1

While discussing archive authenticity/integrity for long term and the problem of potentially re-keying in case of compromise, someone came up with the term "trans-cipher" (or "trans-ciphering").

I have tried to search the entire internet for this, but could not come up with a meaning for this. Can someone enlighten me?

Note: in the context of long term archive authenticity, the term came up while talking signature verification and the problem when a private key is compromised or when the algorithm for the public/private key gets obsolete (e.g. due to hardware raw power reaching a point where what was in the past difficult to factor becomes trivial).

PS: it is possible that the exact term is slightly different than "trans-cipher". On this project digital footprint is being used for digital fingerprint (aka crypto-hash function)...

4
  • 3
    I've never heard that term before. Sounds like a term that someone just made up, possibly to mean "bridging across different algorithms or different keys". Given that they are also using "digital footprint" to mean "digital fingerprint", and the fact that "signatures", "ciphers" and "fingerprints" are all different concepts, I'm not sure I trust this person as an authoritative source of jargon. Commented Sep 14, 2018 at 14:24
  • 3
    Clearly, it's the opposite of a cis-cipher. Commented Sep 14, 2018 at 14:43
  • I would agree with you but check this doc page 2 (last paragraph) eprint.iacr.org/2017/246.pdf they do mention trans-ciphering but I also didn’t know about homomorphic encryption so I’m not understanding there explanation. Does it make sense to you?
    – Huygens
    Commented Sep 14, 2018 at 15:08
  • @MonicaApologistsGetOut I guess any digital signature that has not been re-keyed is a cis-cipher? Commented Jun 6, 2022 at 16:47

1 Answer 1

0

This summary is based on my understanding of the first few minutes of this talk.

The motivation observation is than size(HOMencrypted(text))/size(text) >> 1 while size(classicencryption(text))/size(text) ≈ 1. So in order to minimize data transport instead of sending HOMencrypted(text) and a homorphic program that operates on it, we send classicencryption(text), HOMencrypted(keyphrase), a homomorphic program that implements the decryption of classicencryption(text) given HOMencrypted(keyphrase) into homomorphically encrypted state and a homomorphic program that operates on the "classically unencrypted" but homorphically encrypted state, so that the person providing the compute can never see text without breaking the homomorphic encryption.

This requires that the classical cypher is friendly towards homomorphic encryption while still classically strong enough that the person with the compute can not learn anything from the classically encrypted data.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .