Can “Feature” be used to Model the Changing Access Control Policies?

International Journal of Research in Computer Science
eISSN 2249-8265 Volume 2 Issue 6 (2012) pp. 21-31
www.ijorcs.org, A Unit of White Globe Publications
doi: 10.7815/ijorcs. 26.2012.052

CAN “FEATURE” BE USED TO MODEL THE
CHANGING ACCESS CONTROL POLICIES?
K.Shantha Kumari1, Dr T.Chithralekha2
1
Research Scholar, Dept of Banking Technology, Pondicherry University, INDIA
Email: shanthajayakumar@gmail.com
2
Associate Professor, Dept of Computer Science, Pondicherry University, INDIA
Email: tchitu@yahoo.com

Abstract: Access control policies [ACPs] regulate risks and vulnerabilities without jeopardizing speed or
the access to data and resources in information cost, organizations must bring security into the
systems. These ACPs are framed from the functional development process and this proved to be effective.
requirements and the Organizational security & Hence, the process of integrating the ACPs with the
privacy policies. It was found to be beneficial, when FRs was recommended [2], [3], [4], [5], and [6]].
the ACPs are included in the early phases of the
software development leading to secure development From the design perspective, access control policies
of information systems. Many approaches are provide an insight to the various kinds of threats,
available for including the ACPs in requirements and violations that can be handled in the design phase of
design phase. They relied on U ML artifacts, Aspects the software development. The overall system
and also Feature for this purpose. But the earlier development process is fruitful when the design phase
modeling approaches are limited in expressing the supports integrated modeling of ACPs and FRs. Hence
evolving ACPs due to organizational policy changes defining the ACPs in the same way as the FRs in the
and business process modifications. In this paper, we Design phase is considered as a prudent way. The
analyze, whether “Feature”- defined as an increment modeling process should be expressive and flexible
in program functionality can be used as a modeling enough to accommodate all the different requirements
entity to represent the Evolving Access control that may need to be expressed, while at the same time
requirements. We discuss the two prominent be simple both in terms of use and implementation (so
approaches that use Feature in modeling ACPs. Also that it can be verified with ease).
we have a comparative analysis to find the suitability In some prior works, the existing modeling
of Features in the context of changing ACPs. We methodologies are modified to define ACPs as like
conclude with our findings and provide directions for FRs. These methodologies have to take care of the
further research. proper abstraction of ACPs and the process of
Keywords: Access control polices, Features modeling without losing the consistency of the
Functional requirements, modeling, RBAC functional requirements. Also, the ACPs of today’s
enterprise information systems tend to be complex,
I. INTRODUCTION
14B
dynamic and scalable with respect to the resources
under the protection, top management’s policy changes
Access control is defined as the ability to permit or etc. The Modeling approach has to accommodate the
deny access to a particular resource (object) by a same for providing consistent access control
particular entity (subject). An Access control policy throughout the system development. But the present
defines the (high-level) rules according to which modeling methodologies cannot support these kinds of
access control must be regulated. An ACP may express intricate requirements. There is a need for higher level
conditions that must be satisfied before an access of abstraction to capture the complex characteristics of
request can be granted. ACPs are derived from ACPs.
requirements as well as high-level security and privacy
policies of the organization. Traditionally, ACPs are In recent years, Feature oriented programming and
handled in an ad-hoc way i.e. addressed in an existing Feature modeling caught the attention of the research
system either as an afterthought, or manually injected world. Features- a basic building block that represents
into practices of the organization. Late analysis of the intention of concept can satisfy intuitive user
ACPs can generate conflicts between them and the formulated requirements on the software systems.
FRs of the system. This leads to security failures, Feature modeling also has the capacity to represent
violations from the access control rules, leakage of variations by feature composition. This makes it more
vital information etc [1]. To properly address security suitable for modeling dynamic requirement.

www.ijorcs.org

22 K.Shantha Kumari, Dr T.Chithralekha

In this paper, Feature based ACPs modeling and represents their behavior using anyone of the
approaches are studied and analyzed. A detailed diagram. But the scalable, dynamic and evolving
comparison of each feature modeling approach is natures of the ACPs are not represented in UML.
done. The advantages and the limitations of every Moreover they either focus on specifying security
method are presented. This analysis would help for requirements in UML notation or analyzing UML
possibilities of new areas in further research. In section models against the specified requirements. Still the
2, the modeling approaches prior to feature modeling problem of systematic enforcement of the specified
are presented and their limitations are reviewed. In requirements exists. To overcome this, researchers
Section 3, Features and Feature modeling are took the idea from “Principle of separation of
introduced. Then ACP modeling approaches using concerns”. This principle helps to identify,
features are analyzed and a comparative study is done encapsulate and manipulate those parts of software
in Section 4. The paper concludes with the future that are relevant to a particular “concern” that may
directions of research. crosscut many design elements at the Design level.
4. Aspect oriented Software development technologies
II. LITERATURE REVIEW
15B

