# Polkadot crypto, consensus, and security research directions

Academia vs industry research: Advising and support of Parity teams like parachains, staking, etc. and W3F teams like spec and grants.

## In progress..

#### Redesigns of staking & slashing

(Kian)

#### Time disputes

- Avoid tragedies of the commons by adjusting slashish
- Add billing for time overruns, untangling delays from security

(Robert K)

#### Parachain features

Idea: Issues largely economic. No fast auctions!

- Parathreads
- Short-term leases
- Multi-core parablocks

(Rob)

#### XCMP

Idea: Pass hashes via relay chain, but pass message data off-chain

- On-chain auth locking (ordered vs unordered)
- Off-chain delivery via approval checkers

(?)

#### Arkworks integration

(Achim)

#### Light client proofs

Idea: Prove 2/3rds of validators signed a block.

- Sampling a Merkle tree of signatures
- Bespoke MIPP SNARK on BW6 for adding BLS12-377 public keys

#### Sassafras

Idea: Block production by cards against humanity

- Reduces fork dramatically
- Improves randomness collection and quality (somewhat)
- Anonymizes half of block producers (vs $p < 0.293$ adversary)
- Encrypt to upcoming block producers
- Avoid memepool, saving CPU and bandwidth

(With Davide at Parity)

#### Ring VRFs for Sassafras

Prove an anonymous validator/collator's evaluation of a VRF.

- Bespoke MIPP SNARK for zk choosing JubJub public key
- "Bootleproof" IPP variant (RFP)
- Improved formalization of pseudo-randomness for use in VRFs

#### Ring VRFs for people

Anonymity for identity, blacklisting, and rate limiting

- Zero-knowledge continuations. Woot!
- Optimize ring membership proof
- Optimize revocation
- IRMA style linking?

#### Mixnet incentives & theory

Goal: Reward mix nodes from a distributed system.

And fix some security proofs eventually.

(With Julian Loss and Dimitris Papachristoudis at CISPA)

#### Mixnet implementaiton

(Arkady, David Emmit, ...)

## Not yet in progress

#### Smaller Merkle proofs in Storage

- Radix 16 makes PoVs 4 times larger in dense parts of the tree
- Radix 2 plus optimizations makes sparse parts smaller too

#### zkStorage

How should we present alterative storage schemes in substrate?

- zkSNARK friendly hashes like Poseidon
- Optimization zoo for KZG commitments

#### System parathreads

Almost orthogonal feature sets:
- Validator elections / NPoS PJR tests
- Staking

(?)

#### Slashing resistant certificates

Avoid slashing for many innocent validator mistakes

- Add validator key aka meta-session key?
- Register on-chain before

#### Machine elves

Security proof via Gambler's ruin for the "roll up" that is Polkadot.

#### Elf flipping aka multiple relay chains

- Analyize what happens with weaker randomness.
- Design messaging?

#### State channels

#### Understand optimized gossip

Do we ensure our gossip assumptions in our gossip topologies and protocols, ideally including with politness.

#### Authenticate transport layer

- How should TLS-like certificates treat the chain as the root of trust?
- Are there advantages of TLS?
- Do we need Noise? Should we do nQUIC anyways?
- Is libp2p as bad as Pierre says? Is QUIC slow? WebRTC story? etc.

#### Post-quantum session keys

- Crystals-Kyber (hybrid curve25519)
- Falcon is a blockchain friendly lattice signature selected by NIST (hybrid ed25519).
- Stateful 1 layer XMSS is a VRF, but demands key rotation, no key backups or high-availability.
- Tor friendly ring VRF for sassafras

#### zkDEX

Zero-knowledge continuation applications!

(Some parachain team?)