2020-01-09 Eth2 implementers call 31

2020-01-09-Eth2-impl-call-31.txt

Agenda: https://github.com/ethereum/eth2.0-pm/issues/118

——————
Nimbus:

Holidays

Core:
  - 0.9.4 in sync
  - 0.9.3 branch kept for fuzzing purposes

Networking:
  - Verifying interop with go-libp2p (following the initial version of nim-libp2p 3 weeks ago)
  - Reworking sync strategy, inspiration from Bittorrent
  - Discovery v5, first PR, halfway done, trying to join Prysm/Lighthouse

Tooling:
  - Planning a RPC endpoint for beacon node

—————————
Testing and release updates

v0.10: Breaking changes to facilitate phase 1
Stable target for testnet and audits
Release tomorrow, maybe today
Slight modification to network spec
Extra configuration parameters (for phase 1)
Major breaking change will be BLS
Additional tests on deposits processing

Q: Testing for fork choices
—> Harmony test vectors, integration test
https://github.com/ethereum/eth2.0-spec-tests/issues/17

—————————
Nimbus

Artemis (Shahan)
Eth1Data management and polling Eth1 clients
Sync
Discovery module from Harmony to be put in Artemis
BeaconFuzz integration almost complete
Merged the HashtoG2 curve
Q: how are people syncing Eth1Data to the chain
Q: for beacon-fuzz, currently 0.9.1, what are the plans to update beacon-fuzz

Trinity (Hsiao Wei Wang)
Beacon node and validator separation

Prysm
Syncing time will be a bit slower
v0.9.3 on the spec
optimisations on handling the deposits and Eth1Data
Aggregation strategy on the testnet

Lighthouse
Testnet “released” on 22 Dec
Docker compose to spin up node and validators
Switched to BLS Herumi (not on master yet)
In the process of adding the naive aggregation strategy

Harmony
Test vectors
Dmitry + Anton on Discv5 simulations
Writeup in progress on consensus (?)
TXRX work

Lodestar
Holidays
WIP, updating to 0.9.3
Refactoring SSZ to allow remerkleable (?) (abstract over Merkle Tree)
https://github.com/protolambda/remerkleable

Nethermind (Sly Gryphon)
Validator block signing working
On 0.9.1
Most of the team on Eth1 at the moment
Will be moving some of the team to Eth2

—————————————
Research updates

Vitalik

(?)
Stateless clients

Justin
Supranational (semiconductor?)
Accelerating BLS12-381, currently 10% faster than Herumi’s
Also a formal verification expert
—> formally verifying the code will be difficult due to the custom compiler

Factor of 1000x improvement of perf via small additional card on computers

Will
Eth2 call on Tuesday 14
Ganache like tool for EE on multiple shards
Building a testing suite for phase2
Writeup today on state providers

Leo
Machine Learning EE (see Johny Rhea work)
Eth Barcelona, lining up speakers, contact Leo

—————————————
Networking updates

Allow empty and stream responses
—> will be in 0.10
Next call 2 weeks from now


——————————————
Vitalik: Accelerated Eth1-Eth2 switch, how do you feel about it?

EthDenver: planning cowering space, tell us