are a promising proposal to enable the
Olden year software systems, by their very nature, modularization of these crosscutting concerns [20].
have simple and generic access operations and are not ACPs being a crosscutting concern across the
concerned with the meaning of the resources they functional design and implementation can be
handle. On the other hand, modern Information encapsulated as “aspects”. Subsequently, a weaving
sensitive systems and applications handle quite process is employed to compose core functionality
complex access operations due to their specific user model elements with those aspects, thereby
purposes and high functionality. This complexity generating an architecture design. The dynamic
needs to be translated by generic models and specific nature of ACPs is well addressed by “Aspect
ACPs into design phase understandable processes. oriented modeling” [AOM]. The use of aspects in
Many approaches including the Formal approaches, security is among the most successful uses of
High level languages, Model based approaches aspect-oriented concepts, both at the specification
[UML], Process based approaches are available. and coding levels. The Aspect oriented software
development which uses the aspects as its base unit
1. Formal logic approaches [[7], [8], [9], [10], [11], typically has a programming language character in
[12], [13], [14]] have been a significant part in this order to attract a wide user community.
research area. They proved to be very much useful
− Despite its attractiveness to programmers, this
due to their formal theory. Approaches for
code-level approach has its disadvantages.
specifying and analyzing the ACPs are based on
Generally aspects are stated operationally, that
sophisticated mathematical concepts, that formally
results in unclear specification of it intention. It is
stated allows one to check whether the developed
quite difficult to prove the correctness of an
ACPs enforce the required level of protection.
aspect. Also, it is quite unsure about the
However, in practice, applying mathematically-
application of pointcut to all occurrences of the
based formal specification techniques can be
intended runtime events. Moreover the
difficult because of the high degree of mathematical
complexities of the system and ACPs may require
skill needed. Thus formal-logic based approaches
extending the pointcut and advice description to
which are difficult to use and understand, are
cover any extra cases which requires an
seldom used by application developers.
understanding of the modified set of runtime
2. Another extreme end is High level languages [[15], events being targeted, and what code to execute
[16], [17]]. High-level languages are easy to use in each specific context of occurrence [21].
and understand, but are not amenable for analysis.
5. But a feature represents the intention of a concept.
Therefore, a representation that can be analyzed
A feature is a logically cohesive piece of
without sacrificing understandability and usability
functionality and is present in all phases of software
is desirable.
development. Features may occur at any level, for
3. The UML diagrams [3] [18] [19]and UML profiles example high-level system requirements,
[4] [5] model the static ACPs. The constraints in the architectural level, subsystem and component level.
requirements are captured using OCL. The Thus, it is natural to expect that modularization of
effectiveness of the UML as a standard is software into features can provide a lot advantages.
predicated. Among other things, being a clear, ACPs which tend to be continual can be modeled
precise, and pragmatic semantics for its notations is using feature modeling as the features reflect user
an advantage. But the semi-formal semantics of requirements and can incrementally refine one
UML has ambiguity and inconsistency issues. another.
Moreover UML captures the static requirements

www.ijorcs.org

Can “Feature” be used to Model the Changing Access Control Policies? 23

The available approaches available for access features based approaches are more appropriate for
control modeling in the design phase could be modeling complex and dynamic access control
classified into the following way as shown in figure 1. requirements modeling. Hence, the disadvantages of
all the other approaches as a whole are explained and
However, in this paper, the emphasis is on delving the subsequent discussion progresses with the feature
into the details of Feature Oriented modeling oriented modeling.
approaches only. This is because, as mentioned earlier,

Access control Policy Modeling at Design Phase

Formal Logic HLL Model based approaches
Process –Based Approaches

UML Diagrams UML Patterns UML Profiles

Aspects
OCL and Class Secure Patterns
Diagrams

Features
UML Template classes

Usecase, Object, Class
and Sequence diagrams
Feature Meta-Model Feature as Patterns

