Industrial automation is undergone a tremendous change with the introduction of the Internet of Things (IoT) which leads the way to Industry 4.0. Factory automation and process automation deploy a plethora of IoT sensors and actuators in a wide area. These automation IoT applications require very low latency, high computation and high reliability. Due to the size constraint and remote deployment scenario, computing and storage resources and battery life are limited in these devices. Thus, data processing and control are not feasible in these devices. To overcome these limitations, one of the promising solutions is to offload the processing and control tasks to a remote server and execute them in resource-rich servers. By offloading computing-intensive tasks to a remote server, the battery life of these devices can be improved. Fifth-generation (5G) mobile communication provide native support for Massive Machine Type Communications (mMTC) use case including IoT devices through MEC hosts as remote servers. On the other hand, MEC service providers have limited edge resources in MEC hosts deployed in the existing network. Since MEC hosts can be deployed hierarchically in the network, MEC host service providers can deploy different resources at different levels. For instance, MEC host located deeper in the network can have more resources than the MEC host deployed in the base station.

Once MEC hosts are deployed, maintaining those hosts with the minimal operational cost is a critical challenge for MEC service providers. Selecting an energy-efficient MEC host to provide offloading services while keeping other unused MEC hosts in the idle state improves energy efficiency. Thus, selecting an energy-efficient MEC host in the hierarchical network for processing IoT application data to trigger a real-time control system is a critical challenge for MEC service providers. MEC host section shall consider different requirements of the IoT applications such as latency, uploading data and required CPU resources. We modeled the energy-efficient MEC host selection problem as an energy minimization problem with IoT application requirements as constraints. In our simulation, we deployed IoT devices in a larger area covered by 100 microcells.  We then simulated to find out the energy-efficient MEC host in the hierarchical network that satisfies the IoT application requirements. Our results show a threshold-based approach is possible to select energy-efficient MEC hosts based on the IoT device application requirement. Our results also show that despite the energy efficiency, the latency requirement of the application is a key factor in selecting an energy-efficient MEC host.