[Cryptography] RIP Claude Shannon

[Henry Baker]

At 03:59 PM 2/24/2016, Jon Callas wrote:
>> The 1x pad random keying material is "freeze dried adjacency".  It is a *resource* that you can stock up on in advance -- kind of like fuel or gold, that you can stockpile in your fallout bunker.
>> 
>> If you've got nothing important to talk about with a friend over a beer, share some key material instead.  Later, when you *do* have something to talk about, you can use some of that key material to gain the advantage of adjacency when discussing over long distances.
>> 
>> When you need to send out a confidential message, you sprinkle some of this fairy dust on the message, and you're good to go.
>
>My little thought experiment is saying that if you can securely transmit pads, you can securely transmit data.
>
>The reason for "stockpiling" the pad as you put it is only that you're on a crap network, and don't have crypto.

The establishment of 2 equal "keys", one at each location, is analogous to the establishment of 2 "entangled" sets of qubits, one at each location.

The mere creation and transportation of these shared keys and entangled qubits in the two different locations isn't "communication", per se.

However, these keys and qubits can be *later* used to enable confidential communications.  (BTW, quantum key distribution "QKD" is a really key *amplification* technology, rather than a key "distribution" technology, since there must already exist some initial shared key to protect against MITM attacks.)

If you are depending upon traditional "communication" to transfer the key bits, then you are correct.  However, at least in the traditional use of 1x pads, both of the equal keys are separately transported to their final location prior to being used for confidential communications.

The whole point of the 1x pad is to separate the secure transportation of the random keys -- which have no *current* value -- from the communication of the message -- which presumably has enormous value.  This separation allows us to (in effect) *move the communication backwards in time* to the point where the keys were being transported, and thus transfer the message as securely as those keys were transported.

Note that the amount of this "time travel" can be seconds, days, years, or decades.  The contents of the message cannot possibly have been predicted at the time the keys were transported, so there is no a priori association of a particular key with a particular message.

So yes, you are right, but you have built the equivalent of a time travel "wormhole" (of a particular bit length capacity) for later use for the confidential transmission of messages.  This is a far more powerful ability than the simple secure transportation of a message in real time.