Agenda
- FOCIL has SFI’d
- FOCIL spec updates
- Development updates
- and more
Meeting Time: Tuesday, February 24, 2026 at 14:00 UTC (60 minutes)
Meeting Summary:
The team discussed updates to the FOL Breakout 29 meeting, focusing on changes to the FOCIL spec for Agoda and its implementation in Glowas, along with integration plans for FOCIL into Besu and ePBS implementation on Lodestar. The team addressed rebasing the FOCIL branch against the ePBS devnet and discussed the need for code cleanup and testing. Pelle reported on transaction inclusion rates and data processing issues, with suggestions for improving data collection methods.
Click to expand detailed summary
Jihoon led a discussion on FOL Breakout 29, focusing on updates to the FOCIL spec for Agoda and its implementation in Glowas. He explained a key change in the spec regarding payload reorganization, where a payload for slot n-1 will be reorganized if it is not IR compliant and if a proposer doesn’t extend it. Jihoon noted that he is still discussing with spec maintainers about rebasing and shared a link to the relevant spec. No questions or development updates were raised during the meeting.
The team discussed the integration of FOCIL into Besu, which will be implemented on top of the Amsterdam branch. Jihoon suggested creating a spec test to provide test factors for clients. The team also briefly touched on the ePBS implementation on Lodestar, but no concrete decisions were made regarding this topic.
The team discussed rebasing the FOCIL branch, which is currently based on the unstable branch, against the ePBS devnet. nflaig suggested cleaning up the code and rebasing in about two weeks, potentially before the interop. Jihoon noted that the field devnet or devnet 0 scope should cover everything for Fossil, but PTC implementation might be delayed to devnet 1. The conversation ended with plans to reconvene in two weeks.
Pelle reported that 84-85% of transactions in the inclusion list are already included in the next block, with further analysis showing inclusion rates of about 95% with block-native data. He noted some issues with the configuration and data processing, which has taken several days due to the large volume of data. Jihoon suggested updating the script to get more precise data, and the conversation ended with plans to continue the discussion.
Next Steps:
- Fabio/Besu team: Start integrating FOCIL into Besu this week on top of Amsterdam
- Jihoon: Work on spec test to provide test factors for clients
- nflaig/Lodestar team: Clean up and merge ePBS implementation into unstable branch, then rebase FOCIL branch against it (in approximately two weeks)
- nflaig/Lodestar team: Have FOCIL implementation ready before the interop
- Pelle: Update script to get more precise data and complete processing of block-native data
Recording Access:
- Join Recording Session
- Download Transcript (Passcode:
k=?Y1AQt) - Download Chat (Passcode:
k=?Y1AQt) - Download Audio (Passcode:
k=?Y1AQt)
YouTube recording available: https://youtu.be/7eHb1g1B1OU
FOCIL has been SFI’d for Hegotá.
FOCIL for Hegotá
-
FOCIL has been SFI’d for Hegotá. The CL spec has been rebased onto Heze. execution-specs and execution-apis are still pending rebase, but no meaningful changes are expected.
-
The main spec change is a modification to
get_head(). When an attester for slot N chooses a head by callingget_head()upon receiving a beacon block for slot N or near the attestation deadline, they choose not to extend a payload for slot N-1 if it failed to satisfy the IL constraints.
Development Updates
-
Besu is starting FOCIL implementation on top of Amsterdam.
-
Lodestar has implemented ePBS devnet-0 but needs to have it stabilized before rebasing FOCIL onto it. Lodestar aims to have a FOCIL implementation before the interop at the end of April.
IL Analysis
-
Pelle shared early results from the IL analysis using Block Native data. With the priority fee IL building algorithm, 84% of IL transactions are already included in the next block at zero-slot delay. Moving to a two-slot delay brings this to about 85%, so the difference is minimal.
-
The inclusion rates with Block Native data are about 95%, which is much higher than the previous Xatu data. Pelle is still processing data due to volume and configuration issues.
Links
1 Like
1 Like