Research And Implementation Of Openstack Based Cloud Platform Automated Deployment Mechanism

Cloud computing is a very hot research direction in IT field in recent years. It will be distributed in different geographical location infrastructure resources are virtualized as a huge pool of resources, users only need to through the network connection, you can use cloud computing provides a variety of cloud services, and on-demand access to cloud computing platform in the computing, storage and network resources. OpenStack is the most popular open source IaaS platform in the field of cloud computing. So this paper choose OpenStack open source project as a development platform.With the rapid development of computer information technology, the users of the university laboratory are increasing day by day, which leads to the sharing of hardware and software resources and the contradiction between the individual and the individual. In order to meet the college laboratory hardware equipment of high resource utilization rate and mass customization experimental environment demands, this paper based on the open-source openstack projects build a suitable for University Laboratory of private cloud computing platform, provides on-demand cloud services using computer hardware resources and support user customization of experimental environment and batch to create a virtual machine service for the user. In this paper, the following work has been carried out:(1) Combined with openstack installation source code and documentation design a set of can automatic deployment of cloud computing platform of shell scripts, and then use the script to build a cloud computing platform for laboratories in Colleges and universities, the script is greatly simplifies the deployment of cloud computing platform of complex installation process and save the corresponding installation time. In order to meet the needs of the individualized laboratory test environment configuration are also studied in detail the through the virtual machine configuration files to customize the virtual machine image method, for the subsequent batch create virtual machine solutions to lay a good foundation.(2) According to the requirements of the laboratory environment of university laboratory, a new method based on qcow2 image to generate the virtual machine(the user’s experimental environment) is designed. The scheme mainly through copy and modify the mirror qcow2 automatically generate virtual machine name, MAC address and IP information, and to revise the definition of the virtual machine configuration XML file, then the Shell Scripting batch to create a virtual machine, greatly reduces the experimental cloud platform quantities to create and difficulty of the virtual machine management.(3) By analyzing the deficiencies of the OpenStack default resource scheduling algorithm, an optimized OpenStack resource scheduling strategy is proposed. In the host filtration stage, the compute node for the first time enough to create virtual machine resources to provide second chance to choose; in the stage of weight calculation, the effects of computer of load factor expands from the mere memory to CPU, memory, disk, take a cost function to calculate the host weight; at last, it puts forward the improved openstack scheduling strategy, through the analysis and comparison of experiment to prove the improved scheduling strategy can effectively improve the utilization of laboratory hardware equipment, and reduce the failure probability of the test environment configuration, promote the stability and reliability of the cloud experiment platform.