDs. element will be considered as a CPE. Hence, for each model and for
It is worth mentioning that the IS required domain follows the same each RD, a list of CPEs is elaborated. Third, the list of CPEs is enriched
hierarchical structure provided by the ISO/IEC 27002 because this by analyzing the mapping proposed by existing studies; for instance, the
standard provides a comprehensive approach and detailed processes CCM [29] or ITGI guidance for aligning IT models [61]. Thus, if the
that consider all information security aspects. It is also worth noting analysis of existing studies reveals other process elements that we have
that the objectives of each RC have been elaborated based on the not considered in the first place, these process elements are added to the
overall objectives set for the RDs. This is performed thanks to a func- list of CPEs.
tional breakdown of the overall objective of each RD into a set of sub- For instance, the execution of this first step has enabled us to
objectives. These sub-objectives are further detailed and used to de- identify the following CPEs with regard to the SLA required domain:
velop the objectives of each RC (see Table A2 of the appendix). “SD 3.[3–6]”, “SD 4.2.5 [1–10]”, and “SS 5.4″ from ITIL; “APO 09.
The SLA required domain will be used afterwards to illustrate the [01–05]”, “APO 11.03″, “DSS 01.02″, “BAI04.03″, and “BAI 06.01″
evaluation and integration approaches. Moreover, in order to further from COBIT; “Clauses 15.1.[2–3]”, “Clauses 15.2.[1–2]”, and “Clause
elucidate the models mapping and evaluation phases, the focus will be 8.1.4″ from ISO/IEC 27001/2.
put on the required control “SLA_SNE: SLA Negotiation and We would like to mention that the identifier “SD x.x” or “SD x.x.x.x”
Establishment” (see Table A2 of the appendix). This RC focuses on the (with x integer) is used by ITIL as an abbreviation to identify the pro-
establishment and maintenance of a stable, reliable, and measurable cess elements that belong to the “Service Design” publication. Likewise,
business relationship between the Cloud service provider and the Cloud the identifiers “APO x.x”, “BAI x.x”, and “DSS x.x” (with x integer) are
service user. This RC includes the following RAs: “SLA_SNE_1: CSP used by COBIT as abbreviations to identify the process elements that
discovery”; “SLA_SNE_2: SLA negotiation and establishment” ; respectively belong to the governance domains “Align, Plan and
“SLA_SNE_3: SLA monitoring and reporting”; and “SLA_SNE_4: SLA re- Organise”, “Build, Acquire and Implement”, and “Deliver, Service and
view and enforcement”. Support”.
(ii.2) Process elements validation (1st level mapping):
4.2. Models mapping The second step deals with the process elements validation. This
step is achieved through analyzing the objectives of each CPE with
The second phase of the evaluation method is models mapping. In regard to the RCs’ objectives. Thus, if a CPE contributes or partially
this phase, the models’ process elements are mapped to the identified contributes to achieving the some objectives of a RC, it is mapped to
RAs to determine which of these process elements validate the re- this RC; otherwise, it is eliminated. In order to simplify the mapping
quirements addressed by the RAs. As shown in Fig. 3 bellow, the models between the CPEs and the RCs, all the objectives are considered at the
mapping is performed as follows: same significance level. For each model, the mapping between the CPEs

Fig. 3. Models mapping.

105
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Table 5.
Proposed mapping of ITIL, COBIT, and ISO/IEC 27001/2 for the RAs of the SLA required domain.
ID of RA ITIL mapping COBIT mapping ISO/IEC 27001/2 mapping

SLA_SI_1 SD 3.3, SD 3.4, SD 3.5, SD 3.6, SD 4.2.5.1, SD 4.2.5.9 APO 09.01, APO 09.02 Clause 15.1.1
SLA_SNE_1 SD 4.2.5.1, SS 5.4 APO 09.01, APO 09.02,DSS 01.02 N/A
SLA_SNE_2 SD 4.2.5.2, SD 4.2.5.5 APO 09.03, APO 09.04, BAI 04.03, BAI 06.01 Clause 15.1.2, Clause 15.1.3
SLA_SNE_3 SD 4.2.5.3, SD 4.2.5.6, SD 4.2.5.10 APO09.04 Clause 15.1.2, Clause 15.1.3, Clause 15.2.1, Clause 15.2.2
SLA_SNE_4 SD 4.2.5.4, SD 4.2.5.5, SD 4.2.5.7, SD 4.2.5.8 APO09.05 Clause 15.1.2, Clause 15.1.3,
SLA_ST_1 SD 4.2.5.4, SD 4.2.5.5, SD 4.2.5.8 APO09.04 Clause 8.1.4

and the RCs is performed according to the following mapping criteria: produces a recognizable result.
For each model, the output of this step is a list of KWIRs and their
a When the objectives of a CPE totally or partially match the objec- corresponding contexts. We present in Table 6 below the list of key-
tives of a single RC, a (1:1) mapping is performed. words that are used by each of the three models to describe their re-
b When the objectives of a CPE totally or partially match the objec- quirements. The analysis of the three models has shown that the re-
tives of more than one RC, a (1:n) mapping is performed. quirements may be expressed without the use of KWIRs. This
c If a and b fail, no mapping is performed. corresponds to the two forms of requirements expressing “[verb]…”
and “[verb]… and [verb]”. Although these two forms do not contain
We would like to mention that the mapping is accomplished without keywords, they are considered in the study as KWIRs since their context
considering the RAs included in the RCs. The validated CPEs will then can vary across the models.
establish the list of the Process Elements to Be Mapped (PEBMs). For (ii.4) Mapping performance (2nd level mapping):
each model, the output of this step is a list of RCs and their corre- In this step, the mapping is performed between the RAs and the
sponding PEBMs. We also would like to stress that the first and second PEBMs. The mapping is achieved by accomplishing a comparative
steps of the mapping procedure are iteratively executed until there are analysis between the RAs’ requirements and those of the PEBMs. The
no more CPEs. KWIRs play an important role in the comparison as they enable us to
For instance, the CPEs that have been mapped to the RC “SLA_SNE” easily identify the key requirements of a PEBM. Thus, when the re-
are the ones that appear in rows 2, 3, 4, and 5 of Table 5 below. We note quirements of a PEBM partially or totally meet the requirements of at
that the following CPEs: “SD 4.2.5.9: Develop contacts and relation- least one RA, it is mapped to this RA; if not, it is eliminated. For each
ships” from ITIL and “APO11.03: Focus quality management on cus- model the mapping between the PEBMs and the RAs is performed by
tomers” from COBIT are not mapped to the RC “SLA_SNE” since they do considering the following mapping criteria:
not match its objectives.
(ii.3) Keywords identification: a When the requirements of a PEBM totally or partially address the
The third step focuses on analyzing the models in order to elaborate requirements of a single RA, a (1:1) mapping is performed.
the list of Key Words that Identify Requirements (KWIRs). To identify b When the requirements of a PEBM totally or partially address the
these keywords, we have first relied on existing guidelines to determine requirements of more than one RA, a (1:n) mapping is performed.
the common keywords used in the literature to document the require- c If a and b fail, no mapping is performed.
ments; for example, Bradner guideline of RFC Series [62]. Second, we
have analyzed the three models and we have extracted the KWIRs that We would like to mention that when two RAs or more are validated
each model uses to describe its requirements. The context of each KWIR with the same process element (which corresponds to n:1 mapping
is also identified. For instance, the statement “Should [verb] that” is between the RAs and the PEBM), this process element will be dupli-
used by the ISO 27001/2 to indicate the list of requirements derived cated in all the RAs. Thereby, the n: 1 mapping will be substituted by n
from a particular process, activity, or task; while the same statement is mapping of the type 1: 1; where n is the number of the RAs that are
used by COBIT to indicate the requirements that must be satisfied by a validated by the same PEBM. For each model, the output of this step is a
particular activity. We would like to mention that a process is a logical list of RAs and their corresponding mapped process elements. We pre-
group of activities which uses resources to transform inputs into out- sent in Table 5 the obtained results from mapping ITIL, COBIT, and ISO
puts. An activity, meanwhile, is a logical group of tasks (set of atomic 27001/2 to the RAs included in the SLA required domain.
units of work to be completed) that occurs over time and necessarily For instance, the PEBMs “APO09.03″, “APO09.04″, “BAI04.03″, and

Table 6.
List of KWIRs for ITIL, COBIT, and ISO/IEC 27001/2.
Model KWIRs Context description

ITIL [verb]… / [verb]… and [verb] / Should [verb] / Should [verb] … and Indicates the requirements that must be satisfied by a particular process, activity,
[verb] / Must [verb] / Must [verb] … and [verb] or action.
Should be [verb] / Must be [verb] Specifies the characteristics of a particular process, activity, or action.
Should include / Must include Indicates the details to include in a particular process, activity or action.
COBIT [verb]… / [verb]… and [verb] / Should [verb] / Should [verb] … and Indicate the requirements that must be satisfied by a particular activity.
[verb] / should [verb] that
Should be [verb] / Must be [verb] Specifies the characteristics of a particular activity.
Should include Indicates the details to include in a particular activity.
ISO/IEC 27001/2 [verb]…/ [verb]… and [verb] / Should [verb] / Should [verb] … and Indicate the processes, procedures, activities, or tasks that should be developed
[verb] and maintained.
Should [verb] that Indicates the requirements that must be satisfied by a particular process,
procedure, activity, or task.
Should be [verb] Specifies the characteristics of a particular process, procedure, activity, or task.
Should include Indicates the details to include in a particular process procedure, activity, or task.

106
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Fig. 4. Models evaluation.

