Research And Implementation Of Dynamic Resource Scheduling Method Based On OpenStack

Cloud computing has become one of the most hot topics in IT field, it brought a great changes for the management and usage mode of traditional IT resources.With the extensive use of cloud computing, how to make use of the entire cloud datacenter resources has become one of the most important problems in cloud computing enviro-nment for resource scheduling.OpenStack is the most mainstream open source cloud IaaS platform in cloud computing area,has become the open source products of IT ent-erprise, it encapsulates the underlying huge hardware resource into cloud service,and provides user with all kinds of public resources, such as operating system, billing,com-puting, storage, network, etc.But OpenStack also has many shortcomings in the initial virtual machine allocation and resources dynamic migration,for example,when the new virtual machine is created, or the virtual machine live migration is occurred, it only considers the memory utilization of physical machines,and does not take other resource metrics into account, such as the CPU utilization, hard disk utilization, network band width, etc. In order to improve the resource utilization of cloud data center, the paper focuses on the OpenStack resource scheduling mechanism, the main studies and inno-vations are as follows:(1) Combine with the OpenStack source code for further, analyze the virtual machine initial allocation module and the virtual machine live migration module, then points out the shortage.(2) Put forward the dynamic resource scheduling strategy based on the Openstack cloud platform, namely comprehensive load optimization scheduling strategy. This strategy is mainly aimed at the virtual machine initial allocation and virtual machine live migration mechanism, and elaborates the comprehensive load optimization scheduling strategy execution process, algorithms and pseudo code in detail.(3) Realize OpenStack cloud platform comprehensive load optimization scheduling strategy, the strategy mainly includes two parts:the Python API and global configuration XML file.(4) Deploy comprehensive load optimization scheduling strategy of Python API and XML files into the OpenStack Scheduler module, then constructing the platform to validate it. The experimental results show that the comprehensive load optimization scheduling policy changes the original scheduling policy, and improves the resource utilization of the physical machine.