Canto Core Devs Meeting 11 Agenda

[bebekim]

This week, continuing on B-harvest's Polygon CDK improvement (link)

Design Choice

Replacing CDK Sequencer with CometBFT vs. Complementing Sequencer with CometBFT

• minimise changes to Sequencer and yet, achieve decentralised block creation
• L2 consensus verification by L1 zkEVM proof

          (ABCI)                             (SBCI)
CometBFT <-----> Cosmos-SDK  :::  CometBFT <--------> Sequencer | zkEVM

Components

• Pool & Broadcast: CDK's Sequencer lacks broadcasting so CometBFT will check Tx's, broadcast them (all nodes have Tx's now), and chuck them into Sequencer pool.
• PrepareProposal: CometBFT will fetch Tx's from Pool (CDK Sequencer) that are best fit for proposal to form a batch
  • ExtendedCommit: to be included as part of PrepareProposal containing sign of checkpoints on previous block
• ExtendVote: replacing verifyContract (Sequencer) by aggregating signature on checkpoint
• Sequencer's DAC (Data Availability Committee): no longer needed. SequenceSender used to interact with DAC to aggregate sign but ExtendVote of CometBFT replaces this functionality
• FinalizeBlock

Follow-up

• Proposal will be split in two: MVP and Performance improvements
  • MVP: decentralized consensus, zk-proved consensus
  • Performance Improvements: mem-IAVL, parallel signature verification, optimistic block processing
• Planning to share progress with Polygon in very near future.