Model-Driven Dependability and Power Consumption Quantification of Kubernetes-Based Cloud-Fog Continuum

Abstract

System dependability is pivotal for the reliable execution of designated computing functions. With the emergence of cloud-fog computing and microservices architectures, new challenges and opportunities arise in evaluating system dependability. Enhancing dependability in microservices often involves component replication, potentially increasing energy costs. Thus, discerning optimal redundancy strategies and understanding their energy implications is crucial for both cost efficiency and ecological sustainability. This paper presents a model-driven approach to evaluate the dependability and energy consumption of cloud-fog systems, utilizing Kubernetes, a container application orchestration platform. The developed model considers various determinants affecting system dependability, including hardware and software reliability, resource accessibility, and support personnel availability. Empirical studies validate the model’s effectiveness, demonstrating a 22.33% increase in system availability with only a 1.33% rise in energy consumption. Moreover, this methodology provides a structured framework for understanding cloud-fog system dependability, serves as a reference for comparing dependability across different systems, and aids in resource allocation optimization. This research significantly contributes to the efforts to enhance cloud-fog system dependability.

Keywords

• Cloud-fog continuum
• dependability
• Kubernetes
• stochastic modeling