Docker Swarm cluster resource scheduling optimization method based on load prediction

Abstract

The invention discloses a Docker Swarm cluster resource scheduling optimization method based on load prediction and belongs to the technical field of computer system virtualization. The Docker Swarm cluster resource scheduling optimization method based on load prediction comprises the steps of utilizing an API interface function of a docker daemon to periodically collect resource historical use amount of a container; utilizing an ARIMA-RBF model to model and predict the resource historical use amount, obtaining resource future usage amount and combining resource current use situation to adjust resource use upper limit and resource use authority limit; determining a resource use deviation degree of the container according to the resource historical use amount of the container; choosing a node with the most balanced resource use deviation according to the container and a resource use deviation degree of a node set capable of meeting a container resource requirement to deploy a new container after the container is added when the new container is started in a cluster. The Docker Swarm cluster resource scheduling optimization method disclosed by the technical scheme of the invention improves Docker Swarm cluster resource utilization rate and can improve actual operation performance of all the containers.

Description

technical field [0001] The invention belongs to the technical field of computer system virtualization, and more specifically relates to a load forecast-based Docker Swarm cluster resource scheduling optimization method. Background technique [0002] Docker has attracted high attention in the field of cloud computing since its release, and the community is extremely active, making it one of the hottest open source software. Industry giants such as Microsoft, Google, and IBM have established cooperation with Docker and supported Docker on their respective platforms. Docker's development momentum is quite rapid. Simply put, Docker is an application running container that packages applications and all dependencies into a standard unit, just like a container. The function of Docker is similar to that of a virtual machine, but it is faster to start and stop than a virtual machine and saves computer resources. [0003] The emergence of Swarm, a Docker cluster management tool, fu...

AI Technical Summary

Problems solved by technology

[0009] The spread strategy is currently used the most, but it also has shortcomings: when allocating new containers, it only pays attention to the memory and CPU usage ratio on the node, and does not consider the I / O performance of the node, nor does it consider the container on the node resource-biased features, such as the new container is CPU-intensive, and the nodes that happen to meet the conditions are also CPU-intensive containers, so it is likely that serious CPU problems will occur later.

Images