Figure 1: Classification of Different Approaches in ACP Modeling at Design Phase

In the forthcoming sections, a preliminary future requirements, the integration of domain analysis
introduction to Feature, Feature modeling is given. activities with software development turned out to be
Following is the brief explanation of TWO different necessary, both from a process and from the economic
approaches taken for modeling the ACPs using point of view. Ever since its introduction in 1990,
Feature. From the Literature survey, it is noted that the feature modeling has attracted a great number of
Feature based modeling for ACPs is a new emerging application domains. And it becomes the most popular
area of research and hence limited number of method of domain analysis with the development of
publications are available. The Available publications domain engineering and product line.
are either based on designing a Feature meta-model or
treating a Feature as a Pattern similar to UML. As the Secure Software development is increasingly under
years progress, many research works based on Feature the demands of unpredictable, diverse and growing
would be available. business requirements, as well as limited time-to-
market, meticulous product value and mounting
competition. The ACPs fall under this category can be
III. FEATURES, FEATURE MODELING AND
elegantly modeled with the FRs using feature
16B

ACPS
modeling
Features are basic building blocks that satisfy
intuitive user formulated requirements on the software − Feature Meta-modeling is a standard way of
system. They reflect user requirements and representing the ACPs as features and also provides
incrementally refine one another. Feature model was a conceptual scheme to integrate the semiformal
introduced from the Feature-Oriented Domain requirements with the functional requirements.
Analysis (FODA) methodology [22] and further − In the same lines, Patterns are also proved to be a
developed from a number of approaches. wonderful tool for defining the ACPs.
Feature models are used to represent the variability − Techniques like Ontological models and Case-base
and commonality of software product lines, and permit reasoning is also available as higher abstract
the configuration of specific applications. Hence it can structures.
be defined as the activity of modeling the common and
variable properties of concepts [23]. For planning

www.ijorcs.org


real time. Interactions between features are not
In this paper, Features being an intention of concept
considered in this approach.
is modeled and discussed as Meta-models and
Patterns. As mentioned earlier, being a novel thought,
B. Pattern Based ACP Modeling approaches
Feature based ACP modeling is carried out with these
two approaches only till now. The Following section A design pattern describes a generic solution to a
reviews the related work in both the approaches. recurring design problem. ACPs for any domain can be
captivated in form for Access control patterns and can
A. Meta-Model based ACPs Modeling Approaches be applied along with the software functional
requirements. Whereas Meta-Model represents a high
Meta-modeling in software engineering and
level of abstraction of various models, the Patterns can
systems engineering among other disciplines, is the
be defined as the solution to a repeated problem with
analysis, construction and development of the frames,
respect to context. Just as Meta-model is tailored to
rules, constraints, models and theories applicable and
domain specific requirements; patterns too can be
useful for modeling a predefined class of problems. A
instantiated for a particular application. Features as
meta-model is yet another abstraction, highlighting
Patterns and its structure can be represented using
properties of the model itself. A model conforms to its
Feature Modeling. Any Access control model’s
meta-model.
properties and characteristics can be defined using a
− Using Meta-modeling concept, the ACPs can be declarative language such as OCL, when the Pattern is
presented as a Feature along with their represented using UML.
relationships, dependencies and constraints in a
Research work available: The works defined in [[26],
simple yet powerful way. The complexity and the
[27], [28]] have done modeling of Access control
dynamic nature of the ACPs can be captured and
policies as patterns and configured with respective
represented using composition and decomposition
domain requirements. ACPs are conformed to RBAC
relationships.
model and then they are treated as a Pattern with OCL
defined constrains. The works are briefly explained in
− Any Access control requirement represented in the the following section:
model that conforms to the meta-model ensures the
1. Any Policy is useful only, when it is appropriate at
consistency between the FR and ACPs. For
that particular instant. The work in [26] used this
example, RBAC model can be used to represent the
concept and presents a modeling approach that
ACPs with respect to the meta-model. The Core
configures ACP features on “Need Basis”. This
RBAC and other forms of RBAC can be step-wise
method produces Configured ACP features with
refined. The RBAC requirements can be composed
respect to the Application. ACPs are represented as
with the domain requirements as per the
RBAC features and they are composed with the
composition relations described in the meta-model.
FRs using “Partial Inheritance”. The RBAC
Research work available: In this section, one of the features and the relationships are captured by
premier works done in modeling the access control Feature modeling and UML is used for specifying
requirements using Feature Meta-model is reviewed. the RBAC features in a form of patterns. The partial
The work in [24] presents a feature-based Access inheritance of RBAC features enables step-wise
Control Requirements Modeling approach. The Access composition, which allows verification of
control policy is unfolded as Access control immediate impact of selected features. The authors
requirement in this approach. Access control use the Object Constraint Language (OCL) [29] to
requirement and FRs are represented as AC-Feature define operation semantics. This approach enables
and F-Feature respectively. Any Access control model fine-grained configuration of RBAC at the feature
including RBAC, DAC, and MAC is abstracted as level in a systematic manner, which helps to lower
ACM-Feature. Any Access control requirement is development complexity and reduce potential errors
modeled as a composition of ACM and F-Features. by excluding unnecessary features. . This work
The Composition relations are defined in the Feature didn’t represent all the properties of the Access
meta-model. This helps to capture the model to which control model [RBAC] taken for representing the
the ACPs conform and also ensures the consistency ACPs.
between FR and ACPs. This work adapts the Feature 2. The lacking factor of the previous work is resolved
meta-model defined in [25]. The easy-to-tailor nature in [27]. The authors have extended to include the
of the feature model can be used to elegantly control specification for the static separation of duty feature
the complexity and changefulness of ACPs. The and also updated the Core & General features with
lacking factor of this approach is that, the relationships additional behaviors of RBAC. The complete
modeled are very minimum, when compared to the

