Virtual Dataspace -- A Service Oriented Model for Scientific Big Data

The massive, distributed, heterogeneous and diverse features of big data have raised challenges to traditional data management systems. As the development and innovation of Data Space, virtual data space (VDS) model is proposed for big data management. Local ontologies are created from data sources. Then the local ontologies are mapped and formed a global ontology. Based on this, access log and user feedback are considered for data evolution. At last, a material scientists-oriented service (materials scholar assistant) is introduced as the application case of VDS.