Zerokit development planning

[s1fr0]

This is a tentative dynamic meta-issue which collects required and desirable features for Zerokit with a reasonably clear scope.

Ideally, each item should correspond to one or more issues to be addressed according to agreed monthly development planning.

Items marked with a (*) are currently prioritized.

Implementation

General

• [x] Release v0.2 (select subset of items in this issue) (*):
  • https://github.com/vacp2p/zerokit/issues/100
• [x] Check license compatibility with dependencies and add a license (*):
  • https://github.com/vacp2p/zerokit/issues/97
• [x] Reduce static libraries and WASM sizes (includes revising dependencies):
  • https://github.com/vacp2p/zerokit/issues/37
• [x] Standardize an abstract ZK circuit interface to be used on all modules:
  • https://github.com/vacp2p/zerokit/issues/71
• [x] Address any eventual requirement coming from projects using zerokit modules (nwaku, js-waku, etc.)
• [x] Separate tests from logic (*):
  • https://github.com/vacp2p/zerokit/issues/98
• [x] Embed the circuit loading logic in utils (at the moment we use ark-circom, which uses heavy dependencies)
• [x] Add Makefile for zerokit workspace and for each package (*):
  • https://github.com/vacp2p/zerokit/issues/41

rln

• [x] Code refactoring/optimization/clean-up (chores, todos, clippy, revise serialization, etc.) (*):
  • https://github.com/vacp2p/zerokit/issues/96
  • https://github.com/vacp2p/zerokit/issues/39
• [x] Complete documentation of the modules (ffi, protocol, utils, etc.) (*) :
  • https://github.com/vacp2p/zerokit/issues/69
• [x] Audit FFI (in particular usage of not_unsafe_ptr_arg_deref) (*):
  • https://github.com/vacp2p/zerokit/issues/96 (item)
• [x] Push RLN crate to crates.io (requires license) (*):
  • https://github.com/vacp2p/zerokit/issues/99
• [x] MVP CLI interface:
  • https://github.com/vacp2p/zerokit/issues/70
• [x] ~Add support to nrln, i.e. use degree-n polynomials (requires analysis and agreement on coefficients generation)~
  • not planned since nrln has been deprecated in favor of rln-v2

rln-wasm

• [ ] Generate proofs using multi-threading:
  • https://github.com/vacp2p/zerokit/issues/55
• [x] Explore/implement WASM-friendly DB solutions (to be implemented in pmtree) for tree persistence using browsers
• [x] Ensure all rln module public APIs are exposed by rln-wasm (*) :
  • https://github.com/vacp2p/zerokit/issues/64

semaphore

• [x] Refactor semaphore module (i.e. make it stand-alone similarly as RLN):
  • https://github.com/vacp2p/zerokit/issues/78

utils

Merkle tree (pmtree - in collaboration with PSE)

• [x] Align pmtree APIs with OptimalMerkleTree APIs, i.e. make the two interchangeable
• [x] Properly benchmark pmtree performances:
  • https://github.com/vacp2p/zerokit/issues/89
• [x] Add support in pmtree to database DB configurations in order to keep traits and APIs abstract
  • https://github.com/vacp2p/zerokit/pull/86#discussion_r1044333626

Poseidon Hash

• [ ] Generate Poseidon round parameters and constants:
  • https://github.com/vacp2p/zerokit/issues/50 Only round parameters (easy) and secure MDS matrices verification algorithms (hard) missing

Design/analysis

rln

• [x] RLN V2 (in collaboration with PSE):
  • Optimize circuit evaluation costs, i.e. compute outside the circuit external_nullifier, etc.
  • Implement/audit a single general degree-n zk circuit with n compile-time parameter (falls back to rln when n=1)
• [x] Allow higher messaging rate (i.e. >1 message/epoch):
  • Compare computational/communication/zk-evaluation costs of using n degree-1 RLN credentials vs. 1 degree-n RLN credential

utils

Merkle tree (pmtree)

• [x] #151

[oskarth]

Should we cut up a Zerokit v0.2 release issue and aim to get it out in a few weeks?

Also generic Makefile probably useful https://github.com/vacp2p/zerokit/issues/41