www.ijorcs.org


profile of RBAC is represented as a Pattern. In this IV. DISCUSSION
17B

pattern Core and non-core features are defined to From the early days, feature modeling is used in
represent the basic and additional behaviour of the various application domains. It has become one of the
Access control model. Along with Partial success methods of domain analysis with the
inheritance, “Compatibility principle of design” is development of domain engineering and product line.
used to check the Compatible features. This Also, enormous tool support is available.
approach benefits most in-house development as it
allows feature configuration in full scale to address In spite of their success, feature modeling has not
the specific needs of the target system and seamless seen widespread adoption as a routine part of systems
integration of configured RBAC into architecture. development practice. Previous surveys and our recent
Still the privacy and temporal features have to be review indicate that there have been successes in the
explored and also the systematic integration of application of feature models to problems of secure
dynamic ACPs has to be addressed. software systems; yet have certain limitations too,
which have to be researched further. This section
3. All Access control models can be treated as
provides valuable insights of every approach explained
Patterns and the ACPs can be represented using
earlier, through summarized overview and Qualitative
those patterns. Hybrid Access control models were
analysis. Also this paper tries to bring out the work
proved to be effective in many instances. Hence
still to be explored in ACP Modeling of feature
they can also be represented as patterns to define
modeling.
ACPs of same kind. This thought was materialized
in [28]. RBAC and MAC were considered in this In table 1, we revisit concerns raised in previous
work. Similar to [27], the features are defined and surveys and identify progress, trends, and remaining
configured with the application design on need challenges in the light of more recent projects through
basis. “Homogeneous” and “Heterogeneous summary analysis. The summary includes the design
Composition” are defined as Composition rules for of the features used in the method, the composition
RBAC and MAC features. This approach enables techniques, the verification process, the representation
fine-grained configuration of hybrid models of of ACPs in the proposed models and finally the
RBAC and MAC at the feature level, which helps representation of relationships and dependencies. This
reducing development complexity and errors by summary would highlight the application of Features
excluding unnecessary features in early in representing the characteristics and properties of the
development phase. Access control model taken for representing the ACPs
of an application.
Table 1: Summary of Contributions
Approaches Meta-Model based Pattern Based ACPs modeling
ACPs modeling
Ref: [26] Ref: [27] Ref:[28]
Parameters Ref: [24 ]
How a Feature is Feature Meta- Patterns
defined Model
Access control RBAC RBAC RBAC RBAC &
model used for MAC
ACP
representation.
Definition of ACP ACP is defined as The Access control Model –RBAC is defined as a
Access control Pattern.
feature
Features defined AC-Features Core Feature Core Feature Core Feature
for representing F-Features Hierarchy Hierarchy feature Hierarchy
the ACPs ACM-Features feature General feature feature
General feature Advanced Feature General
Advanced SOD Feature feature
Feature Review feature Advanced
SOD Feature Temporal feature Feature
SOD Feature
Refinement Decomposition, Refinement using Partial inheritance
Relationships Specialization, [No Specialization]
defined Characterization

www.ijorcs.org


Representation of “Relationships” Class diagrams
Refinement from UML core
relationships package
Dependencies Requires, excludes, Cardinality, Containment
defined Complex
constraints,
Protection
Representation of “Relationships” Sequence diagrams
Dependencies from UML core
package
Interpretation of Using the Object Constraint Language
relations, constraints Pre and Post-Conditions
dependencies,
constraints
Feature interaction Behaviour and Not discussed Compatibility relation is a set of (1) or
and behavior Interaction are relations (e.g., (SSD, DSD)) under the
defined using the same branch, (2) relations of features
constraints across branches that can be combined
(e.g., (General, DSD)), and (3)
implication dependencies.
Composition Composition is Step-wise Composition using Partial Homogeneous
techniques done through inheritance. Composition
integration AC 1. Relationship composition Heterogeneous
features with F- 2. Class Composition Composition
Features using 3. Operation Composition
relationships
defined
Representation of Feature Meta- Feature Composition is defined in terms of class
Composition Model diagrams and sequence diagrams. Also, Feature
process composition is the refinement in view of Multiple
inheritances
Verification Reference process Step-wise Composition and Configuration is the
process is described as a verification process
series of steps.
Tool Support Not Available IBM Rational Software architect [RSA] and Eclipse
plugin
UML Support UML Core Meta- UML Class Diagrams and Sequence diagrams
Model
Meta-Model Yes NO
Usage
Reusable Features Yes Patterns can be configured based on the application
Notation of Tree structure Class Diagrams
Feature Model
Role Hierarchies Easily The features have to configured depending upon the
customizable needs
Feature Conflicts They are not They are not analyzed
and Interactions analyzed

