Eth2 Implementers’ Call #57 - 2021-02-11

--- tags: eth2devs description: Notes from the regular Eth2 implementers call image: https://benjaminion.xyz/f/favicon-96x96.png --- # Eth2 Implementers’ Call #57 - 2021-02-11 [Quick contemporaneous notes by Ben Edgington] Agenda: https://github.com/ethereum/eth2.0-pm/issues/203 Livestream: https://youtu.be/z3Gj6TXgcb0 ## Client updates **Teku** - Implemented UPnP for router port configuration. - Gossip scoring is now in, although not enabled by default yet. We've followed the same approach as Lighthouse in implementing this. - Preparing for the upgrade involves a lot of refactoring work. Mostly prototyping good ways to support versioned datastracutures at the moment. - Huge refactor of our SSZ library to bring it into the 21st century and make it more maintainable for the future - We've updated the weak subjectivity calculations as per the latest version in the specs repo. - A lot of work was done to get building the Blst library integrated with our CI process. - Adrian wrote up [an analysis](https://hackmd.io/@ajsutton/balancing_attack) of the work involved in the proposed LMD GHOST balancing attack mitigation - I've made a start on a KZG commitments library in C that we can use for sharding protoyping work until something better comes along. **Lodestar** Refactoring sync and peer management. Preparing for hardfork and datastructure versioning. Multiproof creation and consumption and linking it to the SSZ library. **Prysm** v1.2.0 is out. Has EIP-3076 slashing protection implemented. Continuing to work on beacon API (70% done). Working on caching next slot state. **Lighthouse** Optimisations for block propagation. Found a way to reduce memory usage by around 2/3. Looking at attacks on p2p networks. Hiring! Minor PR merged to networking spec ([#2196 on specs repo](https://github.com/ethereum/eth2.0-specs/pull/2196)). Bootnode IP addresses changed. **Nimbus** Release v1.0.7. Significant performance improvements. Includes "doppelganger" protection - detection of 2 validators running in different places. Dashboard challenge being extended - good interest, but no submissions yet. ConsenSys Dili found a vulnerability in Nim package manager. This is not used in Nimbus. It was fixed yesterday in any case. A user has integrated Nimbus into Rocket Pool. v1.0.8 will have IO optmisation, and EIP-3076 slashing protection. Adding new benchmarking tooling. ## Upgrade 1 https://notes.ethereum.org/@vbuterin/HF1_proposal There is an issue to discuss the naming of Upgrades... Sync committee and accounting stuff is merged. Working o the updates to constants right now. Main decision to make is around the bonus feature, accounting reform (tier 2): per validator inactivity leak accounting, and reduces the overhead of empty epochs by a factor of 64. **Action: provide feedback** [This PR](https://github.com/ethereum/eth2.0-specs/pull/2192), I think. Will merge next week unless there are objections. Planning pre-release mid next week with test vectors. [Fork-choice block--slot pair](https://github.com/ethereum/eth2.0-specs/pull/2197) PR is up. Expect some iteration. Take a look and give feedback. Balancing attack: see [Adrian's analysis](https://hackmd.io/@ajsutton/balancing_attack). Will be incorporating tests for the fork/upgrade process. Hsiao-Wei now owns getting fork choice tests integrated into the repo. Proto is working on extending Hive test framework for Eth2 clients. Will also be useful for testing the merge. Summary: - Expect all the non-bonus changes to go in. - Plus maybe the first of the "bonus" changes, but not the second. ## Merge discussion Mikhail has a [draft spec](https://hackmd.io/QFm6Ih_-Si6_kSLCWTZQyw) for this and is tidying up. Will publish more widely within 1-2 weeks. Conversations are beginning between Eth2 and Eth1 devs. It is on the agenda for next week's All Core Devs call. Then maybe a monthly call after that. ## Sharding discussion [Terence] Is single secret leader election on the sharding roadmap? [V] Yes, but want to focus on scalability features before security features. SSLE is not a hard requirement for sharding to be implemented. [Justin] SSLE R&D is very advanced. Mary M has some formal proofs, and there is a PoC in Rust. Performance is great. Uses a custom zero-knowledge proof, specialised for the task of proving things about the shuffling. But it makes sense to ship MVP sharding first, with SSLE, Proof of Custody, Data availability sampling, VDFs later. Security is already good enough in general. [Lightclient] What’s the status of the validity proof for DAS? [V] Protocol is basically finished. ## Research updates Stateless Ethereum calls have picked up again. Ping Danny to get on the email list. ## Spec discussion Nothing further. ## Open discussion Nope. * * * # Chat highlights From danny to Everyone: 02:17 PM : https://github.com/ethereum/eth2.0-pm/issues/203 https://notes.ethereum.org/@vbuterin/HF1_proposal From Me to Everyone: 02:17 PM : https://hackmd.io/@ajsutton/balancing_attack From Hsiao-Wei Wang to Everyone: 02:18 PM : Eth2 protocol upgrade naming theme brainstorming thread: https://github.com/ethereum/eth2.0-pm/issues/202 From Aditya Asgaonkar to Everyone: 02:19 PM : (block, slot) fork choice PR: https://github.com/ethereum/eth2.0-specs/pull/2197 From Mikhail Kalinin to Everyone: 02:24 PM : https://hackmd.io/QFm6Ih_-Si6_kSLCWTZQyw From lightclient to Everyone: 02:29 PM : what’s the status of the validity proof for DAS? is it fully specified or is there r&d still taking place on the topic? From lightclient to Everyone: 02:34 PM : thanks for putting on the workshop last week, it was very helpful!