In current data warehouse environments there is either no or only insufficient support for a consistent and comprehensive metadata management. Typically, a multitude of largely autonomous and heterogeneously organized repositories coexist. We categorize the major metadata types and their interdependencies within a three-dimensional classification approach. We then investigate how interoperability and integration of metadata can be achieved based on a federated metadata architecture and standardization efforts such as OIM and CWM. In particular, we examine synchronization alternatives to keep replicated metadata consistent. We also give an overview of currently available commercial repositories and discuss interoperability issues to couple data warehouses with information portals.