The Utility and Practicality of Quantifying Software Reliability

TL;DR

This paper discusses the importance and feasibility of quantifying software reliability to demonstrate low system risks, highlighting that while possible in some cases, it may not be practical universally, affecting standard compliance.

Contribution

It shows that quantifying software reliability is feasible in meaningful cases and discusses its implications for risk management and standards.

Findings

Quantification of software reliability is possible in at least one meaningful case.

Universal practicality of quantification is unlikely, affecting standard adherence.

Quantification can reveal risk-reduction opportunities that standards might miss.

Abstract

We argue that quantifying software reliability is important in demonstrating that system-level risks are As Low As Reasonably Practicable (ALARP). Furthermore, we demonstrate that such quantification is possible in at least one meaningful case. It is, however, unlikely to be practical in every case. This means it is unlikely to be included as an explicit objective in standards. Hence, for those cases where software reliability can be quantified, merely following a standard may lead to risk-reduction opportunities being missed.