This thesis provides the conceptual foundations for the extension of federated database technology with the management of versioned data. Database federations are a means for the a posteriori integration of data managed by heterogenous (component) systems. In the engineering domain, these data are typically versioned. The thesis first determines the requirements of a "versioned" federation compared to an "unversioned" one, and then develops a comprehensive set of features to fulfil these requirements. These features include a version model for the federation that allows to integrate versions from the component systems' heterogenous version models, and a mechanism for the propagation of versioning-relevant operations from the component systems to the federation and vice versa. A prototypical implementation for a simplified environment is described to demonstrate the feasibility of the proposed solutions.