Seednote

Federation

Shared identity without a shared database

Independently operated Semantic Spaces connected by shared identity and resolution.

RSM ID Learning Center2 min read

Federation is how independently operated Semantic Spaces work together. Each space keeps its own database, software, content lifecycle, and governance policies; what they share is identity — registered authority and the resolution protocol — so that a reference by RID means the same thing everywhere. A federation does not require one physical database, one application, or one hosting provider.

Shared identity makes cross-space work safe only if trust is verified at every hop. Resolvers check prefix authority, binding provenance, lifecycle, and the active space binding before relying on a remote representation, detect resolution loops, and answer untrusted or unavailable instead of improvising.

Example

A Wellzai formation document references the RSM concept Formation by 451.5VJC3VFV46EVR73V and a Musewoods fieldnote by 451.5Z7MCTTCHRBCBS2S. Each resource stays in its own space, revised on its own schedule; the references keep working because they name RIDs, not application routes.

A common misconception

“Federated search merges matching records.” Search may combine results from many spaces using the RID as the deduplication key, but ranking and relevance stay separate from identity. Similar descriptions are never merged unless they share the canonical RID or a verified equivalence backed by authority and evidence.

Federation connects Semantic Spaces, each perhaps run by Locus and explored by Atlas. The federation Viewgraph draws the topology.

In the specification

Federation model: Document 08 §10.1. Federation query: §10.5. Growth: §10.6. Trust, caching, and failure handling: §18.8.