“BAI06.01″ of COBIT are mapped to the RA “SLA_SNE_2″ as follows: few aspects addressed; some aspects addressed; many aspects addressed;
The requirements description of the RA “SLA_SNE_2″ (see Table A2 of semi-complete coverage; and complete coverage. The rationale behind
the appendix) and those of the PEBMs “APO09.03″, “APO09.04″, using this scale and not a more granular one (for example, a scale that
“BAI04.03″, and “BAI06.01″ are decomposed into multiple independent contains 11 levels of coverage scores with a step equal to 10%) is to
and integrated sub-descriptions of requirements. The sub-descriptions reduce the complexity of how the diverse RAs are evaluated.
of the PEBMs “APO09.03″, “APO09.04″, “BAI04.03″, and “BAI06.01″ For instance, if an 11 levels scale is used, it would be difficult to
are denoted by a numerical value. They are then mapped to the sub- determine with precision the coverage score of the process elements
descriptions of the RA “SLA_SNE_2″ which are displayed on the left side “APO09.03″, “APO09.04″, “BAI04.03″, and “BAI06.01″ from COBIT
of a matrix. Table A3 of the appendix gives the performed mapping with regard to the RA “SLA_SNE_2″ (see Table A3 of the appendix). In
between the RA “SLA_SNE_2″ and the PEBMs “APO09.03″, “APO09.04″, this case, the coverage score of these process elements would be 80 or
“BAI04.03″, and “BAI06.01″. 90% since the RA is almost completely covered by these process ele-
(ii.5) Mapping analysis: ments. If a 6 level scale is used, the coverage score of these process
In this last step, an analysis of the mapping results is performed. The elements will be exactly 80% since the RA is not completely covered,
analysis process focuses on verifying the findings, making the necessary and the mapped process elements would cover more than 60% of the
modifications if errors have been made during the mapping process, RA's requirements.
and synthesizing the results. (iii.2) Required activities evaluation:
For instance, the analysis of the mapping results of the three models The second step focuses on measuring the coverage score of the
shows that all the SLA required activities are validated by ITIL with two mapped process elements for each RA (CS _RAi ). This score gives the
process elements or more. COBIT validates 50% of the SLA required percentage of the CC governance requirements that have been validated
activities with one process element, and the rest is validated with two by the proposed mapping. Based on the elaborated measurement scale,
process elements or more. Concerning the ISO/IEC 27001/2, 16.67% of we present in Table 8 (columns C8, C9, and C10) the coverage scores of
the of the SLA required activities are not validated by any process ITIL, COBIT, and ISO/IEC 27001/2 for the RAs included in the SLA
elements, 33.33% of the SLA required activities are validated with one required domain.
process element, and the rest is validated with two process elements or For instance, the analysis of the mapping results presented in
more. Table A3 of the appendix shows that the process elements “APO09.03″,
“APO09.04″, “BAI04.03″, and “BAI06.01″ from COBIT cover almost all
4.3. Models evaluation the requirements addressed by the RA “SLA_SNE_2″. Based on the scale
presented in Table 7 above, the coverage of these process elements
The third phase of the evaluation method aims to calculate the corresponds to a semi-complete coverage (level 4). Thus, the coverage
coverage score of the models for each RA, RC, and RD. As shown in score of these process elements will be 80%.
Fig. 4 bellow, the models evaluation is performed as follows: We would like to point out that even if the number of the mapped
(iii.1) Design of required activities evaluation: process elements to a RA is high, it does not mean that the coverage
The first step focuses on designing the measurement scale that we score of these process elements will be as high. What allows the cov-
will use to determine the coverage score of the mapped process elements. erage score to increase is the number of requirements that are validated
Given the difficulty to measure the correlation between the descriptions by the mapped process elements. For instance, the “SLA_SNE_3″ is
of the process elements and those of the RAs (due to their qualitative covered by four different process elements from ISO/IEC 27001/2 and
nature), the measurement scale will only contain six levels of coverage the corresponding coverage score is 60%; whereas, this RA is covered
scores. These levels will range from 0 to 100% with a step equal to 20%. by just a single process element from COBIT and the corresponding
As shown in Table 7 bellow, these levels are respectively: Not addressed; coverage score is 100%.

107
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Table 7.
Scale for determining the coverage score of the RAs.
Level Converge score (%) Description Detail

0 0 Not addressed There is no match between the RA's requirements and those of the mapped process elements.
1 20 Few aspects addressed Few requirements are addressed by the mapped process elements.
2 40 Some aspects addressed Some requirements are addressed by the mapped process elements.
3 60 Many aspects addressed Many requirements are addressed by the mapped process elements.
4 80 Semi-complete coverage The RA is almost completely covered by the mapped process elements.
5 100 Complete coverage The RA is completely covered by the mapped process elements.

(iii.3) Required controls evaluation: Table 9.


The third step of the models evaluation focuses on calculating the Coverage score of ITIL, COBIT, and ISO/IEC 27001/2 for the four RDs (CS_RDi).
coverage score for each RC (CS _RCi ) . This coverage score is calculated RDi CS_RDi
using Eq. (1) and gives the average score of the RAs comprised in a
particular RC. To compute the coverage score of a RC, a relative weight ITIL COBIT ISO/IEC 27001/2
CM 85.10 81.05 51.45
is first determined for each of its RAs (RW _RAi ). This relative weight
IS 44.10 79.29 90.33
emphasizes the relation between the RAs and the RCs by considering RM 51.47 95.47 35.73
the number of the supporting studies or the appearance frequency of SLA 80.40 61.87 21.67
each RA ( AF _RAj ). The relative weight of a RA is calculated using Eq.
(2). We can notice that the more the appearance frequency of a RA
increases the more the relative weight follows suit. We give in Table 8 Eq. (3). Calculating the coverage score of RDs.
(columns C6 and C7) the appearance frequencies and the calculated m
CS_RCj
relative weights for the RAs included in the SLA required domain. We CS_RDi =
j=1
*100
also give in Table 8 (columns C2, C3, and C4) the calculated coverage m
scores of ITIL, COBIT, and ISO/IEC 27001/2 for the RCs included in where,
same RD. CS _RDi : Coverage score of mapped process elements for the RDi.
(iii.4) Required domains evaluation: CS _RCj : Coverage score of mapped process elements for the RCj.
The final step of the models evaluation deals with computing the m: Number of RCs included within the RDi.
coverage score for each RD (CS _RDi ). This Score is calculated using Eq. (iii.5) Results analysis:
(3) and gives the average score of the RCs included within a specific In this last step, an analysis of the evaluation results is carried out.
RD. However, since all the RCs are obligatory for each RD, they will The analysis process focuses on verifying the obtained results, making
consequently have the same relative weight when considering the re- the necessary modifications if errors have been made during the eva-
lation between the RCs and the RDs. We give in Table 9 the calculated luation process, and synthesizing the results.
coverage scores of ITIL, COBIT, and ISO/IEC 27001/2 for the four RDs.
Eq. (1). Calculating the coverage score of RCs. 4.4. Findings analysis
n
CS _RCi = CS _RAj *RW _RAj Based on the models assessment, the following can be observed.
j=1 First, as shown in Table 9 above, the CM required domain is validated
by the process elements of ITIL, COBIT, and ISO/IEC 27,001/2 with a
where,
coverage score respectively equal to 85.10%, 81.05%, and 51.45%. We
CS _RCi : Coverage score of mapped process elements for the RCi.
note that ITIL and COBIT have demonstrated a high coverage for the
CS _RAj : Coverage score of mapped process elements for the RAj.
requirements addressed by this RD, while ISO/IEC 27001/2 has shown
RW _RAj : Relative weight of the RAj.
an acceptable coverage. As shown in Table A4 of the appendix, ITIL
n: Number of RAs included within the RCi.
validates 40% of the CM required activities with a coverage score equal
Eq. (2). Calculating the relative weight of RAs.
to 100%, and validates the rest with a coverage score ranging from 40
FA_RAj to 80%. COBIT validates 33.33% of the CM required activities with a
RW_RAj = n
AF_RAk coverage score equal to 100%, and validates 66.67% of the RAs with a
k=1
coverage score ranging from 40 to 80%. All the CM required activities
where, are validated by ISO/IEC 27001/2 with a coverage score inferior or
RW _RAj : Relative weight of the RAj. equal to 80%.
AF _RAj : Appearance frequency of the RAj (number of supporting Second, the IS required domain is validated by the process elements
studies). of ITIL, COBIT, and ISO/IEC 27001/2 with a coverage score respec-
n: Number of RAs included within the RC that contains RAj. tively equal to 44.10%, 79.29%, and 90.33% (see Table 9 above). The

Table 8.
Coverage scores for the RAs and RCs of the SLA required domain (CS_RAi and CS_RCi).
ID of RCi(C1) CS_RCi ID of RAj (C5) AF_RAj(C6) RW_RAj(C7) CS_RAj
ITIL(C2) COBIT (C3) ISO/IEC 27001/2 (C4) ITIL (C8) COBIT(C9) ISO/IEC 27001/2(C10)

SLA_SI 80 60 20 SLA_SI_1 8 1 80 60 20
SLA_SNE 81.2 85.6 26.8 SLA_SNE_1 7 0.22 60 60 0
SLA_SNE_2 9 0.28 80 80 20
SLA_SNE_3 9 0.28 100 100 60
SLA_SNE_4 7 0.22 80 100 20
SLA_ST 80 40 20 SLA_ST_1 6 1 80 40 20

