【摘要】：云計(jì)算是新一代IT模式,它從網(wǎng)格計(jì)算,并行計(jì)算和分布式計(jì)算發(fā)展而來(lái),用戶可以使用它來(lái)方便地按需通過(guò)網(wǎng)絡(luò)訪問(wèn)一個(gè)可配置的計(jì)算資源(如計(jì)算、網(wǎng)絡(luò)、存儲(chǔ)、應(yīng)用和服務(wù)等)的共享池,只需最小化的管理工作量或服務(wù)提供商干預(yù)就可以快速地開通和釋放資源。當(dāng)前云環(huán)境中的資源都是通過(guò)虛擬化技術(shù)將底層的硬件資源進(jìn)行虛擬化,形成一個(gè)龐大的虛擬資源池之后然后通過(guò)動(dòng)態(tài)伸縮的部署方式以服務(wù)的形式提供給用戶。隨著使用云計(jì)算的用戶持續(xù)的增加,云數(shù)據(jù)中心的規(guī)模也不斷的加大,如何讓使云中的虛擬化資源高效的利用并快速的提供給用戶,減少用戶等待時(shí)間同時(shí)提高整個(gè)云數(shù)據(jù)中心的利用率,這已經(jīng)成為目前云計(jì)算環(huán)境中虛擬機(jī)資源動(dòng)態(tài)調(diào)度的重要問(wèn)題。 本文主要研究云數(shù)據(jù)中心虛擬化資源的動(dòng)態(tài)調(diào)度策略,在總結(jié)前人工作的基礎(chǔ)之上,基于當(dāng)前最熱門的開源云計(jì)算平臺(tái)OpenStack展開以下一系列工作和創(chuàng)新之處： (1)分析了當(dāng)前云計(jì)算的基本特征、體系架構(gòu)和關(guān)鍵技術(shù),對(duì)比了幾種開源的云計(jì)算平臺(tái),同時(shí)對(duì)云數(shù)據(jù)中心的資源調(diào)度和管理的需求和關(guān)鍵技術(shù)進(jìn)行詳細(xì)的研究。 (2)基于OpenStack的架構(gòu)對(duì)虛擬化資源進(jìn)行建模,從服務(wù)層面和資源層面分別對(duì)資源進(jìn)行描述,并提出面向計(jì)算資源實(shí)時(shí)監(jiān)測(cè)反饋綜合負(fù)載均衡調(diào)度策略和算法,分別以CPU、內(nèi)存、存儲(chǔ)和網(wǎng)絡(luò)帶寬四個(gè)維度對(duì)云平臺(tái)的計(jì)算資源進(jìn)行綜合負(fù)載均值分析,同時(shí)分別計(jì)算出云平臺(tái)的數(shù)據(jù)中心和物理服務(wù)器的不均衡度。通過(guò)在cloudsim仿真平臺(tái)對(duì)本算法與輪轉(zhuǎn)調(diào)度算法、OpenStack調(diào)度算法以及隨機(jī)選擇算法進(jìn)行對(duì)比實(shí)驗(yàn),結(jié)果表明本文提出的算法能夠使申請(qǐng)的虛擬機(jī)實(shí)例獲得更佳的部署位置,能夠使云數(shù)據(jù)中心的資源達(dá)到更加的負(fù)載均衡,證明了本算法的有效性和穩(wěn)定性。 (3)結(jié)合集群管理工具xCAT并對(duì)其進(jìn)行二次開發(fā)了在OpenStack環(huán)境下的資源自動(dòng)化管理平臺(tái),能夠有效的對(duì)資源進(jìn)行監(jiān)控管理,并能以自動(dòng)化的方式動(dòng)態(tài)擴(kuò)展云環(huán)境下的資源,實(shí)現(xiàn)規(guī)�；淖詣�(dòng)化運(yùn)維以及裸機(jī)部署管理。
[Abstract]:Cloud computing is a new generation of IT model, which is developed from grid computing, parallel computing and distributed computing. It can be used by users to easily access a configurable computing resource (such as computing, network, storage, etc.) via the network on demand. The shared pool of applications, services, etc., can be quickly opened and released with minimal management effort or service provider intervention. At present, the resources in the cloud environment are virtualized by virtualization technology, forming a huge virtual resource pool, and then providing the users with services through dynamic scalable deployment. With the continuous increase in the number of users using cloud computing, the scale of cloud data centers is also increasing. How to make the virtualization resources in the cloud efficient and quickly available to users, Reducing the waiting time of users and improving the utilization of the whole cloud data center has become an important issue of dynamic scheduling of virtual machine resources in cloud computing environment. This paper mainly studies the dynamic scheduling strategy of cloud data center virtualization resources. Based on the most popular open source cloud computing platform OpenStack, the following works and innovations are carried out: (1) the basic characteristics, architecture and key technologies of current cloud computing are analyzed, and several open source cloud computing platforms are compared. At the same time, the requirements and key technologies of resource scheduling and management in cloud data center are studied in detail. (2) based on the architecture of OpenStack, the virtual resources are modeled, the resources are described from the service level and the resource level, and a real-time monitoring feedback comprehensive load balancing scheduling strategy and algorithm for computing resources are proposed, respectively, using CPU, memory. The four dimensions of storage and network bandwidth are used to analyze the average load of computing resources of cloud platform. At the same time, the unbalance of data center and physical server of cloud platform are calculated respectively. By comparing the algorithm with rotation scheduling algorithm, OpenStack scheduling algorithm and random selection algorithm on the cloudsim simulation platform, the results show that the proposed algorithm can obtain a better deployment location for the applied virtual machine instance. It can make the resources of the cloud data center achieve more load balance, which proves the validity and stability of the algorithm. (3) combined with cluster management tool xCAT and secondary development of resource automation management platform in OpenStack environment, it can effectively monitor and manage resources, and can dynamically expand resources in cloud environment in an automatic way. Realize the automatic operation and maintenance of scale and the deployment management of naked machine.
【學(xué)位授予單位】：大連理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2013
【分類號(hào)】：TP308