Following is the Qualitative analysis of the A. Factors that focus on representation of ACPs-
41B

approaches for ACP modeling in table 2. An analytical Applicability
framework is partially adapted from the work [30] and 1. Expressiveness [31]: Refers to the ability of
[31] , that defines the following technical criteria is representing security policies as models. This is an
used for analysis. Certain criteria are newly proposed important criterion that would make the modeling
in this paper that would enhance the qualitative approach successful. Regarding this criterion, any
analysis. specification approach will evaluated in expressing

www.ijorcs.org


the kind of requirements and can take one of the 2. Usability: The usability principle measures how
following values: simple or complex a technique can be used. In the
context of evaluation, the amount of work involved
− Static ACPs: if it allows the specification of the
in applying the technique, gathering information
majority of statically enforceable ACPs.
and processing for output is considered. Once the
− Dynamic ACPs: if it allows the specification of information is collected, how usable is the process
the majority of dynamically enforceable ACPs. of designing a diagram, tree, or table. It can take the
− Evolving ACPs: if it allows the specification of following values:
majority of evolving new ACPs. − Difficult: if the modeling approach consists
− Scalable ACPs: if it supports scalability of the difficult, long procedures/methodology
ACPs. − Medium: if the approach consists of
− All ACPs: if it supports all ACPs intermediate level of methodology with respect
to time and work.
2. Level of formality [30]: Refers to the ability to
check the formality [Procedures] in Modeling − Simple: if the approach consists of simple
approaches, to design and produce clear, reliable, methodology
complete and provable Models for the expressed 3. Analyzability: The analyzability measures how
ACPs. easily a developer can interpret and implement the
− Clear design: if the procedure produce a results provided by the techniques. For example,
unambiguous design whether the conflicts between the ACPs and FRs
− Reliable design: if the procedure supports are modeled? Is it possible to verify the security
consistent design assurance provided by the technique?
− Complete design: if the approach produces a − Easy: Analysis of the Output of the Modeling
complete model approach is Simple
− Provable design: if the approach produces a − Difficult: if the output interpretation is difficult.
model that can be verified.
C. Factors that check the Verifiability of the Models
3. Scope of specifications [30]: Refers to the ability of
43B

produced by the approaches:
representing both ACPs and FRs using same
notations. It can be measured with the value 1. Support for Variations: Refers to the ability in
“complete”. providing representations to show all possible
variations and their output.
4. Comprehensiveness: Refers to the ability to check
whether the particular set of modeling notations and 2. Support for Consistency checking: A Modeling
parameters including relationships are approach should produce a Model that allows
comprehensive to represent the ACPs. This can take checking the consistency between the ACPs and
the following values FRs composed together. Automatic internal
verification supported by a technique may improve
− Complete :if the approach can model all kinds of
the reduction of ambiguity, ensure completeness,
ACPs and relationships
improve consistency, and hence, make
− Incomplete: if the approach fails to capture specifications more verifiable.
certain ACPs and their relationships.
3. Support for Correctness of Specification: It may
B. Factors that Focus on Usability of the Approaches:
42B
ensure correctness of the design produced by the
approach by validating it against requirements
1. Learn ability: Refers to the ability of learning the and/or implementation.
modeling approach and using it. It can take the
following values: Table 2 allows us to relate Modeling approaches
and specification technique criteria. We can see, for
− Difficult : if the modeling approach is difficult instance, that the fulfillment of a technical criterion
to learn and use for designing must generate the fulfillment of all Requirement
− Medium: if the approach is learnable and usable criteria related to that criterion. The degree of
after meticulous study and training fulfillment will be “Y” (Yes), “N” (No) or “P”
− Simple: if the approach is suitable for usage (Partial).
even for new users.

