事務(wù)處理

前言

Why We Wrote this Book The purpose of this book is to give you an understanding of how large，distributed，hetero- geneous computer systems can be made to work reliablv．In contrast to the often complex methods of distributed computing，it presents a distributed system application development approach that call be used by mere mortals．WhY then doesn't the title use a term like dis- tributed systems，high reliability，interoperability,or client-server？Why use something as prosaic as transaction processing。a term that for many people denotes old-fashioned，batch- oriented，mainframe-minded data processing？The point is-and that'S what makes the book SO long～that the design，implementation， and operation of large application systems，with thousands of terminals，employing hundreds of computers。providing service with absolutely no downtime，cannot be done from a single perspective．An integrated（and integrating）perspective and methodology is needed to ap。 proach the distributed systems problem．Our gOal is to demonstrate that ffansactions provide this integrative conceptual framework．a(chǎn)nd that distributed transaction-oriented operating systems are the enabling technology．The client-server paradigm provides a good way of structuring the system and of developing applications．but it stiil needs transactions to con- trol the client．server interactions．In a nutshell：without transactions．distributed systems cannot be made to work for typical real．1ife applications．This is not an outrageous claim；rather it is a lesson many people——system implemen- tors，system owners，and application developers-have learned the hard way．Of course，the concepts for building lurge systems have been evolving for a long time．In fact，some of the key ideas were developed way back when batch processing was in full swing，but they are far from being obsolete．Transaction processing concepts were conceived to master the corn- plexity in single-processor online applications．If anything，these concepts are even more critical now for the SUCCESSful implementation of massively distributed systems that work and fail in much more complex ways．

內(nèi)容概要

本書從系統(tǒng)的角度全面闡述事務(wù)處理的概念和技術(shù)，其中涉及終端上的表示管理、通信子系統(tǒng)、操作系統(tǒng)、數(shù)據(jù)庫、程序設(shè)計語言的運(yùn)行時系統(tǒng)以及應(yīng)用開發(fā)環(huán)境等。本書重點放在事務(wù)處理的基本概念上，主要闡述事務(wù)概念是如何用于解決分布式系統(tǒng)問題的，以及利用這些概念如何能夠在有限的資金和風(fēng)險范圍內(nèi)建立高性能、高可用性的應(yīng)用系統(tǒng)。全書重點講述了事務(wù)處理基礎(chǔ)、容錯基礎(chǔ)知識、面向事務(wù)的計算，并發(fā)控制、恢復(fù)、事務(wù)型文件系統(tǒng)、系統(tǒng)概覽等7個主題，介紹了事務(wù)的ACID特性、并發(fā)的理論和實踐、事務(wù)管理和恢復(fù)技術(shù)等方面的內(nèi)容，最后還介紹了一個非常重要的資源管理器的實現(xiàn)。    本書主要面向計算機(jī)及相關(guān)專業(yè)的高年級本科生和研究生，適合作為事務(wù)處理導(dǎo)論、數(shù)據(jù)庫系統(tǒng)、分布式系統(tǒng)、操作系統(tǒng)等課程的輔助教材，需要了解事務(wù)處理系統(tǒng)的開發(fā)人員也可將其作為基本參考書。

作者簡介

Jim Gray （1944-2007）計算機(jī)科學(xué)大師，因在數(shù)據(jù)庫和事務(wù)處理研究和實現(xiàn)方面的開創(chuàng)性貢獻(xiàn)而獲得1998年圖靈獎。美國科學(xué)院、工程院兩院院士，ACM和IEEE兩會會士：他25歲成為加州大學(xué)伯克利分校計算機(jī)科學(xué)學(xué)院第一位博士。在IBM工作期間參與和主持了IMS、System R、SQUDS、DB2等項目的開發(fā)。后任職于微軟研究院．主要關(guān)注應(yīng)用數(shù)據(jù)庫技術(shù)來處理各學(xué)科的海量信息。2007年1月獨自駕船出海后失蹤。

書籍目錄

PART ONE -The Basics of Transaction Processing1  INTRODUCTION　　1.1  Historical Perspective　　1.2  What ls a Transaction Processing System?　　　1.2.1   The End User's View of a Transaction Processing System      1.2.2   The Administrator/Operator's View of a TP System      1.2.3   Application Designer's View of a TP System      1.2.4   The Resource Manager's View of a TP System      1.2.5   TP System Core Services　　1.3  A Transaction Processing System Feature List      1.3.1   Application Development Features      1.3.2   Repository Features      1.3.3   TP Monitor Features      1.3.4   Data Communications Features      1.3.5   Database Features      1.3.6   Operations Features      1.3.7   Education and Testing Features      1.3.8   Feature Summary　　1.4  Summary　　1.5  Historical Notes      Exercises      Answers　2   BASIC COMPUTER SCIENCE TERMINOLOGY　　2.1  Introduction      2.1.1   Units　　2.2  Basic Hardware      2.2.1   Memories      2.2.2   Processors      2.2.3   Communications Hardware      2.2.4   Hardware Architectures　　2.3  Basic Software---Address Spaces, Processes, Sessions      2.3.1   Address Spaces      2.3.2   Processes, Protection Domains, and Threads      2.3.3   Messages and Sessions　　2.4  Generic System Issues      2.4.1   Clients and Servers      2.4.2   Naming      2.4.3   Authentication      2.4.4   Authorization      2.4.5   Scheduling and Performance      2.4.6   Summary　　2.5  Files      2.5.1   File Operations      2.5.2   File Organizations      2.5.3   Distributed Files      2.5.4  SQL　　2.6  Software Performance　　2.7  Transaction Processing Standards      2.7.1   Portability versus Interoperability Standards      2.7.2   APIs and FAPs      2.7.3   LU6.2, a de facto Standard      2.7.4   OSI-TP with X/Open DTP, a de jure Standard　　2.8  Summary      Exercises      AnswersPART TWO The Basics of Fault Tolerance　3   FAULT TOLERANCE　　3.1  Intreduction      3.1.1   A Crash Course in Simple Probability      3.1.2   An External View of Fault Tolerance　　　……PART THREE-Transaction-Oriented ComprtingPART FOUR-concurrency ControlPART FIVE-RecoveryPART SIX-Transaxtional File System:ASample Resource ManagerPART SEVEN-System SurveysPART EIGHT-AddendaINDEX

