Reliability growth modelling of software for process control systems

Abstract

Reliability growth models are the most widely known and used in practical applications. Nevertheless, the performances of these models limit their applicability to the control and assessment of the attainment of very low reliability levels. One of the reasons for their insufficient predictive quality is the scarcity of information they use for reliability estimation. They do not take into account information that could be easily collected from code observation, such as the code complexity, or during testing such as the coverage obtained. The paper analyzes the problems encountered using reliability growth models for the evaluation of software used in a measurement equipment control system.<>