www.ijorcs.org


Table 2: Qualitative Analysis of Features in ACPs modeling

Technique Criterion Requirement Methodologies
Criteria Meta-Model Pattern Based
based modeling
Modeling
Expressiveness Static ACPs Y Y
Dynamic ACPs P N
Evolvable ACPs P N
Scalable ACPs N N
All ACPs N N
Level of Formality Clear Y Y
Reliable P Y
Complete P P
Provable P P
Scope of Specification Complete P P
Comprehensiveness Complete P P
Incomplete Y Y
Learn ability Difficult P P
Medium Y Y
Easy N N
Usability Difficult P P
Medium Y Y
Easy N N
Analyzability Difficult P P
Easy P P
Support for Variations - N N
Support for Consistency - P P
checking
Support for Correctness - P P
of specification

V. INFERENCE
18B application. Hence the behaviour description is
relevant when the feature is ready to compose with
From Literature survey, Feature modeling using
another in order to produce a model.
Meta-Model or patterns was found to be optimal and
efficient method that supports complex ACPs without 1. In both approaches, the feature is defined a type
much cost and wastage of time. The Summary table that signifies how the feature adds up to the
highlighted the building blocks of each approach. All variability of the system. This confines the
the proposed ideas are very interesting and they prospects to reuse the same feature in a
provide important contributions to solve the security dissimilar perspective with different variability
problem in a methodological approach. Nevertheless, requirements. For E.g.: An Amount transfer
they have a series of limitations. The performance of payment feature may be mandatory in one
each approach is considered by the way, how the context while optional in another [depending
concept of “Feature” is handled in modeling ACPs upon the target application]. As the type is
with FRs. inextricably associated with the feature, it will
not be possible to reuse the feature as it is.
The Comparative analysis is made with factor
concerning to the Usability, Applicability and 2. In both the approaches, the variant behaviour of
Verifiability of the approaches. The results are the feature is not explored completely. The
analyzed with respect to the kind of feature and Feature behaviour is defined statically. This
relations defined in the techniques. reduces the purpose of reusability and also the
support for variation.
A. Feature Behaviour: A feature has the capacity to
behave differently in several instances. This was B. Feature Relationships: Within a feature model the
well illustrated in [23]. Feature behaviour or type features are structured by relations. Common to all
determines the kind of variability provided at the methods mentioned above are hierarchical relations
time of assembling with another feature in an between a feature and its sub features. They control

www.ijorcs.org


the inclusion of features to instances. If an optional VI. CONCLUSION
19B

feature is selected for an instance, then all
mandatory sub-features have to be included as well, This paper explores the Feature modeling
and optional sub-features can be included. approaches used for defining and integrating the ACPs
with the functional requirements. Also it surveys the
1. The “Expressiveness” factor in the qualitative related state of the art. A detailed discussion for the
comparison entirely depends on the kind of explained approaches was presented. This includes the
relationships the approaches support. The concise review and the Comparative analysis of the
Complex ACPs could not be expressed approaches. From this paper, it is understood that
completely with the relationship set defined in features are efficient in modeling the ACPs when they
both the approaches. tend to be complex. Two approaches based on features
2. The dependencies and the refinement were studied and analyzed. They present systematic
relationships should be expanded to include the way for representing ACPs, yet have certain
complex ACPs. limitations. Based on the limitations, there is a need for
complete feature definition including more
C. Feature Conflicts and Interactions: A feature relationships and constraints that represent the real
interaction occurs whenever one feature affects the time security requirements and their complexities.
behaviour of another. The feature interaction Moreover, the Feature definition should also be able to
problem is generally associated with conflicting manage the feature interactions and conflicts to
features causing undesirable effects. The feature support the consistency. And, to increase the
interaction problem is how to rapidly develop and variability options the type of the features has to be
deploy new features without disrupting the defined separately from feature definition. This would
functionality of existing features. So this concept support higher level of reusability and variation
has direct relationship with Consistency between productions. The Step-wise refinement of features can
the ACPs and FRs. From the summary, it was be strongly explored to adapt for complex and
found that feature interactions were not handled in dynamic requirements. The main contribution of this
both the approaches. Hence the support for work is the study and discussion of two important
Consistency check is not completely defined in Feature modeling methodologies used for the
both techniques. development of the software systems with the proper
D. Feature Decomposition: In both the approaches, the and systematic integration of ACPs. The major
Features are represented in Top-down approach limitation of this
[Hierarchical]. This poses a problem that the
Work is the inadequate availability of research
problem domain should be fully understood in
works based on Feature. We reviewed available works
prior, to be able to decompose into smaller
in those methodologies and found the certain
problems. This needs a combination of Top-down
observations that would propel further research.
and bottom up approach adaptation that could help
in composing and reusing features from an existing
and continuously growing storage of features. VII. REFERENCES
20B

