Integração entre redes heterogêneas de ativos digitais: modelo híbrido de metadados baseado em meta-modelos dinâmicos

Abstract

The increased production of digital assets in enterprise environments makes digital asset management an increasingly complex task, since we must make use of well-defined metadata, so that you can easily find and use assets from different repositories. However, obtaining a standardized model that cover the peculiarities of all involved, is usually a highly complex task due to heterogeneity between the models. In addition, is considered the fact that existing solutions specialize in solving part of the problem, leaving a gap in the other phases of the integration process. Thus, for the development of this dissertation was proposed, implemented and evaluated a service framework for metadata interoperability that aims to encompass most of the workflow of digital asset management integration process. This framework aims to bring together and complement existing solutions, providing support to semi-automatic identification of correlations between elements of metadata models and a component-based architecture responsible for the dynamic generation of meta-models, which in turn will aggregate the result of the correlation analysis and serve as a unified language, enabling interoperability between digital assets repositories, regardless the metadata standard used by each one. Lastly, the work validation was conducted by performing an experiment using a proof of concept applied to a scenario involving heterogeneous models of digital asset management real systems. The objective of experiment was perform an assessment of quality of the automatically mapping obtained. The result was accuracy up to 98% within the universe of elements that had at least an correlation suggestion, which occurred in 66% of the elements of the models.