Download serializations:
Workbench:
How to cite: Bueno Momcilovic, T., Gallina, B., Kessler, I., Hendricks, J., \& Balta, D. (2025). OntoGSN: An Ontology-Based Framework for Semantic Management and Extension of Assurance Cases. Retrieved from https://fortiss.github.io/OntoGSN/
About
OntoGSN is an ontology for enabling users to manage (i.e., create, maintain, align, evaluate, explain and display) GSN assurance cases semantically. Key benefits include:
- native storage and querying of graph data;
- human‑readable, visualisable domain representation;
- seamless linking to data, documents and code in one store;
- logic‑based quality verification through rule engines;
- a foundation for advanced methods such as GraphRAG; and
- an active, mostly open‑source tooling ecosystem.
The ontology acts as a "translator" (i.e., middleware) between a system and an assurance case, as well as a validator for all rules under the standard. We have created the ontology in order to reduce the pain for users who have to specify how each system should interface with the assurance case, and/or have to update the information in the assurance case manually using an additional software interface. The added advantage of having the ontology handle the management of updates across interfaces is that one can monitor the status of their system's design, implementation, testing and runtime from the assurance case tree itself, which is well-suited for auditors, safety or security engineers, and other supervisory or monitoring roles.
Acknowledgements
Many thanks to:
- Prof. Barbara Gallina, Ingmar Kessler, and Dian Balta for contributions to the ontology and the paper;
- Jule Hendricks for contributions to the code;
- Yannick Landeck, Radouane Bouchekir, Tewodros Beyene, Carmen Carlan, and Laure Millet for reviewing various iterations of the contributions;
- Will Franks (Adelard / NCC Group), Ewen Denney (NASA Ames) and Laure Millet (Critical Systems Labs) for providing access to the tools of their respective organizations;
- Damir Safin for technical support.
Our preprint (v1):
