Consensus-layer Call 93

[djrtwo]

Consensus-layer Call 93 Agenda

prev: call 92

Meeting Date/Time: Thursday 2022/8/11 at 14:00 UTC
Meeting Duration 1.5 hours
livestream

1. Merge
   • Goerli merge
   • tbh update
   • spec-ing placeholder ttd in engine API
   • Mainnet Bellatrix epoch and TTD
2. mev-boost
   • clarification of builder specs
   • circuit breaker design specify liveness failsafe for builder network builder-specs#47
3. Research, spec, etc
   • points of discussion ([1] and [2]) for 4844
4. Open Discussion/Closing Remarks

[timbeiko]

There were two things that came up in #581 where CL teams' feedback is necessary to unblock things on 4844:

1. Whether the Blob gasprice update rule in the EIP should optimize for stable throughput or "bursts" of blobs (context)
2. Whether blob syncing should be tightly coupled to block syncing. Long term, the desire is to decouple these, but this adds implementation complexity and, given the low number of blobs expected with EIP-4844, it might be best to sync blobs along with their associated blocks.

If time permits, it'd be great to discuss both these things.

[mkalinin]

I'd like to touch base on the status of TBH override implementation with EL client teams

[taxmeifyoucan]

I would like to present estimations of potential TTD values for the mainnet Merge.
Summary in this document: https://notes.ethereum.org/@MarioHavel/merge-ttd

[ajsutton]

And alongside the potential TTD values we'd need to pick a Bellatrix fork epoch. Some options: https://hackmd.io/Xq0HavcIRvufUVvA2ZjoUA

[ralexstokes]

I have two things to discuss w/ associated PRs:

1. clarification to builder specs around orchestrating remote vs local building: clarify local building should happen in parallel builder-specs#46
2. concrete spec for the circuit breaker idea we discussed on the last call: specify liveness failsafe for builder network builder-specs#47

[ralexstokes]

if there is spare time, I would also like to have a discussion about the validator registration part of the builder-specs but this can go at the very end as it is not as timely as everything else

[terencechain]

2. Whether blob syncing should be tightly coupled to block syncing. Long term, the desire is to decouple these, but this adds implementation complexity and, given the low number of blobs expected with EIP-4844, it might be best to sync blobs along with their associated blocks.

I have written down some thoughts on the loosely coupled block/sidecar approach and what the dependencies are like from the client's perspective. We don't have to go through this, I'll link it here for reference: https://hackmd.io/_3lpo0FzRNa1l7XB0ELH7Q?view

[protolambda]

Can't make it to the call today. But regarding EIP4844:

• Tracking excess blobs, instead of total blobs, sounds great, we can merge the current EIP update PR when it includes that.
• Let's ignore the burst vs smaller incremental blob inclusion thing for now, there are more important spec and implementation issues currently.
• Coupling blobs sidecar and beacon block in gossip propagation and/or sync would make things easier, but due to current implementation doing the uncoupled approach, and future compatibility of the uncoupled approach, I'm hesitant to change it. Do the client teams prefer the simplification of coupling in this context?
• I'm curious if there's any progress with the BLST extension to include the building blocks desired for optimized KZG related functionality.
• EIP4844 testnet: first experimental version is running I think, but is missing the new fees approach, and likely needs to be restarted to include more test eth and add a faucet.

[asn-d6]

I'm curious if there's any progress with the BLST extension to include the building blocks desired for optimized KZG related functionality.

We did an initial call with Supranational about this last week. We will get an updated list of delivereables and time estimations either this week or the next one.

[paulhauner]

I would like to raise this: ethereum/execution-apis#285