Font Size: a A A

A Scheduling Plan Of Cloud Platform Virtual Machine Based On Cloud Users

Posted on:2018-11-30Degree:MasterType:Thesis
Country:ChinaCandidate:Q P WangFull Text:PDF
GTID:2348330518461455Subject:Engineering
Abstract/Summary:PDF Full Text Request
With the development of the cloud computing technology,the demand of IT infrastructure has been from the traditional equipment purchase into on-demand leasing.However,people always assessed the resource consumption of their applications unreasonably and lead users to buy virtual machines large or small.Resource utilization was very low.For the cloud provider,they needs to have a certain understanding of the user's resources used before resource allocation to make a better strategy in developing a virtual machine deployment.To solve these problems,this paper proposes a cloud virtual machine scheduling plan based on cloud user application.It is divided into two parts,which are cloud application demand resource estimation and virtual machine scheduling planning.Based on the change of cloud application parameters,this paper presents a prediction method based on parameter variation of cloud application for resources allocation.According to the idea of the interpreter,analysis program source code statically,obtained task execution statement number;through the curve fitting and fewer times program running,draw the trend of running time of the program with the change of the parameters.Finally,according to limit the proportion of CPU occupancy,the relation between parameters,user program running time and CPU resources occupied was obtained.Hanoi Pascal program was used to verify the method,the method was found to be very close to the actual running time and the error rate was very small.By limiting the occupancy rate of CPU,the actual running time and expected operating time error are small under different cpu-MIPS values.This paper proposes a virtual machine resource allocation scheme based on load forecasting,avoid host overload by dual mechanism that combined with the static parameters and dynamic monitoring,filter out the hosts who reached the warning value by forecasting the situation of host load.Use BFD algorithm issued of two-dimensional packing problem to maximize the utilization of resources.The simulation results show that the proposed algorithm has better performance than the polling algorithm.Experiments show that contrasting with the polling algorithm the proposed scheme can improve the resource utilization ratio when the A2 parameter is 0.95.While the host load is relatively stable,the host performance is relatively stable.
Keywords/Search Tags:Cloud application, Resource allocation prediction, Load forecasting, Virtual machine, Resource utilization
PDF Full Text Request
Related items