A Time Dependent Error Detection Rate Model for Software Performance Assessment with Applications.

Abstract
The objective of this study was to develop a parsimonious model whose parameters have a physical interpretation, and which can be used to predict various quantitative measures for software performance assessment. With this objective, the behavior of the software failure counting process (N(t), t equal to or greater than zero) has been studied. It is shown that N(t) can be well described by a non-homogeneous Poisson process (NHPP) with a two parameter exponentially decaying error detection rate. Several measures, such as the number of failures by some prespecified time, the number of errors remaining in the system at a future time, and software reliability during a mission, have been proposed in this report. Models for software performance assessment are also derived. Two methods are developed to estimate model parameters from either failure count data or times between failures. A goodness-of-fit test is also developed to check the adequacy of the fitted model. Finally, actual failure data are analyzed from two DOD software systems. One is a large command and control system and the other a Naval data analysis system. (Author)

This publication has 0 references indexed in Scilit: