TL;DR

The Ethereum Virtual Machine (EVM) was not designed to run in a zk-circuit, and to achieve this, three different strategies have been presented so far:

• Virtual machine optimised for zero-knowledge, which equates to starting a new blockchain from scratch.
• EVM-compatible, which also starts from scratch and adapts the current tooling to the new virtual machine.
• Hermez zkEVM: Implementing OPCODE full compatibility.

Full compatibility with Opcodes has inherent advantages and technical features:

• uVM: A zkCircuit optimised VM with tailor made opcodes to optimise the EVM interpretation.
• Easy to implement variable EVM opcodes: CALL, DATACOPY, EXP, CREATE, etc.
• Multi StateMachine architecture to implement difficult opcodes like MULMOD, EXP or even pairings.
• Fixed uVM assembly code to process all transactions of a block fetching and interpreting all the opcodes.

Inputs: OldState & List of TXs to process

Output: New State

The Hermez team believes that the OPCODE-based approach is the needed solution. Efficiency should be similar to a compiler-based approach, and it will inherit Ethereum’s security model.

Even more importantly, it will be fully compatible with all infrastructure, as specifications of the EVM are very well defined.

The Road to Mainnet

The Hermez zkEVM is perhaps the most ambitious development that the team is working on and, as Jordi said during the presentation, we don’t want to be constrained by a hard roadmap.

Here’s an estimated timeline of the zkEVM developmentAgain, this should be taken as an estimation, not a commitment, as this is a whole new technology and unexpected delays or challenges might arise.

If you want to know more about the current status of Hermez development, you can watch the interview that Bankless did with Jordi at EthCC.