**Federation** in [[Information technology|information technology]] refers to an arrangement in which multiple distinct systems, organizations, or administrative domains interoperate under a common set of agreed standards and protocols while each retaining autonomous control over its own resources and governance. Federated systems allow users, services, or data to interact across organizational or technical boundaries without requiring centralized ownership or a single controlling authority. The concept is applied across numerous IT disciplines including [[Identity management|identity management]], [[Database federation|database management]], [[Cloud computing|cloud computing]], and [[Network architecture|network architecture]]. In the context of [[Identity federation|identity and access management]], federation enables users authenticated by one organization's [[Identity provider|identity provider]] to access resources hosted by another, using trust relationships established through standards such as [[Security Assertion Markup Language|SAML]], [[OpenID Connect]], and [[OAuth]]. This model underpins [[Single sign-on|single sign-on]] (SSO) across enterprise applications and between organizations, allowing credentials issued by one domain to be accepted by another without requiring separate accounts or [[Authentication|authentication]] steps. Major implementations include academic identity federations such as [[Shibboleth (software)|Shibboleth]] and enterprise [[Cloud computing|cloud]] identity services offered by providers including [[Microsoft Azure Active Directory|Microsoft Entra ID]] and [[Okta]]. Federation also applies to [[Data management|data management]], where a [[Federated database system|federated database system]] provides a unified query interface over multiple heterogeneous [[Database|databases]] that remain independently managed and physically distributed. This approach allows organizations to integrate data from disparate sources without migrating or consolidating it into a central repository. In [[Data mesh|data mesh]] architectures, federation is a governing principle, distributing data ownership to [[Domain-driven design|domain]] teams while maintaining interoperability through global standards for data discovery, access, and [[Data governance|governance]]. Similar federated principles are applied in [[Federated learning|federated machine learning]], where [[Machine learning model|models]] are trained across distributed datasets without centralizing the underlying data, preserving [[Data privacy|privacy]] while enabling collective [[Model training|model improvement]].