Planning and certifying software system reliability

Poore, Jesse H.; Mills, Harlan D.; Mutchler, David

doi:10.1109/52.207234

Cited by 87 publications

(30 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Of which the most important statistic, the single use reliability, estimates "the probability of the software executing a randomly selected use without a failure relative to a specification of correct behavior." [15] The low single use reliability observed in this example (0.270545355) was due to the high number of failed test cases (1, 538 out of 2, 248).…”

Section: Automated Statistical Testing and Test Case Analysismentioning

confidence: 99%

An Automated Testing Framework for Statistical Testing of GUI Applications

Lin

Xue

2015

International Conferences on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

show abstract

Section: Automated Statistical Testing and Test Case Analysismentioning

confidence: 99%

An Automated Testing Framework for Statistical Testing of GUI Applications

Lin

Xue

2015

International Conferences on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

show abstract

“…When considering no multiple failures, the model may be interpreted as an elaboration of Cheung's model [Che80]. We emphasize that Cheung's model is integrated in the Cleanroom Reliability Manager [PMM93] whose aim is planning and certification of component-based software system reliability.…”

Section: ) As It Can Be Expected This Model Is a Discretetime Countementioning

confidence: 99%

A Poisson Limit Theorem for Reliability Models Based on Markov Chains

Ledoux

2006

Communications in Statistics - Theory and Methods

View full text Add to dashboard Cite

In this article, we deal with a class of discrete-time reliability models. The failures are assumed to be generated by an underlying time inhomogeneous Markov chain. The multivariate point process of failures is proved to converge to a Poissontype process when the failures are rare. As a result, we obtain a Compound Poisson approximation of the cumulative number of failures. A rate of convergence is provided. Key-Words Discrete-time multivariate point process, time inhomogeneous Markov chain, Compound Poisson approximation, filtering (AMS 2000)Primary : 60F99, 60G55, 60J10 Secondary : 60F17, 60G42, 60K15, 93E11 IntroductionA basic issue in reliability modeling is what happens when the system becomes reliable. In other words, what becomes the model when the failure parameters tend to be smaller and smaller? We discuss such an issue for a class of reliability models which are based on discrete-time Markov chains. Specifically, we consider a failure process which is generated by the dynamics of a non-homogeneous Markov chain. Clumping of failures is also considered. Such a failure process may be thought of as a discrete-time multivariate point process. Then, we investigate its asymptotic distribution when the failure parameters converge to 0. As it can be expected, the asymptotic distribution is of Poisson-type. Under a condition on the rate of the convergence of the underlying non-homogeneous Markov chain to stationarity, we provide a rate of convergence for the distance in variation. As a result, we obtain a Compound Poisson approximation of the cumulative number of failures. The ergodic and irreducible cases for the limit Markov chain are discussed. The derivation * James Ledoux, INSA 20 avenue des Buttes de Coësmes, 35043 Rennes Cedex, FRANCE; email: james.ledoux@insa-rennes.fr 1 of the main results is based on elementary discrete-time stochastic calculus and on basic convergence results for non-homogeneous Markov chains reported in [HIV76,BDI77]. It is not intended to deal with the mildest assumptions for deriving Poisson limit theorems. The assumptions here, are expected to be easily checked and general enough to be of value in the reliability framework.The problem of the Poisson approximation of a flow of rare events is a very old topic. Our purpose is not to review the huge literature on this topic. We limit ourselves to mentionning standards methods for deriving such Poisson approximations and only discuss the closest works to ours. The counting process associated with a discrete-time point process is nothing else but a sum of dependent random variables. There is a vast literature on limit theorems for such processes, especially on the central limit theorem and functional theorems. In a much more general setting, a complete account for such an "limit theorems approach" is in [JS89]. For a martingale approach of the Poisson approximation of point processes, some specific references are [Bro82, Bro83, KLS83] (see also the references therein). The basic tool is the compensator associated with a point ...

show abstract

“…However, the software component failure data available is not sufficient to apply these models. The information about the succeeded and failed executions during the testing are used to find the reliabilities of software components [1], [4], [6], [10].…”

Section: Reliability Of Software Componentsmentioning

confidence: 99%

An Experimental Study on Reliability Estimation of GNU Compiler Components A Review

Chinnaiyan¹,

Somasundaram²

2011

IJCA

View full text Add to dashboard Cite

In today's software arena most of the software's are developed using the component based software development methodology. Hence it is necessary to ensure that the developed software possess high reliability as perceived by the consumer before the software release. Many researchers have proposed various analytical models for assessing the reliability of component based software systems, where the black-box testing procedure is used for reliability assessment for evaluating the software components where internal workings of the software components are not evaluated. So it is necessary for the software developers to use white box testing technique for ensuring high reliability of the software components which yields reliable software system. This paper proposes a renewed reliability engineering approach and it is elucidated with the real software system case study with GNU compiler.

show abstract

Planning and certifying software system reliability

Cited by 87 publications

References 5 publications

An Automated Testing Framework for Statistical Testing of GUI Applications

An Automated Testing Framework for Statistical Testing of GUI Applications

A Poisson Limit Theorem for Reliability Models Based on Markov Chains

An Experimental Study on Reliability Estimation of GNU Compiler Components A Review

Contact Info

Product

Resources

About