Towards a Scalable and Collaborative Information Integration Platform and Methodology

Abstract

The quality and scalability problems in information integration are still not sufficiently solved by current integration approaches. In this paper, we propose an integration approach that adopts a uniform, context-aware, and collaborative methodology enabling efficient reuse to come to a scalable integration platform. We present a two-step methodology that splits up the traditional process of information integration into two separated phases. In the first phase, the Mapping phase, heterogeneous models are matched and mappings are created between the corresponding entities. We further introduce a community and contextual dimension for each mapping. In the second phase, the Commitment phase, a final, application-specific alignment is created in a certain integration format. We argue that this methodology enables a more scalable, efficient and collaborative integration process. We have developed a platform which is based on this methodology and have also done a preliminary evaluation by going through a real use case.