If a protocol has a statistical security parameter, does this mean it is only computationally secure?

HelloWorld123

10/29/22, 1:00 PM

if a cryptographic protocol has a computational security parameter and a statistical security parameter, does this mean it is only computationally secure instead of information-theoretically secure?

I am wondering because this answer says that statistical indistinguishability is when the adversary is computationally unbounded: https://crypto.stackexchange.com/a/11790 That would imply that the presence of a statistical security parameter means that a protocol is information-theoretically secure. Is that true?

Thank you!

1 + 2

provable-security

Ievgeni

10/30/22, 10:25 AM

Does this answer your question? [Statistical security parameter -> information theoretically secure](https://crypto.stackexchange.com/questions/91814/statistical-security-parameter-information-theoretically-secure)

HelloWorld123

10/30/22, 10:33 AM

@levgeni Yes that's my question and the answer helped me a lot. This question right here is one I initially posted on stackoverflow and because it didnt belong there it got moved here it seems. I will try if i can delete this one here because it is basically copied.

Score:1

Crypto

Meir Maor

10/30/22, 9:00 AM

Yes, security parameters imply we are dealing with computational security not with information theoretical security. But there could be other usages for security parameters.

When we use one time pad or secret share we have information theoretical security. It can't be broken with any computational power and there are no security parameters.

When we do computational security, we often have a security parameter, to give e.g key size or iteration count, and we hope as we increase it so do we increase the computational cost for attacker.

However there could be other cases, we could have proofs or signature which are secure even against an unbounded adversary but still have a probability of failing (e.g probability of a forgery not being detected) and these may be controlled with a security parameter as well.

With some quantum systems, there seems to be a guarantee of perfect security regardless of adversary, e.g with Quantum Key Distribution yet the actual system may fail to full meet the requirements for the proof to actually hold. Again we may have a security parameter (though probably not one we control and pick) which could describe e.g the probability of getting more than a single photon. This as well can be a quantifier of system security.

0 + 0

Elon Musk

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

EN: If a protocol has a statistical security parameter, does this mean it is only computationally secure?

TH: หากโปรโตคอลมีพารามิเตอร์ความปลอดภัยทางสถิติ แสดงว่ามีความปลอดภัยทางคอมพิวเตอร์เท่านั้นใช่หรือไม่

RO: Dacă un protocol are un parametru de securitate statistică, înseamnă asta că este sigur doar din punct de vedere computațional?

RU: Если протокол имеет статистический параметр безопасности, означает ли это, что он безопасен только с точки зрения вычислений?

VI: Nếu một giao thức có tham số bảo mật thống kê, điều này có nghĩa là nó chỉ an toàn về mặt tính toán?

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.