First and foremost the most important compendium of papers on DBMS theory that you must read is 'Readings in Database Systems 4E' -- Stonebraker (aka 'The Red Book')
Every paper in this book is a milestone; failing which it wouldn't have made a cut to this book. :-)
Stonebraker also has an excellent survey on DBMS architecture 'Architecture of a Database System' : Foundations and Trends in Databases 1:2 (2007)
THE BOOK for DBMS writers: 'Transaction Processing: Concepts and Techniques' -- Jim Gray (his seminal work)
This could be considered as a one big paper : 'The Relational Model For Database Management V2' -- Codd
Foundations for Object/Relational Databases, The Third Manifesto' --- C. J. Date
Readings in Object-Oriented Database Systems -- Zdonik
Concurrency control and Recovery in Database Systems -- Bernstein
如果我没记错的话，格雷和路透的文本并没有涵盖莫汉的重复历史恢复技术，而这是必不可少的。至少请参阅 ARIES/NT: A Recovery Method Based on Write-Ahead Logging for Nested Transactions 和 ARIES: A Transaction Recovery Method Supporting Fine-Granularity Locking and Partial Rollbacks using Write-Ahead Logging，但 Mohan 的大部分论文都值得一看阅读。
Bernstein 等人所著的《数据库系统中的并发控制和恢复》一书已绝版，但您可以从他的 Microsoft Research 页面下载该书。
The book that zamanbakshi recommends, Transaction Processing: Concepts and Techniques, by Gray and Reuter, is really, really good. I carried it around so much that the cover fell off--and this is a hardcover. It is somewhat outdated on some topics, of course, but it is a much better read than most later books, such as Weikum and Vossen's Transactional Information Systems, which is a good book, but makes my eyes glaze over, I have to admit.
If I recall correctly, the Gray and Reuter text doesn't cover Mohan's repeating histories recovery technique, which is essential. See ARIES/NT: A Recovery Method Based on Write-Ahead Logging for Nested Transactions and ARIES: A Transaction Recovery Method Supporting Fine-Granularity Locking and Partial Rollbacks Using Write-Ahead Logging, at the very least, but most of Mohan's papers are worth reading.
The book Concurrency control and Recovery in Database Systems by Bernstein, et al., is out of print, but you can download it from his Microsoft Research page.
There are also a lot of good publications there from David Lomet and the late (or missing) Jim Gray.
Some important papers that aren't in the 2nd edition of the Red Book (the edition I have):
A critique of ANSI SQL isolation levels (1995) Gray, et al.
The dangers of replication and a solution (1996) Gray and Helland
Generalized Isolation Level Definitions (2000) Adya, et al.
A recent paper that I think deserves more attention is Serializable Isolation for Snapshot Databases (2009) by Cahill, Röhm, and Fekete. It's a really simple technique that works surprisingly well. I hope that it gets implemented in some DBMSes. While searching for related stuff, I came across this interesting reading list. It mostly has flash memory-related stuff, but there are some general papers of interest, too, including some recent Stonebraker papers.
I recommend skipping Date's "Third Manifesto" stuff. I was quite disappointed by it. I don't think he has ever done any object-oriented programming. His earlier books and articles on relational DBMSes are good, if a bit repetitive.
A good description of a main-memory DBMS is The Architecture of the Dalí Main-Memory Storage Manager. The non-action-consisent, non-WAL checkpointing blew my mind at first.
