Enhancing Deployment Variability Management by Pruning Elements in Deployment Models
Miles Stötzner, Floriment Klinaku, Robin D. Pesl, Steffen Becker · 2023
Since applications often need to be deployed in different variants, deployment technologies, such as Ansible and Terraform, support modeling variability. Unfortunately, applications typically need the combination of multiple deployment technologies, which have proprietary and non-interoperable variability concepts. Therefore, variable deployment models have been introduced to model deployment variability across different technologies by assigning variability conditions to elements to specify their presence. However, the manual modeling of these conditions is repetitive, error-prone, and time-consuming. In this paper, we propose to reduce the modeling effort by a pruning concept, i.e., the automated removal of elements due to consistency issues and semantic aspects. To validate the practical feasibility, we implemented a prototype based on Open-TOSCA Vintner. Moreover, we present a case study that shows that the number of conditions to be modeled is significantly decreased.