—————— Nimbus
Core:
- Started on attestation aggregation
- benchmark tool in the zcli vein
Sysadmin:
- Managing a fleet of Nimbus nodes
- Seeing if we can have stack traces with less overhead
Libp2p/networking:
- UPnP fixes
- Integration into testnet worked and was deployed by incident (holding it back will have a mixed daemon+libp2p first)
- Blocker to join other testnet: discv5, target lighthouse
Waiting for zcli genesis state (or pyspec)
Phase 2:
- Jacques (Vyper) is looking into phase 2/WASM
Eth1:
- Started on EVMC, might require heavy refactoring
- Currently storage and sync speed are our blockers for syncing with Eth1
Couple of team members will be off
——— Testing updates (Danny)
Fixes to be released today BLS, probably first week of January Audit of phase 0 by Least Authority, v0.10 as a target
No fork choice tests yet Format proposal + some vectors: ethereum/consensus-spec-tests#17
Beacon Fuzz: fuzzers on 0.9.1 and upgrade to 0.9.3/4 Fuzzing speed of pyspecs and trinity is slow Nimbus integrated Fuzzers published on a fuzzer as a service platform Onboard a new team member from Eth1
———— Client updates
Trinity:
- working toward testnet
- naive attestation aggregation
- update to PySSZ, crazy perf increase for SSZ hashing
- separating validator client
- syncing stability
Artemis(Meredith Baxter)
- (?)
- Harmony/disc v5 integration
Prysmatic
- Mainnet config on local system
- Testnet restart after holiday break will be on main net
Lodestar
- Refactoring to ease maintenance and update JS style
- Currently didn’t validate message,
Lighthouse
- 60k validators testnets,
- Got into a loop which exhausted 2 nodes and took out 60% of the nodes
- Restarted but add the bug again
- don’t store state before block or you will have a pointer to state that doesn’t exist yet (?need blog post)
- a block point to a state, if you crash in between you have a stale state in the DB
Parity/Shasper
- (internet unstable)
Metamind C# (Sly Gryphon & Tomasz Stanczak)
- SSZ, BLS, …
- Danny: libp2p?
———
Research updates
Joseph Delong Harmony and Artemis focusing on phase 1 and phase 2 TXRX team: https://twitter.com/TXRXResearch
Vitalik Shorter Data Availability scheme Cross-shard messages: protocol should guarantee that the shard B receives and execute the message —> protocol level Danny and Will: tightly coupled, need to be demonstrated V: detailed explanation of cross-shard transaction strategies
Will Phase 2 call in mid-Jan, focus on cross-shard transaction and fee market Tooling for contract EE Goal: make them Truffle like Contract EE are slow Vitalik Q: is there someone in charge on the Eth1 to Eth2 bridge —> Michael/Mikhail (?) Joseph: Workshop at Stanford Blockchain Conference
Leo (Barcelona SuperComputing) In May: short event in Barcelona, May 15, 2020 https://ethbarcelona.github.io/
Justin: ZKP The breakthrough from last time was actually a mistake :/ Trying to build an optimal prover and an optimal verifier Currently optimal verifier but prover is not -> Polynomial commitment without FFTs -> requires a new way to design an IOP without FFT -> has an idea with optimal prover but the verifier is not optimal anymore Comment Vitalik, Danny, Dankrad on
- optimality
- making prover sublinear
- recursion
———————— Networking Next one in second week of January
Spec change proposal on Eth1: ethereum/consensus-specs#1537
Other spec items:
- None