Energy-saving job scheduling system

Abstract

The invention provides an energy-saving job scheduling system applied to a computer cluster. The scheduling process comprises the following steps that: step S1, a job submitted by a user is received by a manager; step S2, the job received by the manager is put in a queue, in which the job is required to be submitted, by the manager, the state of the job is changed into an idle state; step S3, a message that a new job comes is sent to a scheduler by the manager, priorities of jobs in the idle states are counted according to information of the jobs and scheduling policies of the jobs, and the job with the highest priority is selected; step S4, a counting node is allocated to the job with the highest priority by the scheduler according to resource requirements of the job with the highest priority, job features, a node state and a node scheduling policy, the manager is informed by the scheduler to start the job with the highest priority on the distributed counting node, and an actuator is informed by the manger to start and execute the job on the distributed counting node; and step S5, idle nodes are located in energy-saving states. According to the energy-saving job scheduling system, the time for awakening each node and enabling each node to enter the energy-saving state is reasonably controlled.

Description

technical field [0001] The invention relates to the technical field of network management, in particular to an energy-saving job scheduling system. Background technique [0002] With the development of the information industry, large-scale computer clusters are becoming more and more popular, and the large increase in the number of nodes inevitably leads to a rapid increase in power consumption. The expansion of the cluster also means the expansion of supporting facilities, which intensifies the energy consumption. A large amount of energy consumption means a large amount of heat energy, which not only increases the investment in supporting cooling facilities in the computer room, but also affects the stability and reliability of the entire cluster as the temperature rises. The importance of energy-saving technology is self-evident. [0003] Existing energy-saving technologies are mainly divided into three levels: [0004] hardware level. The energy-saving technology at ...

AI Technical Summary

Problems solved by technology

But when the job comes, only node A meets the requirements of the job, we have to wake up A

This causes A to change frequently, while B is in an idle state, wasting energy in vain

Similarly, waking up the node in advance will have similar problems

Images