Defence and National Security: IES, HQDM and Allied Standards
Defence is the one domain where ontology stopped being a research topic and became procurement policy. In February 2024 the chief data officers of the US Department of Defense, the Office of the Director of National Intelligence and the Chief Digital and Artificial Intelligence Office agreed on Basic Formal Ontology and the Common Core Ontologies as the baseline standards for formal ontology work across defence and intelligence. In the UK, the Information Exchange Standard has become the common ontology of the National Digital Twin Programme’s Integration Architecture, and a government department is now the named custodian of an OWL file. If you want to know who is actually paying for ontologists, start here.
Exchange standards and modelling standards are not the same thing
The first distinction a defence data officer needs is the one nobody explains. An exchange standard tells two systems what message to send: field names, cardinalities, code lists, a schema to validate against. A modelling standard tells you what the world is made of: what counts as an object, a state, an event, a role, and under what identity conditions two records refer to the same thing. Exchange standards make messages parse. Modelling standards make integrated data mean something when you join it five years later.
NATO’s command and control lineage sits firmly on the exchange side. The Joint C3 Information Exchange Data Model, JC3IEDM, is produced by the Multilateral Interoperability Programme, a consortium of over twenty NATO and non-NATO nations, and is ratified under NATO STANAG 5525. Version 3.1.4 superseded the earlier 3.0.2 baseline. Its successor artefact, the MIP Information Model, keeps the operational concepts of JC3IEDM while restructuring for modularity and semantic strictness. These are relational and UML models with rigorous code lists, not OWL ontologies, and they are excellent at what they do: getting a corps to talk to a battalion across national boundaries. What they do not give you is a formal account of identity and change that survives merging with logistics, personnel, intelligence and sensor data.
The UK route: IES and its 4D commitments
The Information Exchange Standard, usually still called IES, is a 4D ontology for information exchange developed inside UK government with contributions from Dstl, the Ministry of Defence, the Metropolitan Police Service, the Foreign, Commonwealth and Development Office, the Home Office, the Department for Business and Trade and HMRC. Technical support came from Telicent and Aurora Consulting. The original dstl/IES4 repository was archived on 4 March 2025; the authoritative, actively maintained home is now the IES-Org ont-ies repository, with the Department for Business and Trade as repository custodian and the contributing departments as joint decision makers. Code is MIT licensed and documentation carries the Open Government Licence v3.0, so there is no access barrier at all: you can clone the ontology tonight.
What does IES actually do? Three things that matter. First, it commits to a 4D worldview: things persist through time by having temporal parts, so a person’s state as a serving officer is a modelled entity in its own right with its own period of existence, rather than an attribute that silently overwrites history. Second, it treats representation as first-class: a name, a passport number or a call sign is a thing that represents another thing over a period, which is the only sane way to model aliases, cover identities and renamed units. Third, it is deliberately modular and extensible, so domain groups add extensions without forking the core. The National Digital Twin Programme is currently extending it with IES:Building, a 4D extension for the built environment.
The deployment example is unusually easy to point at. The National Digital Twin Programme, run by the Department for Business and Trade, uses IES as the common ontology of its Integration Architecture, an open source data sharing infrastructure released under Apache 2.0 with documentation under the Open Government Licence. The programme has also been procuring ongoing ontologist support to extend and refine IES. That is a public, funded, named demand signal for exactly the skills in this course.
What breaks · coalition integration by name and call sign
Two allied feeds are merged on the strongest signal available: the unit designator. Both describe TASK FORCE ORION, so the join is made. One feed means the formation that stood up in 2019 and was disbanded in 2023; the other means the formation that inherited the designator in 2025, with different constituent units, a different commander and a different area of operations.
The merged picture attributes the older formation’s dispositions, engagements and equipment holdings to the current one, and no field is malformed anywhere: the falsehood lives entirely in the identity assumption, so no schema validator will ever see it. This is the reused-tag failure from the register’s lesson 5 at coalition scale, and it is exactly what the representation commitment in the UK route exists to prevent. In a 4D exchange model, a designator is a thing that represents another thing over a period, so ORION-the-name and ORION-the-formation are different entities and the two feeds’ referents are distinguished by their periods rather than welded by their spelling. Note also what the 2024 US baseline decision does and does not settle here: adopting a top-level ontology fixes the categories, but if the exchange payload still carries a bare designator with no period, the identity ambiguity survives the standard.
HQDM: the other British 4D lineage
HQDM is Matthew West’s High Quality Data Model, the 4-dimensionalist reference model that grew out of Shell’s engineering data work and the same intellectual family as BORO and IDEAS. Where IES optimises for exchange between organisations, HQDM optimises for completeness of the underlying pattern catalogue: individuals and their states, classes and kinds, activities and associations, ordinary physical objects, systems and system components, plus the relationships between them, all under extensional identity.
Access is straightforward. GCHQ publishes a Java implementation of HQDM together with an OWL rendering, and MagmaCore is a GCHQ library that lets HQDM objects and patterns be created and consumed as RDF linked data on top of Apache Jena. Both are Apache 2.0 with Crown Copyright. A national intelligence agency publishing a permissively licensed 4D ontology toolkit is a strong signal about where the UK public sector thinks its data foundations should sit.
