Issues and Challenges of Access Control in the Cloud

Francesca Lonetti and Eda Marchetti

ISTI-CNR, Via G. Moruzzi 1, Pisa, Italy

Keywords:

Access Control, Cloud, Testing.

Abstract:

Cloud computing offers scalable and efﬁcient information sharing and storage resources. However, the risk of

security breaches for personal and private data in this computing environment is very high. Access control is

among the most adopted means to assure that sensible information or resources are correctly accessed. This

paper provides an overview of main access control models in cloud infrastructures discussing most important

challenges. In particular, focusing on access control policy speciﬁcation, analysis, veriﬁcation and enforce-

ment, we also identify some emerging issues and point out some solutions and future research directions for

cloud computing.

1 INTRODUCTION

Nowadasys, the management of (personal) data beco-

mes extremely important in many distributed environ-

ments, especially when it involves data that are sub-

ject to different regulations depending on the context

in which they are stored or accessed.

In particular, the General Data Protection Regu-

lation

only allows the processing of such data on

the basis of speciﬁc measures or safeguards rules that

may vary depending on the European or Member

State laws.

Thus, mechanisms for accessing of personal data

should guarantee the sufﬁcient level of protection: i.e.

the processing is not legitimate without the consent of

the data subject, which must be expressed beforehand

through speciﬁc access control policies. Unfortuna-

tely, in the recent modern multi-cloud environment,

possibly distributed all over the world, problems may

also emerge with respect to the dynamic transmission

and adaptation of (personal) data when access control

policies deal with data belonging to different coun-

tries, which provide different degrees of data pro-

tection, or different local realities, which may have

own data format or speciﬁc data restrictions.

Just as real world common example: consider the

situation in which a seafarer EU citizen is frequently

moving around ports in different countries. In the

unlucky case in which he/she may have health pro-

EUR-Lex Access to European Union law:

http://eur-lex.europa.eu/legal-content/EN/LSU/

?uri=celex:31995L0046

blems, there could be serious difﬁculties in obtaining

the best medical treatment, because doctors visiting

him/her could not have the access to the patient’s me-

dical history or ongoing treatments. This problem

is currently mitigated by the enforcement of the Ge-

neral Data Protection Regulation, because it will set

a uniform legislation across all EU Member States,

but will still imply signiﬁcant issues when transfer-

ring data to non-EU countries. Access control adapta-

tion and interaction are just some of the multiple fac-

tors preventing continuous data storing and sharing in

cloud computing environments. However, conventio-

nal access control models suffer from the lack of ﬂex-

ibility and scalability in attribute management and are

not able to cope with the heterogeneity and variety of

cloud services.

In this paper, we provide an overview of the well-

known access control models, such as Mandatory

Access Control (MAC), Discretionary Access Control

(DAC), Role Based Access Control (RBAC), Attri-

bute Based Access Control (ABAC) (Ubale Swapnaja

et al., 2014), highlighting their limitations in cloud

computing. Then, we present some speciﬁc access

control proposals for cloud environments and review

the most important issues and challenges concerning

the distributed access of data through the multi-cloud

infrastructures.

Testing and veriﬁcation of access control policies

are key aspects to enhance the security level of cloud

environments. Traditional testing and analysis sys-

tems have to face many issues and challenges pre-

venting their adoption in cloud environments. In this

Lonetti, F. and Marchetti, E.

Issues and Challenges of Access Control in the Cloud.

DOI: 10.5220/0006948702610268

In Proceedings of the 14th International Conference on Web Information Systems and Technologies (WEBIST 2018), pages 261-268

ISBN: 978-989-758-324-7

261

speciﬁc context, we discuss in more detail the issues

and challenging concerning the policy speciﬁcation,

analysis, testing and enforcement of XACML-based

access control systems, trying to provide some possi-

ble solutions and future research directions.

The rest of this paper is structured as follows.

Section 2 brieﬂy introduces the main access control

models. Section 3 overviews the main issues and

challenges for access control systems in the cloud in-

frastructures. Section 4 focuses on possible soluti-

ons and future research directions in the context of

XACML-based access control policies speciﬁcation,

reﬁnement, analysis, testing and enforcement whe-

reas Section 5 draws conclusions.

