All Core Devs - Consensus (ACDC) #155, April 17 2025

Protocol Calls & happenings
,
4

ACDC #155 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
  • Add contacts per client team for Pectra mainnet observation
  • Fusaka SFI EIPs
    • EIP-7594 (PeerDAS)
    • EIP-7892 (BPO)
  • Fusaka CFI EIPs
    • EIP-7917 (Proposer lookahead stability)
  • Other PFI Fusaks EIPs are moved to DFI status.

Summary

Pectra

  • See timelines in action items.
  • Testing update from Mario:
    • Teams writing and reviewing new tests
    • All tests running against all clients
    • Increased confidence in test coverage compared to previous weeks
  • Incidence response document needs client team contacts added via PR
    • See link in action items.

PeerDAS

  • peerdas-devnet-6 launched ~1 week ago
  • Some issues with EL processing times, investigation ongoing
  • EIP-7892 (BPO) configuration format:
    • Will use YAML “list of records” structure on CL side.
    • Configuration to be included in main config file, not separate
    • No CLI override functionality planned for now
  • Discussed plans for fusaka-devnet-0 to launch by late May
    • Follows peerdas-devnet-6 spec along with EIP-7892

Fusaka

  • Agreed to move EIP-7892 to SFI as part of the Fusaka blob scaling strategy
  • Then discussed on Fusaka CFI EIP set
    • Everyone agreed to focus primarily on PeerDAS and blob scaling on the CL side
    • Only once Fusaka devnets with the current SFI set is stable, would we consider moving any additional EIPs from CFI to SFI.
  • EIP 7917 (proposer lookahead stability):
    • Early investigation shows implementation is small to medium complexity
    • Potuz highlights this EIP does touch the beacon state, where as the current SFI set does not; given the centrality of this data structure it does imply this EIP would add significant surface area to Fusaka overall.
    • Given this, client teams agreed to DFI at first sign of complications with weak support to CFI today.
  • EIP 7688 (SSZ stable containers):
    • Decided to handle in future fork since no immediate proof impacts with the current SFI set
    • Participants reiterated the focus on blob scaling and so elected to not CFI today.

Fusaka builder API discussion

  • Current concerns with builder API scalability for high blob counts:
    • large response sizes, esp. as blob counts become higher with Fusaka in a synchronous API
    • Concern for resource constraints for smaller nodes when processing large responses
  • Potential solutions discussed:
    • Remove API response entirely (relay handles gossip)
    • Keep status quo (relay returns all data)
    • Hybrid approach (return payload, separate blob distribution)
      • and potentially leverage existing “get blobs” functionality to assist with distribution
  • Teams agreed API needs revision for Fusaka; next step is to get input from builder/relay community
1 Like