108
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

evaluation of ISO/IEC 27001/2 and COBIT has revealed a high cov- score of each RC and RD as described by the evaluation method. It is
erage for the requirements addressed by the IS required domain, while significantly noted that when a process element PE validates more than
the evaluation of ITIL has demonstrated a less than acceptable cov- one RA with different sub-process elements (PEi.1, PEi.2 …), each sub-
erage. As shown in Table A5 of the appendix, only one RA is validated process element will be considered as a normal process element in the
by ITIL with a coverage score equal to 100%, and the rest is validated integration process. This is reclusively performed as long as there are
with a coverage score inferior or equal to 80%. COBIT validates 37.14% more RAs validated with the same upper process element.
of the IS required activities with a coverage score equal to 100%, and To reduce the complexity of integrating the process elements of
validates the rest with a coverage score inferior or equal to 80%. ISO/ multiple models, the integration will be carried out in a group of two.
IEC 27001/2 validates 60% of the IS required activities with a coverage For instance, to integrate the process elements of models A, B, and C
score equal to 100%, and validates 40% of the RAs with a coverage with regard to a specific RA, the process elements of models A and B are
score ranging from 60 to 80%. first integrated to form a hybrid practice AB. This practice will be in-
Third, the RM required domain is validated by the process elements tegrated thereafter with the process elements of model C to form the
of ITIL, COBIT, and ISO/IEC 27001/2 with a coverage score respec- final integrated practice. We would like to reveal that a practice is the
tively equal to 51.47%, 95.47%, and 35.73% (see Table 9 above). We result of integrating various process elements from different models.
note that COBIT has shown a high coverage score for the requirements Two types of practices can be distinguished: The reference and the
addressed by the RM required domain, while ITIL and ISO/IEC 27001/2 complement practices. The reference practices contain the process
have respectively demonstrated an acceptable and quasi-acceptable elements that constitute the basis of models integration. The comple-
coverages. As shown in Table A6 of the appendix, all the RM required ment practices are particularly used when the organization has a pre-
activities are validated by ITIL with a coverage score ranging from 20 to ference for one model instead of another; otherwise, the com-
60%, and validated by ISO/IEC 27001/2 with a coverage score inferior plementary practice is rejected.
or equal to 80%. COBIT validates 28.57% of the RM required activities Let's mention that if a model (say M1) validates a RA (say RA1) (or
with a coverage score equal to 80%, and the rest is validated with a part of a RA – say p1) with n process elements (n ≥ 1), and a model M2
coverage score equal to 100%. validates RA1or p1 with m process elements such that m > n, then M1 is
Finally, the SLA required domain is validated by the process ele- considered more detailed than M2 with respect to RA1 (or p1) because
ments of ITIL, COBIT, and ISO/IEC 27001/2 with a coverage score re- it performs the validation with fewer process elements than M2. This
spectively equal to 80.40%, 61.87%, and 21.67% (see Table 9 above). means that the n process elements through which M1 validates RA1 (or
The evaluation of ITIL has shown a high coverage for the requirements p1) are more detailed, on average, than the m process elements that M2
addressed by this required domain, while COBIT and ISO/IEC 27001/2 requires in order to validate RA1 (or p1). It is important to mention that
have respectively shown a moderate and low coverages. As shown in the detail level considered in this study does not correspond to the
Table 8 above, 17% of the SLA required activities are validated by ITIL abstraction level of the process elements. This is not quantified in this
with a coverage score equal to 100%, while 83% of the RAs are vali- study since the focus of the three models is different. In fact, COBIT and
dated with a coverage score ranging from 60 to 80%. COBIT validates ISO/IEC 27001/2 focus on what should be done while ITIL focuses on
33.33% of the SLA required activities with a coverage score equal to how it should be done. In this way, ITIL will always be the model that
100, and validates the rest with a coverage score ranging from 40 to provides the lowest abstraction level compared with COBIT and ISO/
80%. All the SLA required activities are validated by ISO/IEC 27001/2 IEC 27,001/2. Accordingly, the process elements of ITIL will always be
with a coverage score inferior or equal to 60%. prioritized in the models integration. The detail level that we take into
The evaluation of the three models with regard to the CC govern- consideration corresponds to the number of process elements that va-
ance requirements has revealed several strengths and limitations. These lidate the requirements of a particular RA. The more the number of
vary according to the models. The obtained results from evaluating ITIL process elements decreases the higher the detail level will be. The detail
are quite consistent since the mapped process elements of this model level is quantified using Eq. (4).
provide detailed practices for defining the Cloud migration strategy, Eq. (4). Calculating the level of detail.
planning and managing change as well as IT service transition, and 1
managing service level agreements. The results obtained from evalu- DL p,i =
NPE p,i
ating COBIT are also fairly consistent stressing the comprehensiveness
of this framework that shows a high coverage score for all the RDs; where,
except for the SLA required domain. The evaluation findings of ISO/IEC DLp,i: Detail level of a model that validates a part p of a RA i.
27001/2 are also reasonably consistent since its process elements focus NPEp,i: Number of process elements of a model that validates a part
on information security, which explain the low coverage scores shown p of a RA i.
by the model with regard to the CM, RM, and SLA required domains. For instance, assuming that ITIL validates the same requirements as
the process elements “APO09.03″, “APO09.04″, “BAI04.03″, and
5. Integration method “BAI06.01″ from COBIT with regard to the RA “SLA_SNE_2″ (see
Table A3 of the appendix); but with only one process element. The
The evaluation method described in the previous section has en- process element of ITIL will be considered as more detailed and will
abled us to measure the adequacy of each IT model with regard to the have a detail level equal to one, whereas the process elements
CC governance requirements. However, to better cover these require- “APO09.03″, “APO09.04″, “BAI04.03″, and “BAI06.01″ from COBIT
ments, the models integration is required. This section aims both at will be considered as less detailed and will have a detail level equal to
describing the integration approach and analyzing the integration re- 0.25.
sults of ITIL, COBIT, and ISO/IEC 27001/2. The process elements of two models A and B are integrated based on
the method described in Fig. 5. It is absolutely important to mention
5.1. IT models integration that the “α” parameter figuring in the value of the new coverage score
represents the percentage of requirements that are validated by the
The integration is based on the models assessment. It aims at pro- process elements of A and B. For each RA, the integration is performed
viding an approach to identify the possible combinations among the as follows:
mapped process elements. The integration is performed at the RAs
level. This will enable us to measure the coverage score of integrated i When the RA is not validated by any process element, no integration
process elements at this level first, and then to calculate the coverage is performed.

