Reputation
354
Top tag
Next privilege 500 Rep.
Access review queues
Badges
1 8
Impact
~2k people reached

Nov
15
comment Is TEA considered secure?
@user54609 Sure, but all my non-microcontroller systems have AES-NI :)
Nov
14
comment MD5 Hash outputs not produced
@MaartenBodewes I agree.
Nov
13
comment Where can I get the correct and precise algorithms for elliptic curve cryptography?
Wikipedia! It's a very reliable source. :-P
Nov
13
comment Prevent replay attack without storage
Also, I don't see how this prevents replays of mote data. It authenticates the base, but not the mote. If the base doesn't store R, it has no way of knowing if the mote sent the right R+2 in step 3.
Nov
13
comment Prevent replay attack without storage
Wait, so the base can store received data but not session info?
Nov
13
comment Could eVoting on the bitcoin blockchain be done?
A good way to mitigate this would be to bind the satoshi given to the citizen to a multisig address, with the government's key as the other signature. Then, after physical presence at the polling place, the second signature is given.
Nov
13
comment Is TEA considered secure?
@user54609 One of the reasons that AES is so versatile is that when you need speed, you can treat those unusual field operations as just another lookup table.
Nov
13
comment A “one time pad” can be thought of a Vigenere cipher with
@SimonJohnson They didn't say it had an infinitely long key, just that it can be thought of as a cipher with an infinitely long key.
Sep
21
comment Compact digital signature for noisy data
@fgrieu Can you explain what you mean by "disregard for correlation (if any) between adjacent symbols"?
Sep
21
comment Compact digital signature for noisy data
@fgrieu Well, it would handle a few bitflips, but my bad. I was kind of thinking you might be asking for something that worked for arbitrary metrics because you wanted something like that as your arbitrary metric. I'll keep looking, focusing more specifically on Euclidean distance
Sep
21
comment Compact digital signature for noisy data
@fgrieu Take a look at ssdeep. It might be what you need. In any case, I suggest separating the hash and signature steps; instead of rolling your own special crypto that can detect differences, use standard, vetted crypto against a hash function designed for your purposes. That way you have very specific and testable potential failure modes.
Sep
20
comment Compact digital signature for noisy data
Instead of signing the message, or a hash, why don't you sign a fuzzy hash?
Sep
18
comment Birthday-attack calculation in planning password salting strategy
Never underestimate the number of users whose password is "password".
Sep
18
comment Is it not secure if we use weak passwords to protect exported PKCS #12 files with private keys?
@MaartenBodewes In my research on this stimulating subject, I seem to have found that Mac OS 10.6 introduced kSecAttrIsExtractable, even though it's not exposed in the GUI, so I guess my original point about Windows being the only platform dumb enough to offer an unenforceable guarantee of non-exportability wasn't correct. No way it'll happen on Linux though! crosses fingers
Sep
18
comment Is it not secure if we use weak passwords to protect exported PKCS #12 files with private keys?
@MaartenBodewes Only on mobile would the kernel be considered something that's not under the user's control! Still though, that's a good point. Android not so much, but iOS on devices with an A7 or later chip have a smart-card-like thing called the Secure Enclave that provides hardware protection to keychain elements. Still, on iOS it's not so much that you can mark a key as non-exportable as that there is no API for exporting keys at all. It's more neglect than a security guarantee, and might change in the future.
Sep
18
comment Is it not secure if we use weak passwords to protect exported PKCS #12 files with private keys?
@MaartenBodewes By "platform" I was thinking of an OS's default system for storing keys. JavaCard provides exportability guarantees, so I mentioned smart cards as an exception, but otherwise Windows's CryptoAPI is the only platform I know of that offers an no-exportability guarantee for key storage. Of course, as you say, throw in hardware tokens (or third party software modules) and all bets are off.
Sep
18
comment FIPS 140-2 Compliance Testing
@MaartenBodewes Good catch, I didn't know that. The only problem is it's in Java! (But seriously, C is more portable.)
Sep
18
comment Is it not secure if we use weak passwords to protect exported PKCS #12 files with private keys?
@phan "Marking a key as exportable" is, in general, a Windows-only concept. Most other platforms (smart cards excepted) have no notion of a locally-stored key that can't be exported, because they know they couldn't enforce such a guarantee.
Sep
18
comment Switching between AES and a highly performant algorithm
If you're going to use a less-secure cipher when on battery, I strongly recommend you use it at all times. It would be trivial for an adversary to downgrade the security to the lowest level, so the higher one buys you nothing but complexity, and complexity is the enemy of security.
Sep
18
comment Switching between AES and a highly performant algorithm
@Kar Not to mention the power-consumption side-channel attacks.