Research On Resource Scheduling Algorithm Of Docker Cloud Platform

With the development of information technology and the rapid development of the Internet industry,cloud computing has been applied in various industries and fields such as economy,culture,military,business,etc.With the rise of container technology,Docker has quickly become a new favorite of cloud platforms with its lightweight,standardized and second-level start-up.However,in the application process of Docker container,there is still a problem of low resource utilization.For this reason,this paper mainly studies the resource allocation algorithm of Docker platform.The main work is as follows:Firstly,aiming at the problem of low resource utilization and unreasonable task allocation of Docker container,an ant colony scheduling algorithm is proposed to improve the resource utilization of the system.Firstly,this paper introduces a scheduling algorithm framework for Docker cloud platform,and analyses in detail the process of constructing scheduling algorithm for Docker platform,as well as the functions of each component in the architecture diagram.Secondly,the scheduling algorithm steps for Docker platform are introduced in detail.Then,the length of particle coding in particle swarm optimization is corresponded to the task of container by natural number coding;the position and velocity of NP particles in solution space are randomly initialized by population initialization step,and the parameters of particle swarm optimization are set;in fitness function step,fitness function is used to evaluate the quality of particles.Before satisfying the stopping condition of iteration,the particle optimum value and the total optimum value are updated according to the fitness function.In order to avoid the particle swarm optimization falling into the local optimum solution,the simulated annealing step is introduced and the optimum value is updated by simulated annealing.Finally,the experimental results show that the algorithm can effectively improve the resource utilization of Docker platform.Secondly,aiming at the problem of multi-resource utilization and load balancing of Docker cloud platform,this paper proposes a resource scheduling algorithm for Docker cloud platform to support load balancing,which improves the resource utilization of the system on the premise of ensuring load balancing.This paper considers the total memory,CPU utilization,memory utilization,average network load and I/O load of nodes.Firstly,the current resource balance factor is calculated,and the allocated resource balance factor is calculated according to the resource requested by the application.Then,the nodes are sorted according to the resource balance factor,and the nodes with the highest utilization rate are selected.Finally,the experimental results show that the proposed algorithm improves the resource utilization,improves the efficiency of cluster scheduling,reduces the response time of tasks,and achieves load balancing.The load balancing ability of this method is significantly higher than that of Docker container resource scheduling algorithm.