An Elite Model for COTS Component Selection Process

International Journal of Engineering and Applied Computer Science (IJEACS)
Volume: 02, Issue: 01, January 2017
ISBN: 978-0-9957075-2-8
www.ijeacs.com 18
An Elite Model for COTS Component Selection
Process
Asif Irshad Khan
Department of Computer Science, FCIT
King Abdulaziz University
Jeddah, Saudi Arabia
Abstract—Component-based software development
(CBD) promises development of high-quality
trustworthy software systems within specified budget
and deadline. The selection of the most appropriate
component based on specific requirement plays a vital
role for high-quality software product. Multi-Agent
software (MAS) engineering approach played a crucial
role for selection of the most appropriate component
based on a specific requirement in a distributed
environment. In this paper, multi agent technique is
used for component selection. A semi-automated
solution to COTS component selection is proposed. It is
evident from the result that (MAS) plays an essential
role and is suitable for component selection in a
distributed environment keeping in view of the system
design and testing strategies.
Keywords- Component selection, Component based
development, COTS, Multi Agent Software Engineering
I. INTRODUCTION
Developing high-quality software product within
budget and time is always a challenging concern to
the software industries. Component-based software
development plays a crucial role in addressing
challenges faced by the software industries. CBD
supports development of software using software
components, also, known as commercial of the shelf
system (COTS). COTS components are sold in open
market and manufacturers of COTS define how a
component can plug and play into a system based on
the specified requirement [1].
Ideally, COTS components are like black box
solutions, one has to learn how a component can be
plugged into the system by using its required and
provide interfaces as defined by its manufacturers.
The component is developed by following software
development phases such as System requirement
specification, System Modeling and Design, System
implementation and Testing [10]. COTS Components
are highly reliable and trustworthy as they are
rigorously tested by its manufacturer and most of the
components have a rating that is evaluated by its
customers [12].
However, Selection of component plays a vital
role in developing high-quality software systems
within specified constraints like budget, efforts and
development time. Selection of candidate component
as per specified requirement is a challenging task as
the selection is usually based on multiple criteria such
as functionality support, cost, reliability, security,
adoptability etc.
Multi-Agent System (MAS) is intelligent software
system consists of software components also known
as Agent. These agents can interact with each other or
can coordinate with each other to accomplish assigned
task. For example, agents can be used to identify most
suitable candidates for a particular set of
requirements.
Multi-Agent approach is ideal for this research as
the agent can explore all available servers in the
distributed network and recommend potential
candidate components.
This paper is organized as follows: section I
describes introduction of the paper, sections II
discusses about the related work, section III puts
forward the proposed model, section IV explains the
role of different agents in the proposed mode, and
section V conclusion and future work.
II. RELATED WORK
Software engineering research community has
suggested different strategies and methodologies for
the selection of candidate COTS components.
Rikard et al. surveyed component selection state
of the art in the embedded systems domain and
presented a meta-model for selection method based on
common activities and practices. The authors also
gave suggestions which can be utilized as a schema
when making the technique and methodology for
component selection [2].

