<p>Due to rapid increase in frequency and sophistication of cyber threats along with the complexity of modern IT systems, the Cyber Ranges (CR) have been widely adopted for testing cybersecurity solutions and training professionals. Unfortunately, most CRs operate in isolation, therefore they do not collaborate effectively with other CRs. This isolation causes several issues such as limited resources and expertise availability, low scalability, and inability to truly replicate complex and coordinated cyber threats. To address these issues, new frameworks and technologies that enable the federation of CRs are being explored by researchers and practitioners. This paper maps the FCR landscape and organizes it around a simple taxonomy of federation models (centralized, decentralized, and hybrid). We compile past and ongoing efforts, align them to this taxonomy, and synthesize a layered reference architecture of FCR that clarifies roles, interfaces, and typical flows. We investigate benefits and limitations of FCR platform, and analyze FCR development challenges across interoperability, cross-range operations, security and trust, and governance. From this, we identify gaps which shows the mains issues of slow adoption to FCRs. These issues include the lack of common packaging (e.g., heterogeneous infrastructures and varied technology stack), heterogeneous data exchange formats, and lack of consistent identity/policy models in existing ranges. Finally, we outline practical next steps to help researchers, practitioners, and policymakers design or improve FCR platforms.</p>