Research On Elastic Scaling Strategy In Edge Computing Platform For Wearable Application

Wearable application scenarios continue to emerge,effectively improving the productivity of individuals and teams.In the process of teamwork,wearable devices can provide AR assistance,face detection,data fusion and other functions.These features facilitate team collaboration,but also put forward a great challenge to the computing and battery life of wearable devices.In order to improve the performance of wearable devices,edge-end collaboration schemes have been rapidly developed.By taking advantage of edge computing,energy consumption of wearable devices can be effectively reduced and service quality can be improved.Elastic strategy is a key technique in edge computing,which aims to allocate resources to applications in a timely manner to cope with unexpected workload changes.Edge computing clusters have limited resources and need to meet the computing needs of a variety of wearable applications.Therefore,an elastic strategy should be lightweight,using as few cluster resources as possible while responding effectively to workload changes.Firstly,this paper analyzes wearable application scenarios to determine the necessity of elastic scaling in edge-end collaborative wearable applications.Then the basic concepts and container technologies of the lightweight Kubernetes(K3S)edge computing platform are analyzed,and the open source solution Prometheus is used to monitor the node and application status of edge computing cluster.The native elastic strategy of K3 S is a static threshold response strategy that cannot provide good quality of service in the face of dynamic application workloads.Therefore,after the optimization requirements and feasibility analysis of the elastic strategy of the edge computing platform,the subject adopts the prediction model based on time series analysis and the elastic decision model based on BP neural network to carry out the elastic optimization at the application level.Then,based on the characteristics of K3 S edge computing platform,the implementation architecture of elastic strategy is designed.It mainly includes resource monitoring module,prediction and analysis module,elastic strategy module and container scheduling module.Finally,the edge computing cluster is built on the actual hardware nodes,and the performance of the designed elastic strategy is tested based on the face detection application in the application scene.The experimental results show that the elastic strategy can effectively improve the service quality and avoid the impact of sudden fluctuations on service quality.