Fog computing framework for smart city design

Abstract

Fog computing is a new network architecture and computing paradigm that uses user or near-user devices (network edge) to conduct some processing tasks. Accordingly, this network architecture extends cloud computing with more flexibility than that found in ubiquitous networks. A smart city based on the concept of fog computing with flexible hierarchy is proposed in this study. The aim of the proposed design is to overcome the limitations of previous approaches, which depend on using various network architectures, such as cloud computing, autonomic network architecture, and ubiquitous network architecture. Accordingly, the proposed approach reduces the latency of data processing and transmission with enabled real-time applications, distributes processing tasks over edge devices to reduce the cost of data processing, and allows collaborative data exchange among the applications of a smart city. The design comprises five major layers, which can be increased or merged according to the amount of data processing and transmission in each application. The involved layers are as follows: connection, real-time processing, neighborhood linking, main processing, and data server layers. A case study of a novel smart public car parking, traveling, and direction advisor is implemented using iFog-Sim, and results show that the delay of real-time application, cost, and network usage are significantly reduced compared with that of a cloud computing paradigm. Moreover, the proposed approach increases the scalability and reliability of user access without considerably compromising time, cost, and network usage compared with fixed fog computing.