Fault Tolerance Analysis in IoT Systems Based on Fog Computing

In the context of the rapid growth of scale and criticality of IoT infrastructures, the task of ensuring the fault tolerance of hybrid systems combining fog and cloud computing, capable of maintaining a given level of service in case of failures of nodes and communication channels, becomes especially urgent.

The purpose of the study is to develop and evaluate a mathematical model of fault tolerance for a distributed IoT environment, as well as to develop approaches to dynamic load balancing, taking into account reliability indicators, network delays and node load intensity.

The paper uses methods of continuous Markov chain theory, queuing theory, as well as an experiment on foggy nodes and a cloud server with varying network parameters and failure scenarios. In the course of the study, solutions were obtained for building an integral metric combining reliability, latency and load, as well as algorithmic rules for redistributing tasks between fog and cloud nodes, ensuring the stable functioning of the IoT network with partial infrastructure failures.

Experimental results have shown that using the proposed model for dynamic load balancing reduces the average system downtime by 32 % and reduces delays for critical tasks by 4–6 times compared with purely cloud architectures.

The scientific novelty of the work lies in the development of a mathematical model of fault tolerance of a hybrid IoT system based on continuous Markov circuits, which integrates node reliability, network delays and load factors into a single integrated stability metric, as well as a dynamic load balancing method in a cloud architecture using an integrated metric as a criterion. redistribute tasks between nodes to minimize downtime and delays in case of partial infrastructure failures.

The theoretical significance of the research lies in the development of a mathematical modeling apparatus for fault tolerance of distributed IoT systems and in the proposal of a formalized criterion for assessing their stability in dynamic operating conditions.

The practical significance lies in the possibility of using the developed model and integrated metrics when designing and configuring hybrid architectures of the Internet of Things, selecting redundancy parameters and load redistribution rules, as well as developing recommendations for optimizing data flows to improve the stability and performance of real telecommunications and industrial IoT platforms.

1. Yermakov S.G., Khalil M.M., Khomonenko A.D., Bukharova K.A. Evaluating the efficiency of fog computingon the internet of things using a Non-Markov model. T-Comm. 2022;16(12):46‒52. DOI:10.36724/2072-8735-2022-16-12-46-52. EDN:MVRSBV

2. Glushak E.V. Cloud and fog computing: architecture, modeling, application. Vologda; Moscow: Infra-Engineering Publ.; 2025. 180 p. (in Russ.) EDN:BUZGWB

3. Fletcher M., Paulz E., Ridge D., Michaels A.J. Low-Latency Wireless Network Extension for Industrial Internet of Things. Sensors. 2024;24(7):2113. DOI:10.3390/s24072113. EDN:YATJXO

4. Plotnikov P.V., Vladyko A.G. Analysis of Approaches to Optimization of V2X Systems: Clustering, Edge and Fog Computing. Proceedings of Telecommunication Universities. 2024;10(3):7‒22. (in Russ.) DOI:10.31854/1813-324X-2024-10-3-7-22. EDN:TRWNON

5. Daraselia A.V. Models and analysis of performance indicators of traffic offloading mechanisms in heterogeneous wireless networks. Ph.D. Thesis. Moscow: Peoples’ Friendship University of Russia Publ.; 2022. 106 p. (in Russ.) EDN:VQNBTU

6. Murashkin I.N. Research of latency minimization algorithms in data stream processing systems. Innovations and Investments. 2025;4:356‒359. (in Russ.) EDN:AHKBWE

7. Melnik E.V., Klimenko A.B., Klimenko V.V. Models and reliability analysis of information and control systems based on fog computing. Proceedings of the 5th All-Russian Scientific and Technical Conference at Supercomputer Technologies, SKT ‒ 2018, 17–22 September 2018, Rostov-on-Don, Russian Federation, vol.2. Rostov-on-Don: Southern Federal University Publ.; 2018. p.103‒107. (in Russ.) EDN:YQBQRN

8. Glushak E.V., Klyuev D.S. Development and research of models for the functioning of cloud and fog computing. Radioengineering. 2025;89(3):157‒168. (in Russ.) DOI:10.18127/j00338486-202503-14. EDN:IGUDLR

9. Melnik E.V., Klimenko A.B. A recovery technique of the fog-computing-based information and control system computational process. Izvestiya Tula State University. 2018;9:551‒563. (in Russ.) EDN:YNDROH

10. Alsweiti M. Deep Learning Approaches for Traffic Prediction Forecasting in Multi-Level Cloud Architecture for Autonomous Vehicle Services. Proceedings of Telecommunication Universities. 2022;8(4):89‒99. (in Russ.) DOI:10.31854/1813-324X-2022-8-4-89-99. EDN:CNIDPH

11. Melnik E.V., Ivanov D.Ya., Orda-Zhigulina M.V., Orda-Zhigulina D.V., Rodina A.A. Application of fog computing for monitoring and forecasting system of hazardous natural phenomena. Izvestiya Tula State University. 2019;2:300‒311. (in Russ.) EDN:ZAFGCT

12. Volkov A.N. Routing Task in Dynamic Fog Computing Network. Proceedings of Telecommunication Universities. 2024;10(4): 27‒37. (in Russ.) DOI:10.31854/1813-324X-2024-10-4-27-37. EDN:QWBVQY

13. Volkov A.N. Dynamic Fog Computing Towards Green ICT. Proceedings of Telecommunication Universities 2024;10(3):24‒34. (in Russ.) DOI:10.31854/1813-324X-2024-10-3-24-34. EDN:QOELMJ

14. Glushak E.V., Mikhailova P.D. Overview of the adaptive algorithms for distributing internet of things data flows in cloud and foggy media. Infokommunikacionnye Tehnologii. 2024;22(4-88):15‒22. (in Russ.) DOI:10.18469/ikt.2024.22.4.02. EDN:ALXFGY

15. Volkov A.N. Development and study of a set of models and methods for constructing communication networks based on fog computing and providing telepresence services. DSc Thesis. St. Petersburg: The Bonch-Bruevich Saint Petersburg State University of Telecommunications Publ.; 2024. 54 p. (in Russ.)

16. Kupriyanov D.O. Mathematical modeling of requests flow to cloud compute cluster. T-Comm. 2020;14(10):39‒44. (in Russ.) DOI:10.36724/2072-8735-2020-14-10-39-44. EDN:MWCXTC

17. Melnik E.V., Klimenko A.B., Ivanov D.Ya. A model of the computational load distribution problem for information and control systems based on the fog computing concept. Izvestiya Tula State University. 2018;2:174‒187. (in Russ.) EDN:YXJCRP

18. Glushak E.V., Klyuev D.S., Volovach V.I. Analysis of distribution of tasks in the cloud and fog computing system. T-Comm. 2025;19(7):25–33. (in Russ.) DOI:10.36724/2072-8735-2025-19-7-25-33. EDN:WHDPWF

19. Vorobyov S.P. Mathematical model of optimization of the network infrastructure of a distributed enterprise system on a cloud, misty and edge technologies. Modeling, Optimization, and Information Technology. 2019(7);3-26:4. (in Russ.) DOI:10.26102/2310-6018/2019.26.3.003. EDN:LIIPTF