Virtual Machine Performance Prediction Method Based on Random Forest Regression

Abstract

The invention discloses a virtual machine performance prediction method based on random forest regression. According to the resource characteristics of a virtual environment to be predicted, characteristics which may influence the performance of a virtual machine are extracted from four aspects including underlying hardware characteristics, virtual machine software characteristics, virtual machine resource configuration characteristics and environment characteristics during virtual machine operation, the virtual machines of different virtual machine performance characteristic combinations are configured, required performance index values are obtained through the operation of benchmark test programs to obtain a data sample set; and according to the data sample set, a random forest regression model is constructed, and the random forest regression model is adopted to carry out performance prediction on the virtual machine of a specific performance characteristic configuration. By use of the method, the random forest regression model is adopted to describe a relationship between virtual machine performance characteristics and performance indexes so as to effectively predict the virtual machine performance under the specific configuration.

Description

technical field [0001] The invention belongs to the technical field of cloud computing, and more specifically relates to a virtual machine performance prediction method based on random forest regression. Background technique [0002] The infrastructure-as-a-service model of cloud computing services supports users to rent computing resources with a certain level of computing power on demand in the form of virtual machines. Users pay rent according to the performance of resources and the lease time. When users need to rent a virtual machine, Need to choose a resource provider, and choose a specific configuration of virtual machines, at this time, the performance of the virtual machine is an important factor affecting the decision. Therefore, how to accurately evaluate the performance of virtual machines becomes a key issue. The factors that affect the performance of virtual machines mainly include three categories: one is the software and hardware parameters of the virtualiza...

AI Technical Summary

Problems solved by technology

However, the above studies did not address the impact of the load on other virtual machines in the virtual machine runtime environment on its performance

[0006] There are many features that affect the performance of virtual machines, and the relationship between each feature is complex and nonlinear, and different application performances are sensitive to changes in feature values. Traditional statistical methods are challenged when analyzing these data. Especially for the generalized linear model, it is impossible to obtain a more accurate regression model for the complex relationship between features, and the universality of the model is not good

Images