« first day (3378 days earlier)      last day (1293 days later) » 

Crypto HNQ
Crypto HNQ
2:01 AM
1
Q: is MAC double-hashing enough to prevent length extension attacks?

hanshenrikI know that $$mac=\operatorname{SHA1}(secret\mathbin\|message)$$ is prone to length extension attacks, but what about: $$mac=\operatorname{SHA1}(\operatorname{SHA1}(secret\mathbin\|message))?$$ In order for someone to do a length extension attack of that, won't they either have to bruteforce the ...

hash mac sha-1 length-extension
 
 
10 hours later…
defalt
defalt
11:43 AM
"Don't roll out your own crypto" means to not use your own mathematical algorithm for cryptographic operations, what is the phrase for "Don't even combine existing cryptographic algorithms for your own application"?
 
kelalaka
kelalaka
11:53 AM
Don't roll out your own crypto means first learn the crypto. For beginners, this is very important. They come up with many ideas they think they have clever ideas that no one came up yet...
Combining is also requires knowledge and analysis, again first learn. Simple see Double encryption...
 
Maarten Bodewes
Maarten Bodewes
12:38 PM
@kelalaka Yeah, take for instance MAC-then-encrypt combined with CBC mode leaving you with a padding oracle. Or encrypt-then-MAC where people forget to include the IV.
 
kelalaka
kelalaka
1:22 PM
@MaartenBodewes yes. Maybe we should ask a canonical question for that :)
 
xorhash
xorhash
1:56 PM
“Don't roll your own crypto” has several layers of “don't”.

1. “Don't be clever about crypto”: Boring is good. If there's an industry-standard solution that fits your needs for 98% and you can live without the 2% custom solution, you're going to work with the 98% and throw the 2% under the bus.
2. “Don't implement your own crypto”: Use off-the-shelf libraries wherever possible.
3. “Don't design your own crypto”: Use tried-and-true designs. This applies to both protocols (use Noise or TLS over homebrewing your own authenticated key exchange) and primitives (use ChaCha20, AES, Poly1305, HMAC…
 
kelalaka
kelalaka
2:32 PM
But publishing is another story, you can publish almost everything
 
xorhash
xorhash
Anyone armed with LaTeX and two BibTeX references can at least get something into the IACR ePrint archive, at least.
Doesn't mean anybody's going to read it, but hey, you could have something that almost looks legit to your name
(It's a fun pastime, but I wouldn't really recommend it)
 
kelalaka
kelalaka
2:48 PM
I did not know that it is automatic.
 
xorhash
xorhash
It's not automatic, but the barriers of entry are seriously low
(Conversely, I would drop everything and flee if a paper is on arXiv but not on IACR ePrint archive)
 
kelalaka
kelalaka
Anyone else to submit AES-GCM vs ChaCha20-Poly1305 speed of their (good) system?
 
xorhash
xorhash
Dunno, but that's another one you can trivially do. Maybe throw in something funny as a gimmick, like Iunno, NORX or Xoodoo?
 

« first day (3378 days earlier)      last day (1293 days later) »