109
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

ii When the RA is validated by the process elements of only one model, “DSS 01.02″ from COBIT are integrated since they cover different re-
the process elements of this model are taken as a reference practice. quirements. They are therefore considered as a reference practice.
iii When the RA is validated by the process elements of two models, Third, with respect to the RA “SLA_SNE_2″, the ITIL and COBIT
this implies considering the following: coverage scores are equal and greater than the ISO/IEC 27001/2 cov-
a When the coverage score of A is superior to the coverage score of erage score. The process elements “Clause 15.1.2″ and “Clause 15.1.3″
B (or vice versa), this implies considering the following: from ISO/IEC 27001/2 are thus considered as a complement practice.
• When the process elements of B partially validate the same re- The process elements “SD 4.2.5.2″ and “SD 4.2.5.5″ from ITIL and the
quirements as the process elements of A (B ⊂ A), the process process elements “APO 09.03″, “APO 09.04″, “BAI 04.03″, and “BAI
elements of A are taken as a reference practice and the process 06.01″ from COBIT are integrated since they cover different require-
elements of B are taken as a complement practice. ments. They are therefore considered as a reference practice.
• When the process elements of B validate different requirements Fourth, concerning the RA “SLA_SNE_3″, the ITIL and COBIT cov-
than the process elements of A (A ∩ B = ∅), the process ele- erage scores are equal and greater than the ISO/IEC 27001/2 coverage
ments of A and B are integrated and considered as a reference score. The process elements “Clause 15.1.2″, “Clause 15.1.3″, “Clause
practice. 15.2.1″, and “Clause 15.2.2″ from ISO/IEC 27001/2 are thus considered
• When the process elements of A validate both the same and as a complement practice. The process elements “SD 4.2.5.3″, “SD
different requirements than the process elements of B (A ∩ B ≠ 4.2.5.6″, and “SD 4.2.5.10″ from ITIL are less detailed than the process
∅ and B ⊄ A), the process elements of A and B that validate element “APO 09.04″ from COBIT. The COBIT process element will be
different requirements are integrated into a provisional prac- considered as a reference practice, while the ITIL process elements will
tice. The process elements of A and B that validate the same be considered as a complement practice.
requirements are selected based on the detail level that A and B Fifth, concerning the RA “SLA_SNE_4″, the COBIT coverage score is
provide. Thus, the less detailed process elements are rejected. greater than the ITIL and ISO/IEC 27001/2 coverage scores. The pro-
The more detailed process elements are integrated with the cess element “APO 09.05″ from COBIT is thus considered as a reference
provisional practice and are considered as a reference practice. practice. The process elements “SD 4.2.5.4″, “SD 4.2.5.5″, “SD 4.2.5.7″,
b When the coverage score of A is equal to the coverage score of B, and “SD 4.2.5.8″ from ITIL and the process elements “Clause 15.2.1″
this implies considering the following: and “Clause 15.2.2″ from ISO/IEC 27001/2 are therefore considered as
• When the process elements of A validate different requirements a complement practice.
than the process elements of B (A ∩ B = ∅), the process ele- Finally, regarding the RA “SLA_ST_1″, the ITIL coverage score is
ments of A and B are integrated and considered as a reference greater than the COBIT and ISO/IEC 27001/2 coverage scores. The
practice. process elements “SD 4.2.5.4″, “SD 4.2.5.5″, and “SD 4.2.5.8″ from ITIL
• When the process elements of A validate both the same and and the process element “APO 09.04″ from COBIT are integrated since
different requirements than the process elements of B (A ∩ they cover different requirements. They are thus considered as a re-
B ≠ ∅ and A ≠ B), the process elements of A and B that va- ference practice. The process element “Clause 8.1.4″ from ISO/IEC
lidate different requirements are integrated into a provisional 27001/2 is therefore considered as a complement practice.
practice. The process elements of A and B that validate the same Once the integration is performed, a post evaluation phase of the
requirements are selected based on the detail level that A and B integrated process elements is achieved. This process follows the same
provide. Thus, the less detailed process elements are rejected. evaluation method described in Section 4. It aims at calculating the
The more detailed process elements are integrated with the coverage score of the integrated process elements for each RA, RC, and
provisional practice and are considered as a reference practice. RD. Hence, we give in Fig. 6 the coverage scores of the integrated
• When the process elements of A validate the same requirements practices and their comparison with to the results described in Table 9.
as the process elements of B (A = B), more detailed process The findings have shown the added value to integrate the three
elements of A and B are considered as a reference practice, less models. For all the RDs, we note that the integrated practices have
detailed process elements are, then, considered as a comple- demonstrated a coverage score that is superior to the max coverage
ment practice. score when the models’ process elements are used separately. We note
that the coverage score of the integrated process elements increased the
5.2. Results analysis max coverage by: 10.75% for the requirements addressed by the CM
required domain; 4.9% for the requirements addressed by the IS re-
Following the method described in the previous subsection, the quired domain; 2.4% for the requirements addressed by the RM re-
following results are obtained from integrating ITIL, COBIT, and ISO/ quired domain; and 18.13% for the requirements addressed by the SLA
IEC 27001/2. Table A7 of the appendix gives the integrated practices required domain. We also note that the models integration has led to a
for the required activities included in the SLA required domain. For radical increase in the coverage score of the CM and the SLA required
each RA, we specify the reference and the complement practices. For domains. It has also led to a minor increase in the coverage score of the
each practice, the integrated process elements are specified according IS and RM required domains.
to the model in which they are included. On the basis of the results
provided in Table 8, the six RAs of the SLA required domain are in- 6. Conclusion and outlooks
tegrated as follows:
First, with regard to the RA “SLA_SI_1″, the ITIL coverage score is This paper has proposed a general approach to CC governance based
greater than the COBIT and ISO/IEC 27001/2 coverage scores. The on the unification of IT models and more particularly on the latter's
process elements “SD 3.[3–6]”, “SD 4.2.5.1″, and “SD 4.2.5.9″ from ITIL evaluation and integration. This approach has been particularly applied
and the process elements “APO 09.01″ and “APO 09.02″ from COBIT to ITIL, COBIT, and ISO 27001/2 to elaborate a unified CC management
are integrated since they cover different requirements. They are and governance framework.
therefore considered as a reference practice. The process element The performed state-of-the-art review has shown the existence of
“Clause 15.1.1″ from ISO/IEC 27001/2 is considered as a complement several approaches for CC governance. Despite of the existing ap-
practice. proaches quality, several shortcomings have been spotted and the need
Second, regarding the RA “SLA_SNE_1″, the ITIL and COBIT cov- for a comprehensive and complete approach for CC governance has
erage scores are equal. The process elements “SD 4.2.5.1″ and “SS.5.4″ been highlighted.
from ITIL and the process elements “APO 09.01″, “APO 09.02″, and As ITIL, COBIT, and ISO 27001/2 constitute a promising starting

110
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Fig. 5. Method used to carry out the models integration.

97.87 98.53
100.00 95.85 95.23 95.47
90.33
90.00 85.10
81.05 79.29 80.40
Coverage Score of RDs (%)

80.00
70.00 61.87
60.00
51.45 51.47
50.00 44.10

40.00 35.73

30.00
21.67
20.00
10.00
0.00
CM IS RM SLA
ITIL COBIT ISO/IEC 27001/2 Integrated models

Fig. 6. Coverage score of the integrated process elements for the four RDs.

point for the elaboration of a CC governance framework, their review integration methods that have been developed and described in this
has revealed strengths and weaknesses when used separately to support paper.
the CC governance. The models review has also exposed the need to The proposed evaluation method has been based on the identified
unify these models to better support and back up the new CC govern- CC governance requirements. The evaluation of ITIL has shown a high
ance requirements. To perform the unification several methods have coverage score for the requirements addressed by the CM and SLA re-
been suggested to become the topic of the evaluation and the quired domains, an acceptable coverage score for the requirements

111
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

addressed by RM required domain and a low coverage score for the to be considered and processed in the unification process, a supporting
requirements addressed by the IS required domain. COBIT evaluation platform will be needed so as to automate the activities and tasks to
has shown a high coverage score for the requirements addressed by the achieve the elaborated methods. Moreover, it would facilitate the uni-
CM, the IS, and the RM required domains; and a moderate coverage fication process to become more efficient. Finally, the proposed CC
score for the requirements addressed by the SLA required domain. The governance framework will be tested and validated through multiple
evaluation of ISO/IEC 27001/2 has demonstrated a high coverage score industrial case studies. The validation will focus on implementing the
for the requirements addressed by the IS required domain, an accep- elaborated methods in real cases and demonstrate the added value of
table coverage score for the requirements addressed by the CM required unifying IT models.
domain, and a low coverage score for the requirements addressed by We are aware that our work may have some limitations. The first
the RM and the SLA required domains. concerns the granularity of the measurement scale we have used to
The integration method has been based on the evaluation results determine the process elements coverage score for the RAs. As we have
and has proposed an approach for integrating the mapped process already stated, we couldn't perform the measurement on more granular
elements. The post-evaluation phase of the integrated models has scale than the one we have used. The second limitation is concerned
shown a remarkable increase in the coverage score with regard to the with the objectives significance level when performing the mapping
max coverage score obtained when the models are used separately. between the CPEs and the RCs. In fact, all the objectives have been
Based on the research paper results, future works will focus on de- considered at the same significance level and importance in the com-
veloping the homogenization method, elaborating a supporting tool for parison process. Unfortunately, it was not possible to consider such an
assisting the unification process, and validating the proposed frame- aspect in our approach. The third limitation is linked to the partial
work. First, the development of the homogenization method will focus matching between the requirements when performing the mapping
on resolving the heterogeneities that can exist among the models when between the PEBMs and the RAs. For example, the sub-requirement 17
describing their process elements. This would be done through pro- of the process element “BAI04.03″ (see Table A3 of the appendix)
viding an independent representation of the models concepts and re- partially matches the sub-requirement “The negotiation process should
lationships. This representation, which is in the form of a meta-model, be extended with mechanisms for dynamic negotiation…” of the RA
will be used to rearrange the models’ content into a common process “SLA_SNE_2″. Unluckily, our approach was unable to consider such a
structure in order to solve their multiple structural differences in the partial matching between the requirements. We think that these lim-
first place; and also, to reduce the complexity when dealing with a itations could be a motivation for developing a more complete ap-
multi-model environment. Second, given the large number of elements proach in our future work.

Appendix

Table A1.
List of primary selected studies.
Reference Authors, Title Channel Year

[29] Cloud Security Alliance (CSA), Cloud Controls Matrix Version 3.0.1 CSA 2016
[63] K. Lu, R. Yahyapour, P. Wieder, E. Yaqub, Fault-tolerant Service Level Agreement lifecycle management in clouds using actor system FGCS 2016
[4] S. Karkošková, G. Feuerlicht, Cloud Computing Governance Reference Model BIR 2016
[64] F. Faniyi, R. Bahsoon, A Systematic Review of Service Level Management in the Cloud CSUR 2015
[12] A.S. Saidah, N. Abdelbaki, A new cloud computing governance framework CLOSER 2014
[65] FedRAMP, FedRAMP Baseline Security Controls FedRAMP 2014
[9] J. Becker, E. Bailey, A Comparison of IT Governance & Control Frameworks in Cloud Computing AMCIS 2014
[21] K. Brandis, S. Dzombeta, K. Haufe, Towards a framework for governance architecture management in cloud environments: A FGCS 2014
semantic perspective
[66] T. Labidi, A. Mtibaa, F. Gargouri, Ontology-Based Context-Aware SLA Management for Cloud Computing MEDI 2014
[67] W.A. Ghumman, Automation of the SLA life cycle in cloud computing ICSOC 2014
[68] A. Hammadi, O.K. Hussain, T. Dillon, F.K. Hussain, A framework for SLA management in cloud computing for informed decision Cluster Computing 2013
making
[69] K. Sun, Y. Li, Effort estimation in cloud migration process SOSE 2013
[70] P. Jamshidi, A. Ahmad, C. Pahl, Cloud Migration Research: A Systematic Review IEEE TCC 2013
[71] A.U. Khan, M. Oriol, M. Kiran, M. Jiang, K. Djemame, Security risks and their management in cloud computing CloudCom 2012
[34] C.H. LLP, W. Chan, E. Leung, H. Pili, Enterprise Risk Management for Cloud Computing COSO 2012
[30] ENISA, Cloud Computing: Benefits, Risks and Recommendations for Information Security ENISA 2012
[72] F. Xie, Y. Peng, W. Zhao, D. Chen, X. Wang, X. Huo, A risk management framework for cloud computing CCIS 2012
[73] G. Nie, X. E., D. Chen, Research on Service Level Agreement in Cloud Computing AEE 2012
[74] J. Ding, Z. Zhao, Towards autonomic SLA management: A review ICSAI 2012
[75] N.V. Pavan Kumar Illa, Sushma Nathani, Enterprise Cloud Migration: A Phase Driven Step-by-Step-Strategy intended for Scalability, IJCA 2012
Elasticity, Agility and Reliability
[76] P. V. Beserra, A. Camara, R. Ximenes, A.B. Albuquerque, N.C. Mendonca, Cloudstep: A step-by-step decision process to support legacy MESOCA 2012
application migration to the cloud
[8] S. Mehfuz, G. Sahoo, A five-phased approach for the cloud migration IJETAE 2012
[77] T.-C. Kao, C.-H. Mao, C.-Y. Chang, K.-C. Chang, Cloud SSDLC: Cloud Security Governance Deployment Framework in Secure System TrustCom 2012
Development Life Cycle
[13] ISACA, IT Control Objectives for Cloud Computing: Controls and Assurance in the Cloud ISACA 2011
[78] M. Cochran, P. Witman, Governance and service level agreement issues in a cloud computing environment JITM 2011
[20] R. Ahmad, L. Janczewski, Governance Life Cycle Framework for Managing Security in Public Cloud: From User Perspective CLOUD 2011
[79] W. Jansen, T. Grance, Guidelines on security and privacy in public cloud computing NIST 2011
[80] D. Catteddu, Cloud Computing: Benefits, Risks and Recommendations for Information Security IBIWAS 2010
[81] J.O. Fitó, M. Macias, J. Guitart, Toward business-driven risk management for Cloud computing CNSM 2010
[82] M. Alhamad, T. Dillon, E. Chang, SLA-Based Trust Model for Cloud Computing NBiS 2010
[83] X. Zhang, N. Wuwong, H. Li, X. Zhang, Information Security Risk Management Framework for the Cloud Computing Environments CIT 2010
[15] Z. Guo, M. Song, A Governance Model for Cloud Computing MASS 2010

