Data-Sovereign Architecture for Quality-As-A-Service Across Industrial Supply Chain Partners

The digital orchestration of value chains in Industry 4.0 is crucial for exploiting the potential of globalized partner networks. However, ensuring data sovereignty, standardizing data models, and incorporating recent advances in Artificial Intelligence to increase partner trustworthiness are just a few ongoing challenges to enable flexibility, ensure security, reduce bureaucracy, and increase resilience and scalability among companies. In this context, we propose a Quality-as-a-Service (QaaS) architecture and a reference implementation to enhance trust and automation of product quality assurance in industrial supply chains. Our solution eliminates the need to transfer large datasets or deep-learning models. Instead, it leverages recent advances in Simulation-to-Reality (sim-to-real) domain gap reduction to train supervised deep-learning models using only partners’ design files, enabling in-situ product quality assessments on 2D/3D shop floor images. The generated quality data is standardized according to Industrial Digital Twin Association (IDTA) submodels and securely stored within an Asset Administration Shell (AAS), ensuring data integrity. Furthermore, peer-to-peer data transfers are carried out through Eclipse Data Space Connectors (EDCs), enabling sovereign asset access to industrial partners within the secure dataspace. Our framework enables automated quality insights across companies without transmission overhead, thereby fostering transparency, trust, and efficiency in globalized manufacturing processes.