Cyber-Risk Informatics

8 SOFTWARE RELIABILITY MODELING AND METRICS IN CYBER-RISK

LEARNING OBJECTIVES

• Present an overview of the state of the art and science in software reliability.
• Review historically traditional models and recently discovered advances in modeling and form a basis of comparison with analytical techniques studied in hardware reliability.
• Categorize reliability models: (i) time-between-failures (Poisson and binomial types), (ii) failure count, (iii) Bayesian, (iv) static (nondynamic), and (v) others.
• Study the “reliability growth models” decomposed into (i) negative exponential class of failure times, (ii) Gamma, Weibull, and (iii) other classes of failure times.
• Study Musa–Okumoto, Goel–Okumoto, and Sahinoglu’s compound Poisson failure-count models with examples, as well as the parametric and nonparametric models.

8.1 INTRODUCTION, MOTIVATION, AND METHODOLOGY

Software reliability models do not consider the same statistical techniques as those used in hardware reliability models described earlier [1–3]. Software reliability is defined as the quality measure of a piece of software. It is generally defined as the probability of failure-free performance of computer logic within a specified period of time (mission time) under the predefined conditions and environment (operational profile) encountered. Simplicity and practicality are two key factors in bridging the gap between the state of the art and that of applied software reliability modeling. The assumptions must ...