【摘要】：高可用系統(tǒng)，也稱容災(zāi)系統(tǒng)，對(duì)于IT而言，就是為計(jì)算機(jī)信息系統(tǒng)提供的一個(gè)能應(yīng)付各種災(zāi)難的環(huán)境。當(dāng)計(jì)算機(jī)系統(tǒng)在遭受如火災(zāi)、水災(zāi)、地震、戰(zhàn)爭(zhēng)等不可抗拒的自然災(zāi)難以及計(jì)算機(jī)犯罪、計(jì)算機(jī)病毒、掉電、網(wǎng)絡(luò)/通信失敗、硬件/軟件錯(cuò)誤和人為操作錯(cuò)誤等人為災(zāi)難時(shí)，容災(zāi)系統(tǒng)將保證用戶數(shù)據(jù)的安全性（數(shù)據(jù)容災(zāi)），甚至，一個(gè)更加完善的容災(zāi)系統(tǒng)，還能提供不間斷的應(yīng)用服務(wù)。高可用技術(shù)已經(jīng)有很長(zhǎng)的發(fā)展歷史，在關(guān)鍵性業(yè)務(wù)環(huán)節(jié)，可以說是必不可少。如何快速，準(zhǔn)確的處理故障，并減少故障時(shí)帶來的損失，一直是高可用技術(shù)追求的目標(biāo)。 本論文主要研究并實(shí)現(xiàn)了一種面向集群的高可用中間框架。該框架提供了通用的二次開發(fā)接口，簡(jiǎn)單易用，可以作為高可用系統(tǒng)或容災(zāi)系統(tǒng)的底層框架。在研究和設(shè)計(jì)本框架期間，深入研究和學(xué)習(xí)了國(guó)內(nèi)外集群系統(tǒng)、分布式系統(tǒng)、搜索引擎系統(tǒng)、云平臺(tái)以及大數(shù)據(jù)相關(guān)領(lǐng)域的現(xiàn)有技術(shù)和發(fā)展趨勢(shì)。 本論文對(duì)諸多高可用系統(tǒng)所需要的技術(shù)，進(jìn)行了深入研究和總結(jié)，包括如何支持多個(gè)操作系統(tǒng)平臺(tái)，，諸如Uinx/Linux, Windows, MacOS以及Solaris等，也包括如何支持在不同處理器架構(gòu)上的編程研究，諸如32位和64位編程的總結(jié)。同時(shí)在設(shè)計(jì)中，結(jié)合使用了軟件工程中的面向?qū)ο缶幊趟枷�、面向接口編程思想、面向組件編程思想、設(shè)計(jì)模式的思想以及多種數(shù)據(jù)結(jié)構(gòu)和算法等。通過對(duì)這一系列方法和思想的運(yùn)用，設(shè)計(jì)了本論文所提的框架。該框架包含了多個(gè)子模塊，底層通訊傳輸協(xié)議層，是本框架的網(wǎng)絡(luò)通信模塊，用于多個(gè)節(jié)點(diǎn)間的相互通信以及心跳檢測(cè)；組與成員原子廣播模塊，是框架的成員、組和資源管理中心，同時(shí)也會(huì)負(fù)責(zé)全局狀態(tài)的廣播和維護(hù)；防腦裂模塊，是框架對(duì)于異常情況的處理，用于在框架出現(xiàn)故障時(shí)，保護(hù)系統(tǒng)的數(shù)據(jù)和資源的一致性和完整性；高可用組件守護(hù)進(jìn)程，是節(jié)點(diǎn)上唯一的運(yùn)行實(shí)例，用于統(tǒng)管全局，以及對(duì)外的命令交互；最后，以基于CIFS協(xié)議的高可用Samba文件共享服務(wù)器為例，實(shí)際運(yùn)用本框架，實(shí)現(xiàn)一個(gè)并發(fā)的高可用文件共享服務(wù)器。
[Abstract]:High availability system, also known as disaster recovery system, is a kind of environment for computer information system to cope with all kinds of disasters for IT. When computer systems are suffering from irresistible natural disasters such as fires, floods, earthquakes, wars, as well as human disasters such as computer crime, computer viruses, power outages, network / communication failures, hardware / software errors and man-made errors, Disaster recovery system will ensure the security of user data (data disaster recovery), and even a more perfect disaster recovery system, but also can provide uninterrupted application services. High availability technology has a long history of development, in the key business links, can be said to be essential. How to deal with faults quickly and accurately and reduce the losses caused by faults has always been the goal of high available technology. In this paper, a cluster-oriented highly available intermediate framework is studied and implemented. The framework provides a general secondary development interface, which is simple and easy to use, and can be used as the underlying framework of high availability system or disaster recovery system. During the research and design of this framework, the existing technologies and development trends of cluster system, distributed system, search engine system, cloud platform and big data related fields at home and abroad have been deeply studied and studied. In this paper, the technologies needed by many highly available systems are deeply studied and summarized, including how to support multiple operating system platforms, such as Uinx/Linux, Windows, MacOS and Solaris. It also includes how to support programming research on different processor architecture, such as 32-bit and 64-bit programming. At the same time, in the design, the idea of object-oriented programming, interface-oriented programming, component-oriented programming, design pattern and a variety of data structures and algorithms in software engineering are combined. Through the application of this series of methods and ideas, the framework proposed in this paper is designed. The framework consists of several sub-modules, the underlying communication and transmission protocol layer, which is the network communication module of this framework, which is used for communication and heartbeat detection between multiple nodes. The group and member atomic broadcasting module is the member of the framework, the group and the resource management center, and will also be responsible for the broadcasting and maintenance of the global state. The anti-brain crack module is the processing of abnormal conditions in the framework, which is used to protect the consistency and integrity of the data and resources of the system in the event of failure of the framework. The highly available component daemon is the only running instance on the node, which is used to manage the global situation and external command interaction. Finally, taking the high availability Samba file sharing server based on CIFS protocol as an example, a concurrent high available file sharing server is implemented by using this framework.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2013
【分類號(hào)】：TP311.52

【參考文獻(xiàn)】

相關(guān)期刊論文 前3條

1 王曉東,彭兵,張際平;基于中間件的開發(fā)研究[J];計(jì)算機(jī)應(yīng)用研究;2001年08期

2 段劍弓;存儲(chǔ)系統(tǒng)NAS和SAN的差異和統(tǒng)一[J];計(jì)算機(jī)應(yīng)用研究;2004年12期

3 高靜;段會(huì)川;;JSON數(shù)據(jù)傳輸效率研究[J];計(jì)算機(jī)工程與設(shè)計(jì);2011年07期

相關(guān)會(huì)議論文 前1條

1 姚文琳;姚大鋼;;基于分布式數(shù)據(jù)庫(kù)技術(shù)的高可用容災(zāi)系統(tǒng)[A];2006年全國(guó)開放式分布與并行計(jì)算機(jī)學(xué)術(shù)會(huì)議論文集（三）[C];2006年

本文編號(hào)：2484806