Model Based Systems Engineering

Model-based systems engineering (MBSE) is the practice of doing systems engineering around a formal, shared model of the system rather than around documents.

11 frameworks, from 1970 to today · Systems Engineering · Engineering

still a working view ended · each leads to its entry below

Overview

Model-based systems engineering (MBSE) is the practice of doing systems engineering around a formal, shared model of the system rather than around documents1. In the traditional, "document-based" style, requirements live in a requirements database, the architecture lives in diagrams pasted into a specification, interfaces live in an interface control document, and the analysis lives in a separate simulation. Each artifact is maintained by hand, and consistency between them depends on people noticing when one changes. MBSE replaces that pile of documents with a coherent model—usually built in a modeling language with precise syntax and semantics—from which views, specifications, and analyses can be generated and checked1. The stakes are the classic systems-engineering stakes: large systems with many interacting parts, many stakeholders, and long lives, where a missed interface or an untraced requirement is expensive. What MBSE adds is the claim that a machine-readable, internally consistent model can catch what document review cannot, and can carry engineering intent from concept to disposal.

From diagrams to a discipline: structured analysis and the architecture frameworks

The roots lie in the structured-analysis movement of the 1970s and 1980s23. Structured Analysis and Design Technique (SADT), developed by Douglas Ross at SofTech in the 1970s, represented a system as a hierarchy of functions connected by flows of data and control, drawn as box-and-arrow diagrams with strict rules about what could connect to what. Its descendant, structured analysis for software (as in Tom DeMarco's data-flow diagrams), spread widely. The important idea for MBSE was that a diagram could be more than an illustration: if the notation had rules, the diagram was a partial model whose consistency could be reasoned about.

In parallel, the defense and aerospace world developed architecture frameworks—sets of viewpoints and product descriptions for describing complex systems. The US Department of Defense's C4ISR Architecture Framework (1996), refined into the DoD Architecture Framework (DoDAF), and the UK Ministry of Defence Architecture Framework (MODAF, early 2000s) were the best known. These frameworks told engineers what views to produce (operational, system, technical) but were largely document- and diagram-oriented; the models were often drawn in general-purpose tools with no shared underlying data. The gap—frameworks prescribing views but not a rigorous shared model—is exactly what MBSE set out to close.

Object-oriented modeling and the shift to a single model

The decisive technical shift came from software engineering. Object-oriented modeling, crystallized in the 1990s in the Unified Modeling Language (UML), introduced classes, objects, associations, and inheritance as a way to describe structure and behavior in one notation. Grady Booch, James Rumbaugh, and Ivar Jacobson merged their competing methods into UML, standardized by the Object Management Group (OMG) in 19974. UML's significance for systems engineering was that it offered a language with a metamodel—a model of what models are allowed to contain—so tools could store a model in a repository and generate diagrams from it rather than the reverse.

Systems engineers adopted UML but found it software-centric. The response was the Systems Modeling Language (SysML), developed by INCOSE and the OMG and adopted by the OMG in 2006. SysML reuses part of UML and adds systems-oriented constructs: requirements as model elements with trace relationships, blocks instead of classes, parametric diagrams for binding engineering equations to structure, and allocation relationships between model elements. SysML made the central MBSE promise concrete: a requirement, the function that satisfies it, the component that performs the function, and the analysis parameter that constrains it could all be elements in one model, linked by typed relationships, so a tool could report which requirements had no satisfying element.

Requirements-driven and functional modeling traditions

Two older strands fed into MBSE and remained distinct approaches within it. The requirements-driven tradition descends from requirements management tools and traceability practice: the model's job is to capture requirements, decompose them, and maintain verifiable links from each requirement down to design elements and up to stakeholder needs5. Its strength is auditability; its weakness is that a requirements tree alone says little about how the system behaves.

The functional tradition descends from SADT and from functional decomposition: model the system as a hierarchy of functions, allocate functions to components, and define the interfaces between them62. Functional architecture modeling treats the function structure as the primary artifact, with physical architecture derived by allocation. This strand is strong where behavior and flow matter—control systems, process plants, mission threads—and it is the direct ancestor of the operational and functional views in the architecture frameworks.

Object-Process Methodology

A distinct, more radical answer came from Dov Dori at the Technion, who developed Object-Process Methodology (OPM) from the early 1990s and published it as a book in 20027. OPM's claim was that structure and behavior should not be modeled in separate diagram types at all. In OPM, a system is described with just two kinds of thing—objects (things that exist, possibly over states) and processes (things that happen and transform objects)—and a single kind of diagram, with a matching natural-language sentence generated for each element. A typical OPM model shows, for instance, a "vehicle" object in states such as moving and stopped, and a "braking" process that transforms it from one state to the other. OPM was adopted as ISO 19450 in 2015. It remains a minority but coherent alternative to the UML/SysML lineage, valued for its minimal ontology and its insistence that a model be readable as text as well as diagram.

The INCOSE vision and the institutionalization of MBSE

MBSE became a named movement largely through INCOSE, the International Council on Systems Engineering89. INCOSE's Systems Engineering Vision 2020 (2007) projected a shift from document-centric to model-based systems engineering as a central trend, and subsequent versions of the vision (2025, published 2014) reinforced it. INCOSE published the MBSE Methodologies Survey (2007) and later the MBSE Capability Matrix, and its SysML initiative with the OMG gave the movement its lingua franca. The effect was to turn scattered tool-and-method practice into a recognized discipline with a roadmap, a vocabulary, and a community of practice. Textbooks followed—among them the INCOSE Systems Engineering Handbook's MBSE material and dedicated MBSE texts—and the "MBSE method" (a defined process for building and using the model) became a standard deliverable on large programs.

