An Empirical Investigation of the Effort of Creating Reusable, Component-Based Models for Performance Prediction

摘要: Model-based performance prediction methods aim at evaluating the expected response time, throughput, and resource utilisation of a software system design before implementation. Existing use monolithic, throw-away models or component-based, reusable models. While it is intuitively clear that development requires more effort, actual higher amount effort has not been quantified analysed systematically yet. To study we conducted controlled experiment with 19 computer science students who predicted two example systems applying an established, monolithic method (Software Performance Engineering) as well our own component-based (Palladio). The results show model creation Palladio approximately 1.25 times than SPE in experimental setting, resulting having comparable accuracy. Therefore, some cases, can already be justified, if they are reused least once.