--- tags: eth2devs description: Notes from the regular Eth2 implementers call image: https://benjaminion.xyz/f/favicon-96x96.png --- # Eth2 Implementers’ Call #56 - 2021-01-28 [Quick contemporaneous notes by Ben Edgington] Agenda: https://github.com/ethereum/eth2.0-pm/issues/200 Livestream: https://youtu.be/_WK3k_k-4w8 ## Intro Research workshop coming up next week. Likely to be streamed; will definitely be posted on YouTube. ## Client updates **Prysm** Implementing EIP-3076 slashing interchange format, and Eth2 API conformance. Packing attestations into blocks performance improvements. Design document created for upcoming upgrade. Revamping slasher and slashing protection - database improvements. **Lighthouse** Released v1.1.0. Has more sophisticated validator monitoring. Analysing attestation performance. Late propagation of blocks is causing issues with attestations: optimising propagation of blocks. Adding caches for blocks received early. Looking into the Eth1/2 merge API; plan to incorporate it into LH. Plan to work on attestation performance until it is as perfect as possible. **Lodestar** New sync implementation, should be a lot faster. Keeping 2 week release cycle. Running a validator on a testnet, generally OK. Light client prototyping. Modifying DB to keep blocks from different forks. **Teku** Release v21.1.1 yesterday. Streaming API improvements. UPnP support almost ready. Refactoring to be able to support upcoming forks/upgrades. Enabling gossip scoring. Continuing improvement of sync. Updating to latest Blst library. **Nimbus** Integrated batch verification of BLS signatures for better performance. More multi-threading for signature verification. Resolved performance issue processing blocks with deposits. Solved Eth1 monitoring issue, which could cause an invalid state. Experimenting with peer management in lipp2p - ongoing. ## Upgrade 1 See this [document](https://notes.ethereum.org/@vbuterin/HF1_proposal). Request for engineering input on the second fork-choice fix in particular. Looking to scope whether to include some of the nice-to-haves in the doc in the upgrade. **Action: move the doc structure into a checklist on the repo for async discussion** [Vitalik] Any feedback on doing this exercise around the suggested time? Mid-year for upgrade; testnet for wither sharding or merge by the end of the year. [Jacek] When would we need to put up a new testnet for the upgrade version? [Danny] Min one month for public testnet; smaller networks a month before that. [V] Want to agree the contents of the upgrade asap. [Terence] When do we need to decide about the "Bonus Section"? [V] Decide on main stuff within 2 weeks; Bonus things could be a little later (while teams work on the main things). On the bonus section, Vitalik came up with a different approach recently and made a PR. The other sections have been unchanged for a few weeks. [V] Reiterate: come prepared to make decisions about the HF1 document in two weeks. Discuss async meanwhile. ## Research Updates Workshop is happening on Tuesday Feb 2. Dankrad has been doing the primary work on an implementation of Verkle trees - more relevant to Eth1 than 2. [Dankrad] Migration to Verkle tree rather than binary tree would be good, or for storing state. Uses Kate commitments, which is relevant for Eth2. Will be giving an intro to Kate commitments on Tuesday **Action: come prepared!** Dankrad's [blog post](https://dankradfeist.de/ethereum/2020/06/16/kate-polynomial-commitments.html) is recommended. [Proto] Several people have expressed interest in implementing Kate commitments. Might be good to start a group to discuss. Contact Proto on Discord if interested. [Mikhail] Have started work on specs for The Merge. There is a draft doc; after iterating will add it to the specs repo. [Leo] Have updated the network crawler to get more accurate view of the network. See chart recently posted on the Eth2 Discord. ## Networking Still need to resolve blocks-by-range issue. Danny to progress that today. ## Spec discussion [Jacek] Has been looking at optimisations for the spec. They don't play nicely with the reference tests. Tests are valid, but can lead to invalid states that could never happen. This limits optimisations that could be applied in clients, but also raises a question as to whether the test is useful. Is there any appetite for improving this? Danny is happy to look at this on a case-by-case basis. **Action: Jacek to file an issue** ## Open discussion End * * * # Chat highlights From danny to Everyone: 02:02 PM : https://github.com/ethereum/eth2.0-pm/issues/200 https://notes.ethereum.org/@vbuterin/HF1_proposal From danny to Everyone: 02:26 PM : https://vitalik.ca/general/2021/01/26/snarks.html https://dankradfeist.de/ethereum/2020/06/16/kate-polynomial-commitments.html