HackMDWhat's New in Eth2 - 11 February 2022
- 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 87 at eth2.news
Top picks
No question - last week SuperPhiz and I recorded ourselves doing a walk through of the last edition of WNIE2. It was great fun just chewing it over together, and hopefully it added a bit of colour and depth. This was Phiz's suggestion; I wish I'd thought of it long ago. Anyway, maybe we'll do it again :slightly_smiling_face:
The Beacon Chain
Nothing to see here :eyes:
Client diversity
The client diversity story continues. SuperPhiz put out a call for "critical analyses and data-driven expectations" for how client diversity affects the beacon chain, along with grants and POAPs to encourage submissions.
And the first submission is in - jmcook synthesised a ton of material into Client diversity on Ethereum’s consensus layer. Read it and note it. And, if applicable, send it on over to your staking provider.
The Merge
The Merge readiness checklist is always worth a look. More and more check marks in more and more boxes. I've worked on this thing for so long that it's easy to feel a bit jaded. But looking through this earlier today I had a genuine butterflies-in-my-stomach moment. <insert "it's happening" GIF here> It's all simultaneously slightly terrifying and incredibly exciting.
Kintsugi Testnet
Four hundred thousand slots in, and the Kintsugi Merge testnet continues to run very nicely.
There was some trouble on Kintsugi a while back that we wrote about at the time.
A proper incident report was published last week. Don't miss the important takeaway for validators, infrastructure providers and tooling developers. In summary, don't spec your hardware based only on the happy flow; be prepared for increased resource usage in times of non-finality.
It's definitely not too late to join Kintsugi, or to deploy and test your application on it. The summary information page remains here.
EthStaker made a video on how to join Kintsugi with a Teku/Besu combo, based on Luis Naranjo's guide .
We made a video on how to join Kintsugi with a Teku/Besu combo from our webinar this week, "On the Verge of the Merge" (with Q&A at the end). We used the excellent eth-docker repo for a quick start.
Or you could break out your Raspberry Pi and try out the full current range of client combinations with Ethereum on ARM's incredible PnP Image.
Kiln Testnet
Kiln is the successor to Kintsugi. Kiln is both a spec version and a set of devnets and testnets we plan to run.
Kiln spec v1 is out, and we'll be starting up a first Kiln devnet (merge-devnet-4) to this spec shortly. There will be a Kiln v2 spec imminently that adds two things: authentication for the EngineAPI between the consensus and execution sides, and a (temporary) heartbeat method so that users can be alerted if their execution or consensus client is misconfigured relative to the other.
In a few weeks we will open up the Kiln testnet to all, and all being well, it will be the last bespoke testnet we'll do before we start forking the existing Eth1 testnets such as Rinkeby.
Keep your eyes on the Kiln progress tracker to see how far along things are.
Beyond The Merge
On the consensus devs call this week we took a brief look beyond the Merge.
The Eth1/execution side is planning an upgrade (currently called Shanghai) for later in the year, and much of the last Eth1 ACD call was spent discussing what will go in there. We will very likely do an upgrade on the consensus side at the same time, if only to enable withdrawals which go hand-in-hand with a new Eth1 EIP from Alex Stokes, Beacon state root in EVM.
In addition to withdrawals we are also looking at throwing in Jacek's historical batches clean up, and doing some work to support Vitalik's blob transactions enhancement.
On the topic of withdrawals, there is an EIP proposal with discussion on the Ethereum Magicians' forum that seeks to help anyone who might have had their withdrawal keys compromised.
Staking
Ahead of the Merge, all stakers are being strongly urged to run their own execution clients (Eth1 clients) if they are not already doing so (guilty!). It makes sense to test things out with plenty of time to spare; Eth1 clients are their own special kind of beast. The same client-diversity considerations apply to execution clients as to consensus clients. Bottom line, run something other than Geth.
