blake3-proof-of-work
| Crates.io | blake3-proof-of-work |
| lib.rs | blake3-proof-of-work |
| version | 0.2.1 |
| created_at | 2022-01-17 05:12:57.420724+00 |
| updated_at | 2022-01-17 07:25:03.521532+00 |
| description | A basic proof of work scheme using the blake3 hash function |
| homepage | |
| repository | https://github.com/samuelSchlesinger/proof-of-work |
| max_upload_size | |
| id | 515302 |
| size | 7,006 |
Samuel Schlesinger (SamuelSchlesinger)
documentation
README
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