Data center operating system

Abstract

The present invention provides a data center operating system. The system comprises: a host resource, which comprises a hardware computing resource, an application container engine Docker and a distributed resource management framework Mesos used for managing computing resource information, and which is used for providing a hardware resource and an application container; a core system, which comprises a UI, a RESTful API, a scheduler Swan, a monitoring alarm module Prometheus, a log monitoring module that comprises a LogProxy and an ELK log analysis system, a publishing system Backer and a mirror image repository, and is used for connecting a customization UI system and the host resource and allowing a user to implement management of the host resource by using the customization UI system, including scheduling of the host resource, monitoring and alarming, log recording and analysis, application publishing and management of the mirror image repository; and the customization UI system, which comprises CLI, SSO, an auditing interface and a multi-cluster management interface, and is used for providing a customization operating interface of the data center operating system. The system provided by the present invention can efficiently utilize a server, quickly deploy and migrate an application, and conveniently extend a service and manage the system.

Description

technical field [0001] The invention relates to the field of computers, in particular to a data center operating system. Background technique [0002] The data center operating system is a software stack that provides distributed scheduling and coordination functions for the entire data center and realizes data center-level elastic scalability. It schedules all data center resources as a server. [0003] At present, the performance of a single server limits the deployment scale of applications, and different applications occupy different resources. In this way, some servers in the data center are busy while others are idle, and the average utilization rate of server deployment is relatively low. low, resulting in a great waste of resources. Because the number of servers is too large to manage, the deployment time of new servers and applications is long, and hardware maintenance takes days or weeks to complete, resulting in high management costs. Once the server carrying th...

AI Technical Summary

Problems solved by technology

[0003] At present, the performance of a single server limits the deployment scale of applications, and different applications occupy different resources. In this way, some servers in the data center are busy while others are idle, and the average utilization rate of server deployment is relatively low. low, resulting in a great waste of resources

Because the number of servers is too large to manage, the deployment time of new servers and applications is long, and hardware maintenance takes days or weeks to complete, resulting in high management costs

Once the server hosting the application fails, the application needs to be manually switched to other servers, often causing serious losses to the data center

Even a cluster formed by virtualization technology is difficult to operate and maintain, especially after a fault occurs, it is difficult to analyze the cause and location of the fault

Infinitely scalable services and easy-to-migrate applications cannot be realized

Images