Deriving and validating fault metric for object oriented systems using use case diagram

Abstract

A significant quality factor reliability is an important requirement for all software systems. Reliability analysis and evaluation for a system is necessary for the success of these systems. This paper proposes a fault metric called use case fault metric (UCFM) for object-oriented systems using use case diagram and is built on the theory of Fault Tree Analysis (FTA). The goal is to address an important issue of measuring reliability of object-oriented systems during initial stages of software development. This goal is achieved by extending the model-driven software development process with the concept of Fault Tree Analysis (FTA). The Fault Tree analysis is performed in bottom-up way to compute quantitatively the proposed metric called UCFM. The proposed approach transforms the use case in an equivalent Fault Tree. This FTA is used to compute a metric that quantifies the probability of occurrence of faults in different use cases. The UCFM value is computed at use case, use case diagram and system level. The proposed metric is validated theoretically using measurement theory and practically using Kaner’s framework.