112
Table A2.
Identified CC governance requirements for the SLA required domain.
Required control Required activity
Y. Bounagui et al.

Name of the Objectives of the required control Name of the required Supporting studies Cloud service user requirements Cloud service provider requirements
required control (ID) activity (ID)

SLA initiation (1) Model service level agreements.(2) Define SLA definition and [7,63,66,67,73,74,82,84] Procedure for the definition and modeling of SLAs Cloud Service Providers (CSPs) should analyze
(SLA_SI) service level agreements attributes.(3) Define modeling (SLA_SI_1) should be established and should consider: their business objectives through a business
and Design service attributes. 1. SLA modeling (i.e. QoS model, pricing model) modeling process to optimize their offerings. Based
2. Service level objectives definition (SLOs) on this model, an SLA template in terms of
3. SLA metrics design attributes values should be established.
4. Service design (i.e. dependencies between service
components, elasticity rules, performance estimation,
etc.)
SLA Negotiation and (1) Identify the relevant Cloud service providers. CSP discovery [7,63,67,73,74,82,84] The Cloud Service User (CSU) should locate the CSPs The established SLA template should be published
Improvement (2) Agree about the level of service required. (3) (SLA_SNE_1) that effectively satisfy its business requirements. The by the CSP within a unified registration system e.g.
(SLA_SNE) Formalize and document service level CSU should consider: 1. Procedures for the discovery UDDI to serve for the discovery process. The
agreements. (4) Monitor, measure, report, and and selection of CSPs. These procedure should be template corresponding to the selected CSP will
review the level of services provided. established based on the identified SLOs. 2. Additional serve thereafter for the negotiation and
non-functional requirements as security, scalability, establishment of the SLA.
dynamic changes, heterogeneity, etc.
SLA negotiation and [7,63,64,66,67,73,74,82,84] The CSU and CSP should define the terms of service and agree about the level at which a service will be
establishment provided. The negotiation process should be based on the service level requirements and should be extended
(SLA_SNE_2) with mechanisms for dynamic negotiation and re-negotiation between the CSU and CSP. In addition to the CSU
and the CSP, the negotiation process should also take into account different actors related to Cloud computing,
such as Cloud auditors, Cloud brokers or Cloud carriers. Finally, the SLA document should be created based on
the negotiation strategies and signed by the CSU and CSP. The SLA document should cover CSU and CSP
commitments, synopsis of the services, metrics for service measurement, financial charges, change procedures,

113
security and privacy aspects, level of service, roles and responsibilities, and compensation/penalties for non-
achievement. The SLA should consider additional items such as continuity and disaster recovery, availability,
capacity management, target performance, operating level agreements, system redundancy, maintenance and
support, location of data, seizure of data, and failure of the CSP, etc.
SLA monitoring and [7,63,64,66,67,73,74,82,84] The CSU should continually monitor SLA metrics at The CSP should continually monitor and report
reporting (SLA_SNE_3) different dimension to determine whether SLOs are SLA metrics at different dimension to determine
achieved or violated by the CSP. Monitoring activities whether SLOs are achieved or violated. The CSP
should incorporate mechanisms for automatic detection should provide capabilities for intervening
and reporting of SLA violations. Reporting mechanisms proactively in order to take necessary measures
should also consider changes in the provision of (e.g. providing more virtual machines, provide
services by the CSP. secure end-to-end communications, etc.) that
would avert imminent SLA violations.
SLA review and [63,64,67,73,74,82,84] The CSU should assess and review the reports generated The CSP should provide capabilities for enforcing
enforcement not reaching the agreed SLOs and SLAs. The CSU should SLAs based on CSU enforcement requirements. The
(SLA_SNE_4) also assess the corresponding compensation and CSP should also provide mechanisms for dynamic
penalties for all the SLA violations committed by the SLA enforcement (e.g. ability to recover from an
CSP. The performed reviews and assessments should be SLA violation during the execution).
used to further enforce the SLA. SLA enforcement
activities should also involves taking the necessary
actions to ensure no further SLA violations.
SLA Termination (1) Define capabilities for terminating service SLA conclusion [7,64,67,73,74,84] The CSU should define procedures for terminating the The CSP should use accounting and billing
(SLA_ST) level agreements. (2) Request compensations (SLA_ST_1) SLA when the service delivery is concluded, the SLA mechanisms to provide details on services use to
and penalties for any violations of service level validity period is over or the agreements are violated by the CSU. For any violation, the corresponding
agreements. any of the parties as specified in the SLA. compensations/penalties should be calculated and
Compensations/penalties should be requested by the resolved as specified in the SLA.
CSU for the violations committed by the CSP.
Computer Standards & Interfaces 62 (2019) 98–118
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Table A3.
Mapping of “APO09.03″, “APO09.04″, “BAI04.03″, and “BAI06.01″of COBIT to the RA “SLA_SNE_2″.
Description of the PEBMs “ APO09.03″, “ APO09.04″, “ BAI04.03″, and “ BAI06.01″

APO09.03: Define and prepare service agreements


Activities:
1. Analyse requirements for new or changed service agreements received from business relationship management to ensure that the requirements can be matched1. Consider aspects
such as service times2, availability3, performance4, capacity5, security6, continuity7, compliance8 and regulatory issues9, usability, and demand constraints10.
2. Draft customer service agreements based on the services, service packages and service level options in the relevant service catalogues11.
3. Determine, agree on and document internal operational agreements to underpin the customer service agreements, if applicable12.
4. Liaise with supplier management to ensure that appropriate commercial contracts with external service providers underpin the customer service agreements, if applicable13.
5. Finalise customer service agreements with business relationship management14.
APO09.04: Monitor and report service levels.
Activities:
1. Establish and maintain measures to monitor and collect service level data15.
5. Agree on action plans and remediations for any performance issues or negative trends16.
BAI04.03: Plan for new or changed service requirements
Activities:
4. Adjust the performance and capacity plans and SLAs based on realistic, new, proposed and/or projected business processes and supporting services, applications and infrastructure
changes as well as reviews of actual performance and capacity usage, including workload levels17.
BAI06.01: Evaluate, prioritise and authorize change requests
Activities:
7. Consider the impact of contracted services providers (e.g., of outsourced business processing, infrastructure, application development and shared services) on the change
management process, including integration of organisational change management processes with change management processes of service providers and the impact on contractual
terms and SLAs18.

