International Journal

International journal of of Recent

Research Advances
and Review in in Multidisciplinary
Health Research, May -2015
Sciences, July -2014
sZ

International Journal of Recent Advances in Multidisciplinary Research

Vol. 02, Issue 05, pp.0418-0420, May, 2015

Research Article
STUDY SOFTWARE RELIABILITY USING DEFECT DENSITY IN SOFTWARE ENGINEERING
*1Dr. Ramesh R. Manza and 2Yogini B. Patil
1Department of Computer Science and Information Technology, Dr. Babasaheb Ambedkar Marathwada
University, Aurangabad, 431004(MS) India
2Department of Computer Science, Shantidevi Chavan institute of Polytechnic, Chalisgaon 424101(MS), Jalgaon,

India

ARTICLE INFO ABSTRACT

Article History: To fulfill software requirements and to satisfy the customer, software must be tested by all angles of
th
Received 19 February 2015 users and developer .But the question is that when to stop testing or whether it is enough to test
Received in revised form software, if it not tested properly by software engineer then it is handover to customer, if we want to
21th March, 2015 give quality s/w product then we must calculate software reliability as it one of quality measure. s/w
Accepted 25th April, 2015 reliability is probability of failure free operation of software. usually tester focus on testing functional
Published online 31th May, 2015 or normal behavior of units of s/w and it is in developer’s environment but software reliability focus
on operational behavior of software in natural setting that is in presence of customers and users, it is
Keywords: need of software community to combine both functional and operational approach so that testing
Defect density, should be done effectively and reliability should be accurately measured. In this paper issue is
S/W reliability, highlighted that using metrics defect density and statistical modeling software reliability can be
Functional requirements. measured.

INTRODUCTION
Non-functional requirements define system properties such as
As we know software engineering is i) The application of Reliability, Response time and storage requirements etc. Here
systematic, disciplined, quantifiable approach to development, some requirements are product requirements and some are
operation and maintenance of software that is application of external requirements. Product requirements specify that
engineering to software. ii) ”The study of approaches as in the delivered product must work in a way such as execution speed,
above statement” (IEEE). It contains different phases reliability etc. (as in tree dia.1). The external requirements are
external to system and development process e. g
i) Analysis interoperability requirements, legislative requirements etc.
ii) Design Requirements are validated by testing phase of software
iii) Coding engineering. Every time customer/users executes program it
iv) Support to develop software. tested. System is useless if non-functional requirements are not
met.
Software Requirements
Software reliability
It can be classified as

i) Functional It is amount of time that the software is available for use as

ii) Non-functional. indicated sub attributes such as maturity, fault tolerance,
recoverability etc. It is also defined as probability of failure free
Functional Requirements specify which output should be software operations for specified period of time in specified
produced from given inputs.They describe the relationship environment (ANSI/IEEE, 1991). A simple measure of
between the input and output of the system. reliability is mean time between failure (MTBF). It is
calculated as MTBF=MTTF+MTTR here, MTTF=mean time to
*Corresponding author: Dr. Ramesh R. Manza failure MTTR=mean time to repair. In addition to reliability
Department of Computer Science and Information Technology, Dr. measure, we must develop a measure of availability. It is
Babasaheb Ambedkar Marathwada University, Aurangabad, probability that a program is operating according to
431004(MS) India requirements at a given point in time and is defined as,
International Journal of Recent Advances in Multidisciplinary Research 0419

Availability=(MTTF/(MTTF+MTTR))*100%. S/W Reliability To achieve predictable reliable software traditional software

