Mathematical modeling of replication lags in a distributed MongoDB database
DOI:
https://doi.org/10.34185/1562-9945-5-166-2026-06Keywords:
gradient boosting, prediction, feature іmportance, approximation, coefficient of determination, machine learning, environmental monitoring, mathematical modeling, active probing method, synthetic monitoringAbstract
This paper addresses the topical problem of identifying the factors influencing replication lags and constructing a model of these delays in a distributed MongoDB database under heavy load conditions, using the active probing method. Its application made it possible to overcome the limitations of the discretization inherent in standard monitoring tools—namely, the one–second interval for collecting metrics—and to ensure the acquisition of accurate replication lags values by recording data with precise timestamps on the primary node and reading them from secondary nodes.
Based on the experimental data obtained regarding the impact of the write guarantee level, operation type, processor load, queue wait time, number of concurrent client connections, number of documents in a batch, and the time interval between write operations, an assessment of feature importance was conducted using gradient boosting algorithms. The analysis showed that the number of concurrent client connections, the volume of documents in a batch, and the time interval between write operations have a decisive impact on the occurrence and increase of replication delays. It was also confirmed that queue wait times and processor load do not have a significant impact on the formation of delays.
Based on the study’s results, a nonlinear mathematical model of the cluster’s operation was developed (coefficient of determination R² = 0.9786), which can be used during the design, optimization, and configuration of distributed NoSQL systems to predict replication parameters and select optimal system operating parameters.
References
1. Ncube, M. M., & Ngulube, P. (2024). Enhancing environmental decision–making: a sys-tematic review of data analytics applications in monitoring and management. Discover sus-tainability, 5(1). https://doi.org/10.1007/s43621–024–00510–0
2. Sathyamoorthy, S., Matthew, U. O., Adekunle, T. S., & Okafor, N. U. (2024). Advances and challenges in iot sensors data handling and processing in environmental monitoring net-works. HAFED POLY Journal of Science, Management and Technology, 5(2), 40–60. https://doi.org/10.4314/hpjsmt.v5i2.3
3. Hasan, S., & Mohammed, Y. (2020). Environmental monitoring systems using wireless sensor networks: An overview. Journal of Engineering and Sustainable Develop-ment, 24(Special), 298–303. https://doi.org/10.31272/jeasd.conf.1.33
4. Reznichenko, O., Liashenko, O., & Arkhypova, V. (2025). Database models for storing big data. System Technologies, 6(155), 155–195. https://doi.org/10.34185/1562–9945–6–155–2024–18
5. Jiménez–Galina, A. M., Maldonado–Macías, A. A., Olmos–Sanchez, K. M., Hernandez, I., Estrada–Saldana, F., & Vázquez–Gálvez, F. A. (2024). Framework for heterogeneous data management: An application case in a NoSQL environment from a Climatological Cen-ter. Computación y sistemas, 28(1). https://doi.org/10.13053/cys–28–1–4474
6. Han, M. (2015). The application of NoSQL database in Air Quality Monitoring. 2015 in-ternational conference on intelligent systems research and mechatronics engineering. Atlantis Press. https://doi.org/10.2991/isrme–15.2015.25
7. Andreoli, R., & Cucinotta, T. (2023). Inducing huge tail latency on a mongodb deployment. 2023 IEEE international conference on cloud engineering (IC2E). IEEE. https://doi.org/10.1109/ic2e59103.2023.00020
8. Ahmed, J., Karpenko, A., Tarasyuk, O., Gorbenko, A., & Sheikh–Akbari, A. (2023). Con-sistency issue and related trade–offs in distributed replicated systems and databases: A re-view. Radioelectronic and Computer Systems, (2), 171–179. https://doi.org/10.32620/reks.2023.2.14
9. Gorbenko, A., Karpenko, A., & Tarasyuk, O. (2020). Analysis of trade–offs in fault–tolerant distributed computing and replicated databases. IEEE 11th international conference on dependable systems, services and technologies (DESSERT). IEEE. https://doi.org/10.1109/dessert50317.2020.9125078
