SQL Injection has been a method in the pen tester's toolkit for a long time now.  In the early days, it was very easy because very few people gave thought to security in their web applications but lately it's becoming more difficult.  As more and more people are becoming aware of the dangers of unprotected code, more and more ways of preventing SQL injection are being invented.  Of course, there are still plenty of holes, they just take a bit more technical knowhow to exploit.

Two whitepapers have been released recently describing more complex methods of performing SQL Injections.  The paper entitled Deep Blind SQL Injection by Ferruh Mavituna describes a very complex method of retrieving data from certain database servers, MS SQL and Oracle are named in the paper, though more may fall victim to the same type of attacks.  His method involves using time delay differences to speed up the process of Blind SQL Injection.  This method is much slower than the tried and true ' or 1=1 -- string but it can work in some environments where that method does not.

Discovery and Fuzzing for SQL injections with Web 2.0 Applications from Blueinfy AppSec Labs explores the more specialized world of enumerating databases used by Web 2.0 apps by evaluating the Javascript in the page and traffic to and from the HTTP server.  This method is more about Injecting Javascript and HTTP requests than direct SQL and can provide a new attack vector for web 2.0 applications.

These are both fun papers to read and provide methods you may not already be familiar with.  Good input.