HEURISTIC-GUIDED DETECTION OF RESOURCE EXHAUSTION VULNERABILITIES FOR ENHANCING RESILIENCE IN AUTONOMOUS SYSTEMS
Al Sardy L, Muhammad M, Shaaban AM, Baklizi M (2026)
Publication Type: Conference contribution
Publication year: 2026
Publisher: TRAUNER Verlag + Buchservice GmbH
Pages Range: 135-142
Conference Proceedings Title: IDIMT 2026 - Resilience: ICT, Business and Society: 34th Interdisciplinary Information Management Talks
Event location: Hradec Kralove, CZE
ISBN: 9783991952183
Abstract
Autonomous and industrial control systems increasingly rely on continuous and predictable computation to maintain stable and safe operation. However, some software vulnerabilities do not cause immediate failures but instead degrade performance through excessive resource consumption. CPU-based resource exhaustion is particularly critical because it can delay decision-making processes and reduce system responsiveness, potentially affecting operational resilience. This paper presents a hybrid approach for detecting CPU-based resource exhaustion vulnerabilities. The method combines lightweight static analysis based on LLVM with heuristic-guided dynamic testing to identify input-dependent worst-case execution behaviour. Search-based optimisation techniques generate inputs that maximise resource-oriented fitness metrics, including instruction count, loop iteration frequency, and execution time. The approach was evaluated using representative algorithmic programs and a real-world utility. The results demonstrate that the framework can effectively approximate worst-case execution scenarios and reveal significant performance degradation, including slowdowns of up to 18.3×. These findings show that the proposed method can support the identification of vulnerabilities that may compromise the resilience of autonomous and industrial systems. The framework provides a practical basis for integrating resource-aware testing into existing software development and validation processes.
Authors with CRIS profile
Loui Al Sardy
Computer Science 7 (Computer Networks and Communication Systems)
Mamdouh Muhammad
Computer Science 7 (Computer Networks and Communication Systems)
Involved external institutions
AIT Austrian Institute of Technology
Austria (AT)
University of Petra (UOP) / جامعة البتراء
Jordan (JO)
Austria (AT)
University of Petra (UOP) / جامعة البتراء
Jordan (JO)
How to cite
APA:
Al Sardy, L., Muhammad, M., Shaaban, A.M., & Baklizi, M. (2026). HEURISTIC-GUIDED DETECTION OF RESOURCE EXHAUSTION VULNERABILITIES FOR ENHANCING RESILIENCE IN AUTONOMOUS SYSTEMS. In Doucek Petr, Sonntag Michael, Tkac Michal, Maria Ria Janoskova (Eds.), IDIMT 2026 - Resilience: ICT, Business and Society: 34th Interdisciplinary Information Management Talks (pp. 135-142). Hradec Kralove, CZE: TRAUNER Verlag + Buchservice GmbH.
MLA:
Al Sardy, Loui, et al. "HEURISTIC-GUIDED DETECTION OF RESOURCE EXHAUSTION VULNERABILITIES FOR ENHANCING RESILIENCE IN AUTONOMOUS SYSTEMS." Proceedings of the 34th Interdisciplinary Information Management Talks, IDIMT 2026, Hradec Kralove, CZE Ed. Doucek Petr, Sonntag Michael, Tkac Michal, Maria Ria Janoskova, TRAUNER Verlag + Buchservice GmbH, 2026. 135-142.
BibTeX: Download