Bringing business objects into extraet-transform-load (ETL) technology

Abstract

Business objects represent the key concepts that a business needs to operate such as people, services, products, etc. but transforming these objects to and from existing data models can be difficult. Business objects have traditionally been represented in a backend data store using relational databases, and techniques for transformation must work with these stores. But such access may violate the encapsulation these business objects require, and so conventional approaches may not provide an adequate solution. In this paper, we examine how to use Extract-Transform-Load (ETL) tools to provide business object transformations. We show how to solve some of these issues by using pluggable components and introduce customized operators for ETL tools. We demonstrate our solution using a commercial ETL system that allows access to several Product Data Management systems and illustrate the use of our technique with several case studies drawn from various industries. © 2008 IEEE.