Asif Irshad Khan International Journal of Engineering and Applied Computer Science (IJEACS)
ISBN: 978-0-9957075-2-8
www.ijeacs.com 19
X. Burgu´es et al. suggested a process model for
combined selection of components based on the
features of two levels, a global one matching to the
combined selection itself, and a local level where all
the individual selection processes take place in
specific areas. The main contribution of this study is
to consider the aggregate expense for a system rather
than specifying in advance the individual expenses for
various components [3].
Lawrence et al. proposed an approach (CARE/SA)
which supports the iterative selection (matching and
ranking) of COTS components. The author considered
COTS component's representation within architecture
as an aggregate of their functional and non-functional
requirements with its own set of attributes. The
proposed framework could be viewed as an extension
to previous methodologies with a systematic approach
to matching, ranking, and selection COTS
components as shown in figure 1. [4]
Figure 1. Overview of the CARE Process [4].
Zahid et al. in their work evaluated the recent
practice of evaluation and selection of COTS
components from software houses in Pakistan. Closed
questionnaire methodology was adopted in the study.
The result shows that no criterion is utilized for
evaluating and selecting COTS components by the
majority of respondents. Authors proposed a
framework for COTS selection, integration and
evaluation [5].
Tarawneh et al. [6] suggested a framework support
and improve the COTS software evaluation and
selection processes in industry. To achieve this
objective the authors have shown that specific
objectives have to be addressed:
a) Identify the processes which support COTS
software evaluation and selection.
b) Determine the criteria or requirements
which are important for successful
evaluation and selection process.
c) Propose methods and techniques to address
the mismatch between COTS features and
customer requirements.
d) Develop a repository to manage information
from previous selection cases that support
the decision-making process.
M. Shakeel et al. [7] proposed component
selection methodology for component-based software
engineering. The author claimed that most of the IT
developers in development countries (UDCs) are
unaware about the COTS evaluation and selection
methodology. This methodology will reduce
development times, cost effective and less strenuous
efforts on the basis of software quality model
ISO/IEC 25010 as shown in figure 2.
Figure 2. IROTS Process [7].
Shah et al. proposed a method for component
selection using fuzzy logic. Several software quality
factors like reusability, portability, security,
maintainability etc were considered in the proposed
method. This methodology was evaluated using
hypothetical case study [8].
Agora [9] is a web-based component search
prototype proposed by Carnegie Mellon software
engineering Inst. Agora provides agents that crawl the

ISBN: 978-0-9957075-2-8
www.ijeacs.com 20
Web for components. In this model, the whole process
is divided into two main tasks.
One of the task which is an automated background
task is performed by an agent, i.e. a background agent
is responsible to automatically search the location and
indexes of COTS on the web for component
specific/specification model. While the component
selection and retrieval task is a manual and done by
the expert as shown in figure 3.
The main advantage of agora approach is its
ability to automatically build an index of available
worldwide components on the web [9].
Figure 3: Agora Architecture [9].
III. PROPOSED MODEL
The currently available models for COTS
component selection are complex and difficult to
implement and also its time consuming. The analysis
indicated that there is a need of improvement to
COTS component selection process.
This paper proposed an efficient semi-automated
COTS component selection technique which is easy
to understand and easy to implement as well.
A semi-automated model [ABCS] for COTS
component selection is proposed. This model is based
on the multi-agent technique in which a group of
agents works together with the single objective of
completing a specific task i.e. identifying the list of
suitable candidate components. Figure 4 shows a
conceptual view of this approach.
The whole process of identifying the candidate
component requires multiple agents to work together.
A multi-agent system is a loosely coupled network of
several agents which interact among themselves or
with external environments to solve problems that are
beyond one's individual abilities or knowledge of each
problem solver.
As multi-agent is composed of several
autonomous agents, there is a need of coordination
and cooperation among the agents as shown in the
Figure4.
Firstly the task of searching COTS component as
per defined criteria is broken into several sub-tasks
and each of the sub-tasks is assigned to specific
agents. For example, sub-tasks like matching
requirement against integration complexity, security
features verification, cost-benefit analysis, adaptation,
and quality assurance verification is assigned to
different agents as shown in figure 4.
These agents, after solving their respective sub-
tasks, co-ordinate with the main agent to produce the
result's set of candidate components as per the
selection criteria in the form of a weighted matrix
[11]. Finally, an expert team will decide the candidacy
of the most suitable component available from the list
of choices.
ABCS model take the outline of the user
requirement [User requirements may be in plain
English or in script form or any other defined form]
and application domain [like health, agriculture,
business, education etc] as the input parameter and
select the candidate component from the list of
available components from different repositories in a
distributed network as shown in figure 5.
As it is often seen that based on the requirements,
there might be several COTS products that match the
requirements for different degrees, but in few cases
most likely none of the several candidates would
completely match the user requirements.
The proposed model also has an agent integrating
tool [ABIT] that takes system architecture and list of
candidate components and integrates them to generate
source code for integrated components as shown in
figure 6.
In case of any compatibility issues it is handled by
compatibility test agent. It tests the compatibility
issues while integrating the selected candidate
components with the system architecture.
This agent automatically verify and validate the
compatibility of selected COTS component for any
risk of failure since selected component has to
integrate into the system and there is always a risk of
component failure at later stages of development.
In case the test shows compatibility problem while
integrating the selected candidate components, the
system will select the next possible candidate
component from the ABCS list as shown in figure 6.
Interactions between these agents are shown in figure
7.

