A Multi-Layer Continuous-Time Markov Model for Availability Analysis of Kubernetes Systems

Mankivskyi, Volodymyr; Romanov, Oleksandr; Nesterenko, Mikola; Marinov, Mika

doi:Under indexing

Proceedings of International Conference on Applied Innovation in IT · 2026/04/22 · Vol. 14 · Issue 2 · pp. 59–69

A Multi-Layer Continuous-Time Markov Model for Availability Analysis of Kubernetes Systems

Volodymyr Mankivskyi, Oleksandr Romanov, Mikola Nesterenko, Mika Romanova and Anton Marinov

📄 Download PDF DOI: Under indexing

Abstract

This paper proposes a three-tiered analytical availability model for distributed systems running on top of a Kubernetes platform, where container reliability, pod redundancy, and CNI network component reliability are unified under a single Continuous Time Markov Chain model. Unlike typical methodologies based on end-to-end MTTF/MTTR metrics only, this work introduces a model where failure/recovery behaviors can be directly analyzed across multiple tiers and steady-state end-to-end availability between pods can be evaluated directly. Evaluation results show that the effect of the system’s recovery time is almost linear when MTTR ≪ MTBF and the availability benefits of pod redundancy increase exponentially but diminish their returns when the number of replicas is greater than three. A numerical evaluation of the proposed model is conducted using representative parameters of Kubernetes environments. The results demonstrate the influence of recovery time, Pod redundancy, and network reliability on end-to-end service availability. While redundancy significantly improves baseline availability, network recovery time may become the limiting factor for meeting strict SLA requirements.

Keywords

Availability Markov Chain CTMC Kubernetes Pod Container CNI MTTR MTTB.

References

O. Romanov, V. Mankivskyi, L. Globa, M. Skulysh and A. Romanov, “Enhancing resource availability: Indicators and strategies for optimizing the Kubernetes network,” in Information and Communication Technologies and Sustainable Development (ICT&SD 2022), A. Dovgyi et al., Eds., Lecture Notes in Networks and Systems, vol. 809. Cham, Switzerland: Springer, 2023, , doi: 10.1007/978-3-031-46880-3_2.
The Kubernetes Authors, “Kubernetes API server - Security,” Kubernetes Documentation, 2020, [Online]. Available: https://kubernetes.io/docs/reference/access-authn-authz/. Accessed: May 16, 2020.
L. A. Vayghan, M. A. Saied, M. Toeroe and F. Khendek, “Kubernetes as an availability manager for microservice applications,” arXiv preprint arXiv:1901.04946, 2019, doi: 10.48550/arXiv.1901.04946.
The Kubernetes Authors, “Kubernetes official website,” Kubernetes Documentation, 2020, [Online]. Available: https://kubernetes.io/. Accessed: May 16, 2020.
O. Romanov, M. Nesterenko, L. Veres, R. Kamarali and I. Saychenko, “Methods for calculating the performance indicators IP multimedia subsystem (IMS),” in Lecture Notes in Networks and Systems, vol. 152. Cham, Switzerland: Springer, 2021, pp. 229-256, doi: 10.1007/978-3-030-58359-0_13.
K. Pentikousis, ONOS: Security and Performance Analysis, Technical Report No. 1, Sept. 19, 2017.
C. O’Connor, T. Vachuska and B. Davie, Software Defined Networks: A Systems Approach. Cambridge, MA, USA: Morgan Kaufmann, 2021.
K. Phemius, M. Bouet and J. Leguay, “Distributed multi domain SDN controllers,” Thales Communications & Security, 2013, pp. 198-209, doi: 10.1109/NOMS.2014.6838330.
J. Lam, S. Lee and O. Yustus, “Securing SDN southbound and data plane communication with IBC,” Mobile Information Systems, vol. 2016, art. ID 2016, 2016, doi: 10.1155/2016/1708970.
K. Phemius, M. Bouet and J. Leguay, “ONOS intent monitor and reroute service: Enabling plug and play routing logic,” Thales Communications & Security, 2013.
A. Poniszewska Marańda and E. Czechowska, “Kubernetes cluster for automating software production environment,” Sensors, vol. 21, no. 5, art. 1910, 2021, doi: 10.3390/s21051910.