Here are a couple about non-consistent distributed data management (for very, very big data):
BASE: an Acid alternative (2008) Pritchett
Life beyond Distributed Transactions: an Apostate's Opinion (2007) Helland
发布评论
评论（2）
首先，您必须阅读的最重要的 DBMS 理论论文纲要是“数据库系统 4E 读物”——Stonebraker（又名“红皮书”）
<块引用>
本书中的每一篇论文都是
里程碑;如果失败的话就不会
对这本书做了剪裁。 :-)
Stonebraker 还对 DBMS 架构进行了出色的调查“数据库系统的架构”：数据库基础和趋势 1:2 (2007)
DBMS 编写者的书：“事务处理：概念和技术”——Jim Gray（他的开创性著作）
这可以被视为一篇大论文：“数据库管理 V2 的关系模型”——Codd
对象/关系数据库基础，第三宣言'-- CJ Date
面向对象数据库系统读物 -- Zdonik
数据库系统中的并发控制和恢复 -- Bernstein< /p>
First and foremost the most important compendium of papers on DBMS theory that you must read is 'Readings in Database Systems 4E' -- Stonebraker (aka 'The Red Book')
Stonebraker also has an excellent survey on DBMS architecture 'Architecture of a Database System' : Foundations and Trends in Databases 1:2 (2007)
THE BOOK for DBMS writers: 'Transaction Processing: Concepts and Techniques' -- Jim Gray (his seminal work)
This could be considered as a one big paper : 'The Relational Model For Database Management V2' -- Codd
Foundations for Object/Relational Databases, The Third Manifesto' --- C. J. Date
Readings in Object-Oriented Database Systems -- Zdonik
Concurrency control and Recovery in Database Systems -- Bernstein
由于我是 stackoverflow 的新手，所以我不能发布两个以上的超链接！因此，带有链接的版本位于 http://www.reddit.com/ r/compsci/comments/ghc0w/please_recommend_must_read_favorite_papers_in/c1no849
zamanbakshi 推荐的书《事务处理：概念与技术》，作者 Gray 和 Reuter，真的非常非常好。我带着它到处走，以至于封面掉下来了——这是一本精装本。当然，它在某些主题上有点过时了，但它比大多数后来的书要好读得多，例如 Weikum 和 Vossen 的事务信息系统，这是一本好书，但让我的眼睛变得呆滞结束了，我不得不承认。
如果我没记错的话，格雷和路透的文本并没有涵盖莫汉的重复历史恢复技术，而这是必不可少的。至少请参阅 ARIES/NT: A Recovery Method Based on Write-Ahead Logging for Nested Transactions 和 ARIES: A Transaction Recovery Method Supporting Fine-Granularity Locking and Partial Rollbacks using Write-Ahead Logging，但 Mohan 的大部分论文都值得一看阅读。
Bernstein 等人所著的《数据库系统中的并发控制和恢复》一书已绝版，但您可以从他的 Microsoft Research 页面下载该书。
那里还有大卫·洛梅特 (David Lomet) 和已故（或失踪）的吉姆·格雷 (Jim Gray) 的许多优秀出版物。
红皮书第二版（我拥有的版本）中没有的一些重要论文：
我认为最近值得更多关注的一篇论文是 Cahill、Röhm 和 Fekete 撰写的《快照数据库的串行化隔离》(2009)。这是一个非常简单的技术，但效果却出奇的好。我希望它能在一些 DBMS 中实现。在搜索相关内容时，我发现了这个有趣的阅读列表。它主要包含与闪存相关的内容，但也有一些令人感兴趣的一般论文，包括最近的一些 Stonebraker 论文。
我建议跳过 Date 的“第三宣言”内容。我对此感到非常失望。我认为他从来没有做过任何面向对象的编程。他早期关于关系 DBMS 的书籍和文章都很好，尽管有点重复。
《Dalí 主存存储管理器的架构》是对主存 DBMS 的一个很好的描述。起初，非动作一致性、非 WAL 检查点让我大吃一惊。
以下是关于非一致性分布式数据管理的一些内容（对于非常非常大的数据）：
Since I'm new on stackoverflow, I can't post more than two hyperlinks!! So the version with links is at http://www.reddit.com/r/compsci/comments/ghc0w/please_recommend_must_read_favorite_papers_in/c1no849
The book that zamanbakshi recommends, Transaction Processing: Concepts and Techniques, by Gray and Reuter, is really, really good. I carried it around so much that the cover fell off--and this is a hardcover. It is somewhat outdated on some topics, of course, but it is a much better read than most later books, such as Weikum and Vossen's Transactional Information Systems, which is a good book, but makes my eyes glaze over, I have to admit.
If I recall correctly, the Gray and Reuter text doesn't cover Mohan's repeating histories recovery technique, which is essential. See ARIES/NT: A Recovery Method Based on Write-Ahead Logging for Nested Transactions and ARIES: A Transaction Recovery Method Supporting Fine-Granularity Locking and Partial Rollbacks Using Write-Ahead Logging, at the very least, but most of Mohan's papers are worth reading.
The book Concurrency control and Recovery in Database Systems by Bernstein, et al., is out of print, but you can download it from his Microsoft Research page.
There are also a lot of good publications there from David Lomet and the late (or missing) Jim Gray.
Some important papers that aren't in the 2nd edition of the Red Book (the edition I have):
A recent paper that I think deserves more attention is Serializable Isolation for Snapshot Databases (2009) by Cahill, Röhm, and Fekete. It's a really simple technique that works surprisingly well. I hope that it gets implemented in some DBMSes. While searching for related stuff, I came across this interesting reading list. It mostly has flash memory-related stuff, but there are some general papers of interest, too, including some recent Stonebraker papers.
I recommend skipping Date's "Third Manifesto" stuff. I was quite disappointed by it. I don't think he has ever done any object-oriented programming. His earlier books and articles on relational DBMSes are good, if a bit repetitive.
A good description of a main-memory DBMS is The Architecture of the Dalí Main-Memory Storage Manager. The non-action-consisent, non-WAL checkpointing blew my mind at first.
Here are a couple about non-consistent distributed data management (for very, very big data):