Score:2

Crypto

A perfect key exchange similar to the One Time Pad

undefned

2/2/24, 7:20 PM

The one time pad is unbreakable because each message could be one of many, and each one is equally likely.

Is it possible to do this with a key exchange? (So if someone records the key exchange, there would be a large number of equally possible keys)

I am already aware of Quantum Key Distribution, so please don't put that as your answer.

145

1 + 11

key-exchange

Maarten Bodewes

2/2/24, 7:30 PM

Key agreement algorithms like DH depend on a mathematically hard problem where performing an operation would be computationally infeasible. Those can of course be discounted. So if you require this kind of scheme without a known secret then it is kind of prohibited by default. I don't think that this would be feasible using a symmetric scheme with key derivation either leaving the solution space empty. If it could be done efficiently then we'd of course already been doing it.

Amit

2/2/24, 8:20 PM

If someone ever finds such an algorithm, they should call it "The Zero Time Pad" :)

Amit

2/2/24, 8:30 PM

... This is probably not a satisfactory answer, but just in case you aren't aware of this, Merkle Puzzles have the property where an attacker is faced with the problem of a large number of possible keys: https://en.wikipedia.org/wiki/Merkle%27s_Puzzles

Paul Uszak

2/3/24, 12:18 AM

@Amit Yes, it's called Quantum Key Distribution.

forest

2/3/24, 1:52 AM

@PaulUszak OP said he is not looking for QKD, which is a physical process, not a mathematical algorithm.

Amit

2/3/24, 9:07 AM

QKD is great but I like to say that it's not a part of Cryptography! Because in my mind the whole point of Cryptography is the independence of the technique from the physical medium. The opposite is true in QKD

Paul Uszak

2/3/24, 1:18 PM

@Amit _"QKD is great but I like to say that it's not a part of Cryptography! "_ NATO, China, all the technical universities, all the crypto companies, Switzerland, NSA, CIA, BBC, European Union, USA, Japan, Taiwan etc, etc, etc may disagree. But what do they know?

Amit

2/3/24, 1:37 PM

@PaulUszak - I think you are disregarding the fact that I said it is *great* . My statement was only about how to categorize QKD, no doubt it is useful. Also, whispering in your friend's ear very quietly, is a highly useful way to obtain secrecy. I'm not kidding! So if someone wants to define that as Cryptography, why not? All I am saying is that personally, I don't think it is quite the right category

Paul Uszak

2/11/24, 1:18 PM

@Amit You've already conceded that QKD _is_ a part of cryptography. But now: _"Cryptography is the independence of the technique from the physical medium"_. PUFs, ring oscillators and TRNGs based on physical properties of things like lasers are tied directly to their underlying physical mediums.

Amit

2/11/24, 1:33 PM

@PaulUszak It's not a matter of concession. Who am I to decide what is and what is not part of the field called "Cryptography"? All I did, was try to determine what is the consensus about this. Having understood the consensus, I go along with it. I may disagree with it in some ways, but I go along with it simply in order to have a common language with other people. Nothing more, nothing less.

Amit

2/11/24, 2:32 PM

@PaulUszak - you may find the following bit interesting: https://www.youtube.com/watch?v=N-Xi1VUo-ic&t=1800s&ab_channel=FIDOAlliance You may know the speaker. Point is: opinions are many, convention is singular.

Score:2

Crypto

ckamath

2/3/24, 12:04 PM

Key exchange in the conventional adversarial model (where the adversary can observe and store the whole transcript) is possible only under computational assumptions, and therefore it is not possible to have perfect/information theoretic secrecy.

However, there are restricted adversarial models where this can be guaranteed. One example is Maurer's bounded storage model (also see the introduction in this paper) where the assumption is that the adversary can only store a fraction of the key/transcript.

You can find examples of other such models in the introduction of Maurer's paper.

+ 0

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: A perfect key exchange similar to the One Time Pad

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.