Towards Fog-Relay Selection Strategy for Distributed 5G Networks

THe demand for Cyber-physcal computing requirements beyond 4G spectrum such as increased capacity, low latency, energy efficiency, resource provisioning, reliable data rate, and overall quality of service (QoS) has fuelled the fifth generation (5G) Fog-relay network innovation. The overall benefits of Cloud computing, Internet of Things (IoT), Edge computing (EC) and the recent Osmotic computing (OC) cannot be satisfied without an efficient relay cooperative communication domiciled in distributed 5G Fog network. To meet these demands, deliberate improvements were introduced in existing cellular network infrastructure. First, the paper reported findings on literature survey of 5G relay communication network architecture, relay classification, as well as parameter selection for device-to-device communication (D2D). Second, Single/multiple relays for cooperative communication are discussed while presenting perspectives on typical relay-assisted cooperative network. Third, 5G network relay Use-Cases and its expectations is highlighted. Fourth, 5G Fog network K-clustering relay algorithm for the proposed relay selection scheme is introduced for Fog-data offloading. Monte Carlo simulation model is used to derive the average data values plotted in the OMNeT++ experiment. Furthermore, the performance of a two hop relay transmission scheme in an edge-Fog scenario is discussed. Metrics such as Threshold-distance, threshold SINR and threshold-distance-SINR techniques are used for the relay selection evaluation. Results showed that the accuracy of the proposed relay selection technique has better performance compared to threshold-SINR, threshold-distance and the threshold SINR-distance techniques for all the different numbers of user devices. It was observed that the Root Mean Squared Error (RMSE) of proposed technique has lowest error relative to other techniques regarding data forwarding. The research offers insight into 5G-Fog distributed network for researchers and industry practitioners.

Keywords:

5G-Fog Networks , Cloud Computing , Machine learning , IoT , Cyberphyscial Systems