ISBN: 978-0-9957075-2-8
www.ijeacs.com 21
Figure 4. Conceptual view of multi-agent approach for component section.
Figure 5. Conceptual Model of Component Gathering Agent.
IV. AGENT ROLE IN THE MODEL
A. Gathering Agent
The main responsibility of this agent is to maintain
in-house component meta-data database. Once
requirement is specified to this agent, it travels in
distributed network, collect meta-data details related
to available matching components and classify them
based on the component model to update in-house
database.
Vendor of the component maintain meta-data
details in online repositories, as shown in figure 5.
Following are some of the listed categories of
component classified based on the component model.
 Microsoft’s COM/DCOM/COM+
components.

ISBN: 978-0-9957075-2-8
www.ijeacs.com 22
 Microsoft’s .Net components.
 Object Management Group OMG’s
CORBA.
 Oracle J2EE or JEE 5 components
 Oracle JAVA RMI components
Figure 6. Model of Multi-agent based Component selection and Integration with Testing Tool.
Figure 7. Agent Interaction Sequence diagram for Multi-agent based Component selection.
B. MRC Agent
Matching requirement versus Integration complexity is
addressed by MRC agent. It plays a major role in the
evaluation of every component based on its matching

ISBN: 978-0-9957075-2-8
www.ijeacs.com 23
requirement and its relative complexity to integrate.
Figure 8. MRC Agent.
The component that has highest matching requirement and
lowest effort to integrate will be considered from the
component candidate list as shown in figure 8 and figure 9.
Figure 9. MRC Agent.
C. Security Agent
The role of this agent is to verify the security features in the
selected candidate components. The components should be
verified for any breach of security. This agent has
authentication and authorization manager which checks how
certificate authority, usable security policies etc. are defined
and security mechanisms are implemented and supported.
D. Cost Agent
The main responsibility of this agent is to take part in the
cost-benefit analysis so as to compare the component features
& functionalities with the cost. Costs should be seen in broad
aspects; because low cost components may result in higher
total cost along the product’s life cycle. Buying component
with loads of features also results in a higher cost and most of
the features remain unused by the users.
E. Adoption Agent
The main responsibility of this agent is to verify component
adoption process by studying all pre-requisite in a way that
components can be easily integrated into the system
architecture effectively.
F. Quality Assurance Agent
It is well-known fact that no software can be 100% error-
free. To check component high-quality features, knowing that it
is developed by the vendor company and may have some
issues, it is necessary to check the rating of the component as
most of the vendors maintain a rating of the components.
Also, it is necessary to find out what problems other users
are encountering from vendor site where issue tracker related to
the component are posted. The role of this agent is to validate
and verify above mentioned high-quality attributes and
generate a software quality metric for measurement that is used
to evaluate software quality in a system as shown in figure 10.
Figure10. MRC Agent.
V. CONCLUSION
In this paper, an approach to COTS component selection
was presented. The software component selection is done based
on different attributes like availability of features and
functionalities in the component as per requirements,
component adoption easiness in the architecture, security, cost,
quality of service etc.
The proposed approach attempts to find the best candidate
component list based on the requirements in a distributed
environment. The multi-agent technique is considered in this
approach by breaking a task into several sub-tasks and these
sub-tasks are assigned to specific agents to be solved.
All the rigid conditions are taken into consideration and
based on this; the most optimal component sets are generated.
Finally, domain experts will decide the candidacy of the most
suitable component available from the list of choices.
This approach gives acceptable results. This semi-
automated solution to COTS component selection has the main
improvement in reduction time of COST selection process
which leads to acceleration of development and time to market.
Further, knowledge of available COTS component is another
advantage of the proposed solution.

