Evaluation of safety-critical software
Communications of the ACM
Safeware: system safety and computers
Safeware: system safety and computers
Software Reliability Engineered Testing
Software Reliability Engineered Testing
The Infeasibility of Quantifying the Reliability of Life-Critical Real-Time Software
IEEE Transactions on Software Engineering
| Hi-index | 0.00 |
Given a target probability of functional failure on demand for a system, a corresponding dangerous failure rate for the system can be derived, provided that a proof-test interval for the function is known. IEC 61508, and related standards, requires that this calculation be performed, for certain kinds of systems that are required to provide safety functionality on demand. This paper explains why it is necessary to consider what constitutes a proof-test interval for a function, and then considers what this means for software. We show that there are several problems with the proof-test concept, as applied to software, and describe the problems this presents to practitioners wanting to derive safety integrity levels for system safety functions.