Description of RA “SLA_SNE_2″ Mapping


APO09.03 APO09.04 BAI04.03 BAI06.01

The CSU and CSP should define the terms of service and agree about the level at which a service will be provided. 1, 13, 14- – –
The negotiation process should be based on service level requirements. 1, 13, 14- 17 –
The negotiation process should be extended with mechanisms for dynamic negotiation and re-negotiation between – 17 –
the CSU and CSP.
The negotiation process should also take into account different actors related to Cloud computing, such as Cloud – – – –
auditors, Cloud brokers or Cloud carriers.
The SLA should Cover: CSU and CSP commitments 11 – – –
Synopsis of the services 2, 10 – – –
Metrics for service measurement – 15 – –
Financial charges – – – –
Change procedures – – – 18
Security and privacy aspects 6 – – –
Level of service – – – –
Roles and responsibilities – – – –
Compensation/penalties for non-achievement 8 16 – –
The SLA should consider additional items such as: Continuity and disaster recovery 7 – – –
Availability 3 – – –
Capacity management 5 – – –
Target performance 4 – – –
Operating level agreements 12 – – –
System redundancy – – – –
Maintenance and support – – – –
Data location 9 – – –
Seizure of data and failure of CSP – – – –

Table A4.
Coverage score of the RAs for the CM required domains.
ID of RA ITIL Coverage Score COBIT Coverage Score ISO 27001/2 Coverage Score

CM_PM_1 100 100 20


CM_PM_2 80 100 40
CM_PM_3 80 80 0
CM_PM_4 80 80 40
CM_PM_5 40 40 20
CM_PM_6 100 100 0
CM_PM_7 80 80 60
CM_MEX_1 100 60 80
CM_MEX_2 60 60 40
CM_MEX_3 60 80 20
CM_MEX_4 100 80 20
CM_MEX_5 80 40 80
CM_MEV_1 100 60 60
CM_MEV_2 100 100 60
CM_MR_1 80 100 60

114
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Table A5.
Coverage score of the RAs for the IS required domains.
ID of RA ITIL COBIT ISO 27001/2

IS_ISP_1 60 80 80
IS_OIS_1 60 60 80
IS_OIS_2 0 60 100
IS_HRS_1 0 100 100
IS_HRS_2 40 60 100
IS_HRS_3 60 100 100
IS_AM_1 20 60 80
IS_AM_2 40 80 100
IS_AM_3 60 60 100
IS_AC_1 40 100 80
IS_AC_2 80 100 80
IS_AC_3 80 100 100
IS_AC_4 60 60 100
IS_CR_1 0 80 100
IS_PES_1 40 80 80
IS_PES_2 60 80 100
IS_OS_1 80 80 100
IS_OS_2 0 60 100
IS_OS_3 40 100 80
IS_OS_4 20 80 80
IS_OS_5 20 80 100
IS_OS_6 20 60 100
IS_OS_7 20 100 60
IS_CO_1 60 60 80
IS_CO_2 60 60 100
IS_SADM_1 40 60 60
IS_SADM_2 100 80 100
IS_SADM_3 80 100 100
IS_SR_1 60 100 80
IS_SR_2 80 100 100
IS_ISIM_1 60 80 100
IS_SBCM_1 20 100 100
IS_SBCM_2 20 100 100
IS_CM_1 20 80 60
IS_CM_2 20 100 80

Table A6.
Coverage score of the RAs for the RM required domains.
ID of RA ITIL COBIT ISO 27001/2

RM_RMP_1 60 100 20
RM_RMP_2 60 100 40
RM_RMI_1 20 80 20
RM_RMI_2 80 80 80
RM_RMI_3 60 100 0
RM_RMR_1 40 100 0
RM_RMR_2 40 100 80

Table A7.
Integration of the mapped process elements for the SLA required domain.
ID of RA Reference practice Complement practice

SLA_SI_1L - ITIL: [SD 3.3, SD 3.4, SD 3.5, SD 3.6, SD 4.2.5.1, SD 4.2.5.9] - ITIL: [N/A]
- COBIT: [APO 09.01 APO 09.02] - COBIT: [N/A]
- ISO 27001/2: [N/A] - ISO 27001/2: [Clause 15.1.1]
SLA_SNE_1 - ITIL: [SD 4.2.5.1, SS 5.4] - ITIL: [N/A]
- COBIT: [APO 09.01, APO 09.02, DSS 01.02] - COBIT: [N/A]
- ISO 27001/2: [N/A] - ISO 27001/2: [N/A]
SLA_SNE_2 - ITIL: [SD 4.2.5.2, SD 4.2.5.5 - ITIL: [N/A]
- COBIT: [APO09.03, APO09.04, BAI04.03, BAI06.01] - COBIT: [N/A]
- ISO 27001/2: [N/A] - ISO 27001/2: [Clause 15.1.2, Clause 15.1.3]
SLA_SNE_3 - ITIL: [N/A] - ITIL: [SD 4.2.5.3, SD 4.2.5.6, SD 4.2.5.10]
- COBIT: [APO 09.04] - COBIT: [N/A]
- ISO 27001/2: [N/A] - ISO 27001/2: [Clause 15.1.2, Clause 15.1.3, Clause 15.2.1, Clause 15.2.2]
SLA_SNE_4 - ITIL: [N/A] - ITIL: [SD 4.2.5.4, SD 4.2.5.5, SD 4.2.5.7, SD 4.2.5.8]
- COBIT: [APO 09.05] - COBIT: [N/A]
- ISO 27001/2: [N/A] - ISO 27001/2: [Clause 15.1.2, Clause 15.1.3]
SLA_ST_1 - ITIL: [SD 4.2.5.4, SD 4.2.5.5, SD 4.2.5.8] - ITIL: [N/A]
- COBIT: [APO 09.04] - COBIT: [N/A]
- ISO 27001/2: [N/A] - ISO 27001/2: [Clause 8.1.4]

115
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

Fig. A1. CC Governance requirements architecture.

References news/cloud-computing-service-level-agreements-exploitation-research-results ac-


cessed July 5, 2017.
[8] S. Mehfuz, G. Sahoo, A five-phased approach for the cloud migration, Int. J.
[1] P.M. Mell, T. Grance, The NIST Definition of Cloud Computing, National Institute of Emerging Technol. Adv. Eng. 2 (2012) 286–291, https://doi.org/10.1007/978-3-
Standards & Technology, 2011, https://doi.org/10.6028/NIST.SP.800-145. 642-13645-0.
[2] M. Bayramusta, V.A. Nasir, A fad or future of IT?: A comprehensive literature re- [9] J. Becker, E. Bailey, A comparison of IT governance & control frameworks in cloud
view on the cloud computing research, Int. J. Inf. Manage. 36 (2016) 635–644, computing, Assoc. Inf. Syst. Conf. (2014) 1–16 http://aisel.aisnet.org/cgi/
https://doi.org/10.1016/j.ijinfomgt.2016.04.006. viewcontent.cgi?article=1160&context=amcis2014 accessed July 5, 2017.
[3] D. Chou, Cloud computing: a value creation model, Comput. Stand. Interfaces [10] Y. Bounagui, H. Hafiddi, A. Mezrioui, Requirements definition for a holistic ap-
(2015), http://www.sciencedirect.com/science/article/pii/S0920548914000981 proach of cloud computing governance, Proc. IEEE/ACS Int. Conf. Comput. Syst.
accessed July 5, 2017. Appl. AICCSA, IEEE, 2016, pp. 1–8, , https://doi.org/10.1109/AICCSA.2015.
[4] S. Karkošková, G. Feuerlicht, Cloud Computing Governance Reference Model, 7507245.
Springer, Cham, 2016, pp. 193–203, https://doi.org/10.1007/978-3-319-45321- [11] S.O. Owuonda, D. Orwa, Cloud computing governance readiness assessment : case
7_14. study of a local Airline Company, Int. J. Appl. Inf. Syst. 10 (2016) 33–42, https://
[5] M. Almorsy, J. Grundy, A.S. Ibrahim, Collaboration-based cloud computing security doi.org/10.5120/ijais2016451543.
management framework, Proc. - 2011 IEEE 4th Int. Conf. Cloud Comput. CLOUD, [12] A.S. Saidah, N. Abdelbaki, A new cloud computing governance framework, CLOSER
2011 IEEE, 2011, pp. 364–371, , https://doi.org/10.1109/CLOUD.2011.9. 2014, Proc. 4th Int. Conf. Cloud Comput. Serv. Sci, 2014, pp. 671–678 https://files.
[6] Cloud Security Alliance, Top threats to cloud computing, Security (2010) 1–14 ifi.uzh.ch/stiller/CLOSER 2014/CLOSER/CLOSER/FedCloudGov/Full Papers/
https://www.google.com/search?q=Top+threats+to+cloud+computing+v1. FedCloudGov_2014_2_CR.pdf (accessed July 5, 2017).
+0&ie=utf-8&oe=utf-8&client=firefox-b-ab accessed November 27, 2017. [13] IT ISACAControl Objectives for Cloud Computing, Controls and assurance in the
[7] D. Kyriazis, Cloud computing service level agreements - exploitation of research cloud, ISACA (2011) 1–23 http://dl.acm.org/citation.cfm?id=2829203 accessed
results, Eur. Commun. 61 (2013), https://ec.europa.eu/digital-single-market/en/ July 5, 2017.

