1 Synergies between Cloud-Fog-Thing and Brain-Spinal Cord-Nerve Networks Eren Balevi, Member, IEEE, and Richard D. Gitlin, Life Fellow, IEEE Department of Electrical Engineering, University of South Florida Tampa, Florida 33620, USA erenbalevi@mail.usf.edu, richgitlin@usf.edu Abstract—This paper is directed towards describing the strik- ing similarities and synergies between cloud and fog nodes that constitute the cloud-fog-thing [Fog Network] architecture proposed for 5G networks and the human brain-spinal cord-nerve network model. On the one hand, the central nervous system can be better modeled considering the duality with Fog Networks, and, on the other hand, novel algorithms/protocols inspired from the central nervous system can be developed for throughput and latency performance improvement in Fog Networks. Designing and managing large-scale Fog Networks using stochastic geometry and machine learning is applied to determine the optimum number of fog nodes and their locations that opti- mize throughput and latency for 5G networks. Having observed the close relation between the Fog Networks and the spinal cord, these results may be adapted to increase understanding of the role of spinal cord plasticity in learning and ultimately suggest new means of treating central nervous system disorders associated with the spinal cord plasticity. Inspired by the cooperation between the brain and the spinal cord, a modified coded caching policy is proposed for Fog Networks, that is, the files to be stored at the fog nodes are determined as a result of continuous information flow between cloud and fog nodes through the latent variables assigned to files. Index Terms—Fog networking, stochastic geometry, machine learning, spinal cord, central nervous system. I. I NTRODUCTION Many system advances are achieved by observing analo- gies between systems, that while seemingly disparate, share common properties and knowledge that has been acquired for one system is applied to improve performance of the other. In this paper, the striking similarities and synergies between the cloud-fog-thing architecture based on Fog Networks that have been proposed for 5G networks and the human brain-spinal cord-nerve network model are highlighted, and then possible cross fertilization opportunities are proposed. A. Fog Networking Latency sensitive use cases of 5G networks, such as au- tonomous vehicles, smart cities, and certain Internet of Things (IoT) applications, along with exponentially growing data traffic are driving a paradigm shift in network architecture by using computing and memory resources in the network edge, which is referred to as fog or edge computing, while maintaining cloud resources for appropriate functions [1]. This architecture extends cloud-like functions closer to the end devices so that faster service can be provided to these devices while reducing the load in the network core. Fog computing capable units, i.e., fog nodes with communication, computation and storage capability, along with other resources create a fog network [1]. Fog nodes may be upgraded from the existing nodes in the network such that each node can be a fog node, e.g., a base station, an access point or even a mobile [2]. At this point, it is rather important and appealing to find the number and locations of these fog nodes for a given network while going from theory to practice. Fog networking does not obviate the cloud; on the contrary, the goal is productive cooperation with the cloud. A novel wireless network architecture emerging from this cooperation is the cloud-fog-thing network that can manage the large- scale heterogenous data supporting a wide variety of 5G use cases and IoT applications as shown in Fig. 1 [2]. This architecture has been suggested as being matched to many different cases such as in smart traffic lightning systems, autonomous vehicles, smart grid [1], smart building [3], smart pipeline monitoring [4], augmented reality and real-time video analytics [5]. Despite the popularity of the cloud-fog-thing ar- chitecture, there have been many unexplored questions related to fog nodes, e.g., how many fog nodes should there be in a given area, and what are their locations? Interestingly, the answers of these questions not only improve the performance of fog networking but, as is discussed later, may also help in the treatment of the fundamental disorders in the central nervous system where the spinal cord network is viewed as a fog network. B. Central Nervous System Network Current knowledge about the central nervous system ba- sically comes from experiments; however, these experiments are not sufficient to identify the intrinsic mechanism that leads to inefficiency in the treatment of serious spinal cord injury and other central nervous system disorders [6]. In this regard, modeling the central nervous system in terms of cloud and fog networking can bring a new dimension in which cloud and fog networking technology may be used for further understanding of the central nervous system, and thus this may facilitate future treatment of disorders. The treatment of spinal cord injury and other disorders are related to spinal cord plasticity, which refers to the learning capability ensured by some specialized neurons at the spinal cord similar to the fog nodes, and hence they are denoted as