Prediction of software operational reliability using testing environment factors

A number of software reliability models have been developed to estimate and to predict software reliability. However, there are no established standard models to quantify software reliability. Most models estimate the quality of software in reliability figures such as remaining faults, failure rate, or mean time to next failure at the testing phase, and they consider them ultimate indicators of software reliability. Experience shows that there is a large gap between predicted reliability during development and reliability measured during operation, which means that predicted reliability, or so-called test reliability, is not operational reliability. Customers prefer operational reliability to test reliability. In this study, we propose a method that predicts operational reliability rather than test reliability by introducing the testing environment factor that quantifies the changes in environments.