2 CURRENT RESEARCH ON

ACCESS CONTROL IN THE

CLOUD

Cloud computing models greatly improve the ef-

ﬁciency of information sharing providing scalable

storage resources. Traditional access control mo-

dels that authorize the access to shared data have to

face new challenges to support cloud based applicati-

ons. We brieﬂy describe below the main conventional

access control models and their limitations in cloud

computing (Ausanka-Crues, 2001; Youniand Kifayat

and Merabti, 2014):

• Mandatory Access Control (MAC) model

(Ubale Swapnaja et al., 2014): a central authority

is in charge of giving access decisions to a subject

that requests access to a resource. This model is

very expensive and difﬁcult to deploy in cloud

computing since it does not support separation of

duties, delegation or inheritance aspects.

• Discretionary Access Control (DAC) model

(Ubale Swapnaja et al., 2014): this model grants

the owners of objects the ability to control access

to their objects, according to the identities of users

or their membership. DAC model is generally less

secure than mandatory access control model, so

it is used in environments that do not require a

high level of protection The DAC cannot be used

in cloud computing since there is no mechanism

or method to facilitate the management of impro-

per rights (e.g. risk awareness), which owners of

objects can give to users.

• Role Based Access Control (RBAC) model

(Ubale Swapnaja et al., 2014): it is the most used

in organizations and enterprises. The idea is to

assign rights according to the role or membership

that a subject can have. Deﬁning the right roles

that represent a system and dividing subjects into

categories based upon roles is not an easy task in

cloud computing.

• Attribute Based Access Control (ABAC) model:

it relies on a set of attributes of subjects, resources

and actions in order to allow or deny the access to

the resources. Security policies specify the attri-

butes that have to be considered for making a de-

cision. However, selecting what kind of attributes

should be used, and how many attributes are con-

sidered for making access decisions is a complex

task in cloud computing (Jin et al., 2012). XA-

CML (eXtensible Access Control Markup Lan-

guage) is a standard language that implements

attribute-based access control (OASIS, 2005). It

is a platform-independent XML-based language

for the speciﬁcation of access control policies.

Brieﬂy, an XACML policy has a tree structure

whose main elements are: PolicySet, Policy, Rule,

Target and Condition. The PolicySet includes one

or more policies. A Policy contains a Target and

one or more rules. The Target speciﬁes a set of

constraints on attributes of a given request. The

Rule speciﬁes a Target and a Condition containing

one or more boolean functions. If the Condition

is evaluated to true, then the Rule’s Effect (a value

of Permit or Deny) is returned, otherwise a No-

tApplicable decision is formulated (Indeterminate

is returned in case of errors). The PolicyCombi-

ningAlgorithm and the RuleCombiningAlgorithm

deﬁne how to combine the results from multiple

policies and rules respectively in order to derive a

single access result.

• Risk-Based Access Control (RBAC) model: this

model uses different kinds of risk levels asso-

ciated with environmental conditions for giving

access decisions (Suhendra, 2011). However, this

model is difﬁcult to be deployed in cloud compu-

ting because of the amount of analysis required

and number of systems to be merged to compute

risk levels.

• Usage based access control (UCON) model (Park

and Sandhu, 2002; Bertolino et al., 2014b): The

UCON model is an extension of the traditional

access control models that, besides authorizati-

ons, introduces new factors in the decision pro-

cess, namely obligations, conditions, and muta-

ble attributes. It deals with new requirements for

access control brought by the distributed environ-

ment. The main idea is that attributes related to

subjects and objects could change their values,

while the access is in progress in such a way that

the access right does not hold anymore. Then, the

WEBIST 2018 - 14th International Conference on Web Information Systems and Technologies

262

model allows to interrupt the access to preserve

the system security. For this reason, UCON po-

licies specify whether a decision factor must be

evaluated before and/or during the usage of the

resource (continuous policy enforcement).

We refer to (Cai et al., 2018) for a more detailed

description of these access control models and a per-

formance comparison among them.

Many proposals address access control problem in

cloud computing leveraging and enhancing existing

models (Youniand Kifayat and Merabti, 2014).

