Cipher text only: This attack is almost always open to an attacker. The idea is that based solely on the encrypted message, an attacker tries to deduce the plain text. Brute-force attack on the key is one example of this type of attack.

Known plain text: In some cases, an attacker might know some or all of the encrypted plain text. This knowledge might make it easier for the attacker to determine the key and/or decipher other messages using the protocol. Typical examples of known plain text exposure come when an attacker knows that encrypted content consists of file types that contain standard headers, or when an attacker knows the message concerns a named subject. In other cases, entire messages might get leaked by means other than a break of the encryption, thus helping an attacker break other messages.

Chosen plain text: An attacker might have a way of inserting specially selected plain text into messages prior to their encryption. Initially, this might seem unlikely to occur; but let's look at a plausible example. Suppose Alice runs a mail server that filters out suspected e-mail viruses. Furthermore, she forwards an encrypted copy of suspect e-mails to virus expert Bob. Attacker Mallory can deliberately mail a virus (or something that resembles one) to Alice, knowing that its specific content will appear in a message from Alice to Bob.