SQL Injection Still Causing Trouble

An on-going and important aspect of managing database security is designing your applications to avoid SQL injection attacks. SQL injection is a form of web hacking whereby SQL statements are specified in the fields of a web form to cause a poorly designed web application to dump database content to the attacker.

This type of attack has been known for years now, but still there are new stories where SQL injection was used for nefarious purposes. SQL injection played a role in a hacking incident during the 2016 US presidential election, TalkTalk — a UK-based telecoms company — suffered a data breach in 2015 due to SQL injection, and the hardware manufacturer Archos suffered a SQL injection attack late in 2014.

And remember the Heartland Payment Systems breach from 2009? That SQL injection attack cost $300 million and the hackers that pulled it off were recently sent to federal prison (February 2018).

And these are just a few of the very public instances. The State of Software Security 2017 Report indicates that SQL injection attacks have been steady for years. The percentage of SQL injection attacks has ranged between 28% and 32% between the years 2011 through 2017. And the reality of it all is this: SQL injection attacks work only because of poor coding practices!

What is SQL Injection?

In order for SQL injection to succeed, the application code used by the website must be vulnerable to an injection attack. SQL injection relies upon programs that do not adequately filter for string literal escape characters embedded in SQL statements or where user input is not strongly typed. So instead of inputting data into a form, SQL statements are supplied. The SQL is “injected” from the web form into the database causing it to be executed and access (or even modify) unintended data.

Perhaps it is easiest to comprehend SQL injection by example. Consider a web-based application using dynamic SQL. The website requires users to login with their e-mail address and a password. Almost all sites of this type also offer an option to retrieve your password by supplying your e-mail address. Perhaps the SQL looks something like this:

SELECT userid, password
FROM   uid_pwd_table
WHERE  field = '$EMAIL';

The variable $EMAIL represents the input from the form on the website. A savvy hacker can attempt a SQL injection attack by entering:

   anything' OR '1'='1

If the application does not check the input properly the injection causes the SQL to now look like this:

   SELECT userid, password
   FROM   uid_pwd_table
   WHERE  field = 'anything' OR '1'='1';

Executing this statement causes a complete dump of every userid and password in the database because the OR ‘1’=’1′ component will always evaluate to TRUE. It does not matter what the first part of the injection was, it could be anything, because the second part of the injection gives the hacker everything in the table.

Another form of SQL injection relies upon improper typing, for example not checking whether data that should be numeric is actually numeric. Consider, for example:

   statement := "SELECT * FROM userinfo WHERE id = " + in_var + ";"

In this case, the SQL is being built into the statement variable; in_var is the variable used to supply the input. Let’s assume that the id column is numeric. However, if the program does not check the data type of the in_var variable to ensure that numeric data is supplied, SQL injection can occur. For example, instead of just supplying a numeric value, the hacker can supply something like this:

    4;DROP TABLE customer

If this SQL statement is executed the customer table (if one exists) will be dropped from the database.

SQL Injection Prevention

Using well-designed query language interpreters and coding applications appropriately can prevent SQL injection attacks. When possible use static SQL instead of dynamic SQL to improve the security of your database applications and data. Static SQL is hard-coded into the application and cannot be changed at runtime. Dynamic SQL is flexible and can change at runtime. When the SQL can change at runtime, a sufficiently motivated and skilled hacker can potentially change the SQL or potentially deploy a SQL injection attack to gain access to unauthorized data.

Static SQL is common in mainframe Db2 applications, but not so much for other platforms and database systems. The Db2 bind command “hardens” the SQL and optimizes access to the data.

Always validate user input by testing type, length, format, and range. The program should make absolutely no assumptions about the data that is received. Test the size and data type of input and enforce appropriate limits. Doing so can help to prevent buffer overruns. Test the content of string variables and allow only expected values to be processed. Any input that contain binary data, escape sequences, and comment characters should be summarily rejected.

Avoid concatenating user input that has not been validated. String concatenation is the primary point of entry for SQL injection attacks. Furthermore, consider using stored procedures to validate user input.

Analyze input and reject anything that contains special characters such as the semi-colon (;), the string delimiter (‘), comment delimiters (–, /*…*/), V$ (the beginning of Oracle DBA views), and xp_ (the beginning of SQL Server catalog stored procedures).

With foreknowledge of SQL injection techniques and proper development procedures, all SQL injection attacks can be prevented.

Summary

Understanding SQL injection techniques and coding your database applications appropriately is an important aspect of modern database security.

About craig@craigsmullins.com

I'm a strategist, researcher, and consultant with nearly three decades of experience in all facets of database systems development.
This entry was posted in data breach, Database security, DBA, SQL. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s