Skip to content

Latest commit

 

History

History
104 lines (70 loc) · 4.24 KB

POLYGON.md

File metadata and controls

104 lines (70 loc) · 4.24 KB

Polygon zkEVM

Main documentation found at

The Polygon zkEVM is a Type 2 zkEVM.

Predecessor Polygon Hermez 1.0. deployed on March 2021 -> Proof of Donation.

Smart contracts are written using Solidity

Public testnet launched in oct 2022 -> security audit happenend and they revealed new milestones -> improved transactions proof generation down from 10 to 4 minutes first testnet version will be deprecated on january 2023

Consensus algorithm -> PoE Proof of Efficiency

Technical paper

We can separate zkEVM in 3 big parts :

  • Sequencer / Aggregator
  • Prover
  • Smart contract running PoE consensus

The former zkEVM was running a Proof of Donation consensus. Not it is using a Proof of Efficiency consensus mechanism. Like the previous PoD version it supports permissionless participation of multiple coordinators to produce batches in L2.

You can find a detailled explanation on the differences between the two mechanism here

Sequencers: receives L2 transactions and are in charge to generate batches which are then proposed to smart contract.

Aggregator: receives transaction informations from Sequencer and sends it to prover that generate small zk-proof. Smart contract then verify the proof from the prover is correct.

Synthetised zkEVM transaction flow

The EVM is functionning the given way. Users can can send transactions by the way of having a wallet connnected to zkEVM rpc network. Once these transactions are signed by customer rpc they are pushed to a mempool and then pulled by the sequencer. After this the sequencer batch the transactions to send them to the PoE smart contract. The aggregator is then playing its own role to validate and consolidate transaction batches to the smart contract. To finish the batches of transactions are then committed to L1.

Some more words on zkProofs

To prove transactions off-chain, Polygon zkEVM leverages Stark proof builder and SNARK proof builder. The zkProver is composed of state machines (Main and Secondary) and zk proof builders as mentionned above.

  • Main State Machine : it handles the execution of the EVM bytecode that are interpreted by zkASM language. It also sets up the polynomial constraints that every valid batch transactions must satisfly.
  • Secondary State Machine : is not a subcomponent per se, but a collection of various executors for individual secondary state machine.
  • STARK proof builder : used to prove that batches satisfy polynomial constraints.
  • SNARK proof builder : used to prove the correctness of STARK proofs. They are published as validity proofs.

Graphical representation of the flow can be found there

Here are the connection informations.

Backend language

Mainly javascript.

zkEVM node is backed by go

  • first commit nov 2021

zkEVM proover is mainly c++

  • first commit nov 2021

Project began around Q3/4 of 2021. Around 15 core contributors on the whole project. Althought project has 3 members

Mainnet release should come soon.

Team leader Jordi Baylina for Polygon zkEVM coinvented circom

Opcode example, in the zkASM language

opADD:
    SP - 2          :JMPN(stackUnderflow)
    SP - 1 => SP
    $ => A          :MLOAD(SP--)
    $ => C          :MLOAD(SP)

    ; Add operation with Arith
    A               :MSTORE(arithA)
    C               :MSTORE(arithB)
                    :CALL(addARITH)
    $ => E          :MLOAD(arithRes1)
    E               :MSTORE(SP++)
    1024 - SP       :JMPN(stackOverflow)
    GAS-3 => GAS    :JMPN(outOfGas)
                    :JMP(readCode)

Resources

Detailed Polygon zkEVM flow

zkEVM Flow

Proof of Efficiency research paper