A configuration-driven lightweight hybrid infrastructure platform and data processing method

Abstract

The invention discloses a configuration-driven lightweight hybrid infrastructure platform and a data processing method, which are characterized in that it includes a client server for receiving a resource instance request and sending it to an API server; an analyzer deployed at the sink On the host, it is used to analyze the API of the cloud provider, and form the YAML configuration file of the API according to its resource management capabilities; generate the resource configuration file and function configuration file of the virtual resource according to the API supported by each virtual resource; The converter is used to convert each virtual resource in the input resource configuration file into a Kubernetes-compatible CRD object; the controller is deployed on the API server and is used to analyze the resource instance request received by the API server and obtain resources The request type and resource instance requested by the instance, and then determine to send the resource instance request to the scheduler or executor for processing according to the request type.

Description

A configuration-driven lightweight hybrid infrastructure platform and data processing method technical field The present invention relates to a kind of configuration-driven lightweight hybrid infrastructure platform and data processing method, belong to cloud service business and computing technology. Background technique With the rapid development of information technology, the Internet computing model of cloud computing has been successfully applied to business, government, medical Gartner report shows that the sales share of cloud computing IT infrastructure has exceeded that of traditional IT products, accounting for 58.4%. Cloud computing infrastructure can adopt different computing models, typical models include virtual machines and containers. In recent years, with the rapid development of software and hardware of mobile devices, the edge computing scenario integrating virtual machines and containers has become an industry development. inevitable trend of t...

AI Technical Summary

Problems solved by technology

However, this approach will lead to the following problems: (1) The Pod model can only represent the deployment model of the container, but cannot describe the virtual machine model

Therefore, in the official kubevirt solution at this stage, only virtual machines can be created from the host machine, and manual login is required to start the virtual machine from the host machine, which has poor usability

(2) Manage virtual machines but still need to deploy Docker containers

This is because according to the Kubevirt solution, Pod still needs a guardian container for virtual machine management, which is not easy to use

(3) Kubernetes cannot obtain the real-time resource consumption of virtual machines, because the collection of container information is different from that of VM information, so other monitoring solutions need to be integrated, which is poor in versatility

Images