The foundations of data integration

Abstract

Data integration is concerned with combining data residing at independent and heterogeneous data sources in such a way that the user is presented a single view of the data. Data integration in the narrow sense combines data without changing the data sources and without copying data.<br />In the last fifteen years, data integration has become more and more important. Not only large enterprises which own a multitude of data sources and are in need of data integration projects, but also governments which require them for better cooperation between their agencies; health organizations and scientific research projects. Another field of application for data integration is the World Wide Web where millions of data sources exist which can be combined to offer a good search quality for specific topics. Throughout the research on data integration several basic concepts have been developed. Different architectures for data integration systems have evolved. A piviting point was the idea of data integration systems based on an architectural approach using a global schema presenting the users' virtual view of the system; and a set of sources containing the real data. In this context, different approaches emerged on how to connect the diverse data sources with the single global view of the system presented to the user. One of these approaches, namely the "Local as View" (LAV) approach, is closely concerned with the problem of answering queries using views in database theory.<br />This work, based on literature study, has the goal to introduce the fundamental challenges of, and basic approaches to, data integration as the entry point for further occupation for the topic of data integration. The fundamental issues of, and basic approaches to, data integration are presented in this work to provide the reader a general overview, a basic understanding and the necessary background knowledge to further delve into the subject on their own accord.