Machine learning-based server energy consumption prediction method and system

Abstract

The invention relates to a machine learning-based server energy consumption prediction method and system. The method comprises the following steps: separately collecting system resource utilization rate and real-time energy consumption of a server under a zero load state, a low load state and a high load state, as a zero load training set, a low load training set and a high load training set; respectively, inputting the zero load training set, the low load training set and the high load training set into a machine learning model for training, and generating a zero load energy consumption model, a low load energy consumption model, and a high load energy consumption model; according to the to-be-predicted system resource utilization rate of the server, selecting the zero load energy consumption model, or the low load energy consumption model, or the high load energy consumption model to predict the energy consumption value of the server, wherein the system resource utilization rate comprises the memory utilization rate, the network bandwidth utilization rate, the disk utilization rate, and the CPU utilization rate, thereby improving the prediction accuracy and practicability by expanding the collected system resource utilization rate parameters and the segmentation training models.

Description

technical field [0001] The present invention relates to the field of machine learning, in particular to a method and system for predicting server energy consumption based on machine learning. Background technique [0002] With the popularization of big data and cloud computing technology and the increase in business volume and data volume, the scale of server clusters is increasing day by day, and the energy consumption generated is also increasing, which leads to an increase in energy consumption costs. Therefore, a cloud operating system Or the optimization of server energy consumption in a data center has become a more important issue in the current technical environment. The traditional method of measuring power is to directly measure the electrical parameters of the server through electrical instruments to obtain the actual power of the server, but this physical measurement method can only obtain accurate actual power, and it is impossible to analyze what caused the ser...

AI Technical Summary

Problems solved by technology

However, this method is to monitor the external environment of the server and make predictions based on this. The external environmental factor of the temperature used by it can reflect the performance and energy consumption of the server, which is far less accurate than the internal performance parameters of the server.

Images