cgmossa / blake3

BLAKE3 Cryptographic Hash Function in R

blake3

blake3 is an interface to the BLAKE3 cryptographic hash function.

To quote blake3’s readme:

BLAKE3 is a cryptographic hash function that is:

• Much faster than MD5, SHA-1, SHA-2, SHA-3, and BLAKE2.
• Secure, unlike MD5 and SHA-1. And secure against length extension, unlike SHA-2.
• Highly parallelizable across any number of threads and SIMD lanes, because it’s a Merkle > tree on the inside.
• Capable of verified streaming and incremental updates, again because it’s a Merkle tree.
• A PRF, MAC, KDF, and XOF, as well as a regular hash.
• One algorithm with no variants, which is fast on x86-64 and also on smaller architectures.

BLAKE3 is not a password hashing algorithm, because it’s designed to be fast, whereas password hashing should not be fast. If you hash passwords to store the hashes or if you derive keys from passwords, we recommend Argon2.

If you want to store passwords in R, please also take a look at sodium::password_store.

Installation

remotes::install_github("dirkschumacher/blake3")

Also make sure to have a very recent rust compiler. The best way to install the rust toolchain is to use rustup.

API

It is still a very early version of the package. The api currently just has one function to hash RAW vectors with an optional 32 byte RAW key parameter.

Example

library(blake3)

input <- "This is a string"
hash <- blake3_hash_raw(charToRaw(input))
(sodium::bin2hex(hash))
#> [1] "718b749f12a61257438b2ea6643555fd995001c9d9ff84764f93f82610a780f2"

# you can also hash arbitrary R objects by serializing it first
input <- serialize(LETTERS, NULL, version = 2)
hash <- blake3_hash_raw(input)
(sodium::bin2hex(hash))
#> [1] "7b2fe4e871bf4dfca892e9b46ae237f55103235e635d994f351e4553aced2bee"

# keyed hashes are also supported
key <- blake3_hash_raw(charToRaw("test"))
input <- serialize(LETTERS, NULL, version = 2)
hash <- blake3_hash_raw(input, key)
(sodium::bin2hex(hash))
#> [1] "0021965f7c0ca0233e576670b84d35506129ef4bd94c5c8a07b8bb4bdbfb6ce3"

Benchmark

To get an idea about the speed we hash a random string of 1 million characters and compare it to the openssl implementation of sha1, sha2 and md5.

library(openssl)
set.seed(42)
input <- charToRaw(paste0(sample(LETTERS, 1e6, replace = TRUE), collapse = ""))
microbenchmark::microbenchmark(
  md5 = md5(input),
  sha1 = sha1(input),
  sha2 = sha2(input),
  blake3 = blake3_hash_raw(input),
  times = 1000
)
#> Unit: microseconds
#>    expr      min       lq      mean   median        uq      max neval
#>     md5 1640.739 1679.727 1866.0004 1736.997 1904.4585 5902.874  1000
#>    sha1 1177.078 1206.927 1429.5398 1255.660 1502.0495 6223.994  1000
#>    sha2 2522.645 2589.403 2929.1884 2723.578 3054.4920 6785.628  1000
#>  blake3  389.721  413.038  575.0413  448.384  611.8085 3698.705  1000

License

MIT