Generic Conceptual Framework for Handling Schema Diversity during Source Integration

DataBase Integration Systems (

DIS

) aim at providing a unified view of data stored in different

heterogeneous

local sources through a global schema. The schemas of global and local sources are represented by

a

 − 

priori

known

logical representations. This makes the potential deployment model of the DIS, like for Data Warehouse (

DW

) systems,

rigid

and

inflexible

. To overcome these limitations, we claim that the integration process should be completely performed at the

conceptual

level

independently of any implementation constraint. After studying existing database (

DB

) integration systems, we propose through this paper a generic conceptual framework for

DB

schema integration. This framework is generic as it

subsumes

most important

DB

integration systems studied. It is defined based on the description logic formalism. We then instantiate it through a case study considering a set of Oracle semantic

DB

, that are

DB

storing their own conceptual model, generated using Lehigh University BenchMark (LUBM). These

DB

participate in the construction of a semantic

DW

.