基于IaaS的云计算平台的研究与实现

发布时间：2018-07-12 16:14

本文选题：云计算 + IaaS云计算平台　；参考：《东华大学》2016年硕士论文

【摘要】：随着近几年来虚拟化技术不断进步,网络带宽不断提高以及计算机的能力不断拓展,这三种技术的结合产物——云计算得到了空前的发展。云计算平台通过汇集各处的闲置计算资源,从而形成一个资源池。用户通过平台提供的接口获取资源池中的资源:比如计算资源、存储资源等。而在处理用户提交的任务时,平台会根据任务调度算法将任务分配到选择的资源节点上完成工作。但是,公有云由于其不能自定义任务负载均衡算法,从而会导致任务处理效率低下。如何搭建用户的私有云,并且优化其任务负载均衡算法正逐渐成为当前的研究热点之一。本文以提高私有云平台中的任务负载均衡效率为研究目标,对如何搭建私有云计算平台框架和任务负载均衡算法进行了研究。本文的主要研究内容如下:针对目前的云计算搭建框架,通过对比他们之间的特点,选择最适合要求的私有云计算平台框架。本文重点分析了OpenStack云计算框架,并利用该框架设计和部署了基于计算机集群的私有云计算平台。此外,本文还研究了目前主流的云平台任务负载均衡算法。算法提出了资源效益度的概念,在输入DAG任务数量不太大的情况下,通过牺牲部分任务完成时间,显著提高了系统的任务负载均衡能力。
[Abstract]:With the development of virtualization technology, the improvement of network bandwidth and the expansion of computer capability in recent years, the combination of these three technologies-cloud computing has been unprecedented development. Cloud computing platforms form a pool of resources by pooling idle computing resources everywhere. The user acquires the resources in the resource pool through the interface provided by the platform, such as computing resources, storing resources, etc. When processing the task submitted by the user, the platform will assign the task to the selected resource node to complete the work according to the task scheduling algorithm. However, due to its inability to customize the task load balancing algorithm, the public cloud can result in inefficient task processing. How to set up the private cloud and optimize the load balancing algorithm is becoming one of the research hotspots. In order to improve the efficiency of task load balancing in private cloud platform, this paper studies how to build a private cloud computing platform framework and task load balancing algorithm. The main contents of this paper are as follows: according to the current cloud computing framework, by comparing their characteristics, select the most suitable private cloud computing platform framework. This paper focuses on the analysis of OpenStack cloud computing framework and designs and deploys a private cloud computing platform based on computer cluster. In addition, this paper also studies the current mainstream cloud platform load balancing algorithm. The concept of resource efficiency is put forward in the algorithm. When the number of DAG tasks is not too large, the task load balancing ability of the system is improved significantly by sacrificing some task completion time.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.52

【参考文献】