Envisioning and proposing Data Mesh for Research Data Management in the Engineering Sciences

Mario Moser, Tobias Hamann, Anas Abdelrazeq, Robert Schmitt

In Research Data Management (RDM), data publishing infrastructures play a crucial role for efficient data provisioning and reusage. Data repositories (generic or discipline-specific) serve for this. Nevertheless, they focus rather on technical aspects without including sociological elements; they struggle to cover the heterogeneous nature of research data (formats, sources); and they are typically centralised, leading to increased complexity in operation and maintenance. In industrial data management, the Data Mesh concept as a decentralised and socio-technical approach has been introduced. Data is handled as products for increased usability, ownership is shifted to the respective domains experts, and a federated governance achieves standardisation while allowing discipline-specific decisions. Based on literature review, the distributed characteristics and further requirements of (engineering) research are mapped with the Data Mesh concept. In this envisioning, Data Mesh and its design principles overall appear appropriate as research data publishing infrastructure. A high level architecture is presented leveraging existing RDM components. Although, as differences in details become apparent, items for further adaptions of Data Mesh for RDM are pointed out.