A Practical Approach: Localization-Based Global Metadata Registry for Progressive Data Integration

Until now, many researches based on metadata have been studied for data integration in various fields. However, there exists a problem in that the existing approaches still require much cost to build an initial guideline, which is sometimes called as a registry, global metadata set. The most critical problem of existing researches is that the researches have not been seriously considered on the corresponding domain properties- data level and user level. Most of all, it is difficult to create a standardized guideline on the entire data with the restricted cost. In this paper, we propose a method to progressively integrate databases considering the domain properties under limited cost. For instance, in scientific data, general users are interested in the simple and easy scientific data, which has low complexity. On the other hand, scientists or experts can utilize more detailed and specialized data as well as the general data. It means that an initial data quantity to be integrated is determined by the relationship of the domain properties. We show the implementation results of applying it to scientific data managed by Korea Institute of Science and Technology Information. The proposed method provides a progressive data integration mechanism, and especially it gives successful and efficient output on the creation of a standard guideline in the situation where the given cost is limited.