HackMDWhat's New in Eth2 - 19 November 2021
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Edition 82 at eth2.news
Top picks
No question about it: this week's winner is Trent Van Epps's Ethereum Protocol Update - Nov 2021, originally published on Bankless.
Once you get past some contentious and overly sweeping statements about the use of the "Eth1" and "Eth2" terminology, it is a mine of first-rate information on where we're going with this thing.
The Beacon Chain
By the next time I write to you, the beacon chain will be one year old
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
It all began at 12 noon and 23 seconds UTC on the 1st of December 2020. I'll save the eulogising for next time, but just to remind you to mark the occasion. There is a rumour of at least one party being organised.
On a less joyful note, we've had our first slashing of the Altair era, with the newly increased slashing penalties
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
The Merge
Dev teams are working towards putting up the long-lived Kintsugi Merge testnet towards the beginning of December. There is a progress tracker.
A short-lived (1 week) devnet for client teams was kicked off on Thursday the 18th with participants as follows:
- Eth1, execution clients: Geth, Nethermind.
- Eth2, consensus clients: Prysm, Lodestar, Lighthouse, Nimbus
This first devnet is expected a little rough-and-ready, but is an important step towards having something long-lived and robust up and running. (Update: Marius broke it.)
Most of the Eth2 teams are somewhat blocked by implementing the new optimistic sync algorithm specified by the Sigma Prime team. At Teku HQ, we have been focusing heavily on implementing optimistic sync into our production code and feeding back into its development, and we are skipping the temporary lock-step sync entirely. This is why Teku looks to be a little behind on Kintsugi. But never fear, we will soon be knocking down those milestones!
Marius has catalogued Geth's pre-Merge TODO list. They have the new RANDOM opcode working already.
For those feeling adventurous, you can set up your own Merge testnet with Prysm and Geth.
Tim Beiko posted a place-holder for The Merge fork/upgrade spec into the Eth1 repo. Insert <it's happening!!!> GIF here.
Staking
The Rocket Pool launch on the 10th of November was "a complete success"! Stages 2 and 3 followed on nicely. You can re-live the whole experience with the EthStaker crew.
Stages 1 through 3 have taken limited participants. Stage 4 ignites on 22nd November at 0000 UTC, and is unlimited, so you all can finally pile in.
The Rocket Pool team deserves immense congratulations. They have been on an immense 4 year journey of vision, commitment, and sheer hard work. Eth2 staking badly needs effective decentralised pools like this and I am convinced that Rocket Pool has a stellar future ahead.[1]
Tooling
Péter Szilágyi of the Geth team has been working on minority, an "Ethereum 2.0 node multiplexer between consensus and execution". The goal is to allow post-Merge stakers to run multiple types of Eth1 and multiple types of Eth2 node for increased robustness, with n-of-m agreement enforced before doing anything. It's another approach to the client diversity issue.
The Great Explainers
The excellent PEEPanEIP hosted the equally excellent Sam Wilson from the Quilt team at ConsenSys to talk about The state of Ethereum Execution Layer specs. This is a project to create a full specification of Ethereum 1.0 along similar lines to the Eth2 specs, which are Python-like and executable.
Media and stuff
Matt Leising, author of the brilliant Out of the Ether, interviewed Tim Beiko of the Ethereum Foundation. Some nice background, and lots of Ethereum future stuff. I loved working with Tim at ConsenSys: he is the real deal.
Research
Slides from Dankrad on Gasper High Confidence Fast Block confirmations. This presents a way to determine a "safe head" by using on-chain information to detect whether the beacon chain is under attack. Under normal circumstances the safe head would lag the latest block by up to 4 seconds, but in attack scenarios longer. Under fairly weak assumptions, the safe head will never be reverted, so would act as a decent proxy for finality of transactions. No more waiting for 15 confirmations or whatever it is these days.
This is yet another interesting way in which PoS is superior to PoW. In proof of work, you never know when your chain is under attack. The attacker can simply reveal a longer chain (perhaps with a double spend on it) and you would have had absolutely no idea it was going to happen until it's too late. In proof of stake we have rich information: we know exactly who our validators are, and we get to see all those attestations. Attempted attacks quickly become very obvious.
Also on ethresear.ch:
- Discussion continues on MEV Boost, Flashbots' suggested post-Merge architecture for distributing blocks to validators. I'm surfacing this again as it is a very important conversation, perhaps even at the "existential" level for Ethereum. MEV, and how we respond to it is becoming a critical dimension in the decentralisation wars. I still believe that the best outcome would be a healthy ecosystem of block builders; most of my concerns revolve around Flashbots ending up being the sole block provider on the network, which is not at all far-fetched. See also Two-slot proposer/builder separation.
- For something a little different, here's Vitalik and Casper FFG with backoff. I'm not sure I've really grasped the point of this. It's presumeably a way to deliver some sort of finality even when the network is in meltdown. Perhaps we could also optimistically significantly reduce the time to finality that we currently have (~13 minutes) and then let the back-off mechanism take care of finding the optimum time given prevailing network conditions.
Regular Calls
Implementers
Call #76 took place on the 18th of November.
As per the new pattern we began with Merge office hours, focusing on progress towards the Kintsugi testnet. Geth and Nethermind from the Eth1 world joined for this.
Then a discussion about how to handle terminal total difficulty (TTD) overrides in both Eth1 and Eth2 clients in case of a need for an emergency Merge. It looks like it's heading towards relying on fresh client releases rather than providing CLI overrides or some other mechanism.
Various bits of spec business were covered, including a minor patch to the fork choice rule, and a proposal for a small change to the way the beacon chain stores historical data that would make it easier to verify past blocks.
Finally, the inevitable conversation about naming things…
All Core Devs
Episode 126 (not 127 as per Tim's notes, I think) of the long-running Ethereum All Core Devs call took place on the 12th of November.
Lots of discussion about if and how The Merge will be identified as a fork by execution/Eth1 clients, since it is not scheduled at a specific block, but rather triggered by reaching a terminal total difficulty. Tim's notes are great on this.
Then some further discussion of EIP-4396, which proposes modifications to the EIP-1559 elastic block sizing mechanism to maintain constant throughput post-merge in the face of skipped slots.
Finally, EIP-4444, a proposal to be able to drop historic state after a year. It kind of chimes with Eth2's weak subjectivity assumptions, so is relevant here.
Upcoming events
- November 24th, 1500 UTC: The StakeHouse Community call on the EthStaker Discord voice channel,
stakehose-community-call. - December 1st, the word on the EthStaker Discord is to expect a Beacon Chain 1 Year Anniversary Party 🥳 So look out for that. No place I'd rather be!
- December 3rd, 1400 UTC: Merge Community Call #2
In other news
- Notes from Alex Stokes on light client design.
- Nimbus reminds us that they are also working on an execution client that will complement the Nimbus consensus client.
And finally…
This edition is a little lighter than usual as I am neck deep in a secret project that I'm looking forward to unveiling in a couple of weeks. Nope, nothing to do with NFTs
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Anyway, watch this space!
Learn More →
Learn More →
Advertising on this newsletter.
Disclosure, I hold some $RPL. Talkin' muh bags. ↩︎
