--- tags: eth2devs description: Notes from the regular Eth2 implementers call image: https://benjaminion.xyz/f/favicon-96x96.png --- # Eth2 Implementers’ Call #61 - 2021-04-08 [Quick contemporaneous notes by Ben Edgington] Agenda: https://github.com/ethereum/eth2.0-pm/issues/212 Livestream: TBD - there were some technical difficulties. Recording will be released in due course. Reminder that there are other regular calls to cover Rayonism and the Merge. ## Client updates **Prysm** Altair progress. A bit slow, but making progress. Implemented Altair beaconstate with additional fields. Almost done on sync committee processing. Next is accounting reform. Merge: released demo of Prysm with Catalyst. Did optimisation of slasher DB schema. Weak subjectivity sync UX improvements. Some bug fixes. **Teku** The last little while has been pretty much all about Altair support. Progress is good. We've updated to the alpha.3 reference tests and almost everything is passing. We've also made some optimisations for the workload on the Prater network. This is mostly improvements to peer scoring, updating to the latest Blst library, and doing more batch signature verification of attestations. **Lodestar** New release, 0.19. Big update. Latest LTS Node supported; threaded BLS verification. Much more stable. Done preliminary work for Altair. Naive state transition is working; need to optimise to match current fast state transition. **Lighthouse** Passing Altair alpha.3 consensus tests. Working on Altair networking. Doppelganger protection is in review. Working on memory allocator - ~3-6x memory savings. UI development: should start coding it by end of month. **Nimbus** Altair: merged preparatory changes for Altair, but still evaluating making codebase modular, such as beaconstate. Performance improvements, such as pruning, attestation batching. Added Prater page to Nimbus book. Merged fallback Eth1 provider feature. Finished HTTP server work, which allows secure access to metrics and eliminates JSON-RPC. Moving validator fleet off AWS to save costs. ## Altair ### Alpha.3 Danny is embarrassed that he didn't number alphas from zero! ### Planning Planning target is June/July for the upgrade (not a deadline!). Nimbus requests 2 months lead time for audit - recommend booking asap as auditors are incredibly busy right now. ## Standardisation of metrics [Leo] Proposal to support a set of commonly named metrics. See [this list](https://docs.google.com/spreadsheets/d/1OrToYWl-XeIfTItBM6iqEsWjynmXu0ctyOj-83qzBVM/edit?usp=sharing). Once these are in place and proven, build out from there. Request: one dev from each team to review and work on this. Suggest a meeting to discuss in the next few weeks. **Action: nominate a lead from each team.** Please add the team lead as a comment in the corresponding table (rows 44-48) in the link above. Goal is to avoid too much "client lock-in". Use cases: setting up dashboards for Prater across different clients was very laborious. Note that there was an effort to standardise some metrics a long while back. Lighthouse is keen on this: some community members have made nice dashboards that work mainly with Prysm; focusing on metrics that enable these to be ported easily would be good. ## Research Updates **Aditya** New fork choice spec is ready: [PR2292](https://github.com/ethereum/eth2.0-specs/pull/2292). Please review, comment, feeback. Summary: block tree structure and latest message accounting are changed, but no change to network structure. Has good security and good performance. This seems to be a really good fix, despite some false starts in development. Gives same result as current fork choice if network is not under attack. Some discussion about whether we will need to maintain both old and new alongside each other. General view is that it is safe to do a hard switchover; clients do not need to update simultaneously. **Action: confirm that the conplexity of implementation is not too great.** **Mikhail** The Merge. [Change of terminology](https://github.com/ethereum/eth2.0-specs/pull/2319): "Execution" replaces "Application" everywhere. Work on making the spec executable is underway. There is a [spec for Rayonism](https://notes.ethereum.org/@n0ble/rayonism-the-merge-spec) that is focused on turning the PoW clients into execution engines. Spec is almost complete. Merge implementers' call next week. Contact Mikhail with agenda items. ## Spec discussion No. ## Open discussion There will be "office hours" calls for Rayonism. Hosted on Discord. Very informal. * * * # Chat highlights From Parithosh Jayanthi to Everyone: 03:21 PM : https://docs.google.com/spreadsheets/d/1OrToYWl-XeIfTItBM6iqEsWjynmXu0ctyOj-83qzBVM/edit?usp=sharing From Mamy Ratsimbazafy to Everyone: 03:26 PM : Btw, eth2.0-API was tagged v1 last week. so it can be target as a standard for monitoring? https://github.com/ethereum/eth2.0-APIs/releases/tag/v1 From Cayman Nava to Everyone: 03:26 PM : https://github.com/ethereum/eth2.0-metrics From Mamy Ratsimbazafy to Everyone: 03:26 PM : ah right: https://github.com/ethereum/eth2.0-metrics/blob/master/metrics.md From Hsiao-Wei Wang to Everyone: 03:30 PM : https://github.com/ethereum/eth2.0-specs/pull/2292 From Mikhail Kalinin to Everyone: 03:37 PM : https://notes.ethereum.org/@n0ble/rayonism-the-merge-spec New merge terminology: https://github.com/ethereum/eth2.0-specs/pull/2319