Predicting Reliability of Software Using Thresholds of CK Metrics

Predicting Reliability is one of the key function of a software system. Many of the software fail due to unreliability. The demand of reliable software is increasing day-by-day. In industry, information on reliability is available too late in the software development process. Hence any corrective action becomes unaffordable. A step towards the remedy to this problem is the ability to provide a threshold for the reliability of a software product. Object oriented metrics are most beneficial and reliable for the estimation of the threshold for reliability. In this paper, we use the Chidamber and Kemerer Metrics to assess the threshold values for reliability. A tool is designed and developed called ëJava Class Analyzerí which extracts the values of the metric parameters from the source code. These values are evaluated against the threshold values of the metrics from the literature. It provides a threshold for the software reliability. The result provides a standard against which the software reliability can be evaluated and necessary corrective actions can be implemented.