All Core Devs - Consensus (ACDC) #154

Protocol Calls & happenings
,
1

All Core Devs - Consensus (ACDC) #154, April 3, 2025

  • Apr 3, 2025, 14:00 UTC
  • 90 minutes
  • Recurring meeting : true
  • ACDC
  • bi-weekly
  • Already on Ethereum Calendar : false
  • Need YouTube stream links : true
  • Other optional resources
  • stream

Agenda

  1. Electra
  2. PeerDAS / blob scaling
  3. Fusaka
  4. Research, spec, etc.

Facilitator email: stokes@ethereum.org

GitHub Issue

1 Like
2

Summary

Recordings

https://www.youtube.com/live/RWBhHHrZ48w?t=213s
https://x.com/i/broadcasts/1dRKZYDZemXxB

Writeups

Additional info

3

ACDC #154 summary

Action items

  • Pectra timelines
    • Mainnet date confirmed to be May 7 2025
    • Client releases by April 21 2025
    • Pectra mainnet blog post by April 23 2025
  • PeerDAS implementations should focus on the cell proof computation migration, and getBlobs support; validator custody is another outstanding feature but can be rolled out later so focus should be on the prior features
  • FYI: EIP-7892 (BPO scaling) has been CFI’d for Fusaka
  • Review the set of PFI’d Fusaka EIPs so we can make a final call on Fusaka CFI’d EIPs on ACDC #155.

Summary

  • Pectra
    • Clients discussed attestation performance on Hoodi and potential implications for mainnet.
    • Various clients are working on their own packing algorithms, with analysis and fine-tuning ongoing.
    • Touched on the deposit mechanism and the rate of ingestion from the EL to the CL following EIP-6110. Decided this is expected behavior and will follow up async to determine if anything needs to happen here.
      • Latest thinking is to watch mainnet and can adjust EL ingestion rate in a future fork if needed.
    • Taking the above into account, we turned to mainnet timing where clients agreed to the following dates:
    • Pectra timelines
      • Mainnet date confirmed to be May 7 2025
      • Client releases by April 21 2025
      • Pectra mainnet blog post by April 23 2025
  • PeerDAS
    • Mostly updates here, as clients as busy with Pectra:
    • We started the conversation around how to exactly scale blob counts in Fusaka, following some kind of BPO strategy.
      • One suggestion here: blob/acc in 2025 - HackMD
        • Double blob counts every 2 months, after a two week observation period.
        • Using configuration following EIP-7892
      • Broad agreement in this strategy and that we will want to fine tune the step sizes and cadence pending further PeerDAS R&D
  • Fusaka
    • Next, we turned to finalize a set of CFI’d EIPs for Fusaka.
    • Many client teams shared opinions async, check the agenda for links.
    • EIP-7594 (PeerDAS) is already SFI’d
    • We touched on a number of CL-focused EIPs to move from PFI to CFI.
      • Strong support for EIP-7892 (BPO scaling)
      • Had a discussion around EIP-7917 (Proposer shuffling lookahead)
        • Some interest, but hesitation given how quickly the EIP was introduced relative to Fusaka scoping
      • Anders gave an overview on EIP-7819 that would modify the blob market.
        • The scaling benefits of this EIP were highlighted, and otherwise we agreed this EIP should move to ACDE for the CFI conversation.
      • Gajinder raised EIP-7898 for CFI status which would uncouple the execution payload from the beacon block during broadcast.
        • Clients expressed concerns around the implementation complexity of this EIP especially in light of other methods that seem more promising to achieve the same benefits this EIP suggests. Consensus is that it should not be CFI’d for Fusaka.
      • EIP-7688 was also in the PFI’d list; however given the focus on PeerDAS, consensus was that we would not CFI for Fusaka.
      • We also highlighted deprecation of mplex in the p2p layer as a potential for Fusaka.
    • We agreed to CFI EIP-7892 (BPO scaling), and I had said we would CFI EIP-7917 but am going to take the next ACDC to revisit this EIP’s status given the lack of time for client teams to consider the EIP in-depth.
  • Research, specs, etc.
1 Like