Method for scheduling job in cloud computing environment

Abstract

The invention provides a method for scheduling a job in a cloud computing environment. The method comprises the following steps that: a user submits a task to a system through an application interface of a management node; according to the task which is submitted by the user, the management node allocates computing nodes to execute the task; the management node monitors the operating state of resources; when the resources are insufficient, the management node selects a corresponding application template in an authority range of the user, generates new computing resources and adds the new computing resources into the management node; and when the resources which are obtained by the computing task of the user are surplus, the management nodes selects to return the surplus resources to the system. The method has the advantages that: in the process of generating, adding and configuring new computing nodes, a manager does not need to manually manage configuration items of a job scheduling management node, and the management node of the system can automatically accomplish generation and addition of the new computing nodes according to the basic configuration of the manager during initialization of the system; the computing resources can be flexibly allocated by the system, and different requirements of different users for the scale of the computing resources are met; and the resources of the system can be effectively allocated, so that the utilization rate of the resources of the system is maximized.

Description

technical field [0001] The invention relates to cloud computing resource management, in particular to a method for job scheduling in a cloud computing environment. Background technique [0002] At present, there are many job scheduling algorithms in the field of high-performance computing problems, which can realize the effective management of predefined resources in the system. [0003] In the implementation of the current scheduling algorithm, it is impossible to solve the problem of fast generation of computing nodes on demand. When the scale of the problem that needs to be calculated changes, administrators need to manually add new computing node configurations, and deploy and install different applications on the new computing nodes to meet the computing resource requirements of new computing tasks. [0004] If the application program is deployed in advance to prepare for adding new computing nodes at any time, the problem is that when there are multiple applications i...

AI Technical Summary

Problems solved by technology

[0004] If the application is deployed in advance to prepare for adding new computing nodes at any time, the problem is that when there are multiple applications in the system that need to share the limited computing resources in the system, the allocation of computing resources cannot be effectively planned, because It is impossible to predict the occupancy of computing resources by various applications in different time periods, resulting in an unbalanced distribution of system resources: at a certain moment, some of the pre-occupied computing resources are not used by applications, and need to be replaced. Applications with multiple computing resources cannot obtain more computing resources, resulting in failure to meet the needs of applications