An Overview of Ontology-Driven Data Integration

Introduction

New software requirements have emerged because of innovation in technology, specially involving network aspects. The possibility enterprises, institutions and even common users can improve their connectivity allowing them to work as they are at the same time, generates an explosion in this area. Besides, nowadays it is very common to hear that large enterprises fuse with others. Therefore, requirements as interoperability and integrability are part of any type of organization around the world. In general, large modern enterprises use different database management systems to store and search their critical data. All of these databases are very important for an enterprise but the different interfaces they possibly have make difficult their administration. Therefore, recovering information through a common interface becomes crucial in order to realize, for instance, the full value of data contained in the databases (Hass & Lin, 2002).

Thus, in the ‘90s the term Federated Database emerged to characterize techniques for proving an integrating data access, resulting in a set of distributed, heterogeneous and autonomous databases (Busse, Kutsche, Leser & Weber, 1999; Litwin, Mark & Roussoupoulos, 1990; Sheth & Larson, 1990). Here is where the concept of Data Integration appears. This concept refers to the process of unifying data sharing some common semantics but originated from unrelated sources. Several aspects must be taken into account when working with Federated Systems because the main characteristics of these systems make more difficult the integration tasks. For example, the autonomy of the information sources, their geographical distribution and the heterogeneity among them, are some of the main problems we must face to perform the integration. Autonomy means that users and applications can access data through a federated system or by their own local system. Distribution (Ozsu & Valduriez, 1999) refers to data (or computers) spread among multiple sources and stored in a single computer system or in multiple computer systems. These computer systems may be geographically distributed but interconnected by a communication network. Finally, heterogeneity relates to different meanings that may be inferred from data stored in databases. In (Cui & O’Brien, 2000), heterogeneity is classified into four categories: structural, syntactical, system, and semantic. Structural heterogeneity deals with inconsistencies produced by different data models whereas syntactical heterogeneity deals with consequences of using different languages and data representations. On the other hand, system heterogeneity deals with having different supporting hardware and operating systems. Finally, semantic heterogeneity (Cui & O’Brien, 2000) is one of the most complex problems faced by data integration tasks. Each information source included in the integration has its own interpretation and assumptions about the concepts involved in the domain. Therefore, it is very difficult to determine when two concepts belonging to different sources are related. Some relations among concepts that semantic heterogeneity involves are: synonymous, when the sources use different terms to refer to the same concept; homonymous, when the sources use the same term to denote completely different concepts; hyponym, when one source contains a term less general than another in another source; and hypernym, when one source contains a term more general than another in another source; etc.