The paper in (Youniand Kifayat and Merabti,

2014) proposes an access control model able to en-

sure the secure sharing of resources among potential

untrusted tenants, and to support different access per-

mission to the same cloud user giving to the user the

ability to use multiple services securely.

Task-Role-Based Access Control scheme is anot-

her access control approach which has been proposed

for health care systems in the cloud computing envi-

ronment (). Permissions in this model are activated

or deactivated according to the current task or process

state. This scheme was implemented in the Amazon

Elastic Compute Cloud (Amazon EC2).

Tsai and Shao (Tsai and Shao, 2011) propose

a reference ontology framework using Role-Based

Access Control model allowing multiple roles for a

subject in different sessions as well as role hierarchy

based on domain ontology and transferred between

various ontology domains. This model has to ensure

granting access decisions in a reasonable time and ac-

cording to system requirements and is scalable with

respect to the number of roles, number of permissi-

ons, size of role hierarchy, and limits on tenant-role

assignments.

Another ontological model for access control po-

licies is presented in (Veloudis et al., 2016). It en-

hances the attribute based access control scheme by

capturing a wide range of relevant contextual attribu-

tes, and enabling the policy-related knowledge to be

extended and instantiated to suit the needs of any par-

ticular cloud application, independently of the code

employed by that application. These contextual attri-

butes can be associated at the level of the access con-

trol policy, indicating the contextual conditions that

must be satisﬁed by an entity in order for an access

request to be permitted (or denied)as well as at the le-

vel of the request itself, indicating the actual context

attached to an entity at the time of the request (Velou-

dis et al., 2017).

A recent proposal presents an Access Control

Model for Cloud Computing (AC3) (Youniand Ki-

fayat and Merabti, 2014) that is compliant with the

cloud access control requirements. This model is able

to support different access permissions to the same

cloud user for the different services. The proposed

model takes into account the role and task principles,

and classiﬁes users according to their actual jobs. Dy-

namic and random behaviours of users are controlled

by a risk engine that credits consumers according to

their access behaviours.

3 ISSUES AND CHALLENGES

Considering the problem of accessing distributed

(personal) data, collected though multi-cloud infra-

structures, the following main challenges could be

highlighted:

1. Continuous access: the distributed data need to

be accessible from the owners independently by

the organization or the country where the data are

collected. The data owners must be able to con-

tinuously access the data, or delegate the access

rights to a third party in case of speciﬁc conditi-

ons/problems. Technically, different users should

be able to specify policy rules and those are not af-

fected by when resources or subjects change their

temporal, spacial or status conditions.

2. Different regulations: Different countries in

which the access control entities are located may

have different regulations for accessing data. This

can cause security threats (Sharma et al., 2017).

3. Dynamic access policies: Conventional

access control models in cloud computing would

miss of ﬂexibility in attribute management and

scalability in dealing with a large number of

users, different classiﬁcation, high dynamic

performance, mobility features and changes in

high frequency (Wang et al., 2011). Recently,

the security community has recognized the

importance of dynamicity in security. Indeed,

the distributed management of (personal) data

sharing should be ruled by innovative, ﬂexible,

accessible, dynamically modiﬁable, and legally

compliant access control policies, which may let

the speciﬁcation of precise rules and environmen-

tal conditions under which an entity/person can

access to own data.

4. User-friendly management: The speciﬁcation

of access control policies is very important. Users

(administration, citizen, customers, etc.) should

be able to easily express (and possibly modify)

them at different level of detail avoiding the risk

of errors and without affecting the cloud service

provisioning, code deploying and general mana-

gement. The development of on-line user-centric,

Issues and Challenges of Access Control in the Cloud

263

user-friendly, adaptable and compliant solutions

(dashboard, APIs, interface) able to perform ve-

riﬁcation and testing activity should be integrated

into the Cloud context (Bertolino et al., 2014a).

5. Continuous control: Cloud computing can be

very complex and sophisticated due to the dyn-

amic nature of the clouds resources. Access to

(personal) data cannot be granted once and for all.

