Introducing Ditmara: A modular reference architecture for production-grade digital twins

Ditmara is an open digital twin reference architecture for designing, communicating, and building production-grade digital twin systems. Gary Haus is a Creative Lead with over 30 years in computer graphics and immersive digital experiences, currently serving as a Senior Solutions Architect / Engineer at Unity. His career spans architecture, engineering, AEC visualisation, aerospace, defense, XR, industrial design, games, location based entertainment and enterprise simulation - disciplines where he repeatedly encountered the same pattern: teams building digital twins from scratch, reinventing the same integration problems, with no shared blueprint to build from. Ditmara is the direct result of that experience. It is a practitioner's attempt to give the industry the architectural foundation it was missing.

Digital twin technology has moved well past the proof-of-concept stage. Across manufacturing, smart cities, aerospace, energy, and defense, organizations are committing serious engineering resources to building operational digital twins - systems that mirror physical reality in real time, enable simulation and prediction, and feed intelligence back to the physical world.

And yet, the majority of these projects stall, overshoot budgets, or fragment into isolated silos before they deliver lasting value.

Technology isn't the bottleneck. The absence of a shared architectural blueprint is.

That's why I built Ditmara.

Key takeaways

• The Ditmara framework organizes digital twin architecture into six layers, from physical systems and data ingestion to real-time 3D, services, cloud infrastructure, and MBSE integration. Find it at digital-twin-architectur-haus.web.app
• Ditmara is engine-agnostic, helping teams evaluate how Unity, other engines, cloud platforms, and enterprise systems fit into a digital twin stack.
• The framework is designed for solutions architects, systems engineers, digital transformation leads, and technical decision-makers across manufacturing, aerospace, smart cities, energy, and defense.

What is Ditmara?

Ditmara, the Digital Twin Modular Reference Architecture, is an open, engine-agnostic, six-layer framework for designing, communicating, and building production-grade digital twin systems. It's a living reference document, freely available at digital-twin-architectur-haus.web.app, covering 43 components across six independently deployable layers.

The name itself has a quiet backstory. DTMRA, rendered phonetically, yields Ditmara - its roots are in an Old Germanic name meaning "renowned." It felt apt for a framework built to be referenced, shared, and built upon.

Ditmara is not a product. It's not proprietary. It carries no vendor lock-in. It is a structured vocabulary and integration map that any team using any engine, any cloud, any programming language, can adopt as a starting point.

The problem Ditmara solves

When a digital twin project kicks off, one of the first and most expensive activities is architecture design: deciding how sensors connect to processing pipelines, how CAD/BIM data flows into the 3D engine, where AI inference sits, how the system closes the feedback loop back to the physical asset.

Every team that has done this from scratch has reinvented the same patterns, often incompatibly. The result is a landscape of proprietary, siloed implementations that are difficult to communicate across teams, difficult to extend, and nearly impossible to migrate between vendors.

Ditmara addresses this with a single, composable model that every stakeholder, from the C-suite to the software architect to the procurement team, can read and reason about.

The six-layer architecture

Ditmara organizes digital twin concerns into six vertical layers. Data flows upward from the physical world to the services layer; closed-loop feedback flows back down. A parallel MBSE Integration Layer runs alongside, connecting formal systems engineering models to the operational twin.

Layer 1: Physical

The boundary between physical and digital. IoT sensors, PLCs, SCADA, LiDAR, GPS, edge devices, and industrial protocols (MQTT, OPC UA, Modbus, PROFINET) all live here. This is where reality enters the system.

Layer 2: Data ingestion

The normalization and translation layer. CAD/BIM imports, PLM/PDM connectors, time-series databases, stream processors, MBSE bridges, and GIS integration via Cesium (OGC 3D Tiles) or the ArcGIS Maps SDK for Unity/Unreal (for organizations operating within the Esri ecosystem). These are competing, not complementary, solutions. You choose based on where your geospatial data already lives.

Layer 3: Core processing

The real-time 3D engine. This is where Unity sits - rendering, simulating physics, running AI inference at runtime, and delivering immersive XR experiences. The framework is engine-swappable by design, but Unity's High Precision Framework, Physics Engine, AI inference (Unity Inference Engine / ONNX), and XR capabilities make it a natural fit across the majority of digital twin use cases.

Layer 4: Services

Analytics, machine learning, operator training, and closed-loop command dispatch. This layer generates the actionable intelligence the twin exists to produce.

Layer 5: Cloud infrastructure

Hosting, DevOps, security, CI/CD. Runs on Azure, AWS, or on-premises. Manages version control and multi-platform distribution.

Layer 6: Model-Based Systems Engineering (MBSE) integration

The formal systems engineering bridge. SysML models define system architecture and requirements; the operational twin validates and refines those models through real-world data feedback. Every data artifact maintains full lineage from model to sensor to decision.

The closed-loop principle

What separates a digital twin from a 3D model or a dashboard is the bidirectional, continuous feedback loop between the physical asset and its digital representation. Ditmara treats this loop as a first-class architectural concern, not an afterthought.

The framework maps the full data journey: physical sensors stream to the ingestion layer, which normalizes and timestamps; the core engine renders and simulates; AI inference generates predictions; the services layer surfaces insights and dispatches commands; those commands actuate the physical system; updated sensor data flows back in. The loop is continuous, auditable, and versioned.

Why engine-agnostic matters (and why Unity still leads)

Ditmara does not mandate Unity. It can't; enterprise customers operate across multiple engines, custom stacks, and legacy visualisation platforms. Prescribing a single engine would limit adoption and undermine the framework's value as a neutral reference.

That said, Layer 3 is where the engine choice becomes decisive, and Unity's capabilities in that layer are substantial: the Unity Inference Engine (ONNX runtime), native AR/VR/MR support across every major headset, Addressable Assets for large-scale streaming, physics simulation, the Asset Transformer SDK for CAD/BIM import, and the High Precision Framework for geospatial double-precision coordinates.

For teams evaluating their Layer 3 engine choice, Ditmara gives them the vocabulary to ask the right questions, and Unity consistently answers them well.

Designed to be extended

Ditmara is a community document. Since its initial release, it has grown to incorporate feedback from specialists across disciplines: GIS professionals, systems engineers, BIM practitioners, industrial automation experts. Each contribution has sharpened the framework's accuracy and broadened its applicability.

The interactive ring diagram on the site (43 components across six rings) is a living map of the digital twin component landscape. Each segment can be explored in detail: standards, tools, integration patterns, and cross-layer dependencies.

If you build digital twins, at any layer and in any industry, Ditmara is designed to be useful to you. Not as a prescription, but as a shared language.

Get started

The full Ditmara framework is freely available at:

digital-twin-architectur-haus.web.app

It's an open, evolving reference. Contributions, corrections, and domain expertise from the community are what make it stronger.