field includes to calculate it and develop it to improve software reliability models takes defect density metrics as input and
development process it is calculated by number of field failures produce reliability quantity as the output. In the future function
along with some additional information such as time at which should be developed what should the reliability process for
the failure was found, in which part or module it was found, the what goal and what will be metrics such feedback is useful for
state or mode of software at that time, the nature of failure s/w reliability engineering process (Software Reliability
etc.ISO 9000-3 (ISO, 1991). S/W reliability is one of quality Engineering: A Roadmap, ?). In this research paper data from
factors of ISO 9126 which define reliability as amount of time several projects show a significant relationship between size of
that the software is available for use as indicated by the sub module and its defect density. Here it is shown that defects may
attributes: maturity, fault tolerance, recoverability. be due to interconnections of module or due to complexity of
individual modules. controlling module size distribution defect
Related work and contribution density can be reduced. Exponential distribution shows that
when modules broken into number of modules and bring then
Software reliability is important to achieve a ideal level of together then defect density is minimal (Yashwant K. Malaiya
performance in limited time frame. It is used to measure defects and Jason Denton, 2000).
that cause software failures in different mode and behavior of
software with fixed time. As software is present in different MATERIALS AND METHODS
forms like programs, rules, data, documentation .If defects
occurs it affects on transportation, telecommunication, military, In this research if S/W reliability is to be study then different
entertainment offices, aircraft and business software reliability testing strategies must be addressed specify product
is very important to study. In this paper classification on requirements in a quantifiable manner long before testing
software reliability models based on various dimension is given commences. State testing objectives explicitly in measurable
(Latha Shanmugam and Dr. Lilly Florence, 2012). In this paper terms e.g test effectiveness, test coverage, mean time to failure,
different models are discussed as Time Between Models, Fault the cost to find and fix defects, defect density or frequency of
Count Models, Fault seeding models, Input Domain based occurrence, test work hour per regression test etc..Understand
models, Identical Error Behavior Models, Independently the user of the software (through use cases) and develop a
Distributed inter-failure models. Here it is reflected that testing profile for each user category. Develop a testing plan that
and reliability are different. Testing doesn’t focus how it will emphasizes rapid cycle testing to get quick feedback to control
operate in the field, here test case are designed for boundary quality levels and adjust the test strategy. Build robust software
conditions and software is tried to break or trace than normal that is designed to test itself and can diagnose certain kinds of
condition to find different failure paths .Software reliability on errors. Use effective formal technical reviews as a filter prior to
other side focus on operation profile of software so that testing to reduce the amount of testing required. specify product
accurate reliability estimation and prediction occurs so in future requirements in a quantifiable manner long before testing
testing and reliability should apply together (Software commences. Understand the user of the software (through use
Reliability Engineering: A Roadmap, ?). cases) and develop a profile for each user category. Develop a
testing plan that emphasizes rapid cycle testing to get quick
In this paper it is reveal that collecting metrics data for feedback to control quality levels and adjust the test strategy.
reliability prediction is difficult. It can be obtain from Build robust software that is designed to test itself and can
organization such as NASA, Microsoft, IBM, Cisco etc. diagnose certain kinds of errors. Use effective formal technical
So extracting information of metrics data using statistical reviews as a filter prior to testing to reduce the amount of
and probability foundation, gives right direction for testing required. State testing objectives explicitly in
investigation in software engineering (Cheng et al., 1997). measurable terms as mention above (Glib, 1995).

tree dia :1
International Journal of Recent Advances in Multidisciplinary Research 0420

Defect density=Total number of known defects/KLOC, where Institute of Electrical and Electronics Engineers. ANSI/IEEE
KLOC=LOC*1000.Here relationship between defect density standard glossary of software engineering terminology,
and three repository metrics, i.e number of downloads, software IEEE Std. 729-1991.
size and number of developers (Cobra Rahmani and Deepak ISO 9000-3. Quality management and quality assurance
Khazanchi, 2010). In this paper regression analysis tools used standard- part 3: guidelines for the application of ISO 9001
which offers many statistical options such as Histogram plot, to the development, supply and maintenance of software.
Analysis of Variance, R-squared and the significance of Switzerland: ISO; 1991
observed regression line (p-value). P-value is significance level Volume 2, Issue 10, October 2012. ISSN: 2277 128X.
to accept or reject tests. Here defect density is compared with International Journal of Advanced Research in. Computer
number of downloads, software size, and number of developers. Science and Software Engineering.
Regression analysis doesn’t show relationship between defect Software Reliability Engineering: A Roadmap
density and number of downloads in OSS Project (i. e Open Michael R. Lyu Computer Science and Engineering
Source Software Project) (Cobra Rahmani and Deepak Department The Chinese University of Hong Kong, Hong
Khazanchi, 2010). Kong [email protected]
Cheng, J., D.A. Bell, and W. Liu, 1997. “Learning Belief
Research question Networks from Data: An Information Theory Based
Approach,” Proceedings of the Sixth International
Using non linear regression and different statistical tools like p- Conference on Information and Knowledge Management,
value, R-value, we can explain different trend in OSS Project or Las Vegas, pp. 325-331.
exhibit relationship between Defect Density and number of International Journal of Computational Engineering Research
downloads in OSS Project which is not by linear regression this (ijceronline.com) Vol. 2 Issue. 4 2250-3005(online) August|
may be good research question. 2012 Page 1060 “A Selective Survey and direction on the
software of Reliability Models” By, Vipin Kumar.
Conclusion Latha Shanmugam, Dr. Lilly Florence. Volume 2, Issue 10,
October 2012 ISSN: 2277 128X International Journal of
Using statistical modeling and software reliability theory, Advanced Research in Computer Science and Software
models of software failures as function of execution time can Engineering Research Paper “An Overview of Software
also be developed. It is observed that by collecting metrics Reliability Models “.
during software testing and making use of existing software (GIL 95) Glib, T. “what we fail to do our current testing
reliability models, it is possible to develop meaningful culture” Testing Techniques news letter, Software
guidelines for S/W reliability and it may gives new rule, new Research, Inc. San Francisco, 1995
model, or hypothesis that will be topic of research. Westfall, L. 2010. “Defect Density,” Available: http://www.
westfallteam.com/Papers/defect_density.pdf.
REFERENCES Yashwant K. Malaiya and Jason Denton, 2000. “Module Size
Distribution and Defect Density”, 0-7695-0807-3/00 by
“Software Engineering: A Practitioner’s Approach, 6/e; IEEE.
Chapter 14: Software Testing Techniques”, R.S. Pressman Cobra Rahmani and Deepak Khazanchi, “A Study on Defect
and Associates, Inc., 2005. Density of Open Source Software”, DOI 10.1109/ICIS.
2010.11.

*******