Domain-spanning Maintainability Estimation of Information and Manufacturing Automation Systems

点击翻译按钮获取中文摘要

Information systems and manufacturing automation systems face very similar challenges with respect to evolution. Both struggle with changing requirements and architectures that have grown over time. Frequently, information systems as well as manufacturing automation systems are in operation over decades while they are continuously modified. Modifications may include corrections, improvements, or adaptations of the system to changes in its environment (e.g., hardware or software platform, technology, as well as user properties) and in requirements. Thus, maintainability is an important quality aspect, especially for long-living systems. Accordingly, research in methods to improve maintainability are of scientific interest, e.g. in the DFG Priority Program Design for Future -- Managed Software Evolution (SPP 1593). However, the effect of maintainability improving actions cannot be scientifically empirically validated without the explicit use of case studies. There are two case studies involved in SPP 1593, namely CoCoME and PPU. This proposal has been set up to support the evolution of the case studies in the context of SPP 1593 and to provide service to SPP 1593 projects using these case studies. This support is given in form of a scientific work programme that is superimposed with both case studies. The proposed project targets at the exploration and development of a method to analyse maintainability of software-intensive systems by deriving maintenance task lists, i.e. specific changes to be performed, from architecture descriptions. For assessing a systems architecture, the task list is applied to make estimations about the maintainability of the proposed system design that are of high practical importance. Using the task lists, multiple architecture alternatives can be compared and, by that, change efforts can be identified. In interviews with researchers of the context of SPP 1593, we first analyse requirements and specify maintainability scenarios based on which we prepare evolution scenarios of CoCoME and PPU. We propose meta-models to reflect domain-specific constructs. Moreover, as information systems and manufacturing automation systems are variant-rich, we investigate how models of a systems variability can be used for maintainability estimation. Based on the meta-models, we come up with a novel procedure to maintainability estimation by identifying change efforts for information systems and manufacturing automation systems. The procedure considers the interrelations between processes and the system as well as their effects on maintainability estimation while spanning over all life-phases of the system.