Multi-cloud application framework must be deve-

loped enabling secure and resilient cross-border

continuous control of data access so to promptly

take in place possible predictive and corrective

actions. Instead of waiting for the harmful conse-

quences, predictive security mechanisms and stra-

tegies must be developed or integrated to testing

and verifying possible policy conﬂicts and miti-

gate the risk and ﬂaws of data access conﬁgura-

tion (Calabr

o et al., 2018). Multi-cloud applica-

tion frameworks should therefore allow scalable,

context-aware, secure and resilient access control

systems, able to monitor, control and testing the

data sharing among several entities, through dif-

ferent channels, and by exploiting heterogeneous

means.

6. Scalability: Cloud-based access control sys-

tems have to deal with a large number of cloud

users having different access permissions. These

users must be able to use multiple services accor-

ding to authentication and login time (Wang et al.,

2011; Almutairi et al., 2012).

7. Resource sharing and interoperability:

To allow resources sharing among untrusted

users, access control models deployed in the

cloud have to support transfer of customer’

credentials across layers to access services and

resources (Almutairi et al., 2012). Heterogeneity

and variety of services as well as diversity of

access control policies and various access con-

trol interfaces can cause interoperability issues

(Tianyi et al., 2011).

8. Validation and verification of access

control policies: As access control policies

can be complex and error-prone, it is very

important to validate the implemented policies

(speciﬁed for instance in the standard XACML

notation) against the intended rights, which can

be formally deﬁned (by a model) or informally

expressed, e.g. in tabular form (Bertolino et al.,

2016). Moreover, also solutions aiming to verify

the access control rules against some deﬁned

access control properties have to be put in place

(Hwang et al., 2010).

9. On line tracing of access control

polices execution Continuous tracing of

policy execution allows to detect inconsistencies

in the policy speciﬁcation and provides support

for policies updates if new events occur. Moreo-

ver, it can help to predict security threats due to

policy modiﬁcations or combinations. It gives the

ability to verify correct policy deployment and

activation compliance (Lonetti and Marchetti,

2018).

10. Testing of access control systems

Access control testing is crucial in cloud ba-

sed access control models, since testing facilities

enhance the level of security needed in such envi-

ronments. Policy and rule combining algorithms

are deﬁned in many access control languages.

Testing the right behavior of these algorithms

is important when different policies and rules

are combined in cloud environments and assures

that there are no leaked privileges because of

the syntactic or semantic errors (Bertolino et al.,

2014c).

4 POLICIES SPECIFICATION,

REFINEMENT, ANALYSIS,

TESTING, AND

ENFORCEMENT

In this section, we provide more details about possible

issues, challenges and solutions useful for deﬁning

access control infrastructures able to specify, analyse,

and test networking policies addressing the data ma-

nagement in the cloud environment. In particular, in

the following subsections we target the policy speciﬁ-

cation, reﬁnement, analysis, testing, and enforcement

in the case of XACML-based access control systems.

4.1 Policy Speciﬁcation

Concerning the policy speciﬁcation, currently diffe-

rent solution for policies speciﬁcation and analysis are

available such as (Weimer and Vining, 2017). These

proposals usually rely on the large variety of policy

languages available in literature, which mainly follow

three different speciﬁcation approaches:

• rule-based, as for instance XACML (OASIS,

2005) and FACPL (Margheri et al., 2017), where

policies are basically used to express Event, Con-

dition, and Action (ECA) rules. The target is to

enable or improve the cross-platform interopera-

bility and modelling security aspects;

WEBIST 2018 - 14th International Conference on Web Information Systems and Technologies

264

• logic-based, as e.g. ASL and PDL (Ma et al.,

2015) which let to express authorization policies

based on user’s identity credentials and authori-

zation privileges, to propagate the access rights

among roles and groups of users, and to deﬁne

integrity checks on authorization decisions;

• ontology based,as e.g. KAoS and Rei (Ma et al.,

2015) which permit to represent the contextual

information of an access control system (i.e.,

the knowledge of the system) or Veloudis at al.

(Veloudis et al., 2017) that provide an ontological

template for expressing the relevant contextual at-

tributes associated to an access control policy or

request.

