2025/11/16 by Masoudi, Sepideh, Daly, Mark Edward Michael, Kiesel, Jannis
#Distributed #FOS: Computer and information sciences #Parallel #and Cluster Computing (cs.DC)
paper · doi:10.48550/arxiv.2511.12667
As data mesh architectures grow, organizations increasingly build consumer-specific data-sharing pipelines from modular, cloud-based transformation services. While reusable transformation services can improve cost and energy efficiency, applying traditional cloud design patterns can reduce reusability of services in different pipelines. We present a Kubernetes-based tool that enables non-intrusive, deferred application of design patterns without modifying services code. The tool automates pattern injection and collects energy metrics, supporting energy-aware decisions while preserving reusability of transformation services in various pipeline structures.