Loosely-coupled distributed architecture of the Spatial Data Infrastructure of the URBAN GEO BIG DATA project

Spatial Data Infrastructures (SDIs) are systems in which a network of interconnected but autonomous components is used for the exchange and communication of geospatial information between several stakeholders belonging to distinct organizations. Their successful exploitation requires technologies, policies, common standards and harmonized rules both at the syntactic level, i.e., technical communication protocols for service implementation and at the semantic level, i.e., sharing the geospatial data and their meaning. Within the project "URBAN GEOmatics for Bulk data Generation, Data Assessment and Technology Awareness (URBAN GEO BIG DATA)" (http://www.urbangeobigdata.it/) we are experimenting the definition and application of novel technological solutions for fostering the fruition and synthesis of Big Geo Data (BGD) by public administrators and the citizens of five Italian urban areas (Milan, Naples, Padua, Rome and Turin). Specifically, the project aims to improve the knowledge of urban areas by exploiting the fruition of the vast availability of Earth Observation (EO) data sources on soil consumption for long-term monitoring and IoT data on mobility. An acentric loosely-coupled distributed SDI architecture has been conceived with multiple endpoint nodes to implement such an SDI for multi-source geospatial data sharing on the Web effectively and efficiently, in compliance with interoperable OGC standards. The choice of a distributed architecture offers a more flexible and low-cost solution with respect to a centralized one. In fact, as data increase, it is possible to perform a horizontal scaling by adding new hardware resources and new endpoint nodes when needed, while a centralized solution would require scaling up the hardware which generally is more costly. Within the project, nodes are deemed for the geospatial data sharing (data nodes), which have been set up by partitioning the physical organization of data thematically. This choice allows each project's unit to maintain locally its own data for which they are responsible and for which local expertise is available for their creation, organization and updating. To this end, each data sharing node hosts a geo Web server that manages thematically homogenous geospatial data related to the whole spatial domain of interest, i.e., the five cities. In such a way, for example, all information about the characterization of the long-term behaviour of the deformation mechanisms in all the urbanized areas is managed and shared by the IREA node, those related to Land Use and Land Cover (LULC) are managed by the Politecnico di Milano node, while those related to mobility are maintained by Politecnico di Torino node. Data in each endpoint node are organized according to a suitable data model deemed appropriate for representing a given theme in compliance with OGC standards. Finally, an endpoint node, implemented by the Politecnico di Milano enables the consultation, visualization and analysis of all the distributed geospatial data, by accessing the requested data remotely in a transparent way to the final user who is not asked to be aware of the location of the accessed data. The use of Free and Open Source Software (FOSS) has been a key aspect for ensuring the interoperability, replication and reutilization of the applications while implementing the URBAN GEO BIG DATA SDI.