n South African Computer Journal - A software fault classification model : research article
|Article Title||A software fault classification model : research article|
|© Publisher:||South African Computer Society (SAICSIT)|
|Journal||South African Computer Journal|
|Author||P.K. Kapur, O. Shatnawi and V.S.S. Yadavalli|
|Publication Date||Dec 2004|
|Pages||1 - 9|
|Keyword(s)||Fault severity, NHPP, Non-homogenous poisson process, Software reliability engineering, Software reliability growth model, SRE and SRGM|
This paper presents a software reliability growth model (SRGM) for classification of software faults during testing phase based on a non-homogeneous Poisson process (NHPP). The model assumes that the testing phase consists of three processes namely, failure observation, fault isolation and fault removal. The software faults are classified into three types namely, simple, hard and complex according to the amount of testing-effort needed to remove them. The removal complexity is proportional to the amount of testing-effort required to remove the fault. The testing-effort expenditures are represented by the number of stages required to remove the fault after the failure observation or fault isolation (with delay between the stages). The time delay between the failure observation and the subsequent fault removal is assumed to represent the severity of the fault. The more severe the fault, the more the time delay. The fault is classified as simple if the time delay between failure observation, fault isolation and fault removal is negligible. If there is a time delay, it is classified as a hard fault. If the removal of a fault after its isolation involves an even greater time delay, it is classified as a complex fault. Therefore, the model incorporates a logistic learning-process function during the removal phase of hard and complex faults. Accordingly, the total fault removal phenomenon is the superposition of the three processes. The model has been validated, evaluated and compared with the well-established NHPP models by applying them to actual software reliability data sets cited from real software development projects. The results are fairly encouraging in terms of goodness of fit, predictive validity and software reliability evaluation measures.
Article metrics loading...