Measuring Referential Integrity in Distributed Databases

Distributed relational databases are used by different organizations located at multiple sites that work together on common projects. In this paper, the authors focus on distributed relational databases with incomplete and inconsistent content. They propose to measure referential integrity errors in them for integration and interoperability purposes. They propose local and global referential integrity metrics at three levels: column, table and database. They assume each table can be asynchronously updated at any site and new records are periodically broadcasted to all sites. They explain several distributed query optimization issues.