Resumen
Global sustainability research requires an integrated multi-disciplinary effort underpinned by a cyber infrastructure able to harness big data and heterogeneous information systems across disciplines. Two approaches are possible to achieve the interoperability desired across such systems and data: federating, and brokering. This position paper argues that the former is appropriate to single discipline or domain environments, but that brokering is more scalable and effective in complex multi-disciplinary domains. The paper identifies the principles of brokering, and gives examples of practical implementation relating to data discovery, semantic searching, and data access achieved in the EuroGEOSS project. The value of the EuroGEOSS brokering approach has been demonstrated in extending the data resources available through the Global Earth Observation System of Systems (GEOSS) from a few hundred to over 28 million in a matter of 3 months. Brokering offers therefore a real chance to facilitate truly multi-disciplinary big data science and address the scientific challenges of our time.