Scientific data life cycles are becoming more and more demanding. Data amounts are seen to be ever increasing while at the same time advanced IT infrastructures become more common resulting in a much broader user group demanding a high usability. These challenges are met with a distributed science gateway infrastructure, which in turn is complex in nature. In this situation a main part is missing, a generic metadata management concept. The approach and architecture of such a concept, the current status, an evaluation, and an outlook will be described in this publication. The main goals are to enable scientists to easily manage and use large amounts of data and provide a generic way to enable the fast transfers to a multitude of scientific use cases.