README: Notes on comparison #111
Conversation
|
Doesn't ahash provides HashDOS resistance too? They use various sources of randomness and pull that into their seed and they even say it's HashDOS resistant like SipHash in their README. |
|
I don't think it has been cryptographically proven to be HashDOS resistant... |
|
Yeah that's true |
|
Yeah, to my knowledge aHash has not been subject to any independent cryptographic analysis. Without such an analysis, I wouldn't trust any cryptographic claims. It is undoubtedly much closer to HashDoS resistant than FxHash, but that's not the same thing. |
|
Looks like this note was lost at some point... the README now again advertises hashbrown in a way that does not mention that switching from the standard HashMap to this, one will lose DoS resistance. (I consider any such claims about aHash to be highly speculative, so each user should decide for themselves if they consider this 'good enough'.) |
When hashers with different design goals are compared, that should be noted. Otherwise this compares apples with oranges.