基于虚拟化的云端数据中心的设计及负载平衡的实现
发布时间：2018-08-31 13:32
【摘要】：在当今科学技术快速发展的大背景下，互联网已经融入我们生活的每个角落，大量的数据在网上涌现，，与此同时大量的数据中心也如雨后春笋般涌现，传统的数据中心的模式已经无法完美满足现代大量数据的存储和处理问题，亟需引入新的技术概念来进行改进。云计算的概念在这种大环境需求下诞生了，云计算本身的资源高利用率，能耗合理，良好的平衡体系，以及其稳定的运行保障，给传统数据中心棘手的问题提供了良好的解决办法，以云计算为基础的数据中心将成为新一代数据中心发展的主流趋势。 本文从中小型企业对信息服务的需求出发，注重考虑云计算中的虚拟化和资源分配，海量数据处理以及数据中心的构建模式等众多因素，给出了一个基于虚拟化的云端数据中心的设计，并给出了其中集群化非关系数据库中负载均衡器的实现。 本文主要工作如下： 1.分析现今数据中心的现状和面临问题，通过对云计算的关键技术：虚拟化，海量数据存储，资源调度等的研究，结合传统数据中心的优点，重点考虑了数据中心数据库的建立方式、系统的负载情况，任务的并发访问等方面，给出了一种基于虚拟化的云数据中心的构架。 2.根据已有的非关系数据库原型，对其进行数据库集群的扩展，设计相应的管理节点和数据库节点功能，并给出节点间的通信消息序列。对管理节点进行详细的功能划分。给出其重要模块和重要策略的设计和解决办法。 3.通过采用VMWare虚拟化软件对于计算机物理资源进行虚拟化，给出虚拟机、专用磁盘的建设和加载，实现资源池的构建，在实验环境下给出一个云平台的模拟。 4.设计上述数据库集群中的负载均衡器。在其基础上研究高效的动态负载平衡算法，给出其设计和实现方法，并通过实验对算法的性能进行测试分析。
[Abstract]:With the rapid development of science and technology, the Internet has been integrated into every corner of our lives, a large number of data are emerging on the Internet. At the same time, a large number of data centers are springing up like bamboo shoots. The traditional data center model has been unable to satisfy the storage and processing problems of a large number of modern data, which needs to be introduced urgently. The concept of cloud computing was born under the demand of this big environment. The high utilization of resources, reasonable energy consumption, good balance system and stable operation guarantee of cloud computing provide a good solution to the difficult problems of traditional data center. It will become the main trend of the development of the new generation data center.
Starting from the requirement of information service for small and medium-sized enterprises, this paper considers the virtualization and resource allocation in cloud computing, massive data processing and the construction mode of data center, gives a design of cloud data center based on virtualization, and gives a load balancer in clustered non-relational database. Realization.
The main works of this paper are as follows:
1. Analyze the current status and problems of data center. Through the research of key technologies of cloud computing, such as virtualization, mass data storage, resource scheduling, and combining the advantages of traditional data center, this paper puts emphasis on the establishment of data center database, the load of the system, the concurrent access of tasks and so on. In the framework of virtualized cloud data center.
2. According to the existing non-relational database prototype, we extend the database cluster, design the corresponding functions of management nodes and database nodes, and give the communication message sequence between nodes.
3. Using VMWare virtualization software to virtualize the physical resources of computer, the construction and loading of virtual machine and special disk are given, the construction of resource pool is realized, and a simulation of cloud platform is given in the experimental environment.
4. Designing the load balancer in the above database cluster. On the basis of it, studying the efficient dynamic load balancing algorithm, giving its design and implementation methods, and testing and analyzing the performance of the algorithm through experiments.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP308;TP311.13
本文编号：2215135
[Abstract]:With the rapid development of science and technology, the Internet has been integrated into every corner of our lives, a large number of data are emerging on the Internet. At the same time, a large number of data centers are springing up like bamboo shoots. The traditional data center model has been unable to satisfy the storage and processing problems of a large number of modern data, which needs to be introduced urgently. The concept of cloud computing was born under the demand of this big environment. The high utilization of resources, reasonable energy consumption, good balance system and stable operation guarantee of cloud computing provide a good solution to the difficult problems of traditional data center. It will become the main trend of the development of the new generation data center.
Starting from the requirement of information service for small and medium-sized enterprises, this paper considers the virtualization and resource allocation in cloud computing, massive data processing and the construction mode of data center, gives a design of cloud data center based on virtualization, and gives a load balancer in clustered non-relational database. Realization.
The main works of this paper are as follows:
1. Analyze the current status and problems of data center. Through the research of key technologies of cloud computing, such as virtualization, mass data storage, resource scheduling, and combining the advantages of traditional data center, this paper puts emphasis on the establishment of data center database, the load of the system, the concurrent access of tasks and so on. In the framework of virtualized cloud data center.
2. According to the existing non-relational database prototype, we extend the database cluster, design the corresponding functions of management nodes and database nodes, and give the communication message sequence between nodes.
3. Using VMWare virtualization software to virtualize the physical resources of computer, the construction and loading of virtual machine and special disk are given, the construction of resource pool is realized, and a simulation of cloud platform is given in the experimental environment.
4. Designing the load balancer in the above database cluster. On the basis of it, studying the efficient dynamic load balancing algorithm, giving its design and implementation methods, and testing and analyzing the performance of the algorithm through experiments.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP308;TP311.13
【参考文献】
相关博士学位论文 前1条
1 龚卫华;数据库集群系统的关键技术研究[D];华中科技大学;2006年
相关硕士学位论文 前1条
1 万春;基于Linux的数据库集群系统的研究[D];华中科技大学;2004年
本文编号：2215135
本文链接：https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2215135.html