章節(jié)摘錄

插圖：Key-sequenced placement stores the records sorted in key order．Key sequencing clusters related records together and allows sequential scanning of records in sorted order． Earlier sections of this chapter point out the benefits of this clustering and sequential ac- cess to data．When a new record amves．its key is computed and the record is placed near records witIl related keys．Record insertion is a little expensive．but there are inge- nious algorithms that make it competitive with hashing．Given a key value，it is easy to find the record by using binary search on the file or by using some indexing structure．It is often desirable to associatively access a file via two different keys．For example．it is often convenient to access employees by either name or employee number．Suppose the employee records are stored in an associative file keyed by employee number（empno）．Then a second associative file，keyed by employee name（empname），could store a record of the formfor each record in the employee file．By first looking in this second file under the empname key to find the empno，and then using this empno to associatively access the employee file，the system can fairly quickly find the desired employee record．Such index files are called secondary indices．It is often convenient to think of the direct address of a record as its key．If this is done．then secondary indices can be defined on direct files as well as on associative files．Most systems allow file designers to define many secondary indices on a base file．The file system automatically maintains the records in the secondary indices as records are in- serted into，updated in，and deleted from the primary file．Of course，the definition of the secondary index must be stored in the file descriptor．When a file is first opened． the descriptor is read bv the server．a(chǎn)nd all subsequent record operations on the file cause the relevant secondary indices to be used and maintained．2．5．3 Distributed Files Parts of a file mav be distributed among servers in a computer network．This distribufion can take two forms：The files can be partitioned（fragments of the file are stored in different nodes），or the files can be replicated（the whole contents of the files are stored at several nodes）．The definitions of partitioning and replication are fairly simple．A file is broken into fragments by declaring the key boundaries of each fragment：All records within that key range belong to that fragment．For example，if a file is keyed by sales region and customer number，then the file might be fragmented by region，with each region having a separate fragment．These fragments might be partitioned among the computers of the various regions， with each region storing the fragment for that region．In addition，all the fragments might be replicated at central headquarters．The descriptor of each fragment contaias a complete description of the entire file。When a client opens the file，the file system looks at the descriptor and thereby knows about all the fragments．When the client issues a read-by-key，the request is dispatched to one of the servers managing that fragment．When the client issues a record insert，delete，or update Op． eration，the request is dispatched to all servers managing the fragment that holds the record． Associated secondary index reads and writes are handled similarlv．

編輯推薦

事務(wù)處理廣泛應(yīng)用于數(shù)據(jù)庫和操作系統(tǒng)等領(lǐng)域，對構(gòu)建高性能、并發(fā)、分布式的可靠現(xiàn)代計算機(jī)系統(tǒng)至關(guān)重要?！妒聞?wù)處理概念與技術(shù)(英文版)》是被譽(yù)為“事務(wù)處理圣經(jīng)”的經(jīng)典名著，由圖靈獎得主Jim Gray和世界數(shù)據(jù)庫權(quán)威AndreasReuter合著，是兩位大師數(shù)十年學(xué)術(shù)研究和實踐經(jīng)驗的結(jié)晶。《事務(wù)處理概念與技術(shù)(英文版)》的組織和敘述方法獨樹一幟，作者將事務(wù)作為統(tǒng)一的概念框架．由此出發(fā)，筆鋒所至，縱橫開闔，引導(dǎo)讀者從系統(tǒng)實現(xiàn)者的角度．全面深入地審視了計算機(jī)系統(tǒng)的方方面面．不僅闡述理論，而且針對各種實際問題，詳細(xì)解釋出現(xiàn)的原因，講述大量已經(jīng)在成功的商業(yè)和研究項目中經(jīng)過驗證、行之有效的事務(wù)處理實現(xiàn)技術(shù)，并提供了豐富的C語言代碼。書中處處閃爍著作者對計算機(jī)系統(tǒng)的淵博學(xué)識和真知灼見，無論你是程序員、架構(gòu)師、數(shù)據(jù)庫管理員，還是科研人員和高校師生，都將從《事務(wù)處理概念與技術(shù)(英文版)》中獲益匪淺。

圖書封面

圖書標(biāo)簽Tags

無

評論、評分、閱讀與下載

---

    事務(wù)處理 PDF格式下載

---