Handbook-Based High Unit-Value
Abstract: 2010 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must first be obtained from the IEEE. Modern electronic systems typically contain significant amounts of software. Therefore, for a system reliability prediction to be accurate, it must include software reliability predictions. Many of the commonly used software reliability prediction methods are based on estimation models that require empirical test data. Typically, however, this kind of data is not readily available when a software reliability prediction is needed. For this reason, handbook-based methods, such as the Software Capability Maturity Model (CMMÂ®) developed by the Software Engineering Institute (SEI), are commonly used throughout industry. This paper discusses an enhancement to traditional handbook-based software reliability prediction method that facilitates obtaining repeatable results as well as overcoming limitations in modeling initial reliability of the software. This enhancement results in a handbook-based software reliability prediction method that is appropriate for use on high unit-va! lue software components. As with hardware reliability prediction models, the premise of the software reliability prediction model is that the inherent fault density of the software code can be estimated as a function of the quality of the development processes . Once the inherent fault density is estimated as a function of the quality level of the development processes, it is converted to a failure rate based on the reliability design and operating modes of the software. Software reliability growth characteristics also may be included in the reliability prediction model in the same manner used for hardware. In other words, the change in the rate of reliability growth is assessed and the stabilized failure rate is estimated as a function of time. For the traditional handbookbased software reliability prediction method, 48 months is the default time typically selected for residual faults to stabilize following initial release of the software product. The default stabilization time assumed for subsequent software product releases, such as updated software versions, is typically 24 months. Software reliability growth is a function of the organizational processes that will be performed in the field, which might be different than the organizational processes which developed the software.
Keywords: RAMS 2010 Proceedings - Reliability Analysis/Prediction/Estimation - Software Reliability - Reliability Model - Failure Rate