[1] G Georg, I Ray, and R France, "Using aspects to design a
E. Functional and V ariability Decomposition: In the
secure system," In Proceedings of the International
feature models presented, the top down Conference on Engineering Complex Computing Systems
decomposition is implicitly based on Functional (ICECCS 2002), Greenbelt, MD, ACM Press., 2002. doi:
and Variability Decomposition. But they are not 10.1109/ICECCS.2002.1181504
clearly distinguished, so the modeling process has
become very complex with respect to Consistency [2] T Doan, S Demurjian, C T Ting, and C Phillips,
"RBAC/MAC security for UML," in Research
factor.
Directions in Data and Applications Security XVIII ,IFIP
F. Scalability: Feature meta-model approach have International Federation for Information Processing
significant characteristics to support Scalability of Volume 144. Catalonia, Spain: Springer, 2004, pp. 189-
features, when compared to Pattern based Models. 203.
But the tree structure used in Meta-Model based [3] D Kim, I Ray, R France, and N Li, "Modeling Role-
ACP modeling would lose the added value of a Based Access Control Using Parameterized UML
graphical representation when trees become very Models ," in FASE 2004, LNCS, vol. 2984, 2004, pp.
large and easy over viewing is not possible. There 180–193. doi: 10.1007/978-3-540-24721-0_13
is a need for Suitable Abstraction. [4] J Jurjens, "UMLsec: Extending UML for Secure Systems
Development," in Proceedings. of the 5th International
Conference on t he UML, Dresden, Germany, 2002, pp.
412–425. doi: 10.1007/3-540-45800-X_32

