4 ms·
Their proposed number of rounds is: AES: 11 instead of 14 for AES-256; Blake2: 8 instead of 12 for Blake2b (IIRC, Blake's rounds are derived from Chacha); Ch
by loeg 7y ago
Their proposed number of rounds is:
AES: 11 instead of 14 for AES-256;
Blake2: 8 instead of 12 for Blake2b (IIRC, Blake's rounds are derived from Chacha);
Chacha: 8 rounds instead of 20.
SHA-3: 10 rounds instead of 24 (Keccak).
(§5.3, "How many rounds?")
I'm not any kind of cryptographer, just an engineer who sometimes interacts with cryptographic algorithms, so I cannot comment on the specific recommendations.
On Chacha8 vs 20, see this tweet from DJB (2018): https://twitter.com/hashbreaker/status/1023969586696388613 https://twitter.com/hashbreaker/status/1023969586696388613
- tptacek 7y agoI'm not qualified to say this but because it's HN I'll say it anyways: I'd imagine Aumasson's response to this would be: "we've cut it close on key size by playing around below the 2^128 threshold, and close on block size in the same way, but there's no evidence we've come dangerously close to practical attacks based on iteration count, and some evidence in the other direction; in fact, at this point, a cipher design that wasn't attackable through some number of rounds would likely be rejected for want of cryptanalysis".
- zahllos 7y agoIt's also worth pointing out the NIST LWC has a proposed security level of 112-bit. We're actively designing and analyzing algorithms that can be acceptably below the 128-bit security bound, yet have a sufficient security margin for most people.