Reputation
308
Next privilege 350 Rep.
Access review queues
Badges
1 10
Newest
 Yearling
Impact
~3k people reached

Apr
1
comment Can Poly1305-AES be used with AES-256?
@Maarten - agreed. I'm not really looking to improve theoretical security. Rather, I'm trying to nail down an interface to Poly1305 in the Crypto++ library. In effect, the question becomes how much leeway to allow a programmer when constructing one in Crypto++.
Apr
1
comment Can Poly1305-AES be used with AES-256?
After thinking about it some more, I don't think there's an advantage to using AES-192 or AES-256. Unless, of course, the security level of AES-128 drops below the security provided by r, which is 106 bits. And I don't know what would happen if AES-128 lost 20 or 25 bits of security overnight. It would probably be kind of funny to watch (in a morbid sort of way).
Mar
29
accepted Can Poly1305-AES be used with AES-256?
Mar
29
asked Can Poly1305-AES be used with AES-256?
Feb
28
comment RSA encryption with private key and decryption with a public key
I'm probably splitting hairs, but "encrypt with the private key" is not a valid cryptographic operation. That usually get translated into a Signature Scheme with Recovery (or indicates that's what the OP wants).
Nov
20
awarded  Yearling
Sep
24
awarded  Autobiographer
Jul
28
awarded  Citizen Patrol
Jul
2
awarded  Curious
Jun
29
comment Entropy of Androids motion-sensor data
"... If l is less than 81% of the sensors Peek value the set of values is ignored..." - I think this is less than ideal. When you are trying to get a PRNG in good working order or keep a PRNG in good working order, use whatever you have. Don't throw less-than-perfect ones away. Peter Gutmann said similar: "...mix every entropy source you can get your hands on into your PRNG, including less-than-perfect ones...". From Heml.is - "The Beautiful & Secure Messenger".
Jun
29
comment Entropy of Androids motion-sensor data
On both Android and iOS, I perform similar. I use the sensors to gather a seed; and I don't sample the generator for prolonged periods of time. I grab two or three samples on app startup (or during a reseed), and then feed them to the crypto library's random number generator. You get an {x,y,z}-float triplet from the sensors. Just send them into the PRNG without preprocessing like ((int) x) ^ ((((int) y) ^ ((int) z)) << 16). What's the point of preprocessing and throwing bits away? Let the PRNG handle it. That's what it was designed to do.
Jun
18
revised Trial divisions before Miller-Rabin checks?
Added info on tainting answers.
Jun
18
comment Trial divisions before Miller-Rabin checks?
Thanks @Thomas. I appreciate the help. I would like to accept both answers, but I cannot.
Jun
18
accepted Trial divisions before Miller-Rabin checks?
Jun
18
comment Trial divisions before Miller-Rabin checks?
@fgieu - Forgive my ignorance.... Isn't this log base 2: $t=\lceil{1\over2}\cdot\lg n\rceil$? When I programmed it, I believe I calculated it as ceil( (log(k)/log(2))/2 ). I'm fairly certain k was bits, and not a random k-bit number. Unfortunately, I overwrote the test program already.
Jun
18
revised Trial divisions before Miller-Rabin checks?
deleted 5 characters in body
Jun
18
revised Trial divisions before Miller-Rabin checks?
Added edit about OpenSSL and Crypto++.
Jun
15
revised Trial divisions before Miller-Rabin checks?
Algorithm 4.4 -> Algorithm 4.44
Jun
15
revised Trial divisions before Miller-Rabin checks?
Added that 17863 is the 2048-th prime.
Jun
15
asked Trial divisions before Miller-Rabin checks?