A rule-based approach for estimating the reliability of component-based systems

Reliability is one of the most important nonfunctional requirements for software. Accurately estimating reliability for component-based software systems (CBSSs) is not an easy task, and researchers have proposed many approaches to CBSS reliability estimation. Some of these approaches focus on component reliability and others focus on glue code reliability. All of the approaches that have been proposed are mathematical. However, because reliability is a real-world phenomenon with associated real-time issues, it cannot be measured accurately and efficiently with mathematical models. Soft computing techniques that have recently emerged can be used to model the solution of real-world problems that are too difficult to model mathematically. The two basic soft computing techniques are fuzzy computing and probabilistic computing. In this paper, we focus on four factors that have the strongest effect on CBSS reliability. Based on these four factors, we propose a new fuzzy-logic-based model for estimating CBSS reliability. We implemented and validated our proposed model on small applications, and the results confirm the effectiveness of our model.

► We propose a fuzzy logic based approach for estimating reliability of component based systems (CBSs).
► For formulation of rules we use fuzzy c-means clustering.
► We validate the result by using center of gravity.