本文選題：服務(wù)器集群　切入點(diǎn)：云計(jì)算　出處：《西安電子科技大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：當(dāng)今社會(huì)，Web服務(wù)器系統(tǒng)面臨兩個(gè)顯著變化：(1)隨著網(wǎng)絡(luò)速度的提高，越來(lái)越多的本地任務(wù)處理將依賴遠(yuǎn)程服務(wù)，這就要求Web服務(wù)器系統(tǒng)具有高性能；(2)服務(wù)器所承受的負(fù)載量變化更具突發(fā)性。隨著移動(dòng)互聯(lián)網(wǎng)的發(fā)展，，請(qǐng)求網(wǎng)絡(luò)服務(wù)的方式趨于多樣化。這種情況下系統(tǒng)平均吞吐量并不一定很大，但會(huì)頻繁變化，這就要求Web服務(wù)器具備很好的彈性擴(kuò)展能力。面對(duì)上述需求，通常的解決方法包括兩個(gè)方面。一個(gè)方面是提升Web服務(wù)器的硬件處理速度，然而實(shí)際應(yīng)用中，這種方式首先會(huì)造成資金成本的增加，其次服務(wù)器硬件提升空間也是有限的。另一個(gè)方面是采用基于單一系統(tǒng)鏡像的Web服務(wù)器集群系統(tǒng)，這也是目前應(yīng)用最廣泛的一種方法。然而這種方式也有著其難以克服的缺陷，其中最重要的一點(diǎn)是，應(yīng)對(duì)變化的負(fù)載時(shí)，需要按照之前既定策略或是人工方式調(diào)整服務(wù)器集群規(guī)模，這已經(jīng)無(wú)法滿足當(dāng)前頻繁變化且無(wú)法預(yù)估的網(wǎng)絡(luò)負(fù)載變化。 為了彌補(bǔ)上述缺陷，本文提出了一種基于云技術(shù)的彈性負(fù)載均衡集群式服務(wù)器架構(gòu)（Elastic and Load Balancing Cluster Server Architecture based on CloudTechnology,ELC集群式服務(wù)器架構(gòu)）。本架構(gòu)以Eucalyptus云計(jì)算平臺(tái)為基礎(chǔ)設(shè)施，按照計(jì)算與存儲(chǔ)分離原則整體分為兩部分。其中，計(jì)算服務(wù)系統(tǒng)在Eucalyptus云平臺(tái)的計(jì)算模塊基礎(chǔ)上，通過(guò)虛擬服務(wù)器實(shí)例方式對(duì)外提供服務(wù)，并最終實(shí)現(xiàn)了根據(jù)集群當(dāng)前負(fù)載狀況自動(dòng)調(diào)整集群規(guī)模。數(shù)據(jù)存儲(chǔ)系統(tǒng)在Eucalyptus的存儲(chǔ)模塊基礎(chǔ)上，通過(guò)掛載NAS設(shè)備和創(chuàng)建HDFS分布式文件系統(tǒng)解決了存儲(chǔ)擴(kuò)容和存儲(chǔ)共享問(wèn)題，為系統(tǒng)提升了存儲(chǔ)能力以及存儲(chǔ)擴(kuò)展性。 最后，通過(guò)使用負(fù)載測(cè)試軟件對(duì)本文提出的ELC集群式服務(wù)器系統(tǒng)進(jìn)行測(cè)試和結(jié)果分析，驗(yàn)證了系統(tǒng)能夠按照預(yù)期的方式自動(dòng)調(diào)整服務(wù)器集群規(guī)模，以適應(yīng)當(dāng)前集群負(fù)載狀況，并且實(shí)現(xiàn)了存儲(chǔ)擴(kuò)容和存儲(chǔ)共享。
[Abstract]:As the speed of the network increases, more and more local task processing will rely on remote services. This requires that Web server systems have high performance and the load load on the server changes more suddenly. With the development of the mobile Internet, In this case, the average throughput of the system may not be very large, but it will change frequently, which requires the Web server to have a good flexibility to extend. The common solution includes two aspects. One is to speed up the hardware processing of the Web server. However, in practical applications, this approach first leads to an increase in the cost of capital. Second, the server hardware lifting space is also limited. Another aspect is the use of single-system mirroring Web server cluster system, which is also the most widely used method. However, this method also has its insurmountable shortcomings. The most important point is that when dealing with the changing load, we need to adjust the size of the server cluster according to the established strategy or manual way, which can no longer meet the current frequent and unpredictable network load changes. In order to make up for the above shortcomings, this paper proposes an elastic and Load Balancing Cluster Server Architecture based on cloud technology cluster server architecture based on cloud technology, which is based on the Eucalyptus cloud computing platform. According to the principle of separation of computing and storage, the computing service system is divided into two parts. The computing service system provides service through virtual server instance on the basis of computing module of Eucalyptus cloud platform. The data storage system can automatically adjust the scale of the cluster according to the current load condition of the cluster. On the basis of the storage module of Eucalyptus, the data storage system solves the problems of storage expansion and storage sharing by mounting NAS devices and creating HDFS distributed file system. Improved storage capacity and storage scalability for the system. Finally, by using the load testing software to test and analyze the ELC cluster server system proposed in this paper, it is verified that the system can automatically adjust the scale of the server cluster according to the expected way to adapt to the current cluster load situation. And realized storage expansion and storage sharing.
【學(xué)位授予單位】：西安電子科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP393.09

【參考文獻(xiàn)】

相關(guān)期刊論文 前7條

1 喬靜;屈志強(qiáng);李娟;;Linux服務(wù)器安全動(dòng)態(tài)磁盤管理方案[J];電腦知識(shí)與技術(shù);2010年23期

2 張文;董洛兵;沈沛意;崔治遠(yuǎn);姜新華;;構(gòu)建基于云平臺(tái)的機(jī)頂盒視頻點(diǎn)播系統(tǒng)[J];信息通信;2012年05期

3 張玉芳;魏欽磊;趙膺;;基于負(fù)載權(quán)值的負(fù)載均衡算法[J];計(jì)算機(jī)應(yīng)用研究;2012年12期

4 張文;董洛兵;沈沛意;姜新華;;基于Eucalyptus云平臺(tái)的一種存儲(chǔ)系統(tǒng)架構(gòu)[J];信息通信;2012年06期

5 馬一力,傅湘林,韓曉明,許魯;存儲(chǔ)與計(jì)算的分離[J];世界電信;2004年08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

7 隋坤杰;網(wǎng)絡(luò)數(shù)據(jù)存儲(chǔ)的實(shí)現(xiàn)[J];現(xiàn)代情報(bào);2004年02期

相關(guān)博士學(xué)位論文 前1條

1 王紅斌;Web服務(wù)器集群系統(tǒng)的自適應(yīng)負(fù)載均衡調(diào)度策略研究[D];吉林大學(xué);2013年

本文編號(hào)：1645494