Availability model for edge-fog-cloud continuum: an evaluation of an end-to-end infrastructure of intelligent traffic management service

Our world is being transformed by connectivity and technology as time goes by, which requires continuous improvement of quality of service (QoS) levels in the systems. Currently, many emerging technologies demand latency-aware networks for real-time data processing, and we are becoming more dependent on those technologies day by day. Cloud computing environments provide high availability, reliability, and performance; however, cloud computing may not be suitable for latency-sensitive applications, such as disaster risk minimization, intelligent traffic management, and crime prevention, for instance. Two complementary paradigms, namely edge and fog computing, have been proposed to overcome the latency issues and increase the computing power between the cloud and edge devices (e.g., controllers, sensors, and smartphones). However, evaluating availability aspects is still a significant concern in those distributed computing environments since many challenges must be faced to guarantee the required QoS for those systems. Therefore, this study addresses the edge-fog-cloud continuum’s availability, where we propose a hierarchical availability model using fault tree and Markov chains. Also, we propose analytical availability models for the components in our environment, which may be used to support scalability and capacity planning of edge, fog, and cloud computing environments. Using our proposed hierarchical model, we investigated several scenarios to improve the system’s availability. In one of the case studies, we could investigate how to improve the availability of a baseline intelligent traffic management infrastructure, which was 98.47%, and we improved to 99.91%.