Zero trust is a security paradigm whose fundamental philosophy is that every access to a resource must be explicitly verified, without assuming trust based on origin or identity. In a federated environment composed of multiple domains, ensuring zero trust guarantees for accessing shared resources is a challenge, as information on requesters is generated by their originating domain, yet requires explicit verification from the domain owning the resource. This paper proposes a method for federating zero trust architectures, ensuring the preservation of zero trust guarantees when accessing federated resources. The proposed approach relies on remote attestation, enabling continuous authentication and monitoring of requesters, without requiring intrusive software installations on every device within the federation. Moreover, this paper proposes a proof-of-concept architecture that combines several open-source products, to build an architecture with advanced zero trust maturity level. The feasibility of the proposed federation method is demonstrated through this proof-of-concept, providing detailed information on the federation procedure and its implementation.

Information technology (IT) security has been, and largely is, based on compartmentalization. To implement compartmentalization, system access privileges are granted depending on the topological location of systems, grouped into perimeters, with network mechanisms (firewalls, VLANs, ) enforcing isolation between perimeters, thus implicitly trusting systems based on their location. However, history has shown that such trust is misplaced. This has led to the emergence of an alternative paradigm, called zero trust. After contextualizing the history of IT and the emergence of zero trust for securing networks, this article presents a taxonomy of zero trust models and architectures, summarizing the goals and core principles of zero trust. Furthermore, an in-depth description of state-of-the-art technologies and methods, for transforming perimeter-based architectures to mature zero-trust architectures, is provided. This article presents a formalization of zero trust and of optimal zero-trust architectures, to which traditional architectures migrate, as well as a method for positioning migrating architectures relative to this ideal of zero trust, with as purpose of enabling a clearer understanding of the benefits and risks induced by a migration to zero trust. Finally, this article analyses the benefits, and drawbacks, of zero trust, focusing on the security properties granted by zero trust, as well as the vulnerabilities introduced.