Location of fog nodes mounted on fixed-wing UAVs

Abstract

Fog computing complements cloud computing by enabling a broad range of latency-sensitive applications. To make it feasible, fog nodes with processing capabilities are located in different regions to cope with variable demands in time and space. The fog nodes are usually fixed terrestrial infrastructures providing reliability and easy management; however, these nodes cannot cope with the heavy processing demands even for a limited time in a specific area. To help circumvent these limitations, the fog nodes can be mounted on unmanned aerial vehicles (UAVs) that can be promptly dispatched to cope with a peak of demands. Previous work investigated the use of rotary-wing UAVs as fog nodes. This paper formulates the fog node location problem considering fixed-wing UAVs. Such UAVs have an extended operation time, making them great assets for the improvement of service delivered to end-users. This paper models the location problem as an integer linear programming formulation and presents the Spatio-Temporal UAV Fog Node Location algorithm to solve the problem considering several requests, locations, and UAVs. The results obtained in an evaluation with actual data of mobile users has shown that fixed-wing UAVs could indeed be efficiently located to deal with the posed demands.