Optimizing the Number of Fog Nodes for Cloud-Fog-Thing Networks

Fog networking has recently received considerable attention from a theoretical perspective, but in order for such networks to be practical several open areas need to be addressed. This paper determines the optimum number of nodes that should be upgraded to fog nodes with additional computing capabilities in order to maximize the average data rate and minimize the transmission delay. The optimization is performed for a given set of wireless channel conditions and a fixed total number of network nodes. It will be shown that, having more or less fog nodes than the optimum degrades the data rate. The numerical results indicate that the average data rate can increase nearly an order of magnitude for an optimized number of fog nodes in case of shadowing and fading. It is further shown that the optimum number of fog nodes does not increase in direct proportion to the increase in the total number of nodes. Furthermore, the optimum number of fog nodes decreases when channels have high path loss exponents. These findings suggest that the fog nodes must be selected among those that have the highest computation capability for densely deployed networks and high path loss exponent channels.