Score:2

Equivalent conditions for perfect secrecy of a symmetric crypto system

in flag

I've been reading about perfect secrecy in crypto systems and I've ran across two definitions which turn out to be equivalent.

The first is Shannon secrecy:

A crypto system $(\cal K, \cal M$, $\text{Gen, Enc, Dec})$ is said to have Shannon secrecy if for all distributions $\cal D$ over $\cal M$ and for all $m\in\cal M, c\in \cal C$ $Pr_K[M=m| C=c]=Pr_K[M=m]$

where $K,M,C$ are random variables whose distribution is given by the distribution on $\cal M, \cal K$.

The second is Perfect secrecy:

A crypto system $(\cal K, \cal M$, $\text{Gen, Enc, Dec})$ is said to have perfect secrecy if for all $m_1,m_2\in\cal M, c\in \cal C$ $Pr_K[Enc(K,m_1)=c]=Pr_K[Enc(K,m_2)=c]$

My question is if that is also equivalent to the following:

for all $c_1,c_2 \in \cal C, m\in \cal M$:

$Pr_K[Dec(K,c_1)=m]=Pr_K[Dec(K,c_2)=m]$

Thanks in advance.

kelalaka avatar
in flag
Isn't it clear from the Shannon's? If the probabilities are not equal, then a ciphertext will leave more information to be close to $m$. Writing formal is cumbersome.
mangohost

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.