Preface by Jim Gray to Chinese translation of

Transaction Processing:Concepts and Techniques

 

The fundamental premise of this book is that the transaction concept is a key abstraction for distributed computation. One implements transactions at a one layer, and then all higher layers have a much simpler failure semantics (all or nothing) and much simpler error handling.  

 

This was a radical view when the book first appeared ten years ago. Since then the Internet has blossomed, distributed computing has become more tangible with CORBA, DCOM, and now Web Services. TP monitors have been supplanted by web servers and object request brokers. Database systems have matured enormously.    

 

Today we can see that many have been converted to this view. Transaction techniques are being used within most operating systems, and transactional file systems and transaction managers are being added to most operating systems. Transactional resource managers like queue managers, workflow systems, persistent object systems, and main memory database systems are proliferating now that this enabling technology is in place.

 

Some things about this book are dated – even obsolete – but the core concepts and techniques have moved from the specialty of database systems to the broader context of distributed computing.

 

Andreas Reuter and I are grateful to you for undertaking the arduous task of translating this book from English to Chinese. Chinese, being a more compact language, will make the book shorter (and lighter), and having it in Chinese will make it accessible to a more than a billion people. I hope you find this book informative and helpful. 

 

Jim Gray
Microsoft Research

18 October 20002

Jim Gray,
Microsoft Research,  
455 Market Street, Suite 1690, San Francisco, CA 94105
tel: 415-778-8222 fax 425 706 7329
Gray@Microsoft.com
   http://research.microsoft.com/~gray

 

 

 

 

 

 

 

 

 

 

 

 

 

 前言

--Jim Gray 写给《事务处理概念与技术》中译本

 

本书的基本前提是事务概念是一个分布计算的重要抽象。在某一层次上实现了事务,那么所有高的层次上会有一个简化的失败语义(全做或全不做),并且错误处理也会简单的多。

 

十年前当本书初次面世时,这是一个基本的观点。自此以后,互连网蓬勃发展,分布计算在CORBADCOM,和现在的Web Services的支持下,变得更为切实可行。TP监控器已经被web服务器和对象请求代理所取代。数据库系统也已经成熟多了。

 

今天我们看到,许多技术都已接受了这一观点。事务技术被广泛用于操作系统,事务文件系统和事务管理器被加入到了操作系统中。事务资源管理器如队列管理器,工作流系统,持久对象系统,和主存数据库系统等,现今日益推广流行,这一使能技术是恰如其位。

 

本书中的一些论述有些过时了——甚至是陈腐了——但其核心概念和技术已从数据库系统扩展到了分布计算的范畴。

 

Andreas Reuter和我感谢译者从事本书英译汉的辛勤工作。中文,作为更为紧凑的语言,会使本书更为薄些(也更轻了),中文版还会使十多亿人有机会阅读本书。我希望本书对各位是有价值和有帮助的。

 

 

 

 

 

Jim Gray
Microsoft Research

18 October 20002

Jim Gray,
Microsoft Research,  
455 Market Street, Suite 1690, San Francisco, CA 94105
tel: 415-778-8222 fax 425 706 7329
Gray@Microsoft.com
   http://research.microsoft.com/~gray