The effect of testing on the reliability of single version and 1-out-of-2 software systems

Abstract

Some experimental results related to the effect of testing on the reliability of software systems are presented. The experiments were performed on a simulator designed for behavioral modelling of software systems in the presence of design faults. The topics of the experiments are defined as: the influence of the testing profile on the reliability evolution of a single version system, the influence of the design faults size on the evolution of system reliability in different testing profiles and the effect of the testing strategy on the reliability of 1-out-of-2 fault-tolerant scheme. In the last topic, three testing strategies are considered: individual, in coalition and back-to-back testing.