10,257 reputation
12452
bio website github.com/CodesInChaos
location Frankfurt, Germany
age
visits member for 3 years, 5 months
seen yesterday

1d
comment The purpose of the final xor in Davies–Meyer scheme
It enables a meet-in-the-middle attack, reducing pre-image resistance. Keccak suffers from this attack, which is why its preimage resistance is only half the capacity.
2d
comment What would be the best plain text Cryptography method without the use of a computer?
See Is there any strong enough pen-and-paper or mind cipher?
2d
comment Dancing confusion with Daniel J. Bernstein's stream ciphers
If you have a per session key, you still need a counter to distinguish messages. If you have a per message key (like noise) you don't need a nonce at all. If your protocol fits random nonces, use ciphers which support that. Choosing or designing a good protocol is the important part, choosing between Salsa20 and ChaCha a minor detail.
2d
comment Dancing confusion with Daniel J. Bernstein's stream ciphers
For a random nonce you need XChaCha or XSalsa20. Finding an XChaCha implementation and making sure it's correct will be harder than finding a correct XSalsa20 implementation. So if you need long/random nonces, Salsa20 might be preferable over ChaCha.
2d
comment Dancing confusion with Daniel J. Bernstein's stream ciphers
Deciding between Salsa20 and ChaCha is a matter of preference. Salsa20 has the advantage of having competed in eStream, being part of the NaCl library and a (rather limited) proof of security against differential cryptoanalyis. ChaCha has slightly better security per round, a bit more elegant and it's used in a TLS suite. Personally I'd choose ChaCha if either is equally practical to use but might prefer Salsa20 if I'm already using a library where it's implemented.
2d
comment Dancing confusion with Daniel J. Bernstein's stream ciphers
How to handle nonces should mainly be decided by the requirements of the application. I would avoid keeping nonces between runs of an application, so if you reuse a key beyond application restarts a counter doesn't work well. But if you encrypt a sequence of messages in a communication protocol the key is typically per-connection and a counter works well.
Dec
16
comment What would be the best plain text Cryptography method without the use of a computer?
@tylo I don't know any secure hand cipher that's less terrible than the OTP from a usage point of view. RC4, solitaire etc. have the same key reuse problems.
Dec
15
comment Recovering El Gamal secret key from signatures
@Kairos Since this is a duplicate it has been closed and your bounty refunded. If you want to, you can offer a bounty on the duplicate.
Dec
15
comment Is cm signed signature verification considered “slow”
Which algorithm are you using for signatures? How long are the messages?
Dec
15
comment request for data to test deterministic ecdsa signature algorithm for secp256k1
Thread on bitcointalk that has several test vectors: Deterministic Usage of DSA and ECDSA Digital Signature Algorithms (RFC 6979)
Dec
12
comment Decrypt all RSA ciphertexts if you can break 1% of them
I didn't mean your cross-post, but some older question.
Dec
12
comment Decrypt all RSA ciphertexts if you can break 1% of them
Do we have a duplicate of this question? I thought I already read an equivalent question, but couldn't find it.
Dec
11
comment Probability of factoring keys as a function of bit length
Similar question: How many RSA keys before a collision?
Dec
11
comment RSA or Paillier is good?
I suspect you have a misunderstanding about what homomorphic encryption with RSA or Paillier achieves. Encrypting files like that is almost certainly not useful.
Dec
9
comment Is a strong block cipher usable as a strong sponge function?
What's a bit annoying here is that you need to make relatively strong assumptions about the block cipher, not just that it's a PRP.
Dec
5
comment Is there (this) a case in which salting is dispensable?
Salting prevents many multi-target attacks, not just rainbow tables.
Dec
4
comment SHA1 - SSL/TLS Cipher Suite
@crypto-learner The attacker would get the CA to look the harmless looking one of the colliding pair and then use the malicious one with the same signature.
Dec
3
comment Is this problem based on a known hard problem?
I have never seen a meaningful reduction to knapsack. There are almost always issues around asymptotic vs concrete security and average vs worst case. I wouldn't not use anything knapsack based in practice.
Dec
2
comment Repeating something encrypted and non-encrypted?
As long as your encryption is secure know plaintexts don't help the attacker.
Dec
2
comment Does Ed25519 support cryptographic threshold signatures?
There is always the generic solution: A public key is the concatenation of n individual public keys and the threshold m, a signature is the concatenation of m individual signatures.