Scientific computing requires many and large volumes of complex structured data and metadata that are scattered across data centers. Researchers find it difficult to discover the specific data they need for their research. Traditional search engines, such as Google, are not effective in most of these cases. Moreover, some of the scientific data are confidential and are not publicly indexed. Therefore, it is necessary to develop a custom solution that satisfies the researchers' need for flexible search. This article introduces the Data Discovery Service designed to serve the Virtual Research Environment (VRE) during the VI-SEEM project. The solution is based on a specially configured and upgraded version of the CKAN platform. The main installation procedure is described, as well as the authors' contributions for the purpose of regularly harvesting data from different sources and updating the available content on which user queries are to be executed.