www.ijorcs.org


[5] T Lodderstedt, D Basin, and J Doser, "Secureuml: A Models and Technologies(SACMAT), 2004, pp. 31-40.
UML-based modeling language for model-driven doi: 10.1145/990036.990054
security," in Proceedings of the International [19] I Ray, N Li, D. K Kim, and R. B France, "Using
Conference on t he Unified Modeling Language, parameterized UML to specify and compose access
UML'2002, 2002, pp. 426-441. control models," in Proceedings of the 6th IFIP TC-11
[6] T Priebe, E Fernandez, J Mehlau, and G Pernul, "A WG 11.5 Working Conference on Integrity and Internal
Pattern System for Access Control," in Proceedings of Control in Information Systems, IICIS'03, Lausanne,
Conference on Data and Application Security, 2004, pp. Switzerland, (2003)., 2003. doi: 10.1007/1-4020-7901-
22–28. doi: 10.1007/1-4020-8128-6_16 X_4
[7] Steve Barker, "Security Policy Specification in Logic," [20] R Filman, T Elrad, S Clarke, and M. Aksit, Aspect-
in Proceedings of the International Conference on Oriented Software Development.: Addison
Artificial Intelligence,ICAI'2000, Las Vegas, NV, 2000, Wesley.,ISBN-10: 0321219767 | ISBN-13: 978-
pp. 143-148. 0321219763., 2000.
[8] Steve Barker and Arnon Rosenthal, "Flexible security [21] D R Smith, "A Generative Approach to Aspect-
policies in SQL," in Proceedings of the fifteenth annual Oriented Programming," in Proceedings of the Third
working conference on D atabase and application International Conference on G enerative Programming
security, Niagara, Ontario, Canada, 2001, pp. 167-180. and Component Engineering (GPCE’04), p. 2004. doi:
[9] Elisa Bertino, Piero Andrea Bonatti, and Elena Ferrari, 10.1007/978-3-540-30175-2_3
"TRBAC: a temporal role-based access control model," [22] Kyo C. Kang et al., "FORM : A feature-oriented reuse
in RBAC '00 Proceedings of the fifth ACM workshop on method,Volume 5," Annals of Software Engineering ,
Role-based access control, Berlin, Germany, 2000, pp. pp. 143 - 168, 1998.
21–30. doi: 10.1145/501978.501979 [23] L Abo Zaid, F Kleinermann, and O De Troyer, "Feature
[10] Fang Chen and Ravi S. Sandhu, "Constraints for role- Assembly Framework: Towards Scalable and Reusable
based access control," in RBAC '95 Proceedings of the Feature Models," in Proceedings of the 5th Workshop
first ACM Workshop on R ole-based access control , on Variability Modeling of Software-Intensive Systems,
1995, p. Article No. 14. doi: 10.1145/270152.270177 Namur, Belgium, 2011, pp. 1-9.
[11] R J Hayton, J M Bacon, and K Moody, "Access control [24] Lianshan Sun and Gang Huang, "Modeling Access
in open distributed environment," in In IEEE Control Requirements in Feature Model," in APSEC '09
Symposium on Security and Privacy , Oakland, CA, Proceedings of the 2009 16th Asia-Pacific Software
1998, pp. 3–14. doi: 10.1109/SECPRI.1998.674819 Engineering Conference, Penang, 2009, pp. 241-248.
[12] Michael Hitchens and Vijay Varadharajan, "Tower: A doi: 10.1109/APSEC.2009.21
Language for Role-Based Access Control," in POLICY [25] Hong Mei, Wei Zhang, and Haiyan Zhao, "A
'01 Proceedings of the International Workshop on Metamodel for modeling system Features and their
Policies for Distributed Systems and Networks, Bristol, refinement, constraint and interaction relationships,"
U.K. , 2001, pp. 88 - 106. doi: 10.1007/3-540-44569- Software and Systems Modeling 5(2), pp. 172-186,
2_6 2006. doi: 10.1007/s10270-006-0004-1
[13] S Jajodia, P Samarati, and V S Subrahmanian, "A [26] Dae-Kyoo Kim, Lunjin Lu, and Sangsig Kim, "A
Logical Language for Expressing Authorizations," in Verifiable Modeling Approach to Configurable Role-
IEEE Symposium on Security and P rivacy, pages , Based Access Control," in Proceedings of Fundamental
Oakland, CA, 1997, pp. 31–42. doi: Approaches to Software Engineering (FASE/ETAPS
10.1109/SECPRI.1997.601312 2010), Paphos, Cyprus, 2010, pp. 188-201. doi:
[14] R Ortalo, "A Flexible Method for Information Systems 10.1007/978-3-642-12029-9_14
Security Policy Specification," in Proceedings of the [27] S Kim, D. K Kim, L Lu, S Kim, and S Park, "A
5th European Symposium on Research in Computer Feature-Based Approach for Modeling Role-Based
Security, Louvain-la-Neuve, Belgium, 1998. doi: Access Control Systems;," Journal of Systems and
10.1007/BFb0055856 Software Vol. 84, No. 12, pp. 2035-2052, 2011. doi:
[15] J A Hoagland, R Pandey, and K N Levitt, "Security 10.1016/j.jss.2011.03.084
Policy Specification Using a Graphical Approach," [28] S Kim, D.-K Kim, L Lu, S Park, and S Kim, "A
Computer Science Department, University of Feature-Based Modeling Approach for Building Hybrid
California, Davis., Technical Report 1998. Access Control Systems," in 5th International
[16] OASIS. (2002). Available: http://www.oasis- Conference on Se cure Software Integration and
open.org/committees/xacml. Reliability Improvement (SSIRI), Jeju, Korea, 2011, pp.
88-97. doi: 10.1109/SSIRI.2011.16
[17] C Ribeiro, A Zuquete, and P Ferreira, "SPL: An Access
Control Language for Security Policies with Complex [29] Gail-Joon Ahn and Michael E. Shin, "Role-Based
Constraints," in roceedings of the Network and Authorization Constraints Specification Using Object
Distributed System Security Symposium, San Diego, Constraint Language," in WETICE '01 Proceedings of
CA, 2001. the 10th IEEE International Workshops on E nabling
Technologies: Infrastructure for Collaborative
[18] I Ray, N Li, R. B France, and D. K Kim, "Using UML Enterprises, Washington, DC, USA, 2001, pp. 157 –
to visualize role-based access control constraints," in 162.
Proceedings of the Symposium on A ccess Control

www.ijorcs.org


[30] A Khwaja and J Urban, "A Synthesis of evaluation
criteria for software specifications and specification
techniques," International Journal of Software
Engineering and Knowledge Engineering, vol. 12 , no.
5, pp. 581–599, 2002. doi:
10.1142/S0218194002001062
[31] C Talhi et al., "Usability of Security Specification
Approaches for UML Design: A Survey," Journal of
Object Technology, vol. 8, no. 6, pp. 103-122, 2009.
doi: 10.5381/jot.2009.8.6.a1

How to cite
K.Shantha Kumari, Dr T.Chithralekha, "Can “Feature” be used to Model the Changing Access Control
Policies?". International Journal of Research in Computer Science, 2 (6): pp. 21-31, November 2012.
doi:10.7815/ijorcs.26.2012.052

www.ijorcs.org

Can “Feature” be used to Model the Changing Access Control Policies?

More Related Content

What's hot (20)

Viewers also liked (9)

Similar to Can “Feature” be used to Model the Changing Access Control Policies? (20)

More from IJORCS (20)

Recently uploaded (20)

Can “Feature” be used to Model the Changing Access Control Policies?