Even if there are proﬁcient ﬁelds of research tar-

geting policy speciﬁcation, domain IT skilled experts

are still required for correctly and extensively deﬁne

the required rules (policies). The current state of the

art could be improved by the introduction of expres-

sive, ﬂexible, and user-friendly policy speciﬁcation

languages deﬁned by a simple, yet rich, syntax and a

rigorous semantics, amenable for veriﬁcation and va-

lidation. In particular, due to the widespread diffusion

of access policy usage, existing proposals, should be

enriched by features supporting also the policy speci-

ﬁcation from non IT expert. This should requires the

implementation of automated and reliable applicati-

ons to map the user-friendly (natural language) rules

into processable and machine readable policies.

4.2 Policy Reﬁnement, Monitoring, and

Testing

Once policies are speciﬁed in high-level language,

they need to be automatically transformed into po-

licies in the languages of the various enforcement

points and then ﬁne-tuned to the operational con-

straints of the devices where those are enforced.

The action reﬁnement theory (Rensink and Gorrieri,

2001), is typically used in formal methods for con-

verting the speciﬁcation of an (abstract) action into a

(concrete) process. In the context of security, some

proposals like (Martinelli and Matteucci, 2011) ad-

dress this theory and provide a mechanism for trans-

forming high-level primitives/actions into lower level

processes, in such a way that some security proper-

ties are preserved within the transformation. After

the policy reﬁnement, due to the complexity of lan-

guages used from the various enforcement points for

specifying the access policies and to the growing size

of these policies, testing that the enforced policies

properly implement the intended regulations becomes

a compelling and challenging task (Bertolino et al.,

2016).

In this ﬁeld, a critical issue is the generation of an

efﬁcient test suite. Many methodologies and tools for

the automated test cases generation from an XACML

policy have been developed (Bertolino et al., 2013).

Existing test cases generation tools rely on combi-

natorial approaches of policy values and have been

proven to be effective in the automated generation of

access requests (Hu et al., 2011).

Another important challenge, in the context of

cloud-based access control policy testing, is to gua-

rantee the compliance between the security require-

ments the policy authors intend to specify and those

that the policies actually state. Moreover, combinati-

ons and integration of policies coming from different

cloud based organizations increase the risk of incon-

sistencies. A main research direction envisages the

speciﬁcation of access control policies by means of a

model and the usage of model-driven approaches for

the generation and testing of the automatically deri-

ved security policies (Bertolino et al., 2014a). These

approaches have the main goal of overcoming the dif-

ﬁculties of directly writing XACML code and allow

the testing of XACML policies against the intentions

expressed in the model.

In such a context innovative testing approaches

able to discover inconsistencies between the proces-

sable policy speciﬁcation and the true intentions of

the users in granting/denying accesses are still neces-

sary. In particular solutions should focus on:

• providing innovative testing strategies aiming at

verifying the constraints, permissions and prohi-

bitions deﬁned in the processable policy;

• providing tool support for fault tracking and dis-

covering so to promptly identify the errors in the

processable policy causing inconsistencies, pro-

blems and/or security risks;

• providing facilities for verifying the correct and

compliant integration of access policies coming

from different areas (for instance: hospitals, le-

gal context, generic environment and or organiza-

tion constraints) with the aim of warning the user

whether the ﬁnal processable policy speciﬁcation

could allow access rights not reﬂecting the user

original intentions;

• providing supports for checking the compliance

of the processable policy with data protection and

access data regulations.

4.3 Policy Analysis

The increasing spread of policy-based speciﬁcations

has prompted the development of multiple analysis

Issues and Challenges of Access Control in the Cloud

265

techniques like, e.g., property checking and behavi-

oural characterisations. Such techniques have been

implemented by means of different formalisms, va-

rying from multi-terminal binary decision diagrams

(MTBDD) to different kinds of logics. Current pro-

posals include (Arkoudas et al., 2014): change-impact

analysis of XACML policies to study the consequen-

ces of policies modiﬁcations; logic-based analysis

to verify structural properties of XACML policies:

policy redundancy to increase the evaluation perfor-

mance of access control systems.

However, available semantic-based approaches

should be leveraged so to better analyze the proper-

ties and the structure of policies. The use of these

properties will increase the conﬁdence on the grained

or forbidden system behaviours and could represent

a bases for more complex analyses, as e.g. change

impact analysis and redundancy minimisation. Furt-

hermore, speciﬁc approaches enabling effective and

efﬁcient automated veriﬁcation of these properties so

to provide a concrete support to the security and struc-

tural analysis should be also provided. For instance,

there is the necessity of speciﬁc formalisms that, on

one hand, permit to collapse hierarchical policies into

a single-layered representation and, on the other, are

sufﬁciently ﬂexible to deal with multiple domain va-

lues for attribute assignments.

