The following reflects the views of L2BEAT’s governance team, composed of @kaereste, @Sinkas, and @Manugotsuka, and it’s based on their combined research, fact-checking, and ideation.
We asked L2BEAT’s research team to review PIP-64 and the accompanying PIP-65 for the VEBloP economic model. Their findings informed the comment below.
In summary, we understand the proposed changes to be the following:
- We’ll be moving away from the current ~105 block producing validators, centralizing block proposals to a few validator-elected proposers (VEBloP).
- The existing ~105 validators will validate all blocks ‘statelessly’, which keeps their resource requirements low but requires more input data than just the transactions (state fragments called witness or trace). They also elect and churn the VEBloPs.
- Forced transactions are introduced to detect censorship and enforce censorship resistance against the new centralized block proposers. They will be forced from Ethereum L1
- To keep validator incentives, transaction fees are redirected from the centralized block producers to the ~105 validators that do stateless block validation.
Gas limits and potential transactions per second (TPS) will increase based on the above changes. But there are a few things we’re unclear on, which we’d like to see clarified.
Centralizing block proposals will lead to more potential for MEV. Even as things are right now, MEV is not solved or addressed in protocol, and this proposal doesn’t offer any anti-MEV measures either, except manually rotating out VEBloPs by validators.
Given that, validators will need to monitor for malicious block proposers and especially MEV. How will that be done in practice? Is there any tooling available for the validators to use, assuming that validators cannot run a full node?
The forced transaction mechanism is interesting and would also benefit from more details or specifications.
Lastly, stateless execution / validation is not yet in production anywhere in the Ethereum ecosystem. There’s no information available on how the witnesses would be generated, distributed and validated apart from the quoted research from outside the Polygon ecosystem.