Research On Load Balancing Mechanism Of Energy Saving In Fog Computing Scene

With the increasing in the number of IoT devices and user demands,the problems caused by the traditional Cloud-IoT,e.g.time delay and the pressure of network load,have restricted the popularization of some real-time applications.The concept of fog computing provides a good solution to the above problems.As a supplement to the cloud mode,fog computing shortens the transmission distance of data by deploying fog nodes at the edge of the network,reducing the network delay and load and meeting the real-time application requirements.Since the computational power of the fog node is relatively weak,how to rationally utilize the computing resources of the fog node,how to realize the load balancing in the fog computing has become one of the hot issues among researchers.The paper introduces the related concepts and theories of fog computing modeling and load balancing.The paper constructs a system consisting of IoT device,and analyzes the quality of the fog computing services,focusing on the overall delay of the node,the energy consumption and other performance.Considering the single task delay and cost need,another system consisting of the cloud-fog collaboration is also modeled and analyzed in this paper.The paper is developed as follows:(1)The delay and energy consumption of IoT devices caused by the task offloading strategy of IoT devices in the IoT device-fog node architecture is modeled.Considering the decisions made by each IoT device,which will affect the delay and energy consumption of other devices in the system,game theory is applied.Considering the privacy protection of IoT devices,an iterative-based distributed load balancing algorithm is designed.The simulation results show that the algorithm can improve the system capacity and achieve the load balancing state for each IoT device in the system under the premise of protecting user privacy.(2)The task fine-grained architecture of the cloud-fog-IoT device integration is modeled.Considering each task provides a corresponding strategy of resource allocation,the paper propose a task grading strategy aiming at task demand for delay and energy consumption.Meanwhile,considering load balancing problem,this paper proposed a cost-based energy consumption-delay model under the cloud-fog cross-collaborative situation.Based on the traditional three-tier architecture consisting of cloud-fog-IoT device,the model adds resource scheduling layer to collaborative scheduling of cloud and fog resources.This paper proposes a particle swarm optimization algorithm with faster convergence rate to address the load balancing problem.Simulation results show that on the basis of meeting the task delay,the cost based energy consumption and delay model also reduces the energy consumption of fog layer.When solving the load balancing problem,the number of iterations of the particle dispersion cosine inertia factor-population contraction particle swarm optimization algorithm presented in this paper is significantly lower than that of the traditional particle swarm optimization algorithm.Finally,this paper summarizes the research results,discusses the issues that are overlooked in the research process to simplify the model architecture,and proposes some issue which is worth studying in fog computing resource management research.This paper uses 22 figures,7 tables,52 references.