Anyway, that's all by way of introduction to Ladislaus's article on running execution and consensus clients with Ubuntu and Docker, focusing on a Besu and Lighthouse combination.
Here's the latest report from the excellent Ethereum Pools account on Twitter: "No single #eth2 staking pool will be offline without us noticing".
And while we're on staking pools, take a look at Rated[1] developed by Elias Simos and Aris Koliopoulos. Rated is "an experiment in coordination" around providing transparency on the beacon chain. There is a front-end that groups validators by deposit address and collectively reports their performance, and documentation with some good info[2]. This looks very interesting and promising. I firmly believe that only good things can come out of having better, more transparent data available to the community.
The Great Explainers
SuperPhiz explores how distributed validator technology (DVT fka SSV) relates to staking solutions like Rocket Pool and Lido.
The Book progresses slowly but surely, the latest addition being Simple Serialize. Magical Merkleization coming at you next. Many thanks to AtHeartEngineer for helping improve readability on mobile!
Research
I noted the Balancing Attack: LMD Edition research post last time, but had not had time to look into it then. Here are some observations:
- It is a theoretical way to split the beacon chain and delay finalisation. But it does rely on the attacker having a very high degree of power over the network. (Not inconceivable, but also not very likely.)
- All the attacking validators will end up slashed, so it would be very expensive to pull off.
- There is a very simple defence that we should probably have put in place previously: if the fork choice sees a validator equivocating then remove all of its votes.
- The point is that the LMD fork choice doesn't take account of validator slashings while the chain is not finalising. It bases its decisions on the state at the last justified checkpoint. Discarding equivocations is a workaround for not taking slashings into account.
- It is unfortunate that these kinds of issues keep showing up with fork choice.
- I am super glad to have researchers like these (Joachim Neu, Ertem Nusret Tas, David Tse) continuing to closely scrutinise the beacon chain and openly share their findings.
Regular Calls
Implementers
Call #81 took place on the 10th of February.
- Agenda
- Video
- My quick notes.
Mostly planning and updates for the Kiln testnets. We also took a brief look beyond the Merge. All done in a nice tight 35 minutes.
Merge Community Call
Merge Community Call #3 took place today. Here's the video. It was mostly Q&A led, so the info wasn't presented in a very ordered way, and lots of the action was happening in the chat, which is missing from the video. Nevertheless, there was lots of good advice and important info dropped. Worth checking in to test your understanding of how the Merge will affect you as a user or developer on Ethereum (tl;dw: very little indeed).
Upcoming events
- Monday 14th February, 1430 UTC: Danksharding educational workshop. Details TBA, but it will be recorded and live streamed (presumably via the Ethereum Foundation YouTube channel). Homework:
- Original Danksharding proposal
- Explainer from Polynya
- 11-20 February, EthDenver. The schedule is enormous. I did a very quick scan through, and tbh, I didn't see a lot in there that interests me as a protocol engineer. A few I would make the effort to see were I there (sorry if I missed your thing):
- Colfax on The UX of staking
- Aditya on Distributed Validator Technology
- Carl on Distributed Validators
- Danny on Ethereum Proof-of-Stake and Our Solar Punk Future
- 18-25 April, DevConnect Amsterdam. Plenty going on. See you there :wink:
In other news
- Danny's Finalized no. 33 update: Kiln latest, and client diversity metrics.
- EthStaker received a grant from the Ethereum Foundation to run two validators :tada:
- Stereum's first newsletter, Under the Surface #001: "Update & Upkeep Challenge, our iOS app release and that we’re going to the Devconnect.
And finally…
I got busted. My rank hypocrisy is exposed. Please don't cancel me 😭
Learn More →
Learn More →
Advertising on this newsletter.
The link was correct worked an hour ago and has now stopped working. Hopefully just a temporary issue with Mirror and it will reappear again by the time you see this. Meanwhile this thread is a good backup. ↩︎
Hate to be boring, but this could be usefully added to the resources page. ↩︎
