Note: Big refactoring pending
The spq-sigs project is currently undergoing a bit refactoring into what has been renamed to coinZdense. Check out the technical deep-dive documents to see what's comming. The C++ part of the project is currently on hold untill we reach MVP for the Python library, after what it will be evaluated if the code in this repo is going to get reused, or if a full rewrite is in order.
spq-sigs
C++ Simple (hash-based) Post-Quantum Signing
This is to be a header-only template library running on top of the BLAKE2b function of the sodium library. The design is roughly based on that of XMSS and LMS. The implementation favors simplicity and leveraging of the BLAKE2b hashing primatives over maximum signature size efficiency.
This C++ library started out as a port of the Python PySpqSigs library. It is now being refactored and extended after attempts to make this a funded HIVE project ended up failing. Once multi-tree and wallets have been implemented in this C++ library, the refactored and extended C++ implementation will get backported to Python. An important part of the Python backport will be moving from hashlib to libsodium.
The most up-to-date design info for spq-sigs can currently be found in this blog post
Done
- A signing key has been implemented.
- A signature validator has been implemented.
- The C++ signing key has been tested to work with the validator.
- Basic compiler flag based code quality has been enforced.
- Added some long overdue comments to the code.
- Implemented multi-tree signing (experimental) up to four levels.
- Replace seperate multy-tree templates with a variadic template solution.
- Implement (variadic) multi-tree signature validation.
- Updatate API to work with optionally omitted sub-key signing signatures.
- Add multi-tree signature serialization/deserialisation where known sub-key signing signatures can optionally be ommitted.
- Use crypto_kdf_derive_from_key at multiple layers
Todo for Minimal Viable Product
- Fix intermediate-layer out-of-range bug
- Signature serialization & deserialization.
- Private key serialization and de-serialization (for wallets and persistence).
- Private key password protection (use libsodium).
- Code cleanup.
Todo post-MVP
- Add multi-threading.
- Factor in something close to just in time pre-calculation of replacement keys.
- Work on const-correctness.
- Document usage.
- Add a sample project with cmake and stuff.
- Profile and performance improve where possible.
todo post Python backport
- Test signing/validation cross language compatability and fix if needed.
- Test serialization/deserialization python/c++ cross-compatibility.
other spq-sigs repos
Currently the C++ repo is where all development happens. Here are some links to (mostly empty) repos for languages spq-sigs will eventually (likely) be ported to.
- pysqpsigs (Python) Please note, the current code is from a proof-of concept. The C++ implementation design will soon get backported to Python.
- js-spq-sigs (JavaScript) Nothing there yet. This one is an esential one to get done after C++ and Python.
- m-spq-sigs (Monte) This is one that might or might not come to be. Need to look into this one deeper before we know if it will even become possible, but as a cool ocap language, I think it is worth a try and a relatively high place on the priorities list.
- r-spq-sigs (Rust) Nothing there yet. Rust is an amazing language, haven't done much serious projects with it yet, on my list for that reason.
- c-spq-sigs (Clojure) Nothing there yet. This one is on the list for one reason only and it may disapear for the same reason. At my day job we use FlureeDB a lot, and I would really like to help make FlureeDB support Simple Post-Quantum Signatures.
- e-spq-sigs (Elixir) Nothing there yet. Only here because Elixir is a cool language with libsodium bindings that I feel I should learn.