【參考文獻(xiàn)】

相關(guān)期刊論文 前6條

1 華夏渝;鄭駿;胡文心;;基于云計(jì)算環(huán)境的蟻群優(yōu)化計(jì)算資源分配算法[J];華東師范大學(xué)學(xué)報(bào)(自然科學(xué)版);2010年01期

2 孫瑞鋒;趙政文;;基于云計(jì)算的資源調(diào)度策略[J];航空計(jì)算技術(shù);2010年03期

3 張前進(jìn);齊美彬;李莉;;基于應(yīng)用層負(fù)載均衡策略的分析與研究[J];計(jì)算機(jī)工程與應(yīng)用;2007年32期

4 李強(qiáng);郝沁汾;肖利民;李舟軍;;云計(jì)算中虛擬機(jī)放置的自適應(yīng)管理與多目標(biāo)優(yōu)化[J];計(jì)算機(jī)學(xué)報(bào);2011年12期

5 陳全;鄧倩妮;;云計(jì)算及其關(guān)鍵技術(shù)[J];計(jì)算機(jī)應(yīng)用;2009年09期

6 岳冬利;劉海濤;孫傲冰;;IaaS公有云平臺(tái)調(diào)度模型研究[J];計(jì)算機(jī)工程與設(shè)計(jì);2011年06期

相關(guān)碩士學(xué)位論文 前4條

1 葛新;基于云計(jì)算集群擴(kuò)展中的調(diào)度問(wèn)題研究[D];中國(guó)科學(xué)技術(shù)大學(xué);2011年

2 張先哲;分布式系統(tǒng)中的負(fù)載平衡檢測(cè)與優(yōu)化策略研究[D];河南大學(xué);2009年

3 趙春燕;云環(huán)境下作業(yè)調(diào)度算法研究與實(shí)現(xiàn)[D];北京交通大學(xué);2009年

4 劉鵬程;云計(jì)算中虛擬機(jī)動(dòng)態(tài)遷移的研究[D];復(fù)旦大學(xué);2009年

，

本文編號(hào)：2307038