# A Model-Driven Methodology for Automotive Cybersecurity Test Case Generation

**Authors:** Stefan Marksteiner, Peter Priller  
**arXiv:** [2107.06024](https://arxiv.org/abs/2107.06024)  
**Date:** 2021-07-13  
**Summary page:** https://www.summarizepaper.com/en/paper/2107.06024v1/

## Key points

- Increasing need for cybersecurity in automotive systems recognized by international regulations such as UNECE regulation
- Current manual testing techniques not sufficient to cope with higher rate and comprehensiveness of testing demanded by regulations
- Proposal of a model-driven approach to (semi-)automate automotive cybersecurity testing based on attack tree modeling
- Aim to facilitate comprehensive and efficient testing at all stages of the automotive life cycle, including third-party testing and real-world conditions
- Addressing supply chain parts that are not at band
- Considerations for automatically deriving black box models for use in attack modeling
- Importance of methodologies for testing the cybersecurity of vehicular systems as a black box
- Novel methodology leveraging model-driven techniques to enhance automotive cybersecurity testing in compliance with international regulations, aiming to improve effectiveness and efficiency

## Summary

The paper "A Model-Driven Methodology for Automotive Cybersecurity Test Case Generation" by Stefan Marksteiner and Peter Priller addresses the increasing need for cybersecurity in automotive systems. This need has been recognized by international regulations such as the latest UNECE regulation, which mandates higher efforts for cybersecurity engineering. These efforts must be verified and validated through testing, but current predominantly manual techniques are not sufficient to cope with the significantly higher rate and comprehensiveness of testing demanded by these regulations. To tackle this challenge, the authors propose a model-driven approach to (semi-)automate automotive cybersecurity testing. This approach is based on attack tree modeling and aims to facilitate comprehensive and efficient testing at all stages of the automotive life cycle. It also enables efficient third-party testing and testing under real-world conditions. Additionally, it addresses supply chain parts that are not at band. The paper also discusses considerations for automatically deriving black box models for use in attack modeling. It highlights the importance of methodologies for testing the cybersecurity of vehicular systems as a black box. Overall, this paper presents a novel methodology that leverages model-driven techniques to enhance automotive cybersecurity testing in compliance with international regulations. By automating aspects of the testing process and incorporating attack tree-based approaches, this methodology aims to improve the effectiveness and efficiency of cybersecurity engineering in automotive systems.