In this activity, main challenges are the hierar-

chical structure of policies, the presence of conﬂict

resolution strategies, the intricacies deriving from

the many involved controls and the difﬁculties in

checking whether a given security property is pro-

perly enforced.

Moreover, it is also necessary to improve and ex-

tend the list of relevant properties, concerning the

three main security principles of conﬁdentiality, inte-

grity, and availability, and devise a general approach

for rendering them in terms of policy-based speciﬁca-

tions. To this purpose, formal methods and techniques

for deﬁning the semantics of policy languages in ri-

gorous ways could be adopted and modiﬁed in order

to provide a precise formalization of the conditions

under which a policy properly enforces the intended

properties.

4.4 Policy Enforcement

Many works address policy enforcement aiming to

provide authorization framework and conﬂict resolu-

tion approaches. Usually, the available solutions in-

volve strategy for policy precedences, cryptographic

solutions or combining algorithm for solving conﬂict

at run-time.

Runtime testing and monitoring of the policy en-

forcement enable to detect inconsistencies, or security

ﬂaws in the implementation of the access right enfor-

cements (Bertolino et al., 2014a). Indeed, a runtime

testing framework should: let the combination of dif-

ferent testing strategies; provide coverage and perfor-

mance measures; let simulations on the policy evalu-

ation engine; provide a continuous control during the

runtime execution. To this purpose, in order to make

easier the control of policy enforcement, a possibi-

lity is represented by the use of obligations (Margheri

et al., 2017), i.e. additional actions speciﬁed in the

policies that must be successfully executed at enfor-

cing time.

However, automatic framework equipped with

continuous authentication and authorization mecha-

nisms are still necessary. In particular, research

should be focused on development of strategies and

tools for:

• analyze the weaknesses of the today’s popular

access rights languages, such as XACML and

SAML, when applied to highly dynamic environ-

ments;

• provide effective and efﬁcient extensions of such

languages so to be adapted to satisfy the cloud en-

vironmental requirements;

• analyze whether today’s identity management me-

chanisms are sufﬁciently mature to meet the

privacy-speciﬁc requirements;

• investigate dynamic policy management infra-

structure that will allow on-the-ﬂy update of aut-

horization policies according to the user localiza-

tion;

• put in place enforcement mechanisms of autho-

rization policies taking into account the legal as-

pects;

• improve existing solutions so to be easily adapted

to different speciﬁc domains;

• allow possible conﬂicts resolution of multiple po-

licies written in different policy languages as well

as different policy instances written in the same

policy language.

4.5 Mapping Issues and Challenges

Considering the issues and challenges identiﬁed in

Section 3, we provide in Table 1 the solutions for po-

licies speciﬁcation, reﬁnement, analysis, testing, and

enforcement, listed in the previous subsections, that

are currently addressing these challenges. The tar-

get is to highlight how the current state-of-the practice

for deﬁning access control infrastructures is replying

WEBIST 2018 - 14th International Conference on Web Information Systems and Technologies

266

Table 1: Policy Speciﬁcation, Reﬁnement, Testing, Monitoring, Analysis and Enforcement vs Cloud Challenges.

Speciﬁcation Reﬁnement Testing Monitoring Analysis Enforcement

Continuous access X X

Different regulations X

Dynamic access policies X X

User-friendly management X

Continuous control X X X X

Scalability X

Resource sharing and interoperability X X

Validation and veriﬁcation of access control policies X X

On line tracing of access control policies execution X

Testing of access control systems X X

to the several critical peculiarities of the cloud-based

environment.

As in the table, monitor and testing are currently

the most adopted solutions for replying to the listed

challenges. Speciﬁcally, the possibility of on-line log-

ging and tracing of access control systems seems to

be the most effective way for behavioral control so to

avoid security ﬂaws and assure interoperability.

Moreover, still according to the table, most of the

proposals are focused on the continuous control evi-

dencing its peculiarity and speciﬁcity in the complex

and heterogeneous cloud computing environment.

