Resource Scheduling Algorithm Optimization for Creating Containers for Users Under Large Data Experimental Platform
Current resource scheduling algorithms for container management in large data experimental platform can not meet user-orien-ted requirements.A resource scheduling algorithm for user-oriented container management,Cuckoo Search For User(CSFU),is proposed by using the Docker Swarm container management tool and improving the cuckoo algorithm.The core idea is to use Levy-flight to find the fittest physical node from the user's perspective,and then put the container created by the same user into the same physical node.Finally,compared with three native scheduling algorithms of Docker Swarm,the CSFU algorithm performs better in load balancing and cluster deployment speed.The KNN and k-Means algorithms are run.Experiments show that there is less network latency and faster convergence between the same cluster container under the CSFU algorithm.
万方数据
维普
