Execution-history based reliability estimation for component-based software: considering reusability-ratio and interaction-ratio

Component-based software engineering emphasizes ‘development by means of reuse’ and ‘development meant for reuse’. Whether the system is simple or complex one, the estimation of better reliability remains a crucial concern. The main purpose of this work is to propose a method for reliability estimation and the computation of execution time of component-based software. In this work a metric named ‘reusability-ratio’ is introduced as a factor of reliability estimation. We focus on assessing and exploring reusability of components by defining reusability-ratio for newly developed, mutated (fully-qualified as well as partially-qualified) and off-the-shelf components. On the basis of interactions among components, one more metric is defined called ‘Interaction-ratio’. Interaction-ratio is used as another factor of reliability estimation. Based on the interactions made by components, a graph is constructed, namely ‘Component-Interaction Graph’. The structure of the component-interaction graph depends on the probability of interaction of components as well as the probability of the selection of different path execution-histories. Results obtained through experimental case study conclude that the reusability compete imperative function in the reliability of the component-based applications. Pre-tested, qualified and pre-configured artefacts consume lesser time and are more reliable as compared to the new component constructs. Metrics proposed in this work are well suited to estimate the reliability of component-based software and therefore proved promising to analyze the performance of the software.