Software Safety Analysis to Identify Critical Software Faults in Software-Controlled Safety-Critical Systems

Abstract

Modern systems are increasingly software intensive because of the progress of technology and theproliferation of computers in everyday life. Computers control everything possible from microwave ovens to complex weapon systems. However, software can have a severe impact on the safety of systems, as some high profile accidents like Therac-25 and Ariane5 have shown. Despite the risks, software increasingly is making its way into safety-critical systems. A general purpose software engineering process is insufficient by itself to produce safe and reliable software. While traditional testing and other dynamic analysis techniques are best for uncovering functional errors they are inadequate whenever a computer-based system can cause injury or death. Therefore, software for safety-critical systems must deal with the hazards identified by safety analysis in order to make the system safe, risk-free and fail-safe. Certain critical software faults in critical systems can result in catastrophic consequences such as death, injury or environmental harm. The focus of this paper is a new approach to software safety analysis based on a combination of two existing fault removal techniques. A comprehensive software safety analysis involving a combination of Software Failure Modes and Effects Analysis (SFMEA) and Software Fault Tree Analysis (SFTA) is conducted on the software functions of the critical system to identify potentially hazardous software faults. A prototype safety-critical system - Railroad Crossing Control System (RCCS), incorporating a microcontroller and software to operate the train on a track circuit is described.