Adaptive Agent-based Load Balancing in Fog Computing

As an extension of cloud computing, fog computing push computing power to the edge of the network, making full use of computing resources close to users to reduce communication load. However, the computing power of fog computing devices are relatively low, so the execution time of tasks take a bigger portion of the user-aware latency delay. The servers that make up the fog computing nodes are diverse, and each server has different computing capabilities. The performance of these diverse servers is difficult to predict and estimate precisely. Therefore, this paper proposes a value-iteration load balancing algorithm to optimize the task execution time in the fog computing environment. The results show that our algorithm outperforms the Round Robin, Min-min and active monitoring algorithms in a simulator environment, and outperforms the Weighted Round Robin, min-min, active monitoring and throttled algorithms in a real fog computing environment implementation.