116
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

[14] J. Morin, J. Aubert, B. Gateau, Towards cloud computing SLA risk management: [40] ISACA, CLOUD GOVERNANCE, (2013) https://www.google.com/search?q=Cloud
issues and challenges, 2012 45th Hawaii Int. Conf. Syst. Sci. 2012, pp. 5509–5514, , +Security+Standards%3AWhat+to+Expect+%26+What+to
https://doi.org/10.1109/HICSS.2012.602. +NegotiateVersion+2.0&ie=utf-8&oe=utf-8&client=firefox-b-ab accessed
[15] Z. Guo, M. Song, A governance model for cloud computing, 2010 Int. Conf. Manag. November 23, 2017.
Serv. Sci, IEEE, 2010, pp. 1–6, , https://doi.org/10.1109/ICMSS.2010.5576281. [41] AXELOS, IT Service Management and Cloud Computing, (2014) https://www.
[16] IT Governance Institute, COBIT 4 ISA, 2007, p. 1 http://dl.acm.org/citation.cfm? google.com/search?client=firefox-b-ab&ei=lgMXWvjaCdDDkwXp0Y_oDw&q=
id=1534415 accessed July 5, 2017. IT+service+management+and+cloud+computing&oq=IT+service
[17] Y. He, The lifecycle model for cloud, https://www.google.com/search?q=THE +management+and+cloud+computing&gs_l=psy-ab.12..0i19k1j0i22i30i19k1.
+LIFECYCLE+MODEL+FOR+CLOUD+GOVERNANCE&ie=utf-8&oe=utf-8& 309891.309891.0.310872.1.1.0.0.0.0.239.239.2-1.1.0 accessed November 23,
client=firefox-b-ab, (2011) accessed October 30, 2017. 2017.
[18] R. Dautov, S. Veloudis, I. Paraskakis, S. Distefano, Policy Management and [42] J. Becker, E. Bailey, IT controls and governance in cloud computing, Proc. Twent,
Enforcement Using OWL and SWRL for the Internet of Things, Policy Management 2014 https://pdfs.semanticscholar.org/b802/
and Enforcement Using OWL and SWRL for the Internet of Things, Springer, Cham, 6fa4e58bd193fdfc0e6d37bd2ebe2c6722a2.pdf accessed July 5, 2017.
2017, pp. 342–355, https://doi.org/10.1007/978-3-319-67910-5_28. [43] M. Al Mourad, M. Hussain, The impact of cloud computing on ITIL service strategy
[19] M.H. Hugos, D. Hulitzky, Business in the Cloud : What Every Business Needs to processes, Int. J. Comput. (2014), http://search.proquest.com/openview/
Know About Cloud Computing, Wiley, 2011, https://books.google.co.ma/books? 14fe77b57588ee16024bcaf0ca749616/1?pq-origsite=gscholar&cbl=2027424 ac-
id=Qn9OA8JVmscC&pg=PA2&lpg=PA2&dq=adapt+easily+to+change cessed July 5, 2017.
+cloud&source=bl&ots=0EQJJ9Btxj&sig=dWXLHbh_lV61J0nhsE01iKnOu7A& [44] M. Nieves, Best Practice in the Cloud: An Introduction, (2014) Retrieved http://
hl=fr&sa=X&ved=0ahUKEwjPx9CzgpvXAhXD6xQKHeNuAusQ6AEIRTAI#v= ibpi.vanharen.net/Player/eKnowledge/best_practice_in_the_cloud_an_introduction.
onepage&q&f=false accessed October 31, 2017. pdf accessed July 5, 2017.
[20] R. Ahmad, L. Janczewski, Governance life cycle framework for managing security in [45] ISACA (Serving IT Governance Professionals), Security, Audit, and Control
public cloud: from user perspective, 2011 IEEE 4th Int. Conf. Cloud Comput. IEEE, Features : Oracle Database, ISACA, 2009.
2011, pp. 372–379, , https://doi.org/10.1109/CLOUD.2011.117. [46] R. Sheikhpour, N. Modiri, An approach to map COBIT processes to ISO/IEC 27001
[21] K. Brandis, S. Dzombeta, K. Haufe, Towards a framework for governance archi- information security management controls, Int. J. Secur. Appl. (2012), https://
tecture management in cloud environments: A semantic perspective, Future Gener. www.researchgate.net/profile/Razieh_Sheikhpour2/publication/292833500_An_
Comput. Syst. 32 (2014) 274–281, https://doi.org/10.1016/j.future.2013.09.022. approach_to_map_COBIT_processes_to_ISOIEC_27001_information_security_
[22] S. Sahibudin, M. Sharifi, M. Ayat, Combining ITIL, COBIT and ISO/IEC 27002 in management_controls/links/582eaf6108ae102f072e6dbd/An-approach-to-map-
order to design a comprehensive IT framework in organizations, 2008 Second Asia COBIT-processes-to-ISO-IEC-27001-in accessed July 5, 2017.
Int. Conf. Model. Simul. IEEE, 2008, pp. 749–753, , https://doi.org/10.1109/AMS. [47] ISACA, CLOUD GOVERNANCE, Questions Boards of Directors Need to Ask, (2013)
2008.145. https://www.isaca.org/Knowledge-Center/Research/ResearchDeliverables/Pages/
[23] Office of Government Commerce, ITIL v3: Service Strategy - Service Design - Service Cloud-Governance-Questions-Boards-of-Directors-Need-to-Ask.aspx accessed
Transition - Service Operation - Continual Service Improvement, (2007) UK https:// October 31, 2017.
www.itil.org.uk/all.htm accessed July 6, 2017. [48] O. Illoh, S. Aghili, S. Butakov, Using COBIT 5 for Risk to Develop Cloud Computing
[24] Isaca, Cobit 5, ISA, 2012, https://dl.acm.org/citation.cfm?id=2361913 accessed SLA Evaluation Templates, Springer, Cham, 2015, pp. 236–249, https://doi.org/10.
March 5, 2018. 1007/978-3-319-22885-3_21.
[25] ISO/IEC, ISO/IEC 27001:2013 – Information Technology – Information Security [49] S. Zhang, F. Le, An Examination of the Practicability of COBIT Framework and the
Management Systems – Requirements 23 (2013) https://www.iso.org/standard/ Proposal of a COBIT-BSC Model, J. Econ. (2013), https://openaccess.leidenuniv.nl/
54534.html accessed July 5, 2017. handle/1887/24618 accessed July 5, 2017.
[26] ISO/IEC, ISO/IEC 27002 - Information Technology - Security Techniques - Code of [50] G. Ridley, J. Young, P. Carroll, COBIT and its utilization: a framework from the
Practice for Information Security Controls 80 (2013) https://www.iso.org/ literature, Syst. Sci. (2004) 2004 http://ieeexplore.ieee.org/abstract/document/
standard/54533.html accessed July 5, 2017. 1265566/ accessed July 5, 2017.
[27] Z. Enslin, Cloud Computing : COBIT-Mapped Benefits, Risks and Controls for [51] ISO/IEC, ISO/IEC 17788:2014 - Information Technology – Cloud Computing –
Consumer Enterprises, Thesis Present. Partial Fulfilment Requir. Degree Masters Overview and Vocabulary, 1st ed., (2014) https://www.iso.org/standard/60544.
Commer (Computer Audit. Stellenbosch Univ.), 2012. html accessed July 5, 2017.
[28] CSA, Security Guidance for Critical Areas of Focus in Cloud Computing v3.0, [52] ISO/IEC, ISO/IEC 17789:2014 - Information Technology – Cloud Computing –
(2011). Reference Architecture, 1st ed., (2014) https://www.iso.org/standard/60545.html
[29] Cloud Security Alliance, Cloud Controls Matrix Version 3.0.1 (2016), https:// accessed July 52017.
cloudsecurityalliance.org/group/cloud-controls-matrix/#_overview accessed [53] ISO/IEC, ISO/IEC 27017:2015 - Information Technology – Security Techniques –
November 5, 2017. Code of Practice for Information Security Controls Based on ISO/IEC 27002 for
[30] ENISA, Cloud Computing: Benefits, Risks and Recommendations for Information Cloud Services, 1st ed., (2015) https://www.iso.org/standard/43757.html accessed
Security, (2012) https://resilience.enisa.europa.eu/cloud-security-and-resilience/ July 5, 2017.
publications/cloud-computing-benefits-risks-and-recommendations-for- [54] W. Fumy, M. De Soete, E.J. Humphreys, T. Chikazawa, J. Amsenga, K. Rannenberg,
information-security accessed February 7, 2018. ISO/IEC 27018:2014 - Information Technology – Security Techniques – Code of
[31] D. Catteddu, G. Hogben, Cloud Computing, Information Assurance Framework, Practice for Protection of Personally Identifiable Information (PII) in Public Clouds
(2009) https://www.enisa.europa.eu/publications/cloud-computing-information- Acting as PII Processors 8 (2014), p. 23 https://www.iso.org/standard/61498.html
assurance-framework/ accessed March 3, 2018. accessed July 5, 2017.
[32] D. Catteddu, Security & Resilience in Governmental Clouds–Making an Informed [55] L.M. Jeannine Siviy, P. Kirwan, J. Morley, Maximizing your Process Improvement
Decision, Security & Resilience in Governmental Clouds–Making an Informed ROI through Harmonization, (2008) http://resources.sei.cmu.edu/library/asset-
Decision 146 ENISA, 2011 ENISA ReportJanuary 2011 https://www.enisa.europa. view.cfm?assetid=28907 accessed July 5, 2017.
eu/publications/security-and-resilience-in-governmental-clouds accessed March 3, [56] C. Pardo, F.J. Pino, F. García, M. Piattini Velthius, M.T. Baldassarre, Trends in
2018. Harmonization of Multiple Reference Models, Springer, Berlin, Heidelberg, 2011,
[33] ITU-T, Focus Group on Cloud Computing Technical Report, (2012) http://wwwa. pp. 61–73, https://doi.org/10.1007/978-3-642-23391-3_5.
itu.int/pub/T-FG-CLOUD-2012-P5/fr accessed March 3, 2018. [57] C. Pardo, F.J. Pino, F. García, M. Piattini, M.T. Baldassarre, An ontology for the
[34] C.H. LLP, W. Chan, E. Leung, H. Pili, Enterprise Risk Management for Cloud harmonization of multiple standards and models, Comput. Stand. Interfaces 34
Computing, (2012) https://www.coso.org/Documents/Cloud-Computing-Thought- (2012) 48–59, https://doi.org/10.1016/j.csi.2011.05.005.
Paper.pdf accessed February 25, 2018. [58] C. Pardo, F.J. Pino, F. Garcia, M.T. Baldassarre, M. Piattini, From chaos to the
[35] R.R. Moeller, COSO Enterprise Risk Management: Establishing Effective systematic harmonization of multiple reference models: a harmonization frame-
Governance, Risk, and Compliance Processes, (2011), https://doi.org/10.1002/ work applied in two case studies, J. Syst. Software 86 (2013) 125–143, https://doi.
9781118269145. org/10.1016/j.jss.2012.07.072.
[36] G.A. Lewis, Role of standards in cloud-computing interoperability, 2013 46th [59] Y. Bounagui, H. Hafiddi, A. Mezrioui, COBIT evaluation as a framework for cloud
Hawaii Int. Conf. Syst. Sci, IEEE, 2013, pp. 1652–1661, , https://doi.org/10.1109/ computing governance, Int. J. Cloud (2016), http://www.igi-global.com/article/
HICSS.2013.470. cobit-evaluation-as-a-framework-for-cloud-computing-governance/173772 ac-
[37] ITGI/OGC, Aligning COBIT 4.1, ITIL V3 and ISO/IEC 27002 for Business Benefit, cessed July 5, 2017).</.
(2008) http://www.isaca.org/Knowledge-Center/Research/ResearchDeliverables/ [60] P. Brereton, B.A. Kitchenham, D. Budgen, M. Turner, M. Khalil, Lessons from ap-
Pages/Aligning-COBIT-4-1-ITIL-V3-and-ISO-IEC-27002-for-BusinessBenefit.aspx plying the systematic literature review process within the software engineering
accessed July 5, 2017. domain, J. Syst. Software 80 (2007) 571–583, https://doi.org/10.1016/j.jss.2006.
[38] Y. Ozdemir, H. Basligil, P. Alcan, B.M. Kandemirli, Evaluation and comparison of 07.009.
COBIT, ITIL and ISO27K1/2 standards within the, Int. J. Tech. Res. Appl. 11 (2014) [61] G. Hardy, J. Hesch, Aligning CobiT® 4.1, ITIL® V3 and ISO/IEC 27002 for Business
22–24 https://scholar.googleusercontent.com/scholar.bib?q=info:rlnm- Benefit, IT Gov. Inst., 2008, pp. 1–130 www.isaca.org.
XRs3B0J:scholar.google.com/&output=citation&scisig= [62] S. Bradner, RFC 2119 - key words for use in RFCs to indicate requirement levels
AAGBfm0AAAAAWe9MdiGjYCra1fG_00naDrXRLFaX4LCX&scisf=4&ct=citation& status, Netw. Work. Gr. (1997), pp. 1–3 http://www.ietf.org/rfc/rfc2119.txt.
cd=-1&hl=fr&scfhb=1 accessed October 24, 2017. [63] K. Lu, R. Yahyapour, P. Wieder, E. Yaqub, M. Abdullah, B. Schloer, C. Kotsokalis,
[39] P. Năstase, F. Năstase, C. Ionescu, Challenges generated by the implementation of Fault-tolerant service level agreement lifecycle management in clouds using actor
the IT standards CobiT 4.1, ITIL v3 and ISO/IEC 27002 in enterprises, Econ. system, Future Gener. Comput. Syst. 54 (2016) 247–259, https://doi.org/10.1016/
Comput. (2009), http://www.ecocyb.ase.ro/articles 3.2009/Pavel Nastase.pdf j.future.2015.03.016.
(accessed July 5, 2017. [64] F. Faniyi, R. Bahsoon, A systematic review of service level management in the

