Sequencer

The sequencer has the role of posting chunks onchain and can be seem similar to the role of a sequencer for L2s.

Optimization with BLS signatures

If the consensus is offchain, then we expect the implementation of the sequencer to gather signatures offchain, and aggregate them through BLS signatures. In this scenario, the BLS signature would be gas and networking bandwidth optimization. On Ethereum and EigenLayer, this role is called as the aggregator. But on Overload, an sequencer does not need to utilize BLS signatures given that consensus can be onchain, hence we drop the aggregator term in favour of a more generalized term, being the "sequencer" (due to the similarties to an L2 sequencer).

The EVM supports BLS signatures through the alt_bn128 curve, now called bn254. This has been proven to not give 128-bits of security, although it's the only pairing-friendly elliptic curve that has a precompile available at the moment. When bls12-381 precompiles are added to the EVM, then we suggest AVSs to upgrade to that curve instead. To read more about bls12-381, we refer to EIP-2537.

Decentralizing the sequencer

The effort to decentralize the sequencer is an on-going effort in the Overload team at the moment. We do not believe this to be a feature that will be supported in the initial rollout of AVSs, although will keep researching, as this is an area of interest.

Our current thoughts is to utilize the block.prevrandao native parameter inside of the EVM. On the L2 chains, the prevrandao value is not an actual RANDAO value, instead it is a value that's set pseudorandomly, noted in Optimism documentation (this can still be useful for decentralizing the sequencer for consensus contracts in the future).

Usage of Data Availability blockchains for larger chunks

If a chunk is of larger size, we recommend AVSs to use DA blockchains such as Celestia to post data on. It is considerably cheaper for gossiping data and would only require the sequencer to post the data once, and validators would be able to fetch it through RPCs or running light nodes.