Architecture frameworks meet the model: UAF and the framework-model synthesis

The architecture frameworks did not disappear; they were re-founded on models. The Unified Architecture Framework (UAF), developed under the OMG and adopted in 2017, unified DoDAF, MODAF, and the NATO Architecture Framework into a single metamodel (the UAF Domain Metamodel) with standardized viewpoints, and it is built to be implemented in SysML-based tooling. UAF's significance is that it turned the older "produce these views" frameworks into "populate this model, then generate the views," which is the MBSE proposition applied to enterprise and system-of-systems architecture. Similar convergence occurred in the defense world's shift toward model-based acquisition requirements.

Digital engineering and the digital thread

The most recent expansion is from the system model to the whole engineering enterprise. Digital engineering, articulated in US Department of Defense strategy documents in the 2010s (notably the DoD Digital Engineering Strategy of 2018), treats authoritative models and their data as the primary product of engineering, with the "digital thread" as the linked flow of that data across disciplines and lifecycle phases—requirements, design, analysis, manufacturing, test, and sustainment. MBSE is the systems-engineering component of this larger picture: the system model becomes a node in the thread, connected to CAD, simulation, and logistics models. The claim is that a change anywhere propagates with traceability rather than being rediscovered by document review.

SysML v2 and the present

SysML's first version carried UML's legacy: a complex metamodel, diagram-centric tooling, and semantics that were sometimes ambiguous. SysML v2, developed by the OMG with INCOSE participation and adopted by the OMG in 2023, is a ground-up redesign with a formally specified metamodel, a textual notation alongside the graphical one, and precise semantics intended to support interchange between tools and analysis by machine. Its textual notation lets an engineer write model elements as structured text—definitions, usages, and requirements with explicit relationships—which makes models diffable and reviewable like code, addressing a long-standing complaint that graphical models hide their content in proprietary repositories.

Today the field's main approaches coexist: SysML-based MBSE (v1 widely deployed, v2 arriving), OPM for those who want a single-diagram ontology, UAF for enterprise and defense architecture, and requirements-driven and functional modeling as the process patterns layered on top17. The open questions are practical and unresolved: how to keep models and simulations synchronized, how to make model interchange between vendors' tools genuinely lossless, how to scale models to systems of systems without the model becoming as unwieldy as the documents it replaced, and how to teach and certify modeling competence as a core engineering skill rather than a tool specialty.

References

  1. Model-based systems engineering (introduction, History), Wikipedia
  2. Structured analysis and design technique (History, introduction), Wikipedia
  3. Structured analysis, Wikipedia
  4. Unified Modeling Language, Wikipedia
  5. Requirements management, Wikipedia
  6. Systems modeling, Wikipedia
  7. Object Process Methodology (introduction, History), Wikipedia
  8. Systems engineering (History), Wikipedia
  9. International Council on Systems Engineering, Wikipedia

The frameworks of Model Based Systems Engineering

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1970-1990

    Structured Analysis and Design Technique

    How can a system be decomposed into functions and flows?

    Fly to its tower in Nousfarer

  2. 1980-Present

    Functional architecture modeling

    What functions must a system perform and how are they allocated?

    Fly to its tower in Nousfarer

  3. 1990-Present

    Requirements-driven modeling

    How can every requirement be traced to design and verification?

    Fly to its tower in Nousfarer

  4. 1996-2017

    Architecture Frameworks

    Which views must be produced to describe a complex system?

    Fly to its tower in Nousfarer

  5. 1997-Present

    Object-Oriented Modeling

    How can structure and behavior be modeled in one notation?

    Fly to its tower in Nousfarer

  6. 2002-Present

    Object-Process Methodology

    Can one minimal ontology describe both objects and processes?

    Fly to its tower in Nousfarer

  7. 2006-Present

    Systems modeling language

    What language lets requirements, structure, and analysis share one model?

    Fly to its tower in Nousfarer

  8. 2007-Present

    INCOSE MBSE vision and methodology programme

    How should an organization adopt model-based systems engineering?

    Fly to its tower in Nousfarer

  9. 2017-Present

    Unified Architecture Framework

    How can competing architecture frameworks be unified into one model?

    Fly to its tower in Nousfarer

  10. 2018-Present

    Digital engineering and the digital thread

    How can authoritative models and data flow across the whole lifecycle?

    Fly to its tower in Nousfarer

  11. 2023-Present

    SysML v2

    What precise, textual, formally specified language should replace the first systems modeling language?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Structured Analysis and Design Techni…Functional architecture modelingRequirements-driven modelingArchitecture FrameworksObject-Oriented ModelingObject-Process MethodologySystems modeling languageINCOSE MBSE vision and methodology pr…Unified Architecture FrameworkDigital engineering and the digital t…SysML v2

Beyond the field

Where Model Based Systems Engineering's ideas came from in other fields, and what they led to there: the strongest 12 of 45 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Model Based Systems Engineering is a planet in Nousfarer

Its city has a tower for each of these 11 frameworks. Land, walk in, and climb one concept at a time.

Land on Model Based Systems Engineering

Related fields

Fields its ideas are linked to

More in Systems Engineering

See where Model Based Systems Engineering sits on the Map of Human Knowledge