Nimbus
Support:
- Team blog post on our meeting in Berlin: https://our.status.im/nimbus-berlin-update/
- Documentation: https://nimbus-libs.status.im/#get-started
- Revamped build system, can directly use nim-beacon-chain instead of going through Nimbus
Networking:
- Focus on interop with Rust, minimal gossipsub working
- TCP transport done
- Minimal libp2p implementation via SecIO in progress
Eth1:
- Validators/deposits from Eth1 almost done
Consensus:
- A couple of justification and finalization issues
Tests:
- so far relied on simulations but merged several unit tests and mocking routines based on pyspec to initialize arbitrary state, attestations, blocks.
- Focus: Parsing and running the new SSZ tests for interop
- CI takes a long time
Perf:
- benchmarks on crypto
- scaling of attestation processing/justification/finalization on a single node: 2k validators on a laptop.
Testing
D: v0.8.3 release + tests (~the same as 0.8.2)
Client updates
... benchmarking BLS
New DB initial sync Networking spec ... WIP: updating codebase to 0.8.2 Looking into Slashing detection and prevention algorithm
debugging Go<->Rust libp2p gossipsub cold/hot databases ergonomics of lighthouse for interop
Passes 0.8.3 specs easy to pass but hope it's the last change of test formats. libp2p integration and discovery missing at the moment Hopefully by next week. New attestation pool
Skipped 0.8.2 tests, WIP 0.8.3 tests (...)
Progress on sha256 AssemblyScript impl --> fastest JS impl for certain types of inputs Quickstart for interop BLS optimizations
Research updates
V: (...) immediate scalability similar to plasma without the complexity Might be relevant to Eth2 Exec Environment
J: Review of phase 1
- fork choice fix
- universal slashing condition (reduce complexity) Phase 1: complete polished and tested by Sept 31 (but not frozen)
- SNARK proof system: Plunk (old construction SONIC has huge performance hit) Aztec project
Nicolas: Proof for Handle protocol, paper will be updated (no change in the protocol)
Musab: (...)
Deposit contracts - Vyper merged the compiler fix
Networking
Raul (Protocol Lab)
- contributions:
- decrypting SecIO traffic
- dissectors for (...) to debug introp issues
- handshakes
- visualizations
wireshark dissectors: demo: https://drive.google.com/file/d/1B6pqFZFqC3uuScNBkpE6MEoFgXGxXg-n/view https://github.com/michaelvoronov/secio-dissector https://github.com/michaelvoronov/go-libp2p-secio
meshsim visualisation with gossipsub: demo: https://www.youtube.com/watch?v=jRBnead8Qss https://devpost.com/software/libp2p-gossip-pubsub-visualisation https://github.com/valer-cara/meshsim-node-p2pd
noise handshakes in go: https://devpost.com/software/go-libp2p-noise-the-new-security-transport-for-eth2-0 libp2p/go-libp2p-noise#2
gossipsub profiling: https://github.com/protolambda/go-libp2p-gossip-berlin/ https://devpost.com/software/go-libp2p-gossip-berlin
Paul:
- Data API: ethereum/beacon-APIs#3
Preston:
- Use cases/requirements document:
Interop
Survey: Zak Cole & Proto no issue with version 0.8.2 but issues in testing
everyone is close to a working libp2p stack probable bottleneck syncing
Starting from arbitrary state? Disk format? SSZ, Json, YAML --> would help if you have the possibility to dump SSZ state
Q Raul: Did other clients tried to connect to one another
Hackmd: for people to be able to run each other clients ahead of time