KerML (Kernel Modeling Language)
KerML (Kernel Modeling Language) is the foundation language standardised by the OMG (Object Management Group) on which SysML v2 is built. It defines the core modelling concepts and the formal semantics that languages built on top of it reuse and specialise — above all SysML v2, but the kernel is open to other languages as well.
Why a kernel language?
SysML v1 was defined as a profile of UML and inherited a foundation designed for software modelling, including its semantic gaps. With v2, the OMG separated the concerns: KerML provides a small, formally defined core; SysML v2 adds the systems engineering vocabulary on top. This layering keeps the language stack clean and opens the foundation for reuse: domain-specific extensions and custom languages can build on the same kernel and remain interoperable with SysML v2 models.
Three layers
KerML is organised in three layers that build on each other:
| Layer | Content |
|---|---|
| Root | Basic syntactic elements: elements, relationships, annotations — the raw material every model consists of |
| Core | The formal basis: types, classifiers, features and specialisation, with precisely defined semantics |
| Kernel | General modelling constructs: classes, structures, behaviours, functions, associations and connections |
SysML v2 specialises these kernel constructs into systems engineering concepts: a part def is ultimately a specialised KerML classifier, a connection a specialised association. The definition/usage principle that shapes all of SysML v2 originates here.
Formal semantics
The decisive difference from the UML foundation is semantic precision. KerML grounds the meaning of its concepts mathematically, via a semantic model library and formal logic, instead of describing it in prose. Model elements therefore have a machine-interpretable meaning: tools can check consistency, derive views and reason about models instead of merely storing diagrams. This precision is what SysML v2’s automated analysis capabilities and its cross-tool interoperability rest on.
KerML in practice
Most systems engineers never write KerML directly; they model in SysML v2 and benefit from the kernel indirectly, through a language that behaves consistently across constructs and tools. KerML becomes directly relevant for tool vendors implementing SysML v2, for language engineers extending it, and for teams considering a domain-specific language: the kernel offers a standardised, formally grounded foundation to build on instead of starting from scratch.