Challenges in Developing a Software Architecture for a National Research Data Infrastructure in Earth System Sciences

Abstract

Research data infrastructures (RDIs) are evolving and driven by diverse initiatives ranging from international to local, e.g., institutional scales, who try to build sustainable and useful services. The German national data infrastructure for Earth System Sciences (NFDI4Earth) aims to support researchers in 1) discovering and exploring relevant data sources, 2) data publication and curation, 3) solving research data management problems and 4) creating and publishing information products. In the development of the  software architecture for NFDI4Earth, we face challenges of computational, social, cultural, and strategic nature. Here, we are going to present an overview of these challenges and early outcomes and to reflect  first lessons learned from the initial concept and development phase of the NFDI4Earth architecture.Starting with the (meta) data layer, the landscape of existing ESS services and repositories is diverse and features various metadata and data, like governmental (meta) data following INSPIRE, OGC, and ISO19xxx. This diversity demands harmonising and linking concepts that fit to standards for metadata, data and services, such as OGC APIs or ISO19xxx, as well as to Semantic Web concepts, e.g., FAIR Digital Objects, and provide extension points for (newly developed) specific formats. At the same time, the software stack and technologies of the business layer should consider interoperability, openness and sustainability aspects while providing a flexible solution to manage the distributed metadata. Moreover, in our case, activities on developing (meta) data and business layer concepts also include coordinating a software developer team with different scientific and technological backgrounds spread across several institutions.NFDI4Earth is located in a dynamic landscape of ESS services and repositories which are often not sustainably funded. Hence, we need to implement  practices and collaborations to link or integrate further software, services, and information products so that an up-to-date living and evolving architecture serves the needs of researchers.