Constant time hashing to the curve
- Dominant language
- Python
- Stars
- 231
- Forks
- 88
- PR merge metrics
- No merged PRs in 30d
Description
### What is wrong?
For `bls12-381` we want to map messages to points on the curve. The current implementation uses an expensive "hash and check" method via `hash_to_g2`. It is desirable to have a more performant method of mapping a given message to the curve.
### How can it be fixed
Luckily, there was recent work on a constant time hashing method w/ implementation here:
https://github.com/kwantam/bls12-381_hash
There are some BLS standardization efforts under way that will likely involve making this method the canonical "hash to the curve". If this method makes its way into the standard, then we will definitely want to include it.
The linked repo has a C implementation and a pure Python implementation. We can add both, in the appropriate places, to get the functionality we want.
Contributor guide
Research direction
Start by reading the current bls12-381 hash_to_g2 implementation and compare it with the C and pure Python implementations in kwantam/bls12-381_hash. Identify the appropriate locations for both implementations and review the relevant BLS standardization context. Done means providing a more performant constant-time message-to-curve method for bls12-381.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- c, python
- Domain
- cryptography
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100