Résumé : This paper presents a framework for assessing data and metadata quality within Open Data portals. Although a few benchmark frameworks already exist for this purpose, they are not yet detailed enough in both breadth and depth to make valid statements about the actual discoverability and accessibility of publicly available data collections. To address this research gap, we have designed a quality framework that is able to evaluate data quality in Open Data portals on dedicated and fine-grained dimensions, such as interoperability, findability, uniqueness or completeness. Additionally, we propose quality measures that allow for valid assessments regarding cross-portal findability and uniqueness of dataset descriptions. We have validated our novel quality framework for the German Open Data landscape and found out that metadata often still lacks meaningful descriptions and is not yet extensively connected to the Semantic Web.