Finally, the table suggests also that research acti-

vity should be more focused on the possibility to inte-

grate different regulations, to provide a user-friendly

management, to improve scalability and to promote

validation and veriﬁcation of access control policies,

because only few proposals are currently available.

5 CONCLUSIONS

Cloud computing infrastructures provide a cost ef-

fective mechanism to share hardware and software re-

sources in a scalable and distributed way. Security is a

primary concern in modern cloud interconnected dis-

tributed systems and access control represents an im-

portant security aspect specifying which subjects can

access which resources under which conditions.

In this paper, an analysis of the traditional access

control models as well as of new existing solutions for

access control in the cloud has been provided. Then,

many important challenges related to the distributed

data access in cloud systems, such as continuous con-

trol, different access regulations, scalability, dynamic

access policies, user friendly management of access

control policies, have been discussed.

Speciﬁcally, testing and analysis of access cont-

rol systems has been considered as a key feature to

enhance the security level of data in the cloud. Cur-

rent state of the practice solutions for policies spe-

ciﬁcation, reﬁnement, analysis, testing and enforce-

ment have been depicted and mapped on the identiﬁed

access control challenges. Possible improvements as

well as future research directions have been also high-

lighted.

It is out of the scope of this paper to provide an

exhaustive survey of issues and challenges of access

control in the cloud. However, besides the existing

proposals, the research activity concerning the access

control in the cloud is not keeping the pace with

the quick, dynamic and continuous cloud evolution.

There is still the need of innovative and advanced so-

lutions targeting the many evolving issues evidenced

in this paper.

As future work, we plan to have an in-depth inves-

tigation of the research directions in this context and

provide concrete solutions for addressing some of the

identiﬁed challenges.

ACKNOWLEDGEMENTS

This work has been partially supported by the GAUSS

national research project (MIUR, PRIN2015, Con-

tract2015KWREMX).

REFERENCES

Almutairi, A., Sarfraz, M., Basalamah, S., Aref, W., and

Ghafoor, A. (2012). A distributed access control

architecture for cloud computing. IEEE software,

29(2):36–44.

Arkoudas, K., Chadha, R., and Chiang, J. (2014). Sophisti-

cated access control via smt and logical frameworks.

ACM Trans. Inf. Syst. Secur., 16(4):17:1–17:31.

Ausanka-Crues, R. (2001). Methods for access control: ad-

vances and limitations. Harvey Mudd College, 301:20.

Bertolino, A., Busch, M., Daoudagh, S., Lonetti, F., and

Marchetti, E. (2014a). A toolchain for designing and

testing access control policies. In Engineering Secure

Future Internet Services and Systems: Current Rese-

arch, pages 266–286.

Bertolino, A., Daoudagh, S., Lonetti, F., and Marchetti, E.

(2016). Testing access control policies against inten-

ded access rights. In Proceedings of the 31st Annual

Issues and Challenges of Access Control in the Cloud

267

ACM Symposium on Applied Computing, Pisa, Italy,

April 4-8, 2016, pages 1641–1647.

Bertolino, A., Daoudagh, S., Lonetti, F., Marchetti, E., Mar-

tinelli, F., and Mori, P. (2014b). Testing of polpa-

based usage control systems. Software Quality Jour-

nal, 22(2):241–271.

Bertolino, A., Daoudagh, S., Lonetti, F., Marchetti, E., and

Schilders, L. (2013). Automated testing of extensible

access control markup language-based access control

systems. IET Software, 7(4):203–212.

Bertolino, A., Traon, L., Lonetti, F., Marchetti, E., and

Mouelhi, T. (2014c). Validation of access control sys-

tems. In Engineering Secure Future Internet Services

and Systems - Current Research, pages 210–233.

Cai, F., Zhu, N., He, J., Mu, P., Li, W., and Yu, Y. (2018).

Survey of access control models and technologies for

cloud computing. Cluster Computing, pages 1–12.

Calabr

o, A., Lonetti, F., and Marchetti, E. (2018). Monito-

ring of access control policy for reﬁnement and impro-

vements. In Proc. of Software Quality: Methods and

Tools for Better Software and Systems, pages 17–36.

