Generation and evaluation of design alternatives for software architectures (GenEDA)

• Applicants: Professor Dr. Michael Goedicke; Professorin Dr. Maritta Heisel
• Subject Area: Software Engineering and Programming Languages
• Term: from 2009 to 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 141396433

We aim at deepening and broadening our methodology developed in the ongoing project GenEDA by taking into account further research questions. In our work in GenEDA, we successfully treated security and performance as external quality attributes. For the proposed project, we strive forinvestigating evolvability as an internal quality attribute and integrating it in our method.In GenEDA we have developed a method to generate design alternatives with regard to quality requirements. To select a suitable architecture design alternative, interactions and conflicts among quality requirements such as security, performance, and the new evolvability requirements should be detected and resolved. In the proposed project, we want to provide a method for interaction and trade-off analysis to detect and resolve such conflicts.Third, we aim at providing support for software engineers in their decision-making process when choosing among different design alternatives. For this purpose, we make use of optimisation problems and state-of-the-art techniques to select the optimal design alternative with respect to a certain fitness function.To satisfy the needs of different user groups, different variants of software products are developed. This leads to the research area of software product line engineering (SPLE). Current SPLE approaches focus mainly on functional requirements and pay less attention to quality requirements. We want to provide variability with focus on quality requirements. Introducing quality requirements adds another dimension of variability to the software product family, which we want to deal with in an enhancement of the GenEDA method.All in all, the results obtained when addressing the above-mentioned topics will lead to a much more powerful support for designing quality-aware software architectures. We first plan to investigate the above-mentioned topics one by one, relating them to the current GenEDA methodology. In the second phase of the proposed project, the interplay between the investigated areas will be analyzed, and we will come up with a proposal how to balance the different approaches. During the third project phase, we apply the enhanced GenEDA method to a real-life scenario and develop a benchmark, which supports future research by a broad research community. In parallel to the other phases we will integrate our existing GenEDA tools into one single tool and extend it to provide support for the enhanced GenEDA method.

• DFG Programme: Research Grants