The Database Hacker's Handbook: Defending Database Servers [Paperback]

Here is a book in which you will probably only be interested in 1/7 of the pages. That means that instead of reading 528 pages you only need to read about 70. But, you may really, really need that 70 pages. The reason for this is that the book covers seven of the most common databases: IBM DB2, Oracle, MySQL, PostGreSQL, SQL Server, SyBase, Informix. These programs are so different that what applies to one does not generally apply to the others.

Each section of the book covers one of the databases. It usually begins with some history of both the database and attacks on it. For instance the Slammer worm compromised more than 75,000 SQL Server databases within ten minutes of its release in January 2003.

After that there is a discussion on the database, its architecture, how it handles things like authentication and so on.

Finally it goes into how to defend the database against attack. This includes information on how to remove unncecessary features and services that might serve as gateways to attacks, and talks about how to use the databases own internal security systems to their maximum effectiveness.

As I said, you really need the 70 or so pages that refer to your own database.

PS - What's the most secure database - PostGreSQL, and it goes into why.

David Litchfield is arguably the foremost expert and evangelist when it comes to database security. He, and his team of compatriots from Next Generation Security Software, have written a book that any database or security administrator should be familiar with.

Even if some of the attacks or exploits described in the book were previously obscure or unknown, the fact that they have been outlined in this book means that administrators need to know about them and defend against them before the "bad guys" read this book and take advantage of them.

One of the best aspects of this book is the way it is organized. Splitting the book into sections devoted to specific database systems makes it exceptionally simple and convenient to use. If you only use MySQL, you can skip all of the information regarding Oracle or Microsoft SQL Server, and just focus on the section of the book that applies to you.

Within each section, the authors provide a tremendous wealth of knowledge. Aside from describing weaknesses, potential exploits and protective measures to defend against them, they also look at the general architecture and the methods of authentication used by the database.

Any database admin should have a copy of this on their desk.

The Database Hacker's Handbook (TDHH) is unique for two reasons. First, it is written by experts who spend their lives breaking database systems. Their depth of knowledge is unparalleled. Second, TDHH addresses security for Oracle, IBM DB2, IBM Informix, Sybase ASE, MySQL, Microsoft SQL Server, and PostgreSQL. No other database security book discusses as many products. For this reason, TDHH merits four stars. If a second edition of the book addresses some of my later suggestions, five stars should be easy to achieve.

The first issue I would like to see addressed in a second edition of TDHH is the removal of the 60 pages of C code scattered throughout the book. The code is already provided on the publisher's Web site, and its appearance in a 500 page book adds little. The three pages of characters (that's the best way to describe it) on pages 313-315 in Ch 19 are really beyond what any person should be expected to type.

The second issue involves general presentation. Many chapters end abruptly with no conclusion or summary. Several times I thought "Is that it?" Chapters 2, 5, 7, 10, 13, 15, 18, 21 and 22 all end suddenly. The editor should have told the authors to end those chapters with summaries, as appear in other chapters. On a related note, some of the "chapters" are exceptionally short; Ch 9 and 12 are each 3 pages, for example. Chapters that short are an indication the book is not organized well.

The final issue involves discussion of various databases. I preferred the "Hacking Exposed" style of the 2003 book SQL Server Security, which included Dave Litchfield and Bill Grindlay as co-authors. That book spent more time introducing the fundamentals of database functions before explaining how to break them. For example, more background on PL/SQL would be helpful. With 60 pages of code removed, that leaves plenty of room for such discussion in the second edition.

On the positive side, I thought TDHH started strong with Ch 1. The Oracle security advice was very strong. I thought the time delay tactic for extracting bit-by-bit information from the database was also exceptionally clever.

Although I have not read it, I believe Implementing Database Security and Auditing by Ron Ben Natan might be a good complement to TDHH. Natan's book appears to take a functional approach, whereas TDHH takes a product-specific approach. The drawback of the product-centric approach is repetition of general security advice, such as enabling encryption, disabling default accounts, etc.

At the end of the day TDHH is still a revealing and powerful book. Anyone responsible for database security should refer to the sections of the book covering their database. I also recommend keeping an eye on the Next Generation Security Software Web site for the latest on database security issues. You should also see the authors speak at security conferences whenever possible.