ISBN: 978-0-9957075-2-8
www.ijeacs.com 24
REFERENCES
[1] M. Elammari and R. Ali, “Towards a Multi-Agent Mechanism for
Software Component Selection”, The International Arab Conference On
Information Technology (ACIT2014), Zarqa University in Jordan.
[2] Rikard Land, Laurens Blankers,Michel Chaudron, Ivica Crnković,
“COTS Selection Best Practices in Literature and in Industry”,
Mälardalen University, School of Innovation, Design and Engineering,
Västerås, Sweden.
[3] X. Burgués, C. Estay, X. Franch, J.A. Pastor, and C. Quer, “Combined
selection of COTS components”, Proceedings of ICCBSS, February,
Orlando, Florida USA, 2002, pp. 54–64.
[4] Lawrence Chung, Kendra Cooper, “COTS-Aware Requirements
Engineering and Software Architecting”, Department of Computer
Science, University of Texas at Dallas.
[5] Zahid Javed, Ahsan Raza Sattar, Salman Afsar, Muhammad Shakeel
Faridi, “An Empirical Study of COTS components Persuasion,
Evaluation & Selection and Integration in software houses Faisalabad,
Pakistan”, IJCSI International Journal of Computer Science Issues, Vol.
9, Issue 6, No 2, November 2012.
[6] Feras Tarawneh, Fauziah Baharom, Jamaiah Hj. Yahaya and Faudziah
Ahmad, “Evaluation and Selection COTS Software Process: The State of
the Art”, International Journal on New Computer Architectures and
Their Applications, The Society of Digital Information and Wireless
Communications, (IJNCAA) 1(2): 344-357, 2011 (ISSN: 2220-9085)
[7] M. Shakeel Faridi & Zahid Javed, M. Haris Abid, Mudassar Ahmed, Dr.
Md Asri Bin Ngadi, “IROTS: A Proposed COTS Evaluation & Selection
Methodology for Component Based Software Engineering in Under-
Development Countries”, Department of Komputing Universiti
Teknologi 2nd International Conference on Advances in Computer
Science and Engineering (CSE 2013), Malaysia, Johor Bahru, Malaysia
[8] S Nazir, MA Khan, S Anwar, H Khan, M Nazir “A novel fuzzy logic
based software component selection modeling”, 2012, International
Conference on Information Science and Applications, 1-6.
[9] R.C. Seacord, S.A. Hissam, K.C. Wallnau, "Agora: A Search Engine for
Software Components", IEEE Internet Computing, vol. 6, no. 2, pp. 62-
70, 1998.
[10] Khan, A.I., Khan, U.A., 2012. "An Improved Model for Component
Based Software Development." Software Engineering 2, no. 4 (2012):
138-146.
[11] Khan, A.I., Alam, M.M., Shariq, M., May 2015. " A Perspective Study
of Intelligent System for Component based Development". International
Journal of Computer Applications 117(4):11-17.
[12] A. Barnawi, M. Rizwan Jameel Qureshi and A. IrshadKhan, “A
framework for next generation mobile and wireless networks application
development using hybrid component based development model”, Int. J.
Res. Rev. Next Gener. Netw. (IJRRNGN), vol. 1, no. 2, (2011), pp. 51-
58.
AUTHOR PROFILE
Asif Irshad Khan, Ph.D., is working as a faculty
member in the department of Computer Science,
FCIT, King Abdulaziz University, Jeddah, Saudi
Arabia. Twelve years of experience as a
professional academician and researcher. Dr.
Khan received Ph.D. in Computer Science and
Engineering from Singhania University,
Rajasthan, India and Master & Bachelor degrees
in Computer Science from the Aligarh Muslim
University (A.M.U), Aligarh, India. He has
published several research articles in leading
journals and conferences. He is a member of the editorial boards of
international journals and his current research interest includes Software
Engineering with a focus on Component Based and Software Product Line
Engineering.
© 2017 by the author(s); licensee Empirical Research Press Ltd. United Kingdom. This is an open access article
distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license.
(http://creativecommons.org/licenses/by/4.0/).

An Elite Model for COTS Component Selection Process

More Related Content

What's hot (18)

Similar to An Elite Model for COTS Component Selection Process (20)

Recently uploaded (20)

An Elite Model for COTS Component Selection Process