Hu, V. C., Kuhn, D. R., Xie, T., and Hwang, J. (2011).

Model checking for veriﬁcation of mandatory access

control models and properties. International Journal

of Software Engineering and Knowledge Engineering,

21(01):103–127.

Hwang, J., Xie, T., Hu, V., and Altunay, M. (2010). Acpt:

A tool for modeling and verifying access control poli-

cies. In Proc. of International Symposium on Policies

for Distributed Systems and Networks, pages 40–43.

Jin, X., Krishnan, R., and Sandhu, R. (2012). A uniﬁed

attribute-based access control model covering dac,

mac and rbac. In IFIP Annual Conference on Data

and Applications Security and Privacy, pages 41–55.

Lonetti, F. and Marchetti, E. (2018). On-line tracing of

XACML-based policycoverage criteria. IET Software.

Ma, Z., Yang, Y., and Wang, Y. (2015). A security po-

licy description language for distributed policy self-

management.

Margheri, A., Masi, M., Pugliese, R., and Tiezzi, F. (2017).

A rigorous framework for speciﬁcation, analysis and

enforcement of access control policies. IEEE Tran-

sactions on Software Engineering.

Martinelli, F. and Matteucci, I. (2011). Preserving security

properties under reﬁnement. In Proceedings of the 7th

International Workshop on Software Engineering for

Secure Systems, SESS ’11, pages 15–21.

OASIS (2005). eXtensible Access Control Markup

Language (XACML) Version 2.0. http://docs.oasis-

open.org/xacml/2.0/access

control-xacml-2.0-core-

spec-os.pdf.

Park, J. and Sandhu, R. (2002). Towards usage control mo-

dels: beyond traditional access control. In Proc. of

the seventh ACM symposium on Access control mo-

dels and technologies, pages 57–64. ACM.

Rensink, A. and Gorrieri, R. (2001). Vertical implementa-

tion. Information and Computation, 170(1):95 – 133.

Sharma, P. K., Kaushik, P. S., Agarwal, P., Jain, P., Agarwal,

S., and Dixit, K. (2017). Issues and challenges of data

security in a cloud computing environment. In Proc.

of 8th Annual Conference on Ubiquitous Computing,

Electronics and Mobile Communication, pages 560–

566.

Suhendra, V. (2011). A survey on access control de-

ployment. In International Conference on Security

Technology, pages 11–20. Springer.

Tianyi, Z., Weidong, L., and Jiaxing, S. (2011). An efﬁcient

role based access control system for cloud computing.

In 11th International Conference on Computer and In-

formation Technology, pages 97–102.

Tsai, W.-T. and Shao, Q. (2011). Role-based access-control

using reference ontology in clouds. In Proc. of 10th

International Symposium on Autonomous Decentrali-

zed Systems, pages 121–128.

Ubale Swapnaja, A., Modani Dattatray, G., and Apte Su-

labha, S. (2014). Analysis of DAC MAC RBAC

Access Control based Models for Security. Internati-

onal Journal of Computer Applications, 104(5):6–13.

Veloudis, S., Paraskakis, I., Petsos, C., Verginadis, Y., Pa-

tiniotakis, I., and Mentzas, G. (2017). An ontologi-

cal template for context expressions in attribute-based

access control policies. In Proc. of the 7th Interna-

tional Conference on Cloud Computing and Services

Science, pages 151–162.

Veloudis, S., Verginadis, Y., Patiniotakis, I., Paraskakis,

I., and Mentzas, G. (2016). Context-aware security

models for paas-enabled access control. In Proc. of

the 6th International Conference on Cloud Computing

and Services Science, pages 202–212.

Wang, W., Han, J., Song, M., and Wang, X. (2011). The de-

sign of a trust and role based access control model in

cloud computing. In Proc. of 6th International Confe-

rence on Pervasive Computing and Applications, pa-

ges 330–334.

Weimer, D. L. and Vining, A. R. (2017). Policy analysis:

Concepts and practice. Taylor & Francis.

Youniand Kifayat, K. and Merabti, M. (2014). An access

control model for cloud computing. Journal of Infor-

mation Security and Applications, 19(1):45–60.

WEBIST 2018 - 14th International Conference on Web Information Systems and Technologies

268