Research on Docker Cluster Scheduling Based on Self-define Kubernetes Scheduler

Abstract

To address the issue that native scheduler in kubernetes cannot realize load balancing and cannot give full play to the overall performance of the cluster, we put forward a scheduling strategy based on Docker cluster of self-define kubernetes scheduler. The kubernetes scheduler can improve the native kubernetes scheduling strategy using optimized predicate algorithm model and priority algorithm model. Moreover, it can also be used to conduct contrast experiment between the native scheduling strategy of kubernetes and that of self-define kubernetes scheduler. It is in VMware Workstation environment where Centos system is installed, and where Docker container service and kubernetes container orchestration engine are deployed to establish Docker container cluster. Experimental results show that cluster scheduling strategy of the self-define kubernetes scheduler is better than the native one, because it can save more system resources to improve the fairness of the cluster scheduling and the scheduling efficiency.