A service-oriented data integration architecture and the integrating tree

Data integration is the problem of combining data residing at distributed sources. Despite, its importance, there is no universal architecture available because of the limitations of available technologies. The emerging of SOA and the application of ontology provide a new starting point for solving the problem. By making use of services and ontology technology, the paper presents a novel architecture: the system heterogeneity is solved by services; the global model is constituted by two levels of ontologies to solve the semantic heterogeneity, to provide, the semantic reasoning, to facilitate the query, and to make automatic query answering. Key features and the anatomy of the architecture are provided in the paper, the comparisons with related work are also given. Based on the architecture, a formal integrating tree model of data integration is presented. By studying the tree, many related problems can be studied in depth and quantitatively.