Microservice Instances Provisioning for IoT Applications in Fog Computing

With the advent of fog computing and microservice-based architecture, various latency-sensitive IoT applications can leverage their true potential by moving them from the cloud to fog nodes closer to the end devices. To achieve this, an optimal microservice provisioning process is needed to decide where and how to place or migrate application components considering the heterogeneity of users' workload in terms of needed instances (replicas) and seeking to optimize multiple performance objectives. In this paper, we established a multi-objective optimization model for microservice provisioning in fog computing by considering three conflicting objective models, including the deadline violation, the system's resource consumption, and balancing the load between fog nodes. The simulation results show that our proposal achieved better results in terms of the defined objective model.