Résumé: The software industry is increasingly facing the need for continuous deployment of systems. This leads to the adoption of continuous activities, including planning, integration, and deployment (a.k.a. Continuous Software Engineering (CSE)). At the same time, systems should exhibit high-quality architectures, which are often achieved through architecture evaluation methods. However, there is little insight of how such evaluation happens in the context of CSE. To cover this gap, we investigate in this work the state of the art of continuous evaluation of software architectures in CSE, including agile processes like SCRUM. For this, we systematically examine the literature to collect and summarize evidence. Our results show a diversity of means for evaluating architectures in continuous mode to support the continuous evolution of systems. We also found how such evaluation has been incorporated within continuous development processes and agile processes like SCRUM and Crystal. We finally derive the main trends and open issues in the area, aiming to support the community to better understand and further consolidate the field of continuous evaluation of software architectures.

Langue: Anglais