Fair Task Offloading among Fog Nodes in Fog Computing Networks

Abstract

Fog computing is expected to cope with the long latency and heavy link burden existing in cloud- based networks. Computing tasks of the terminal node can be offloaded to nearby fog nodes thus achieving much lower processing delay than that of cloud-based networks. Existing researches for energy consumption in fog computing networks mainly focus on the total energy consumed by processing a task. However, fair offloading among multiple fog nodes while maintaining a low task delay is of great significance especially for the battery-powered fog nodes. This paper proposes an analytical framework of the fair task offloading for fog computing networks. Task delay and the corresponding energy consumption are formulated. Then, a fairness scheduling metric is constructed for each fog node. A two-step Fair Task Offloading (FTO) scheme is proposed finally, which selects offloading fog nodes according to the fairness metric and then offloads tasks to the selected nodes based on a rule that minimizes the task delay. Numerical simulations and comparisons indicate the satisfactory performance of the proposed task offloading scheme for maintaining a relatively high fairness index for energy consumption and low task delay in the fog computing networks.