The paper reports on an innovative approach to the management of Cognitive Radio Networks (CRNs). As part of achieving this functionality, a new architecture implemented at the application layer is presented. The research challenges for developing a CRN management solution are on addressing, sensing and prediction, routing, decision making and middleware. Cognitive radio networks are expected to address and to solve several important challenges like, e.g., opportunistic spectrum access, spectrum heterogeneity, network heterogeneity and the provision of diverse Quality of Service (QoS) to different applications. Consequently, a number of management functions are necessary to address the associated challenges with the management of CRNs. The paper provides a brief introduction to these components while focusing on the CR routing, addressing and content representation.