10. MongoDB. (n.d.). Alert conditions: Replication lag. MongoDB Ops Manager Documenta-tion. Retrieved July 27, 2026, from https://www.mongodb.com/docs/ops–manager/current/reference/alerts/replication–lag/
11. Hows, D., Membrey, P., Plugge, E., & Hawkins, T. (2015). The definitive guide to mon-godb. Apress. https://doi.org/10.1007/978–1–4842–1182–3
12. Bartlett, G., Heidemann, J., & Papadopoulos, C. (2007). Understanding passive and active service discovery. The 7th ACM SIGCOMM conference. ACM Press. https://doi.org/10.1145/1298306.1298314
13. Yang, Y. (2025). Research on site reliability optimization technology based on synthetic monitoring in cloud environments. 5th international conference on artificial intelligence, au-tomation and high performance computing (AIAHPC) (p. 729–732). IEEE. https://doi.org/10.1109/aiahpc66801.2025.11290592
14. Honcharenko, Y., Priadko, V., Poleshchuk, I., Synkivskyi, V., & Ostra, N. (2022). Ap-proaches of probing signal type determination of means local probing for distributed power grids technical condition diagnostic. Herald of Khmelnytskyi National University, 305(1), 145–151. https://doi.org/10.31891/2307–5732–2022–305–1–145–151
15. Swargo, R. M., & Arifuzzaman, M. (2024). Deploy–Efficient and fast network probing with time–series foundation models. IEEE international conference on big data (big-data) (p. 8846–8848). IEEE. https://doi.org/10.1109/bigdata62323.2024.10824943
16. Rish, I., Brodie, M., Ma, S., Odintsova, N., Beygelzimer, A., Grabarnik, G., & Hernan-dez, K. (2005). Adaptive diagnosis in distributed systems. IEEE Transactions on Neural Net-works, 16(5), 1088–1109. https://doi.org/10.1109/tnn.2005.853423
17. Liao, J., Wang, J., Li, T., & Zhu, X. (2011). Multipath probing and grouping in multi-homed networks. IEICE Transactions on Information and Systems, E94–D(3), 710–713. https://doi.org/10.1587/transinf.e94.d.710
18. Cito, J., Gotowka, D., Leitner, P., Pelette, R., Suljoti, D., & Dustdar, S. (2015). Identify-ing web performance degradations through synthetic and real–user monitoring. Journal of Web Engineering, 14(5–6), 414–442. Retrieved from https://journals.riverpublishers.com/index.php/JWE/article/view/3845
19. Fu, X., Graham, B., Xuan, D., Bettati, R., & Zhao, W. (2004). Empirical and theoretical evaluation of active probing attacks and their countermeasures. Information hiding (p. 266–281). Springer Berlin Heidelberg. https://doi.org/10.1007/978–3–540–30114–1_19
20. Johnson, M., Carter, E., Brooks, S., Bennett, A., & Yusuff, M. (2020). Proactive monitor-ing strategies for high–availability database systems.
21. Badia, A. (2003). Active database systems for monitoring and surveillance. Intelligence and security informatics (p. 296–307). Springer Berlin Heidelberg. https://doi.org/10.1007/3–540–44853–5_22
22. Dhir, K. (2024). Synthetic monitoring as a quality gate in continuous integration and de-ployment (CI/CD) and AI augmentation. International Research Journal of Modernization in Engineering Technology and Science, 6(4), 14. https://doi.org/10.56726/IRJMETS95831
23. MongoDB. (n.d.). Deploy a replica set for testing and development. MongoDB Manual. Retrieved July 27, 2026, from https://www.mongodb.com/docs/manual/tutorial/deploy–replica–set–for–testing/
24. Hawkins, T., Membrey, P., Plugge, E., & Hows, D. (2013). Definitive guide to mongodb: A complete guide to dealing with big data using mongodb. Apress L. P.
25. MongoDB. (n.d.). Journaling. MongoDB Manual. Retrieved July 27, 2026, from https://www.mongodb.com/docs/manual/core/journaling/
26. MongoDB. (n.d.). WiredTiger Storage Engine. MongoDB Manual. Retrieved July 27, 2026, from https://www.mongodb.com/docs/manual/core/wiredtiger/#std–label–storage–wiredtiger/
27. Gotelli, N. J., & Ellison, A. M. (2013). A primer of ecological statistics (2nd ed.). Sinauer Associates, Inc.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 System technologies
This work is licensed under a Creative Commons Attribution 4.0 International License.
