Research On Key Technologies Of Service Discovery And Load Balancing Based On LiteOS-A Container Cluster

The purpose of this paper is to conduct research on key technologies of service discovery and load balancing in container clusters of domestic real-time operating systems.Especially on the container cluster of the LiteOS-A operating system,the service discovery and load balancing model of the container cluster is designed and implemented.The main work is as follows:(1)In the research on collaborative scheduling of container resources,because the container environment is relatively independent,there is a lack of effective means for information exchange and resource sharing wi th other containers.To solve this problem,a container scheduling method is proposed,which combines interrelated containers into a container group,and the whole as the smallest unit of scheduling.Through the special container in the container group,it provides the basis for information exchange and resource sharing for associated containers.(2)When studying container cluster communication,because the container environment is relatively independent,the communication in the container cluster will become difficult.In order to solve this problem,on the basis of container group,a network model of container cluster is designed.The model is divided into node network layer,Cgop network layer,service network layer and external access network layer from bottom to top to solve problems b etween containers,between Cgops,between Cgop and services,and between clients inside and outside the cluster.communication needs.Finally,based on the network model,the service discovery module of the container cluster is designed and implemented.(3)When executing a large number of jobs,the cluster needs a load balancer for scheduling to fully utilize the overall computing power of the cluster.However,commonly used load balancing algorithms lack the consideration of resource types consumed by jobs.To solve this problem,a dynamic load balancing algorithm based on job type and job priority is proposed.According to the type and proportion of resources consumed by the job,the algorithm divides the job into different types,and selects the server t ype suitable for executing the job.At the same time,according to the priority of the job,the execution order of the job is determined.Finally,a load balancing algorithm for container clusters was obtained,and on this basis,comparative experiments were carried out to verify that the algorithm can effectively improve the resource utilization of the cluster and balance the load of the cluster.