Research On Resource Scheduling Method Based On OpenStack Cloud Platform

Developing from grid computing,parallel computing and distributed computing,cloud computing has become a new IT computing pattern which serves as a easily extended way for enabling users to gain the computing resources on-demand through Internet,e.g.,computation,storage and network resources.Leveraging the virtualization technology,clouding computing virtualizes the underlying physical hardware and constructs a huge virtual resource pool,ng,posing new challenges to cloud computing which includes how to allocate physical resources and how to select appropriate resource scheduling of virtual machines for improving the performance of cloud data center and cloud application,as well as reducing its processing time.Based on the OpenStack platform,this thesis studies the resource scheduling problem of cloud data center with regard to the virtual machine scheduling algorithm of cloud computing.The main works are as follows.(1)This thesis first presents the basic concepts and related characteristics of cloud computing,and analyzes its three main application frameworks.In addition to a brief introduction to virtualization and the mainstream virtualization technology,this thesis elaborates the organizational structure and logical architecture of the OpenStack platform and analyzes its core components.(2)This thesis proposes a novel particle swarm-based virtual machine placement algorithm which aims at the high communication delay problem of applications in complex network.Taking four factors,i.e.,CPU,memory,bandwidth and network traffic,into account,this algorithm constructs the delay model and applies the particle swarm optimization to reduce the application delay for better operation efficiency.(3)This thesis deigns the virtual machine scheduling module with the Nova component of the OpenStack platform,whose system architecture and function module are further designed according to the WFPSO algorithm.This thesis implements the deployment of the OpenStack platform and integrates the virtual machine scheduling module on it.(4)With respect of functional testing,this thesis tests the new virtual machine scheduling module-customized OpenStack platform both in its usability and validity.To test the performance of the WFPSO algorithm,this thesis adopts the CloudSim simulation platform for large-scale cluster testing.Experimental results validate that the WFPSO algorithm can greatly improve the operation efficiency of applications in the cloud environment.