Software Reliability Modeling for Dynamic Development Environments
Nestor Ruben Barraza · 2019
The last advances in software engineering introduced dynamic processes in software development and testing that improved the classical waterfall methodology. This dynamic environment proposed by the Agile, Rational Unified Process and test-driven development methodologies among others, demands software reliability models that take into account an iterative coding and testing stages. This is quite different from the large waterfall testing stage where code is constantly introduced in order to fix failures, which leads to the well-known software reliability growth models mainly based on non-homogeneous Poisson processes. This chapter discusses a new model based on a pure birth process. Pure birth processes have been largely applied in software reliability, see for example H. Okamura and Dohi, and the well-known non-homogeneous Poisson processes are a special type of pure birth processes where the failure rate is a non-linear function of time.