Managing multisource databases: between theory and practice (Practice Oriented)

摘要

The need of managing multiple data sources in information systems is increasing and managing integrated access to information that is spread over multiple, disparate, and heterogeneous sources is being among the critical problems encountered in modern information systems. In fact, modern applications and processes used in contemporary companies often solicit distributed sets of data. These data are most of the time characterized by a high overlapping rate as referring to the same or similar entities, entailing, therefore, the necessity of consolidating it. This process of integrating and sharing scattered data is known, in the industrial domain, as the critical part of entity resolution and Master Data Management (MDM) projects, the core business of the Data Quality business unit of A.I.D.. Our article describes the best practices observed by A.I.D. after a decade of managing multisource information systems. This practice-oriented study is preceded by a theoretical overview of the numerous techniques and solutions of handling multisource information systems that underlines the gap existing between the industrial world and the academic one.