PeerDAS Breakout Call #24, Apr 8, 2025

[will-corcoran]

Apr 8 (Tues), 2025, 10:00 UTC
Duration: 60 minutes
Zoom link
Ethereum Protocol Call calendar
PeerDAS-PM.md
Google Doc Notes
Previous call
Facilitator email: will@ethreum.org
Facilitator Telegram: @corcoranwill

Agenda Overview

A. Client Updates
B. Devnet / Testing Updates
C. Spec / EIP Discussion
D. Open Discussion

Agenda Details

Client Updates

PRIORITY: All teams to focus on implementing getBlobsV2 and running across clients

CL Client Teams

Team	Link	Previous Update	Timely Discussion Topics
Lighthouse	link	- Implemented sync fixed for peerdas-devnet-5 - Focused on Validator Custody	-
Prysm	link	- DB resdesign to fix inode issue on devnet-5 - Worked with Francis on EL getBlobsV2 and getPayloadV5 integrations	-
Teku	link	- Tested Validator Custody internally - Worked with Katya on new PeerDAS metrics	-
Nimbus	link	- Paused validator custody and prioritizing work on getBlobsV2 - Tested new column syncer, found a runtime crash in a given scenario, working on that, otherwise CL syncs reliably upto head and backfills correctly in devnet 5, and is able to serve range requests	-
Lodestar	link	- Refactoring, onboarding a new contributor, and reviewing possible peer mgmt improvements	- see discussion item 01 below (nflaig)
Grandine	link	- Started implementing getBlobsV2	- see discussion item 02 below (hangleang)

EL Client Teams

Team	Link	Pervious Update	Timely Discussion Topics
Geth	link	- Open PR (add ComputeCells functionality)	-
Nethermind	link	- PoC branch w/ EL methods implemented - Tx propagation working - Working on multi-client testing environment	-
Besu	link	- No update provided	-
Erigon	link	- No update provided	-
Reth	link	- Types implemented - Awaiting finalized spec, implementation ~1-2 days	- finished impl on Fri (paradigmxyz/reth#15534) - fixed blob schedule issues yesterday (paradigmxyz/reth@b54b5a3) - fixed a bug handling txs with >9 blobs today (paradigmxyz/reth@1a4d585) - very much in favor of consolidating the breakout time to a single time slot

Wins

• Positive feedback on BPO / blob increasing schedule
• EIP-7594: PR's 9378 & 9587 merged last week.
• Consensus Spec: v1.5.0-beta.4 pre-released last week.
• Beacon API Spec: Add EIP-7594 merged.

Devnet / Testing Updates

• peerdas-devnet-6 (specs)
• Kurtosis Interop Update
• See Discussion Item 02 below

Devnet-6 Open Questions

• Do we need a peerdas-devnet-7 for BPO Forks and/or a pre-determined blob schedule?

Sunnyside Updates

• voiceover

Probelab Updates

• Finalizing getBlobsV1 testing on mainnet.

Spec / EIP Discussions

P0 : Open Spec PRs

• cell proof computation | Add EIP-7594 (PeerDAS) related changes (builder-specs) (engine-api)
  NOTE: beacon-metrics Derek G "I'd be interested to see how frequently getBlobsV2 retrieves all blobs requested."

beacon-api

• beacon-api | produceBlock and publishBlock: Remove blobs and KZG proofs (link)
• See Discussion Item 01 below

P1 : Draft Spec PRs

• EIP-7594: PR's 9378 & 9587 merged last week. New PR, 9588 minor edits opened (link)

Spec Release Updates:

• Consensus Specs: v1.5.0-beta.4 has been pre-released.

Open Discussion

Discussion Item 01: Validator Custody API Registration

• Consolidating prepareBeaconProposer and registerValidator APIs for validator custody (reference issue)
• Current implementations (e.g., Prysm, Lodestar) repurpose prepareBeaconProposer with a 1-hour eviction, but concerns exist about separation of concerns and optimal eviction timing (1 hour vs. 2 epochs).
• Need to discuss: standardizing eviction in the spec, potential API changes, and cache consolidation across clients.

Discussion Item 02: engine_getBlobsV2 Integration Issue with Geth-Grandine Pair

• Error encoutered while trying to test engine_getBlobsV2 integration on Grandine w/ multiple EL pairs(Geth, Nethermind, Reth)
• Tests revealed Geth-Grandine fails to include blobs in proposed blocks, unlike other pairs.
• Geth logs show eth_sendRawTransaction errors due to lack of support for blobv2 tx via RPC yet...
• Currently, cell proofs are not stored in the blobpool (but, rather recomputed whenever txs are pulled from the blobpool, which takes longer than 2s)
• The fix (in progress) is to store them properly and not recompute them.

Discussion Item 03: Meeting Time

• Proposal to revert back to 2pm UTC calls every week.
• Flagging conflicting time with FOCIL and History Expiry breakouts

[github-actions]

Note: Meeting already exists for this call series. Skipping event creation.

⚠️ Discourse Topic Error

• Failed to create/update Discourse topic: 422 Client Error: Unprocessable Entity for url: https://ethereum-magicians.org/posts.json

[rkrasiuk]

unfortunately, got a conflict and can't make it in person. update from reth:

• finished impl on Fri (feat: peerdas support paradigmxyz/reth#15534)
• fixed blob schedule issues yesterday (paradigmxyz/reth@b54b5a3)
• fixed a bug handling txs with >9 blobs today (paradigmxyz/reth@1a4d585)
• very much in favor of consolidating the breakout time to a single time slot

[barnabasbusa]

Could EL teams please comment whether they want to support v0 wrapper post osaka activation, or we can assume to only send v1 wrapper post fulu?

[marcindsobczak]

In Nethermind, we have a flag to support the v0 wrapper after Fusaka. It's optional for the user - if TxPool.ProofsTranslationEnabled is set to true, we'll accept old proofs and translate them to the new format. If it's set to false, we won't. The current plan is to have this flag as false by default, but we're open to discussion.