Russian Federation
The article presents the development and analysis of a stochastic model of operating system resilience in 2026 to rare race conditions based on Markov chains with absorbing states and vectorized race injection. The authors classify races, formalize stability phase transitions and implement a fuzzing algorithm with importance sampling for scheduler, filesystems and network stack verification. The uniqueness of the work lies in detecting hidden kernel vulnerabilities in hours instead of months, increasing resilience by orders of magnitude with minimal overhead and predictive verification before production. The article is relevant because in multitasking OS nondeterministic races cause catastrophic failures, while traditional testing misses extremely rare scenarios.
rare races, operating systems, stochastic verification, Markov chains, kernel resilience
1. Formalizaciya verifikacii topologii i elektricheskoy shemy dlya sistem avtomatizirovannogo proektirovaniya / T. V. Skvorcova, K. V. Zol'nikov, A. M. Plotnikov, I. V. Skorkin // Modelirovanie sistem i processov. – 2024. – T. 17, № 3. – S. 61-70. – DOIhttps://doi.org/10.12737/2219-0767-2024-59-68. – EDN DUYQHJ.
2. Ivanov A. A. Konfiguriruemyy metod poiska sostoyaniy gonok v operacionnyh sistemah s ispol'zovaniem predikatnyh abstrakciy // Izvestiya SPIIRAN. – 2016. – T. 13, № 6. – S. 65-78. – URL: https://www.ispras.ru/proceedings/docs/2016/28/6/isp_28_2016_6_65.pdf (data obrascheniya: 17.02.2026).
3. Sidorov V. V. Metod legkovesnogo staticheskogo analiza dlya poiska sostoyaniy gonok // Trudy ISP RAN. – 2024. – T. 47, № 2. – S. 45-56. – URL: https://cyberleninka.ru/article/n/metod-legkovesnogo-staticheskogo-analiza-dlya-poiska-sostoyaniy-gonok (data obrascheniya: 17.02.2026).
4. Geneticheskie algoritmy v modelirovanii povedeniya robota / A. S. Logvinenko, L. Yu. Smilyanskiy, A. I. Zarevich, F. V. Makarenko // Novye aspekty modelirovaniya sistem i processov : materialy Mezhdunar. nauch.-prakt. konf. / otv. red. V. K. Zol'nikov, A. I. Zarevich. – Voronezh, 2023. – S. 94–103.
5. Sintez bol'shih dannyh i korrekciya disbalansa klassov dlya uluchsheniya modeli klassifikacii lesnyh pozharov / A. I. Zarevich, F. V. Makarenko, A. V. Poluektov // Lesnoe hozyaystvo v usloviyah global'nyh vyzovov: novye paradigmy ustoychivogo razvitiya. Forestry – 2025 : materialy Mezhdunar. les. foruma, priuroch. k prazdnovaniyu 95-letiya VGLTU. – Voronezh, 2025. – S. 773–785.
6. Modelirovanie kolebatel'nyh processov v pakete MVStudium / A. V. Poluektov, K. V. Zol'nikov, V. I. Anciferova // Modelirovanie sistem i processov. – 2021. – T. 14, № 4. – S. 139–148. – DOIhttps://doi.org/10.12737/2219-0767-2021-14-4-139-148. – EDN PPOUDI.
