Software Component Dependability Assessment

Quantifying the quality of a logical system in a probabilistic manner is very difficult. Our goal is to do so for each software component before it is linked into a system with other software components. A secondary goal of our work is to develop a system-level theory that calculates the dependability of a system solely from the dependability scores of the components that comprise it. Our contribution to the problem of assessing component dependability employs a unique combination of testing and fault injection techniques. Also, our approach uses a unique probabilistic method called the "Squeeze Play" that is profile independent, a characteristic needed for solving this problem. Assessing the dependability of a software component that will operate in many different environments is extremely difficult. Our research addresses this issue. The last twenty years have produced a series of papers stating why traditional software reliability models are flawed [1, 2, 8, 7, 9, 3]. The underlying..