Combining the Best of Global-as-View and Local-as-View for Data Integration

there are two main basic approach es to data integration: Global-as-View (GaV) and Local-as-View (LaV). However, both GaV and LaV ave th eir limitations. In a GaV approach , ch anges in information sources or adding a new information source requires revisions of a global schA7 and mappings between th global sch ema and source schIW s. In a LaV approach automating query reformulation has exponential time complexitywith respect to query and source sch ema definitions. To resolve these problems, we o#er BGLaV as an alternative point of view that is neither GaV nor LaV. The approach uses source-to-target mappings based on a predefined conceptual target schIz , whzh is specified ontologically and independently of any of th e sources. Th e proposed data integration system is easier to maintain th an both GaV and LaV, and query reformulation reduces to rule unfolding. Compared with othO data integration approach es, our approach combines th e advantages of GaV and LaV, mitigates th e disadvantages, and provides an alternative for flexible and scalable data integration.