Because IES and HQDM share a 4D ancestry but differ in granularity and naming, the practical question in UK government projects is not “which one” but “how do these two line up”. That is a crosswalk problem, and it is why our team published the IES-HQDM crosswalk as an open artefact: explicit mapping predicates, machine-checkable tests, and a record of where the two models genuinely disagree rather than a table of hopeful equivalences. The lesson generalises: in defence you will spend more time bridging standards than authoring them.
The US route: BFO plus CCO as baseline
Basic Formal Ontology is a small realist top-level ontology, standardised as ISO/IEC 21838-2:2021, with roughly three dozen classes and a full first-order axiomatisation alongside the OWL release. It is 3D at heart: continuants that persist wholly at each moment, occurrents that unfold in time. The Common Core Ontologies sit beneath it as a mid-level suite of eleven modules covering agents, artefacts, events, information entities, units of measure, geospatial concepts and more. CCO was created in 2010 by CUBRC with IARPA funding, is available under a BSD-3 licence, and is documented in arXiv 2404.17758.
The 2024 adoption decision matters because it changes procurement language. If you are bidding into a DoD or IC data programme, “BFO-conformant” and “CCO-aligned” are now scoreable requirements rather than intellectual preferences. It also creates a real technical tension for allied programmes: BFO’s 3D continuant model and the IES/HQDM 4D model answer the identity-through-change question differently, which is precisely the material covered in the 3D versus 4D lesson earlier in this course.
| Standard | What it is and does | Maintainer | Licence and access | Real deployment |
| IES | 4D OWL/RDF ontology for cross-organisation information exchange; states, events, representations, periods of existence | Department for Business and Trade as custodian, with Dstl, MOD, Met Police, FCDO, Home Office, HMRC | Public GitHub; MIT for code, Open Government Licence v3.0 for documentation | Common ontology of the NDTP Integration Architecture; IES:Building extension in progress |
| HQDM | 4D reference data model and pattern catalogue with extensional identity; Java plus OWL rendering | GCHQ publishes the implementation; model originated with Matthew West | Public GitHub; Apache 2.0, Crown Copyright | MagmaCore used to publish and consume HQDM patterns as RDF linked data over Apache Jena |
| BFO | Realist top-level ontology, 3D continuant and occurrent split, OWL plus FOL axiomatisation | BFO community, standardised through ISO/IEC | ISO/IEC 21838-2:2021 for the standard text; OWL artefacts openly available | Baseline standard for DoD and IC ontology work from 2024 |
| CCO | Eleven-module mid-level suite under BFO: agents, artefacts, events, information, measurement, geospatial | Originally CUBRC with IARPA funding; open GitHub community | BSD-3, public repository | Adopted with BFO as DoD and IC baseline; described in arXiv 2404.17758 |
| JC3IEDM 3.1.4 | Command and control information exchange data model with strict code lists; message-level interoperability | Multilateral Interoperability Programme with the MIP-NATO Management Board | Ratified under NATO STANAG 5525; distribution through national defence channels | Multinational C2 exchange from corps to battalion; MIP Information Model as successor |
Check yourself
A coalition partner delivers a payload that validates cleanly against the agreed exchange schema, yet the integrated picture is wrong. Which failure is most likely, and which control catches it?
- The payload conforms structurally while carrying identity claims with no temporal qualification; only a modelling standard plus content validation catches it.
- The payload uses an unratified schema version; schema validation catches it.
- The payload violates a code list; schema validation catches it.
- The payload is inconsistent as an OWL ontology; a reasoner catches it.
Show the answer, and why each wrong option is wrong
A is correct. This is the sector’s characteristic failure and the reason the lesson separates exchange standards from modelling standards: a message can parse perfectly and still assert that two different formations are one, because parsing constrains form and identity is a modelling commitment.
B and C are real defects and both are caught upstream by ordinary schema validation, which is precisely why they are not the interesting risk. D inverts the diagnosis the lesson gives explicitly: the failure mode here is rarely an inconsistent ontology, it is a conformant-looking payload with missing or under-qualified mandatory content, which is a shapes problem rather than a reasoner problem. An assurance story that names only the reasoner is running at half coverage, exactly as the register’s testing lesson found.
What a defence data officer actually needs
Five things, in this order. A declared top-level commitment, so that identity and change are decided once rather than per project. A published extension policy, because every programme will need terms the core lacks and uncontrolled extension is how standards die. Crosswalks with evidence, meaning mapping files that carry predicates, provenance and test results rather than a spreadsheet of guesses. Validation in the pipeline, because in this sector the failure mode is not an inconsistent ontology, it is a conformant-looking payload with missing mandatory fields, which is a SHACL problem and not an OWL problem. Finally, sovereign tooling: air-gapped environments do not get to call a hosted API, so favour standards with open implementations you can run offline.
That last point is why the licensing column above is not administrative trivia. IES, HQDM, MagmaCore, BFO and CCO can all be pulled into a classified enclave and run without a vendor relationship. That combination of formal seriousness and permissive licensing is rare, and it is the reason defence has quietly become the most advanced ontology market in the world. If you want to practise the crosswalk and certification workflow that this sector buys, our standards crosswalk course works through the same method end to end on public data.
