Eth2 Implementers’ Call #60 - 2021-03-25

[Quick contemporaneous notes by Ben Edgington]

Agenda: https://github.com/ethereum/eth2.0-pm/issues/210
Livestream: https://youtu.be/Q0EbnFViJFk

Client updates

Teku

Mostly working on Altair. Versioned datastructures mostly done. Spec functions in progress. Started to work on API changes. Updating to latest reference tests and fork choice tests. Prater is supported; working on optimisations to run more efficiently.

Nimbus

Released v1.0.12. Only difference is Prater support. Preparing for a major release with several performance improvements, peer management fixes, and slashing DB refactor (will be enabled by default). Phasing out -Dinsecure flag: will now use secure API by default.

Started work on Altair. Completeing API to be in sync with Eth2 API spec.

Prysm

v1.3.4 released. Added embedded Genesis state to reduce sync time. Added Prater support. New slasher design. Eth2 API work. For Altair, working on supporting multiple specs. Looking at how to integrate with Catalyst for the Merge.

Lighthouse

Got a Merge chain running with Lighthouse and Catalyst. New release with improvements to block production times; should reduce late blocks on mainnet. Nodes on Prater look to be doing well. Progress on fork handling logic; not running test vectors yet, but will work on that. Altair network changes.

Lodestar

New release v0.18 with Prater support. Found a host of issues to tackle on Prater to be more efficient. Working on support for Altair. Profiling and memory leak fixes. Found and fixed memory leak issue in remerklable library.

Altair

alpha.2

Small spec release this week to clean up some naming and fix test generators.

base reward fix

Barnabé has done some analysis and found a minor error in base rewards to be fixed.

Fork choice test fix to come - will be an alpha.3 coming in a week or so.

Prater status

Launch went well: only minor issues around Genesis. Now at ~98% participation.

[Mamy] Prater is initially intended mainly for dev teams and a few others for now. Should we more clearly advise normal users to continue with Pyrmont? [Danny] Probably we should encourage users to migrate over to Prater over the next month or so. Maintaining two testnets is overhead. Pyrmont could be used to test a live fork, then run a non-finality test on it. Action: start advising people to move over to Prater over the coming weeks.

Note that Pyrmont sometimes doesn't finalise. This is likely due to client teams experimenting and upgrading. Client teams and the EF still make up 80% of the validators on Pyrmont.

Research Updates

There will be a call next week to discuss the Merge. This will take place fortnightly, the day before the Ethereum all core devs call (which will be discussing Merge planning).

Mikhail

New PR replaces executable beacon chain PR. It only swaps the consensus from PoW to PoS. Main difference: no RLP string beacon block, but some structured data instead.

Vitalik

Proof of Custody construction. Looking at how easy/hard it is to MPC. Needs a few more days work.

Another proposal to limit the total number of active validators. It's a simplification.

Aditya

Fork choice changes. There is now a better proposal to mitigate the attacks identified last year (balancing attack). This is in the outer fork choice only, so it does not require a hard fork and can be done any time, even incrementally.

EthGlobal Hackathon.

Proto would like an engineer from each team to participate in getting a Merge testnet running, with stretch goals around sharding and roll-ups. Action: nominate a team member

This is mid-April to mid-May.

---

Chat highlights

From danny to Everyone: 02:02 PM

https://github.com/ethereum/eth2.0-pm/issues/210

From Mikhail Kalinin to Everyone: 02:22 PM

https://github.com/ethereum/eth2.0-specs/pull/2257

From Vitalik to Everyone: 02:25 PM

https://notes.ethereum.org/@vbuterin/validator_rotation_proposal