Part I. A Quick Tour of Systems Modeling with SysML v2
Traditional Systems Engineering
- ... is document-based: engineers from various domains (electrical, mechanical, etc.) create a multitude of documents for capturing requirements, design, realization and analysis.
- The involved engineers also create digital artifacts (architecture block diagrams, 3D models, simulations, etc.) using their domain-specific tools.
- But: some of these digital artifacts (like block diagrams) have no precise meaning, and the others are not based on a common data format with a well-defined semantics,
- ... so they cannot be integrated, processed and analyzed by computer programs.
Model-Based Systems Engineering (MBSE)
A system model, as an integrating digital artifact, is
- incrementally built by all engineers involved,
- used as the single source of truth, and
- managed throughout the system lifecycle.
Allows/facilitates:
- shared stakeholder understanding,
- automated analysis with simulation and formal methods,
- automated documentation and reporting.
History of MBSE
- 1997: UML 1.1
- 2005: UML 2
- 2007: SysML v1
- 2008: Foundational UML (fUML)
- 2017: Unified Architecture Framework (UAF)
- 2025: SysML v2
Issues of UML and SysML v1
- UML has three different, but overlapping, behavior modeling languages (state machines, activity diagrams and sequence diagrams), which are not conceptually aligned/integrated with each other.
- UML behavior modeling is not conceptually integrated with state structure modeling.
- UML does not have a precise (formal) semantics for its behavior modeling languages.
- Since SysML v1 simply extends UML 2, it has the same issues as UML.
SysML v2
- ... is a domain-independent systems modeling language, which allows modeling all kinds of systems, including socio-cyber-physical systems such as information systems, spacecrafts or smart cities.
... builds upon 30 years of UML and 20 years of SysML v1 and improves them by
- defining a unified behavior modeling language based on the concepts of actions and states,
- integrating behavioral with structural modeling,
- providing a formal semantics, and
- supporting further SE concepts, such as Analysis Cases, Verification Cases, Views, and Variations.