BB84 protocol what do they do once they have the key? In the BB84 protocol Alice and Bob share a key via a method using both quantum and classical channels. I understand how they do this. But I don't understand what they then do with the key? I.e. How do they use the key to transfer information?
 A: The key is a "one time password" - like a "one time pad". In the good old days of manual cryptography, embassies would be sent these one time pads which they could use to encrypt a message. Using modular addition / subtraction, with a key as long as the message, you can make an unbreakable code - as long as the key remains secret. By rotating keys all the time, you could make an unbreakable messaging system. But getting the key (one time pad) to the embassies (making sure nobody made copies along the way) is extremely difficult. And either everyone uses the same pad (in which case a single security breach compromises the entire system) or you multiply your problems by having to keep track of which embassy (or secret agent) used which pad.
The famous Enigma cryptography machine was an example of a one-time pad system. You had to know the initial settings of the wheels of the machine at the start of the message in order to decode it: these settings changed daily, and they were hard to guess with methods available in the 40's. The cleverness of Bletchley Park was the development of early computers (thanks Alan Turing) and exploitation of certain patterns in the messages to find the daily settings, and thus be able to interpret the intercepted code. It was not a true "one time pad" in the sense that the key was relatively short - but the wheels rotated in accordance with the message, so the first part of the message became, in some sense, part of the key. And since you did not know the message without knowing the key, it was "like having a key as long as the message" - but not quite. Which is why it could be broken.
In modern cryptography, encryption typically happens in two steps. In the first step, Alice and Bob exchange a unique (one time) key, using the most secure electronic method they have available. This ranges from a courier carrying the equivalent of a one-time pad (on a USB stick) to the more common SSL protocol where a public/private key pair is used for secure communication of a key. This is a computationally intensive step - but only needs to apply to the handful of bytes that comprise the key (4096 bit keys are common). This step could be done using quantum cryptography.
Once Alice and Bob have the key, they can encrypt a message using a simple symmetrical algorithm - apply the key once and the message becomes encrypted; apply it again, and the message is decrypted. Unless you have the key, the message remains garbled. The length of the key is often quite a bit shorter than the message, meaning that it is not "as safe" as encrypting the entire message with a key as long as the message; but once the keys get this long, they become pretty much impossible to guess. Of course there is always a race between the good guys and the bad gusy - and this is why crypto code which is capable of using longer keys is export controlled. There are agencies who can do some pretty powerful cryptanalysis "up to a point". They would like it if nobody used keys stronger than that...
