【摘要】：由于計(jì)算機(jī)在21世紀(jì)的大規(guī)模應(yīng)用與普及,硬件和軟件以幾何倍數(shù)的發(fā)展,極大推進(jìn)社會(huì)的信息化前進(jìn)的步伐。可是人們漸漸發(fā)現(xiàn)單純硬件設(shè)備的升級(jí)確實(shí)可以使計(jì)算機(jī)的計(jì)算能力越發(fā)強(qiáng)大,但大部分的超性能計(jì)算機(jī)因其龐大的身軀和昂貴的價(jià)格只能應(yīng)用于自然科技和尖端技術(shù)領(lǐng)域,離普通商用化卻漸行漸遠(yuǎn)。而商用化的基于C/S結(jié)構(gòu)的服務(wù)器集群在互聯(lián)網(wǎng)高速普及的今天面臨海量數(shù)據(jù)的存取和計(jì)算在效用和利用率提升上卻越發(fā)捉襟見肘,這便催生了云計(jì)算的出現(xiàn)和發(fā)展。 云計(jì)算正是將計(jì)算和存儲(chǔ)分布在服務(wù)器集群中,讓處在集群中的服務(wù)器協(xié)同工作完成運(yùn)算任務(wù),這樣可以降低能耗、提高計(jì)算利用效率。雖然云計(jì)算的初衷是惠及大眾,但因?yàn)槠淙孕枰^高端的技術(shù)水平和強(qiáng)大的服務(wù)器集群支持,發(fā)展至今仍只有如Google、Amazon等少數(shù)大型企業(yè)在嘗試并投入實(shí)際應(yīng)用�；诖�,本文結(jié)合自身工作環(huán)境中對(duì)中小企業(yè)實(shí)際需求的分析,考慮在滿足不額外采購硬件設(shè)備、提高已有設(shè)備的利用率和實(shí)現(xiàn)企業(yè)自身定制化應(yīng)用的基礎(chǔ)上,提出了是否能夠結(jié)合云計(jì)算已有核心技術(shù)理念MapReduce和分布式文件系統(tǒng)設(shè)計(jì)一個(gè)小型系統(tǒng)原型來滿足以上的需求。正文通過理論分析和設(shè)計(jì)開發(fā)和實(shí)際應(yīng)用環(huán)境部署測試論證了該想法的可行性。 本文可以分為三個(gè)部分： (1)首先給出目前主流的云計(jì)算定義及商用應(yīng)用實(shí)例。并結(jié)合中小企業(yè)的環(huán)境特征和應(yīng)用要求分析與云計(jì)算相關(guān)技術(shù)結(jié)合的必要性。 (2)給出參照的云計(jì)算開源架構(gòu)Hadoop核心技術(shù)MapReduce和分布式文件存儲(chǔ)的理論依據(jù),并根據(jù)此依據(jù)衍生分析與實(shí)際應(yīng)用結(jié)合的可行性 (3)根據(jù)實(shí)際需求,設(shè)計(jì)與開發(fā)小型系統(tǒng)原型并在實(shí)際環(huán)境中測試,論證論點(diǎn),總結(jié)不足之處與改進(jìn)設(shè)想。
[Abstract]:Due to the large-scale application and popularization of computer in the 21st century, the development of hardware and software with geometric multiple has greatly promoted the pace of social informatization. But people have come to realize that the upgrading of simple hardware really makes computing power more powerful, but most super-performance computers can only be used in natural technology and cutting-edge technology because of their huge size and expensive price. From the ordinary commercial but gradually away from the line. However, the commercial server cluster based on C / S structure is facing the huge data access and computing in the rapid popularization of the Internet today. However, the utility and utilization of the cluster are increasingly overstretched, which leads to the emergence and development of cloud computing. Cloud computing is to distribute computing and storage in the server cluster, so that the server in the cluster can work together to complete the task, which can reduce the energy consumption and improve the efficiency of computing. Although the original intention of cloud computing is to benefit the public, but because it still needs more advanced technology and powerful server cluster support, only a few large enterprises such as Google,Amazon are trying and putting it into practice. Based on this, this paper analyzes the actual needs of small and medium-sized enterprises in their own working environment, and considers not to purchase additional hardware equipment, to improve the utilization of existing equipment and to realize the customized application of the enterprise itself. This paper proposes whether it is possible to design a prototype of a small system based on the core technology concept of cloud computing MapReduce and distributed file system to meet the above needs. The paper demonstrates the feasibility of this idea through theoretical analysis, design and development and practical application environment deployment test. This paper can be divided into three parts: (1) the definition of cloud computing and commercial applications are given. Combined with the environmental characteristics and application requirements of small and medium-sized enterprises, the necessity of combining with cloud computing related technologies is analyzed. (2) the theoretical basis of reference Hadoop core technology MapReduce and distributed file storage is given. According to the feasibility of combining derivation analysis with practical application (3) according to the actual demand, design and develop the prototype of the small system and test it in the actual environment, demonstrate the argument, sum up the deficiency and the improvement assumption.
【學(xué)位授予單位】：復(fù)旦大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2012
【分類號(hào)】：TP3

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 陳康;鄭緯民;;云計(jì)算:系統(tǒng)實(shí)例與研究現(xiàn)狀[J];軟件學(xué)報(bào);2009年05期

，

本文編號(hào)：2273149