Q-balance: A load balancing scheme based on a Multilayer Perceptron for Fog resources on a Smart Grid
Smart Grid; Fog Computing; Cloud Computing; Balanceamento de Carga;
Data processing in Smart Grids applications can use cloud computing. However, this infrastructure can lead to increased response time in such applications due to the distance between cloud data centers and Smart Meters. In this way, we propose a neural network based approach to manage the computational resources in the fog. In this environment, Q-balance aims to reduce the average response time of applications that use data from Smart Meters through the use of a MultiLayer Perceptron -- (MLP) neural network. MLP predicts the time that a computational resource will process a request from the Smart Meter application. Q-balance uses this information about the forecast to balance the load between available resources, reducing the average response time obtained. The performance evaluation of the experiments showed that the Q-balance reduced the average response time by up to 65% and 79% compared to the algorithms in the literature for fog and cloud respectively.