Load resource demand assessment method based on linear relationship

Abstract

The invention discloses a load resource demand assessment method based on a linear relationship. In view of the problem that the load resource demand in an energy-aware cluster system is difficult toassess because of load fluctuation, a general linear relationship is constructed between load characteristics and the load resource demand, the relationship is used to adopt an assessment method withlinear regression as a basic algorithm, and finally, through the assessment method, accurate load resource demand assessment is made. In a system operation stage, correct load characteristics and thedemand information are constantly collected. The specific values, which are consistent with the real situation, of variable parameters in the general linear relationship are calculated by the assessment method based on the demand information through adopting a linear regression mode. After the specific parameter values in the general linear relationship are determined, a linear relationship applicable to the load condition is determined. Further, the assessment method substitutes the current load characteristics into the determined relationship, and finally, the current load demand assessmentvalue is calculated.

Description

technical field [0001] The present invention relates to the field of computer system structure, in particular to a linear relationship-based load resource demand evaluation method, and more specifically to constructing a linear relationship between load characteristics and load resource demand, based on a general linear relationship The actual situation obtains a definite relationship consistent with the actual situation, and finally realizes an accurate assessment of the demand for load resources based on the definite relationship obtained. Background technique [0002] In the energy-aware cluster system, due to the variability of the load, the resource demand of the load (resource demand is defined as: the minimum amount of hardware resources required by the load under the premise of meeting the quality of service requirements) also changes accordingly . In order to ensure the quality of service of the system under the premise of saving energy, the amount of resources pro...

AI Technical Summary

Problems solved by technology

However, most of these technologies have two disadvantages: first, the process of modeling the actual system is complex and time-consuming, and second, there is a large difference between the established model and the actual system

For the first shortcoming, simplified models are used in the implementation of some technical methods, which leads to a large difference between the evaluation results and the actual

For the second shortcoming, some technical methods set a variety of parameters for system modeling, and finally form a complex model, but it is difficult to truly perform parametric modeling in practical applications and it is difficult to use

Images