【摘要】：隨著云計算的發(fā)展，云環(huán)境下的數據中心規(guī)模不斷增大，極大增加了計算部署的工作量。云數據中心環(huán)境部署包括了硬件資源（服務器）、軟件資源（中間件或數據庫等)、網絡資源和存儲資源，其中包括了虛擬化設備和不能被虛擬化的設備。目前云數據中心主要關注部署速度、部署的自動化程度以及部署的友好界面，并沒有關注數據中心中部署服務的可靠性、部署服務器的瓶頸問題，這將影響系統(tǒng)業(yè)務及時上線，造成不必要的損失。如何在云環(huán)境下數據中心中進行資源自動部署以及如何保證部署服務的可用性就成為了一個重要的研究課題。 針對上述問題，提出了一種面向云環(huán)境下數據中心的計算環(huán)境自動部署系統(tǒng)。基于Petri網絡的工作流引擎可以有效處理部署流程的依賴關系，同時將kickstart技術與增強部署服務可用性、部署服務器可靠性技術相結合，，設計并實現云數據中心自動部署系統(tǒng)。云數據中心自動部署系統(tǒng)具有具有可靠性、可用性和負載均衡特性。系統(tǒng)引入了網卡綁定技術、獨立冗余磁盤技術和Heartbeat技術，使之與kickstart之間無縫連接，解決了大規(guī)模云環(huán)境下節(jié)點安裝過程中部署服務器的單點故障問題，同時提出了基于配置文件和基于DNS的負載均衡的算法，解決了節(jié)點部署過程中服務器的瓶頸問題。 功能測試表明，基于Petri網的工作流引擎很好完成了云環(huán)境下跨地域物理資源的自動部署，為云數據中心的管理員提供有效的管理方式。性能測試表明系統(tǒng)能夠減少節(jié)點的安裝時間和提高節(jié)點部署成功率，相比于傳統(tǒng)的數據中心，其部署成功率在90%以上。
[Abstract]:With the development of cloud computing, the scale of data center in cloud environment is increasing, which greatly increases the workload of computing deployment. Cloud data center environment deployment includes hardware resources (servers), software resources (middleware or database), network resources and storage resources, including virtualization devices and devices that cannot be virtualized. At present, the cloud data center mainly pays attention to the deployment speed, the degree of automation and the friendly interface of the deployment, but not the reliability of the deployment service in the data center, the bottleneck problem of the deployment server, which will affect the system business to get online in time. Cause unnecessary losses. How to automatically deploy resources in the data center in the cloud environment and how to ensure the availability of deployment services have become an important research topic. In order to solve the above problems, a computing environment automatic deployment system for data center in cloud environment is proposed. The workflow engine based on Petri network can deal with the dependency of deployment process effectively. At the same time, the automatic deployment system of cloud data center is designed and implemented by combining kickstart technology with enhanced availability of deployment service and reliability technology of deployment server. Cloud data center automatic deployment systems have reliability, availability, and load balancing characteristics. The network card binding technology, independent redundant disk technology and Heartbeat technology are introduced in the system, which makes it connect seamlessly with kickstart, and solves the single point fault problem of deploying server in the process of node installation in large-scale cloud environment. At the same time, a load balancing algorithm based on configuration file and DNS is proposed, which solves the bottleneck problem of server in the process of node deployment. The functional tests show that the workflow engine based on Petri net can accomplish the automatic deployment of physical resources across regions in the cloud environment and provide an effective management method for the administrators of the cloud data center. The performance test shows that the system can reduce the installation time of nodes and improve the success rate of node deployment. Compared with the traditional data center, the success rate of deployment is more than 90%.
【學位授予單位】：華中科技大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：TP308

【參考文獻】

相關期刊論文 前3條

1 董小社;孫發(fā)龍;李紀云;胡雷均;;基于映像的集群部署系統(tǒng)設計與實現[J];計算機工程;2005年24期

2 劉青昆;陰元友;鄭曉薇;劉紹海;;基于Kickstart的Linux網絡安裝新模型[J];計算機工程;2008年21期

3 陳誠;陳海濤;;一種自適應的容災系統(tǒng)心跳檢測算法[J];計算機工程與科學;2008年05期

本文編號：2284403