SEAL
SEAL copied to clipboard
Use Blake3?
Blake3 (https://github.com/BLAKE3-team/BLAKE3) was announced in January 2020 at Real World Crypto, and claims to be ~5x faster than Blake2b.
Would Blake3 be appropriate to replace Blake2 in SEAL? Or is it not widely-enough accepted yet? Perhaps Blake2 is not a bottleneck, so it's not worth the effort?
Very good suggestion again from you. We'll need to test this, do some benchmarks, and figure out whether it makes sense. Changing the hash function is in principle really easy, but I'm not sure if we should support multiple options, what would be the best way to specify them, and where should information about the hash function be stored (EncryptionParameters, or maybe Ciphertext?).