Within the scientific landscape, every discipline has big data management problems due to the heterogeneity and/or to the large volume of scientific data. NFDI provides new FAIR approaches to overcome these problems within domains and cross-cutting among different domains. An integral part of the solution space is to provide scalable methods for integrating data and metadata to foster an easy data exchange among users and services. In particular, this working group examines physical and virtual integration methods, schema matching, semantic annotation, and knowledge graphs. Our envisaged work plan follows an agile approach with an iterative development of results and their validation against an increasing number of use stories obtained from various NFDI consortia. In addition, this project develops a recommendation for an architecture and provides the next steps for a reference implementation.