A comparison of methods for calculating the duration of software reliability demonstration testing, particularly for safety-critical systems

There are several methods for calculating the required testing duration for demonstrating a certain level of reliability at a given level of confidence, with or without failures; or, alternatively, for estimating the demonstrated reliability after a certain number of tests, with or without any failures. The methods are based on different models and assumptions and provide different results. The purpose of this short communication is to review and compare eight relevant methods, some of which have never been used for software reliability demonstration.