A Scalable Heuristic on Energy Consumption of Fog-RAN

This work focuses on the energy analysis of fog computing based-RAN (F-RAN) and its suitability to support the evolution path for 5G and beyond 5G (B5G) network design. As the number of devices continues to grow exponentially, there is an ever-increasing demand for low latency, real-time and energy-efficient network solutions for the upcoming 5G and B5G scenarios. F-RAN has surfaced as a potential architecture for 5G and B5G networks. In F-RAN, a fog computing layer is placed between the end-users and the cloud computing layer. Fog computing is a distributed computing paradigm and consists of different fog nodes with limited computing capabilities. By this characteristic of fog nodes, F-RAN helps share the network load by fulfilling a set of service requirements with an interface with cloud servers via fronthaul links. F-RAN avoids large-scale radio signal processing in a centralized cloud BBU pool. In F-RAN architecture, the baseband processing unit is hosted in the cloud, and conventional high-power remote radio heads (RRHs) can be substituted with the low-power RRHs and fog-access points (F-APs). In such networks, low-power RRHs and F-AP are connected to the BBU cloud through transport links. The power consumption of the transport links, low-power RRH, and F-APs play an essential role in F-RAN infrastructure. An optimization framework for F-AP selection and energy minimization in the F-RAN network is formulated and has been solved using a heuristic algorithm. We have shown that the proposed algorithm provides an optimal performance regarding the F-RAN network's power consumption compared to cloud-RAN.