Model-driven Engineering Techniques for Dependable Adaptive Software Systems

点击翻译按钮获取中文摘要

***Software systems are nowadays the backbone of our modern society diverse activities, including transportation, communication, health, and entertainment. In addition to their inherent complexity, these systems are more and more deployed on highly dynamic runtime platforms (e.g. Cloud Computing) and they interface with an uncertain environment (e.g. Cyber-Physical Systems). In this context, the software systems should be designed with an adaptation capacity and should continuously verify stringent dependability requirements. ******Model-driven software engineering (MDE) is now a well-established approach, which is centered on design-time modeling to support coping with the software system complexity. With the emergence of the models at runtime concept, MDE techniques can be extended to support the engineering of adaptive software systems. However, using these runtime models to support and manage the dependability of adaptive software systems still faces several challenging issues. First, we point to the lack of specific framework to support runtime dependability modeling in contrast to the numerous standard frameworks for design-time modeling. Having the framework to build such models, there is a need for efficient analysis techniques, which enable using the dependability runtime models to generate optimal adaptations. Third, the dependability requirement is a multidimensional concept encompassing several dimensions including availability, reliability, safety, and performance to mention just a few. These are inter-dependent, inter-related and potentially conflicting dimensions. As a consequence, keeping the dependability runtime model consistent on one hand and also continuously synchronized with the running system throughout the system adaptions is challenging. ******In this research program, we propose to leverage the model-driven software engineering approach combined with the models at runtime concept in order to define, develop and validate modeling techniques to support the development and management of dependable adaptive software systems.**