Jelinski Moranda Model
- Predicts how frequently a software system will fail over time to inform testing and maintenance.
- Applied during development and ongoing maintenance to plan testing effort and update schedules.
- The model assumes the failure process follows a negative binomial distribution.
Definition
Section titled “Definition”The Jelinski-Moranda model is a statistical model used in software reliability engineering. It is used to predict the failure rate of a software system over time. The model is based on the assumption that the failure rate of a software system follows a negative binomial distribution, which means that the number of failures follows a certain probability distribution.
Explanation
Section titled “Explanation”The model provides a probabilistic prediction of software failures over time. By modeling the failure process (under the assumption of a negative binomial distribution), it yields estimates of how often failures will occur. These estimates can guide decisions about how much testing is required and how often software updates or maintenance should be scheduled to maintain reliability.
Examples
Section titled “Examples”Development of a new software system
Section titled “Development of a new software system”During the development phase, the software is tested extensively to identify and fix any bugs or defects. The Jelinski-Moranda model can be used to predict the failure rate of the software once it is released to the market. This can help the development team determine how much testing is needed and how often the software should be updated to ensure its reliability.
Maintenance of an existing software system
Section titled “Maintenance of an existing software system”As the software is used over time, it is likely to encounter more failures. The Jelinski-Moranda model can be used to predict the failure rate of the software over time, allowing the maintenance team to plan for regular updates and maintenance to ensure the software continues to operate reliably.
Use cases
Section titled “Use cases”- Estimating required testing effort and update frequency during development.
- Planning maintenance and update schedules for deployed software based on predicted failure rates.
Notes or pitfalls
Section titled “Notes or pitfalls”- The model relies on the assumption that the failure rate follows a negative binomial distribution; results depend on this assumption being appropriate for the software and data.
Related terms
Section titled “Related terms”- negative binomial distribution
- software reliability engineering