FedDW global schema architect: UML-based design tool for the integration of data mart schemas

Extending analytical decision making beyond the boundaries of a single organization is a key challenge of modern Business Intelligence systems. Federated Data Warehouses (FDWs) are an important cornerstone to this end, offering new opportunities for business collaboration and similar scenarios. The FedDW approach provides such a federated architecture with a mediated multidimensional schema over autonomous data marts---with advantages for both, OLAP users and data warehouse administrators. Users comfortably access the global mediated schema with traditional OLAP applications while administrators retain full schema and data management autonomy. Although the underlying concepts are mature, comprehensive design tools for FDWs remain an open issue. To tackle the challenge, this paper presents FedDW Global Schema Architect (GSA), a visual design tool for federations of autonomous ROLAP data marts. FedDW integrates data marts at the schema level which avoids the laborious and error-prone physical DW integration. GSA manages all metadata--the global mediated schema, the import schemas, and the semantic mappings repairing multidimensional heterogeneity among the data marts--within one and the same tool. Its implementation employs the extension mechanisms of Eclipse and is based on the UML and CWM (Common Warehouse Metamodel) standards. Thus, the tool is extensible, intuitive for its users, and supports DW platforms of multiple vendors.