Developing and analyzing high-level designs for safety-critical software systems

Abstract

Due to the rapid increase in computer-related technologies, industry is continuing to put higher demands on software-controlled systems. These demands are sometimes placing software in total or partial control over critical system functions such as shutting down nuclear reactors, navigating planes, determining radiation dosages, and identifying military targets. Unfortunately, certain faults in such a system can result in catastrophic consequences such as death, environmental harm, or damage to the equipment. In order to help ensure that these software systems do not cause accidents, software engineers are concentrating on techniques and methodologies for improving various life-cycle phases such as the software's high-level-design phase. This paper discusses background issues relating to software safety and then presents a framework for developing and analyzing safety-critical high-level designs. The framework provides a design-independent approach and includes activities such as high-level-design hazard analysis, developing verificative information, and conducting a critical review.

A review of research and methods for producing high-consequence software

“…N-Version programming can be applied to multiple fault classes including design faults, compiler faults, language complexity faults, and implementation faults (see Table 5) [20].…”