When does it pay to make software more reliable?

Substantial work has been done in the area of software fault analysis, prediction and evaluation. Some work considers the relationship between failures and cost. Unfortunately, current work does not address the issues of revenues, return on investment, or time-value of money. These shortcomings limit the usefulness and acceptance of software reliability within industry. To address these shortcomings, we have developed an economic reliability analysis (ERA) framework at the University of Virginia, that fuses reliability engineering methods with economic analysis. We introduce the ERA framework to the software reliability community and fuse it with the Musa/Okumoto NHPP software reliability model to estimate the economic impact of operational software faults. This framework defines a set of economic vectors that are calculated using the software reliability model along with a few additional financial elements. This extended software reliability model is then used to evaluate an operational software system with three different proposed software upgrades. All four software systems are analyzed and compared. This analysis provides a more complete method to apply software reliability modeling techniques to existing software systems and proposed design changes.