#proof-of-work #crypto

blake3-proof-of-work

A basic proof of work scheme using the blake3 hash function

3 unstable releases

0.2.1 Jan 17, 2022
0.2.0 Jan 17, 2022
0.1.0 Jan 17, 2022

#1089 in Cryptography

MIT/Apache

7KB
89 lines

Proof of Work

A simple proof of work algorithm using the Blake3 cryptographic hash function.

let cost = 22;
let bytes = b"Hello, world!";
let meter = 10000000;
let nonce = proof_of_work::search(bytes, cost, meter);
assert!(proof_of_work::verify(bytes, nonce, cost));

The main point is: we present some bytes and we say that a "proof of work" for some cost is a nonce : [u8; NONCE_SIZE] such that the hash of nonce concatenated to bytes has cost leading zeros.

To verify such a proof, we compute the hash and check if it has cost leading zeros. To search for such a proof, we continually generate random nonces until we guess one which constitutes a proof of work. That is to say, we randomly guess until we get it right. Given that this could go on forever, we pass in a meter : u32 in order to stop after a certain number of attempts.

Use Cases

When you want to expose functionality to the outside world without allowing bots to take advantage of it at any frequency, you must meter usage somehow. By requesting that API calls come affixed with a proof of costly work associated with the particular request, you can acheive this in a stateless way.

Why Blake3?

  • Efficient on consumer hardware
  • No known ASIC implementations
  • Awesome team behind it
  • Inverting seems incredibly hard to me, though that hardly counts as a security review

References

Dependencies

~2MB
~46K SLoC