Developing Highly-Reliable Software: The MAP Approach

Software reliability is becoming the dominant concern in software development. This particularly holds for the development of safety-critical control systems. Any failure of these systems can result in catastrophic loss of life and property. Hence, it is essential to ensure with a high degree of confidence that these systems meet their reliability requirements prior to deploying them in the field. In this paper, we present a unified approach to the development, reliability assessment and run-time safety enhancement of process-control systems. Evolutionary program design facilitates incremental construction of software reliability MAP (Measured Assurance Prediction). MAP confidence estimates are used as run-time triggers for switching to a reduced capability backup version. The backup version is invoked if its MAP provides a higher degree of confidence in correctness than that for the primary version of the software.