Reputation
38,570
Next tag badge:
384/400 score
81/80 answers
Badges
1 37 84
Newest
 Excavator
Impact
~575k people reached

Mar
10
reviewed Close Known plaintext, unknown 128 bit block cipher
Mar
10
comment Transport SA bundled inside a tunnel with encryption applied before authentication?
The standard way to do what you're doing is to use ESP in tunnel mode, with the ESP performing both encryption and authentication transforms (or alternatively a combined mode transform). Is there a specific reason why this isn't an adequate solution?
Mar
9
comment What happens to entropy after hashing?
@KarlZorn: I have no idea what you mean by 'entropy computed at byte or character level'. In addition, as how E(H(X))/E(X) acts depends a great deal on the precise distribution X (for example, consider a distribution which is defined such that only values $x$ with $H(x)=0$ have nonzero probabilities), it's not clear that your $f(x)$ is actually well-defined.
Mar
5
answered Will SHA1 or Other Hash Functions Ever Contain Quotes
Mar
4
comment Which characters to take into account when calculating unicity distance?
@GeorgeG: in addition to Ilmari's correct comments, there's another issue with unicity distance -- it's not very useful. Not only does it vary depending on the plaintext and the capability of the attacker (as Ilmari points to), in practice (and with real plaintext distributions) it's tiny -- somewhat larger than the key size. When we use AES to encrypt something, we're generally interested in encrypting more than 24 or 32 characters. Hence, we cannot depend on unicity distance; instead, we must rely on the computation complexity of breaking AES.
Mar
4
comment Find the prime factorization in the DLP
@RobertNACIRI: actually, I misspoke; it's $3x-1 \bmod (p-1)(q-1)/4$ which is 0 with high probability. In any case, if we have a value $k(pq-p-q+1)/4$ (for modest $k$), and $pq$, how can we rederive $p, q$?
Mar
4
comment Find the prime factorization in the DLP
@RobertNACIRI: I assumed that this was the question crypto.stackexchange.com/questions/24251/… (asked better); there, we assume that, given $g$ and $h$, we can find $x$ (and are asked to show that that implies an easy factorization method)
Mar
4
comment Find the prime factorization in the DLP
@mikerussel: no, your first statement was correct; there exists an integer $k$ such that $3x-1 = k(p-1)(q-1)/2$. Now, how big can $k$ be? Hint: $x < (p-1)(q-1)/2$, as it's the minimal $x$ value.
Mar
4
comment Find the prime factorization in the DLP
@mikerussel: What is $3x-1 \bmod {(p-1)(q-1)/2}$ with good probability?
Mar
4
comment Find the prime factorization in the DLP
@mikerussel: Actually, $3^{-1}$, not $-3$.
Mar
4
answered Find the prime factorization in the DLP
Mar
3
revised What is the advantage of using the socialist millionaire protocol to authenticate vs HMACing some random number?
added 152 characters in body
Mar
3
revised What is the advantage of using the socialist millionaire protocol to authenticate vs HMACing some random number?
added 293 characters in body
Mar
3
answered What is the advantage of using the socialist millionaire protocol to authenticate vs HMACing some random number?
Mar
3
answered One-one correspondance complete function
Mar
2
revised Is format preserving encryption suitable for use with words or names?
added 3 characters in body
Mar
2
answered Is format preserving encryption suitable for use with words or names?
Feb
27
comment What are the well known protocols that offer perfect forward secrecy?
One terminology nit: IPsec (RFC4301) itself doesn't provide perfect forward secrecy, as IPsec just handles the packet encryption (and other packet handling); perfect forward secrecy is a property of the keying protocol. Hence, it is IKE (both IKEv1 and IKEv2) that offers PFS; however IPsec keyed with a different protocol might not.
Feb
27
reviewed Close Exponentiation with fully homomorphic encryption
Feb
26
comment How to choose an algorithm
"The most important and difficult part is checking to see if there are known vulnerabilities"; actually, no, that's the easy part. The hardest (and most important) part is to make sure that you're using them securely, and in a way that covers all the security concerns in your attack model.