So far, we've discussed Windows hashes as password equivalents, what I like to call naked hashes. Those hashes never hit the network, though. The hash becomes the shared secret in an encrypted challenge-response mechanism. In NTLMv1, once the client connects to the server, a random 8-byte number is sent to the client – this is the challenge. The client takes the naked hash, and after adding some padding to the end, splits it into three and DES encrypts the three pieces, separately, with the challenge – this forms a 24-byte response. As the response is created with the challenge and a shared secret (the hash), the server can authenticate the client. NTLMv2 adds a client-side challenge to the process. Password crackers are aware of these protocol differences, so you can simply import the results of a capture and get to cracking. As a rule of thumb, the more sophisticated algorithms require more time to crack their passwords.

So you can either steal passwords from the SAM within Windows, or you can listen for encrypted network authentication attempts. The first option gets you naked hashes, but it requires a compromise of the target. We'll be looking at post-exploitation later in the book, so for now, let's see what happens when we attack network authentication.