Privacy-Aware Load Balancing in Fog Networks: A Reinforcement Learning Approach

Fog Computing has emerged as a solution to support the growing demands of real-time Internet of Things (IoT) applications, which require high availability of these distributed services. Intelligent workload distribution algorithms are needed to maximize the utilization of such Fog resources while minimizing the time required to process these workloads. These load balancing algorithms are critical in dynamic environments with heterogeneous resources and workload requirements along with unpredictable traffic demands. In this paper, load balancing is provided using a Reinforcement Learning (RL) algorithm, which optimizes the system performance by minimizing the waiting delay of IoT workloads. Unlike previous studies, the proposed solution does not require load and resource information from Fog nodes, which makes the algorithm dynamically adaptable to possible environment changes over time. This also makes the algorithm aware of the privacy requirements of Fog service providers, who might like to hide such information to prevent competing providers from calculating better pricing strategies. The proposed algorithm is interactively evaluated on a Discrete-event Simulator (DES) to mimic a practical deployment of the solution in real environments. In addition, we evaluate the algorithm's generalization ability on simulations longer than what it was trained on, which, to the best of our knowledge, has never been explored before. The results provided in this paper show how our proposed approach outperforms baseline load balancing methods under different workload generation rates.