Fog Computing Application for Effective Fronthaul Management in Fifth Generation Networks

Abstract

Fog computing is in limelight nowadays because of its decentralized nature. It is a promising extension of cloud computing, which is deemed to empower the paradigm of fifth generation (5G) networks. This work exploits the application of fogging in order to effectively manage incoming air traffic of new radio (NR) for the north bound interface (NBI) of cellular fronthaul from distribution unit (DU) to central unit (CU). The study instigates by highlighting key fronthaul design challenges and opportunities in radio access network (RAN) architecture, followed by the proposed front haul topology using fog nodes. OPNET Modeler is used for illustrating a segment of proposed fronthaul topology within fog networking cloud. Modeling and simulation is carried out for two distinctive network topologies. The first topology is simulated without enabling any fog nodes. The second topology provisions fog computing by defining two unidirectional label switched paths (LSPs) between the remote radio head (RRH) node and the base band unit (BBU) switch of the CU. Corresponding forwarding equivalence classes (FECs) are also defined for NBI fog nodes. Considering the varied nature of functional split in 5G, the study assumes it to be at the packet data convergence protocol (PDCP) layer, which is the highest layer of protocol stack. This splitting tradeoff also adheres to the 5G latency requirements. The simulation results yield low latency and the throughput of each node, which are in perfect harmony with the published criteria for distributed management of fronthaul in fifth generation networks.