Software Failures: An Overview

Abstract

Since major software failures often result in disasters ranging from financial loss to loss of lives, preventing their recurrence is absolutely necessary. A post-mortem investigation is required to identify their root cause and implement appropriate countermeasures. Current approaches to software failure investigations are limited and often result in returning the software system back to normal execution as quickly as possible. In the process of doing so, the software system is left vulnerable to a reoccurrence of the same type of software failures. This chapter defines the concept of a software failure and then reviews the problems of major software failures. The aim is to determine how to improve the accuracy of their root-cause analysis in order to prevent the reoccurrence of major accidents. A review of recent cases of major software failures from different industries, such as the medical domain, is given to demonstrate the reality and seriousness of software failures. These software failures are then analysed so as to identify limitations and establish requirements for improvement of the software investigation process. These requirements form the basis for the design of a near-miss management system (NMS).