基于OpenStack的天文数据处理私有云的开发研究

发布时间：2018-03-23 05:26

本文选题：云计算　切入点：天文数据处理　出处：《昆明理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：天文数据处理是天文学研究中的一个重要环节。随着新一代望远镜功能与观测能力的快速发展,在观测地点构建高性能实时计算平台、快速完成数据的分析与处理已经是一种趋势。OpenStack是一个高效灵活并且开源的云平台技术,它的功能丰富,完全能够满足天文数据处理的需求。本文就是以OpenStack云平台技术在天文数据处理领域中的应用为研究背景,详细分析了 OpenStack的总体架构,实现了基于OpenStack的天文数据处理私有云,主要完成了以下工作:一、通过研究OpenStack的架构,实现了基于OpenStack私有云的自动化部署,分为控制节点的部署和计算节点的部署。这种部署方式可以有效地减少人工搭建私有云的时间,节省人力物力资源的消耗;二、设计实现了私有云计算资源的自动管理模式,该模式可以根据私有云中计算集群当前的计算量自动增减计算资源。实验表明该模式完全可以满足天文数据处理的需求,并且比传统的静态分配计算资源的数据处理方法更具高效性;三、使用镜像技术为私有云实现了 UVFITS文件的自动生成;基于Swift组件设计了符合科研人员使用习惯的天文观测数据的存储功能。整体来看,本文将OpenStack云平台技术与天文数据处理相结合,设计实现的私有云功能可以有效地节省资源开销、降低观测成本,对未来天文台站高性能计算平台的建设有一定参考价值。
[Abstract]:Astronomical data processing is an important part of astronomical research. With the rapid development of the new generation of telescopes' functions and observational ability, a high performance real-time computing platform is constructed at observation sites. Fast data analysis and processing is a trend. OpenStack is an efficient, flexible and open source cloud platform technology, which is rich in functionality. This paper takes the application of OpenStack cloud platform technology in astronomical data processing field as the research background, analyzes the overall structure of OpenStack in detail, and realizes the private cloud processing based on OpenStack astronomical data processing. The main works are as follows: firstly, by studying the architecture of OpenStack, the automatic deployment based on OpenStack private cloud is realized. It is divided into control node deployment and computing node deployment. This deployment method can effectively reduce the time of manually building private cloud and save the consumption of manpower and material resources. Second, the design and implementation of private cloud computing resources management model, The model can automatically add or subtract computing resources according to the current computing amount of the cluster in the private cloud. Experimental results show that the model can fully meet the requirements of astronomical data processing. And it is more efficient than the traditional data processing method of static allocation of computing resources. Thirdly, the UVFITS file is automatically generated for private cloud by mirror technology. Based on the Swift component, the storage function of astronomical observation data is designed, which conforms to the scientific researchers' usage habits. On the whole, this paper combines the OpenStack cloud platform technology with astronomical data processing. The design and implementation of the private cloud function can effectively save the cost of resources and reduce the observation cost. It has a certain reference value for the future construction of the high performance computing platform of the observatory station.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P111;TP393.09

【相似文献】