117
Y. Bounagui et al. Computer Standards & Interfaces 62 (2019) 98–118

cloud, ACM Comput. Surv. 48 (2015) 1–27, https://doi.org/10.1145/2843890. +cloud+migration%3A+a+phase+driven+step-by-step-strategy+intended


[65] FedRAMP, FedRAMP Baseline Security Controls, (2014) https://www.google.com/ +for+scalability%2C+elasticity%2C+agility+and+reliability&ie=utf-8&oe=
search?client=firefox-b-ab&ei=mil6Wrr5J8KoUdPqi_AC&q=fedramp+baseline utf-8&client=firefox-b-ab accessed January 31.
+xls&oq=fedramp+baseline+xls&gs_l=psy-ab.3..33i160k1.8535.10767.0. [76] P.V. Beserra, A. Camara, R. Ximenes, A.B. Albuquerque, N.C. Mendonça, Cloudstep:
11015.4.4.0.0.0.0.141.485.0j4.4.0....0...1c.1.64.psy-ab..0.3.376... a step-by-step decision process to support legacy application migration to the cloud,
0i19k1j0i22i30i19k1j33i21k1.0 accessed February 6, 2018. 2012 IEEE 6th Int. Work. Maint. Evol. Serv. Cloud-Based Syst. MESOCA 2012, IEEE,
[66] T. Labidi, A. Mtibaa, F. Gargouri, Ontology-Based Context-Aware SLA Management 2012, pp. 7–16, , https://doi.org/10.1109/MESOCA.2012.6392602.
for Cloud Computing, Springer, Cham, 2014, pp. 193–208, https://doi.org/10. [77] T.-C. Kao, C.-H. Mao, C.-Y. Chang, K.-C. Chang, SSDLC Cloud, Cloud security gov-
1007/978-3-319-11587-0_19. ernance deployment framework in secure system development life cycle, 2012 IEEE
[67] W.A. Ghumman, Automation of the SLA Life Cycle in Cloud Computing, 11th Int. Conf. Trust. Secur. Priv. Comput. Commun. IEEE, 2012, pp. 1143–1148, ,
Automation of the SLA Life Cycle in Cloud Computing, (2014), pp. 557–562, https://doi.org/10.1109/TrustCom.2012.106.
https://doi.org/10.1007/978-3-319-06859-6_51. [78] M. Cochran, P. Witman, Governance and service level agreement issues in a cloud
[68] A. Hammadi, O.K. Hussain, T. Dillon, F.K. Hussain, A framework for SLA man- computing environment, J. Inf. Technol. Manage. 22 (2011) 41–55 https://www.
agement in cloud computing for informed decision making, Cluster Comput. 16 google.com/search?client=firefox-b-ab&ei=5VV_WrnFEsvsUrCJpfgF&q=
(2013) 961–977, https://doi.org/10.1007/s10586-012-0232-9. GOVERNANCE+AND+SERVICE+LEVEL+AGREEMENT+ISSUES+IN
[69] K. Sun, Y. Li, Effort estimation in cloud migration process, Proc. - 2013 IEEE 7th Int. +A+CLOUD+COMPUTING+ENVIRONMENT&oq=GOVERNANCE+AND
Symp. Serv. Syst. Eng. SOSE 2013, IEEE, 2013, pp. 84–91, , https://doi.org/10. +SERVICE+LEVEL+AGREEMENT+ISSUES+IN+A+CLOUD
1109/SOSE.2013.29. +COMPUTING+ENVIRONMENT&gs_l=psy-ab.3..0i3 accessed February 10, 2018.
[70] P. Jamshidi, A. Ahmad, C. Pahl, Cloud migration research: a systematic review, [79] W. Jansen, T. Grance, Guidelines on Security and Privacy in Public Cloud
IEEE Trans. Cloud Comput. 1 (2013) 142–157, https://doi.org/10.1109/TCC. Computing, (2011), https://doi.org/10.6028/NIST.SP.800-144 Gaithersburg, MD.
2013.10. [80] D. Catteddu, Cloud Computing, Benefits, Risks and Recommendations for
[71] A.U. Khan, M. Oriol, M. Kiran, M. Jiang, K. Djemame, Security risks and their Information Security, Springer, Berlin, Heidelberg, 2010, https://doi.org/10.1007/
management in cloud computing, 4th IEEE Int. Conf. Cloud Comput. Technol. Sci. 978-3-642-16120-9_9 pp. 17–17.
Proc, 2012, pp. 121–128, , https://doi.org/10.1109/CloudCom.2012.6427574. [81] J.O. Fitó, M. Macias, J. Guitart, Toward business-driven risk management for Cloud
[72] F. Xie, Y. Peng, W. Zhao, D. Chen, X. Wang, X. Huo, A risk management framework computing, Netw. Serv. Manag. (CNSM), 2010 Int. Conf. 2010, pp. 238–241, ,
for cloud computing, 2012 IEEE 2nd Int. Conf. Cloud Comput. Intell. Syst. IEEE, https://doi.org/10.1109/CNSM.2010.5691291.
2012, pp. 476–480, , https://doi.org/10.1109/CCIS.2012.6664451. [82] M. Alhamad, T. Dillon, E. Chang, SLA-based trust model for cloud computing, 2010
[73] G. Nie, X. E., D. Chen, Research on Service Level Agreement in Cloud Computing, 13th Int. Conf. Network-Based Inf. Syst. IEEE, 2010, pp. 321–324, , https://doi.org/
Springer, Berlin, Heidelberg, 2012, pp. 39–43, https://doi.org/10.1007/978-3-642- 10.1109/NBiS.2010.67.
28744-2_5. [83] X. Zhang, N. Wuwong, H. Li, X. Zhang, Information security risk management
[74] J. Ding, Z. Zhao, Towards autonomic SLA management: a review, 2012 Int. Conf. framework for the cloud computing environments, 2010 10th IEEE Int. Conf.
Syst. Informatics, IEEE, 2012, pp. 2552–2555, , https://doi.org/10.1109/ICSAI. Comput. Inf. Technol. IEEE, 2010, pp. 1328–1334, , https://doi.org/10.1109/CIT.
2012.6223574. 2010.501.
[75] N.V. Pavan Kumar Illa, Sushma Nathani, Enterprise cloud migration: a phase driven [84] L. Wu, R. Buyya, Service level agreement (SLA) in utility computing systems, ArXiv
step-by-step-strategy intended for scalability, elasticity, agility and reliability, Int. J. Prepr. ArXiv1010.2881. (2010) 27. doi:10.4018/978-1-60960-794-4.ch001.
Comput. Appl. 2 (n.d.) (2018) 6 https://www.google.com/search?q=Enterprise

118

You might also like