Background: Software Reliability Growth Models (SRGMs) are the most widely used mathematical models to monitor, predict and assess the software reliability. They play an important role in industries to estimate the release time of a software product. Since 1970s, researchers have suggested a large number of SRGMs to forecast software reliability based on certain assumptions. They all have explained how the system reliability changes over time by analyzing failure data set throughout the testing process. However, none of the models is universally accepted and can be used for all kinds of software.
Objectives: The objective of this paper is to highlight the limitations of SRGMs and to suggest a novel approach towards improvement.
Methods: We have presented the mathematical basis, parameters and assumptions of the software reliability model and analyzed five popular models, namely Jelinski-Moranda (J-M) model, Goel Okumoto NHPP model, Musa-Okumoto Log Poisson model, Gompertz Model and Enhanced NHPP model.
Conclusion: The paper focuses on challenges like flexibility issues, assumptions, and uncertainty factors of using SRGMs. It emphasizes considering all affecting factors in reliability calculation. A possible approach has been mentioned at the end of the paper.
Keywords: SRGMs, NHPP model, software reliability, software quality, testing, software engineering.