SQL injection may be the most common Web application attack. It is based on inserting Structured Query Language (SQL) commands into text boxes, such as the username and password text fields on the logon screen. To understand how SQL injection works, you have to first understand the basics of how a logon screen is programmed. According to the most recent Verizon Terremark Data Breach Investigations Report, SQL injection was once one of the most common types of web attacks, but its use has fallen off in recent years thanks to better website coding practices. Still, it is remarkable how many sites remain susceptible.

A logon screen requires the user to enter a username and password, and these then have to be validated. Most likely they are validated by checking a database that has a table of users to see if the password matches that user. All modern databases “speak” Structured Query Language SQL. If the programmer who created the logon is not careful, the webpage may be susceptible to SQL injection. Here is how that attack works. SQL looks a lot like English. For example, to check a username and password, you might want to query the database and see if there is any entry in the users’ table that matches the username and password that was entered. If there is, then you have a match. The SQL statements in the programming code for the website have to use quotation marks to separate the SQL code from the programming code. A typical SQL statement might look something like this:

SELECT * FROM tblUsers WHERE USERNAME = '" + txtUsername.Text
+' AND PASSWORD = '" + txtPassword.Text +"'

If you enter the username 'thisuser' and the password 'letmein', this code produces the SQL command

SELECT * FROM tblUsers WHERE USERNAME = 'thisuser' AND
PASSWORD = 'letmein'

This is fairly easy to understand even for nonprogrammers. Plus, it is effective. If there is a match in the database, that means the username and password match. If no records are returned from the database, that means there was no match, and this is not a valid logon.

SQL injection is basically about subverting this process. The idea is to create a statement that will always be true. For example, you enter into the username and password boxes the following ‘ or ‘1’ = ‘1. This causes the program to create this query:

SELECT * FROM tblUsers WHERE USERNAME = '' or '1' = '1' AND
PASSWORD = '' or '1' = '1'

So you are telling the database and application to return all records where username and password are blank or if 1 = 1. The fact is that 1 always equals 1, so this works. Now if the programmer wrote the application properly—to not allow any extra characters—this does not work. But in all too many cases, it does work. And then the intruder has logged on to your web application and can do whatever any authorized user can do.

The example just described is the simplest version of SQL injection. There are more advanced SQL injection techniques. In fact, SQL injection is limited only by your own knowledge of SQL. Obviously, SQL injection attacks can be very sophisticated, involving a complex array of SQL statements, or they can be as simple as the example just discussed.

There are even tools that make the process of executing this attack—or testing your website to see if it is vulnerable to this attack—even easier. A few are listed here:

• Several tools from Database Security at http://www.databasesecurity.com/sqlinjectiontools.htm

• Sqlmap at http://sqlmap.org

• SQL Ninja at http://sqlninja.sourceforge.net

This is another very common attack. In this attack, the perpetrator seeks out someplace on the target website that allows end users to post text that other users will see. Product reviews are a great place for this. But instead of posting a review or other text, the attacker will post JavaScript. If the website does not filter the user input before displaying, then when other users navigate to this review, the script will execute. The attacker is only limited by his or her knowledge of JavaScript. One popular thing to do is to redirect the user to a phishing site.

To better understand this attack, let us walk through a scenario. Before we do, I want to point out that to the best of my knowledge, none of the major e-commerce or auction sites are vulnerable to this. In this scenario the attacker sets up a website that looks very much like the target site. He or she gives this site a name similar to the target site, off by only one letter. That makes it likely an end user will not notice the minor difference. Then the attacker infects the target site with cross-site scripting, inserting a script that will redirect the user to the phishing site. Once at the phishing site, the malicious site simply displays a notice stating “Your session has expired, please log in again.” When the user logs in, the malicious site captures the login information, then sends the user back to the real site. Now the attacker has that user’s login information. To the end user, it would appear as if they have always been on the same website.

Probably one of the most basic tools for physically accessing a Windows machine is oph-crack. Ophcrack can be downloaded from http://ophcrack.sourceforge.net. It is a tool to crack the local passwords on Windows systems. It is usually pretty effective. In fact, if an attacker can sit at a Windows machine on your network with an Ophcrack compact disc (CD) or USB flash drive for 10 minutes or less, chances are that the attacker is going to get the local passwords. This is a significant problem on corporate networks that meet the following criteria:

• It is usually possible to find an unoccupied desk.

• If physical security is lax, an outsider can get in the building, often as a cleaning or maintenance person or temporary employee.

• Focus is on domain accounts, not local. There is a good chance that the local accounts for one Windows machine are the same throughout. The systems are usually just imaged from a base system.

How does Ophcrack work? First, let’s discuss how Windows passwords work. When you choose your local Windows password, the password you choose is hashed and stored in the SAM file, which is found in the WindowsSystem32 directory. The hash of the password is not the password itself, but it is created from the password using a hashing algorithm that makes two identical hashes for different passwords very unlikely.

To make this work, all you have to do is put the CD into the system and reboot. During the boot-up process, press the key needed to enter the system’s BIOS, often F2 or the Delete key, to change the boot order. Additionally, pressing the F10 or F12 key might bring up a boot menu, where you can then choose Boot from CD. Note that in FIGURE 2-2, the passwords that Ophcrack found are blocked out.

There’s a lot to know about hashes and cryptography, but for now you just need to know that a hash is one-way; you don’t “unhash” something. When you log on to Windows, the system hashes whatever you type in and compares it with whatever hash is in the SAM file. If there is an exact match, then you are given access.

A rainbow table is a table of precomputed hashes. The simplest, but largest, rainbow table is created by the software creating tables of all the possible keyboard combinations that could be used for a password, then creating hashes of these passwords. There are methods to make smaller, more focused, rainbow tables, but this is essentially the process. All you have to do is search for the hash that is in the SAM file, and the column matching it will have the password. This is called a rainbow table because it contains every conceivable combination of keyboard characters under the rainbow and their associated hashed versions. These tables get huge very fast, so most are no more than 8 to 10 characters in length, which works for most passwords.

The problem is getting the SAM file to start with. When Windows boots up, long before you even get the logon screen, the system locks the SAM file—preventing you from copying or editing it. Well, that is where Ophcrack comes in. It boots to a Linux Live CD and then scans its rainbow table searching for matches. It displays all the passwords it finds in an easy-to-use graphical user interface.

Once the attacker has a valid logon account, particularly an administrator account, he or she can log on to that computer. This doesn’t let the attacker join the domain, but he or she now has a foothold in your network. You can see this in Figure 2-2.

The following is a simple trick, and one that is a follow-up to using Ophcrack to break local accounts. After the attacker has gained access to a local account, he or she will really want to get domain admin privileges. The command net user can help do this. First, the attacker writes the following two-line script:

net user /domain /add localaccountname password
net group /domain "Domain Admins" /add localaccount

The attacker then saves that script in the All Users startup folder. The next step is to get a domain admin to log on to this machine. If that happens, the script runs—in the background and not visible—and your local account is now a domain admin. But how do you get a domain administrator to log on? Well, it just so happens that in many organizations, the tech support personnel are in the domain admin group. So the attacker merely needs to do something to render the machine not fully operational. When a tech support person logs on to fix that problem, the script is run.

Entire books have been written on hacking techniques, and entire certification courses focus on such skills. The purpose here is to introduce you to some techniques that an attacker might use. To go into all the common hacking techniques attackers use is beyond the scope of this text. However, this section illustrates an important point: computer crime investigators should have a strong working knowledge of hacking techniques. Your forensic investigations will be drastically improved if you have an understanding of the techniques that attackers use.