BLAKE3 1.0

#1

Post by ta988 » Mon, Jul 26, 2021, 1:24 PM UTC

Re: BLAKE3 1.0

#2

Post by ta988 » Mon, Jul 26, 2021, 1:27 PM UTC

BLAKE3 is a successor to BLAKE2 (used by wireguard for example) a really fast crypto hash function. It uses binary trees and other approaches that allow to use SIMD and multithreading.

Re: BLAKE3 1.0

#3

Post by leo_bloom » Mon, Jul 26, 2021, 2:35 PM UTC

If I had a BLAKE3 implementation available in the programming language of choice, is there any reason to still prefer the SHA family over it (for integrity checks, not for password hashing, as mentioned in the readme)?

Re: BLAKE3 1.0

#4

Post by garmaine » Mon, Jul 26, 2021, 2:38 PM UTC

leo_bloom wrote: post #3

If I had a BLAKE3 implementation available in the programming language of choice, is there any reason to still prefer the SHA family over it (for integrity checks, not for password hashing, as mentioned in the readme)?

Hardware implementation.

Re: BLAKE3 1.0

#5

Post by api » Mon, Jul 26, 2021, 2:39 PM UTC

leo_bloom wrote: post #3

If I had a BLAKE3 implementation available in the programming language of choice, is there any reason to still prefer the SHA family over it (for integrity checks, not for password hashing, as mentioned in the readme)?

Security-wise they are roughly equivalent. While SHA has had more eyes on it I doubt either construction will ever be practically broken at hash sizes like 384 or 512 bits. Someone may find an "academic break" at some point.

BLAKE3 is faster, sometimes a lot faster, on hardware without SHA instructions. On hardware with SHA instructions SHA may be faster. Same as the AES story where AES is faster than ChaCha on CPUs with dedicated hardware but slower otherwise.

Re: BLAKE3 1.0

#6

Post by Foobar8568 » Mon, Jul 26, 2021, 2:39 PM UTC

And I thought that JPA was slacking all this time drinking beers :o

Re: BLAKE3 1.0

#7

Post by loeg » Mon, Jul 26, 2021, 2:42 PM UTC

leo_bloom wrote: post #3

If I had a BLAKE3 implementation available in the programming language of choice, is there any reason to still prefer the SHA family over it (for integrity checks, not for password hashing, as mentioned in the readme)?

Federal government (FIPS) compliance.

Re: BLAKE3 1.0

#8

Post by fsiefken » Mon, Jul 26, 2021, 2:43 PM UTC

Hopefully, like BLAKE2 it makes it into OpenSSL.

One thing I was wondering about, OpenSSL 1.3 dropped support for SSL compression. Would that mean the compression function in BLAKE2 and future BLAKE3 integration couldn't be used or is this a different layer? https://en.wikipedia.org/wiki/BLAKE_(hash_function)

Re: BLAKE3 1.0

#9

Post by Croftengea » Mon, Jul 26, 2021, 2:44 PM UTC

Is this 1.0 ready indeed?

a comment in blake3_neon.c: // TODO: This is probably incorrect for big-endian ARM. How should that work?

Re: BLAKE3 1.0

#10

Post by chronogram » Mon, Jul 26, 2021, 2:49 PM UTC

Could someone explain how these versions in hash algorithms work please? As I understand it version 0.3.7 from last year would produce the same hashes as version 1.0, so is the algorithm itself finished long ago and it's just the tooling that got updated over time?