Research And Optimization Of Resource Scheduling Strategy Based On Kubernetes Container Cloud

As the need for flexible,efficient deployment continues to grow,more and more companies are leveraging containers to build and deploy applications.Kubernetes is one of the most popular container orchestration tools today because of its automated deployment and scaling,cross-data migration,and flexibility.Resource usage is automatically monitored and adjusted to ensure that the application has adequate resources within the container.However,the default scheduling policy of Kubernetes still has the following problems: slow scheduling time,unreasonable resource load allocation,and unbalanced load.In order to further improve cluster scheduling time,resource allocation efficiency and node load balancing,this paper optimizes its default scheduling strategy as follows:(1)In view of the long execution time and task allocation time of Kubernetes resource scheduling strategy in the pre-selection and optimization stage,as well as the large difference in the utilization rate of various resources of Node nodes,a new monitoring module,improved static resource scheduling strategy and dynamic resource scheduling strategy are designed for Kubernetes resource scheduling strategy.(2)In order to solve the problem that Kubernetes adopts greedy thinking as its default resource scheduling strategy in multi-POD task scenario,a static resource scheduling strategy based on improved WOA-3 is proposed.The proposed algorithm improves the three stages of the traditional whale algorithm: the encircling strategy with nonlinear variation,the spiral position change based on AOS,and the search and prey with adaptive weights.The improved strategy enables individual whales to make adaptive adjustments in the iteration,enabling whale populations to have strong global detection ability in the initial stage and better convergence ability in the later stage,so as to quickly find the optimal resource allocation scheme,and finally improve the efficiency of Kubernetes cluster resource allocation in the multi-POD task scenario.(3)A dynamic resource scheduling strategy based On improved ON-PCA algorithm was proposed to solve the problem that Kubernetes cluster could not cope with the change of Pod resource demand and the cluster load was unbalanced.The improvement of this algorithm includes multi-objective function based on CE and improved randomized SVD algorithm.By monitoring the real-time load of each load node in the cluster,node information is processed in blocks On the basis of ON-PCA algorithm,and the optimal resource allocation scheme is obtained.At the same time,the resource scheme is analyzed and the load is calculated,and the threshold is set according to the load evaluation measure.The Node is divided into light load,normal and overload levels,and the dynamic scheduling trigger condition is set to ensure the load balance of the cluster.The experimental results show that the proposed static resource scheduling strategy based on the improved WOA-3 algorithm and dynamic resource scheduling strategy based on the improved ON-PCA algorithm have been effectively applied in the container cloud platform,which can not only improve the resource utilization and load balancing,but also significantly improve the performance.