What this is.
Position in the architecture · data residency
Specification · No standalone specification is deposited for this framework. Its authoritative treatment is Chapter 9 of the Enterprise Playbook, ISBN 978-1-0678960-1-0. This page and registry entry REG-09 are the most complete normative statement available in citable form.
Boundary and optimization.
Residency is the boundary. The topology is the optimization. The four patterns are four optimizations under one uncrossable clause, which is why this framework sits under the Boundary Invariant rather than under a MESA altitude, and it is the only one of the nine that does.
Four patterns.
Sovereign Silo keeps everything within one jurisdiction. Federated keeps data in place and moves computation. Regional Hub consolidates within a permitted region. Hybrid mixes them per workload. Selection runs by enumerating the boundary set per jurisdiction and classifying each workload against it, and the BOE Declaration at the fourth step is what makes routing evidence usable during a reconstruction rather than merely available.
What this does not claim.
Asserted, and time-sensitive. Jurisdiction-specific transfer rules age. Every regulatory assertion must be re-verified against primary sources at the time of use, and a boundary set carried forward from a prior deployment without re-verification is not a boundary set. This page names no jurisdiction-specific rule, deliberately.
Asserted. The worked cases in the source treatment are disclosed as composites. They illustrate the selection method and are not evidence that any named institution operates any of the four patterns.
Missing. No decision instrument exists, so the method cannot be applied consistently by two architects working independently.
Citation.
Cite this framework. No standalone specification is deposited. Its authoritative treatment is Chapter 9 of the Enterprise Playbook, ISBN 978-1-0678960-1-0, and its normative statement in citable form is registry entry REG-09 in 10.5281/zenodo.22170112.
In the practice.
- Governed production AI, the discipline this framework belongs to
- Defensible AI, the family and the register
- AI Governance and Compliance Frameworks for the Middle East, the source treatment
- Model risk, the service this framework is applied through
- AI governance in the wiring, on where governance actually lives