Computing resource allocation method

Abstract

The invention provides a computing resource allocation method. The computing resource allocation method comprises the following steps: step 1, assessing all computing resources required by a task according to the assessment factors which comprise computing resource factors and limiting factors, judging whether the assessment is valid, if yes, actuating step 2, and if no, actuating step 4; step 2, if the assessment is valid, achieving pre-allocation of the computing resources according to the pre-allocation strategy, and mapping and adding the computing resources which are successfully pre-allocated to a task resource plan; step 3, actually allocating the computing resources, mapping and actuating the task according to the pre-allocated computing resources in the task resource plan; step 4, if the assessment is null and void, giving out assessment feedback and enabling the task required by the computing resources to belong to a computing pool. The computing resource allocation method realizes computing resource allocation as required, and ensures that a current cluster resource is effectively assessed before computing resource pre-allocation to expose the resource bottleneck as early as possible.

Description

technical field [0001] The present invention relates to an allocation method, in particular to a calculation resource allocation method. Background technique [0002] The computing resource allocation method of the cloud desktop cluster plays a crucial role in the entire cluster, and its advantages and disadvantages directly determine the utilization rate of the entire cluster resource and the load balancing status of each server under it. [0003] At present, in the existing technologies, resource allocation methods mainly have the following defects: [0004] 1. The existing computing resource allocation methods mainly only involve resource factors: CPU (Central Processing Unit) and memory. However, due to the lack of quantitative real-time statistics of resources in general cluster management systems, it is impossible to accurately schedule cluster computing resources. distribute; [0005] 2. When computing resource allocation, only consider whether the computing resourc...

AI Technical Summary

Problems solved by technology

[0004] 1. The existing computing resource allocation methods mainly only involve resource factors: CPU (Central Processing Unit) and memory. However, due to the lack of quantitative real-time statistics of resources in general cluster management systems, it is impossible to accurately schedule cluster computing resources. distribute;

[0005] 2. When computing resource allocation, only consider whether the computing resources in the target server are sufficient. If it can meet the requirements, the virtual machine operation task will be distributed to the node for processing. Otherwise, it will prompt "insufficient resources" and the execution will fail. This time the task will fail and you can only wait for resources. After the readjustment is satisfied, re-initiate the task, and the on-demand resource allocation cannot be realized;

[0006] 3. After the execution of the task fails for various reasons, the prompt provided to the administrator is simply "insufficient resources", which cannot correctly reflect the bottleneck in the process of computing resource allocation, which is not conducive to the administrator's subsequent resource expansion or policy adjustment, etc. Behavior

Images