Spectrum Management Scheme in Fog IoT Networks

Abstract

With the rapid growth of communication devices in number and diversity, there is now increasing interest in the Internet of Things (IoT). There are various challenges, in terms of the spectral efficiency, data rates, average cell throughput, data latency, and connectivity density, on largescale IoT networks. To deal with those challenges, we explore the novel IoT network architecture and efficient spectrum management schemes to unlock 5G spectrum potential. In this article, we introduce an overview of the design principles for IoT architecture in relation to adaptability, performance, reliability, and scalability. Based on all of the design principles, a logical architecture is proposed, where both fog computing and network functions virtualization are taken into consideration for fog IoT networks. Furthermore, we provide two channel access schemes, as well as two spectrum allocation schemes based on the multi-armed bandit algorithm to increase spectrum efficiency and maximize economic revenues in unlicensed spectrum. Simulation results verify that the proposed spectrum management schemes significantly improve IoT network performance. Finally, several challenges in IoT networks are discussed, including network architecture, network technology, network security, and economic revenue.