Optimization And Implementation Of Scheduling Policy Based On Kubernetes Scheduler

With the rise of cloud computing technology and a large number of applications,container technology led by Docker has now become an important part of cloud technology.In the face of managing a large number of containers,how to allocate container resources for the cluster more reasonably to achieve load balancing and improve cloud platform service quality is a key issue.As a tool for container orchestration,Kubernetes has been widely used in cloud platforms today.This paper takes Kubernetes as the research object,outlines the Docker container technology and the system architecture and design of Kubernetes,and studies the default scheduling mechanism of Kubernetes.Due to the strong randomness and real-time nature of scheduling tasks during the operation of the actual cloud platform,the resource requirements of containers are constantly changing.At present,the inherent scheduling allocation mechanism of Kubernetes can no longer meet the needs of a large number of container scheduling,and there is a serious waste of resources.Therefore,in response to these problems,this paper proposes the following optimization strategies for resource scheduling.The current default resource scheduling strategy of Kubernetes remains in the stage of sequential scheduling of a single container,For Pods in Kubernetes,after the Pod is scheduled to an node until the life end,the Pod will node migrate between other nodes.However,with the progress of time and the change of the node,the initial scheduling choice may not be optimal.This paper designs a dynamic resource scheduling strategy.By establishing a monitoring module,the resource usage of the Pods in the cluster nodes is periodically collected and the load is evaluated;then different load queues are established according to the different load conditions to screen out the Pods to be migrated on the high-load worker nodes;Migrating Pods are migrated using the optimized resource scheduling strategy to achieve dynamic load balancing of the overall cluster resources.Finally,in order to verify the effect of the improved resource scheduling strategy,this paper builds a Kubernetes experimental cluster environment and conducts experiments on the optimized scheduling strategy to prove the effectiveness of the optimized scheme.The experimental results show that compared with the dynamic scheduling strategy,the proposed batch scheduling strategy improves the load balance of the overall cluster and improves the service quality of the platform.