453 reputation
17
bio website
location Madrid, Spain
age 23
visits member for 1 year, 10 months
seen Sep 4 at 14:05

May
21
comment According to RFC 5246, are nonces in TLS useless (in terms of security) and if yes, why aren't they only random data?
Countermeasures of replay attack en.wikipedia.org/wiki/Freshness_(cryptography)
May
21
comment According to RFC 5246, are nonces in TLS useless (in terms of security) and if yes, why aren't they only random data?
Freshness is used to guarantee that a party is currently involved in a communication and not just replaying messagse. Freshness may be achieved by sequence numbers, timestamps or nonces (like in SSL). If you encrypt something that outputs always the same value, you can re-use that potentially. If the input has some sort of randomness (like a nonce), the encryption output will always be different and you cannot replay it.
May
21
comment According to RFC 5246, are nonces in TLS useless (in terms of security) and if yes, why aren't they only random data?
It provides freshness once again. If you do Finish prf(kcs, hash(messageonewithnonce 1)) and Finish prf(kcs, hash(messageonewithnonce 2) the results are different. You cannot replicate this without knowing kcs. The results are different ONLY because of the nonces.
May
21
comment According to RFC 5246, are nonces in TLS useless (in terms of security) and if yes, why aren't they only random data?
My bad. It is actually mutual Finish messages that you receive, as I said in my main answer. Both depend on the messages sent(including nonces) and the master key.
May
21
comment According to RFC 5246, are nonces in TLS useless (in terms of security) and if yes, why aren't they only random data?
Messages 4 and 5 provide mutual entity authentication. It is like a mutual challenge-response protocol but not only challenges, but the whole messages (including the challenges that are inside the response)
May
21
comment According to RFC 5246, are nonces in TLS useless (in terms of security) and if yes, why aren't they only random data?
It is not sent in plaintext. 7.4.8 is digitally signed, and 7.4.9 is a prf(mastersecret,messages). It is not that complicated if you are not obsessed on breaking an RFC. You cannot forge a digital signature, nor you know the master secret, so you cant complete the handshake without steps 4 and 5. The nonces form part of the encrypted data and allow you to provide freshness to both these messages avoiding replays.
May
21
comment Practical ways to prevent Denial Of Service attacks
MACs are calculated with a preshared key amongst two parties. If you dont have the key, you cannot forge it. These keys require secure distribution of course, but it was a bit out of scope of this question
May
21
comment According to RFC 5246, are nonces in TLS useless (in terms of security) and if yes, why aren't they only random data?
It is the same I said. My message 4 they call 7.4.8. Certificate Verify and message 5 is 7.4.9. Finished.
May
20
comment Using a TPM with Linux
Dont know if this is exactly what you want. But have you tried TrueCrypt? It can encrypt any volume.
May
20
comment Practical ways to prevent Denial Of Service attacks
Never heard of egress filtering. Does it have anything to do with ingress filtering?
May
18
comment Serpent cipher technical details in-depth
The block size for AES is always 128. Modes of operation address how to add message dependancy or position dependancy to more than one block of ciphertext.
May
15
comment Are Telegram secret chats secure assuming MTProto isn't?
For end to end they use AES-256 as well. The fingerprint looks kind of lame though (and re-uses the encryption key). digest = md5(key + iv) fingerprint = substr(digest, 0, 4) XOR substr(digest, 4, 4)
May
12
comment How does one know when a message is successfully decrypted?
Asymmetric cryptography does not ensure nor integrity nor non-repudiation by itself. You need to bind the identity somehow. Confidentiality does not imply data origin authentication. What you need as I said below, is a digitally sign the message (to provide non-repudiation plus data origin authentication) or simply use a MAC function (data origin authentication)
May
12
comment Is breaking a looping one-time pad practical? If so, how?
The key length is the same or bigger than the plaintext length. So practical... If you think that managing this size of keys is practical then fine, but you have other options that are less a pain in the ass.
May
8
comment Can NSA generate all hashes for rainbow table to break md5?
@Gumbo Correct. Ill edit it now.
May
8
comment Can NSA generate all hashes for rainbow table to break md5?
The Birthday Paradox. Read more on it if you like. The final conclusion is that given n numbers there is a probability of sqrt(n) that there is a collision. So for MD5 (128 bits), we would obtain a collision on 2^(128/2) on average - 2^64. My definition of breaking would be obtaining a second preimage for a given hash.