What's New in Eth2 - 12 December 2020

Image Not Showing Possible Reasons

The image file may be corrupted
The server hosting the image is unavailable
The image path is incorrect
The image format is not supported

Ben Edgington (Eth2 at ConsenSys — all views expressed are my own)

Edition 58 at eth2.news

Gratitude

Thank you, so much, everyone who has contributed to my Gitcoin Grant! Incredibly, over 100 people have contributed. I am blown away by this, and I promise to continue to make progress on the annotated spec.

We don't have a grants page for Teku–-I am sometimes asked about this–-if you appreciate Teku, please consider donating to some other Eth2 supporting projects in this round.

Phase 0: The beacon chain

It's been a wonderfully dull eleven days since genesis: apparently it all just works

Image Not Showing Possible Reasons

The image file may be corrupted
The server hosting the image is unavailable
The image path is incorrect
The image format is not supported

Learn More →

The genesis event went as smoothly as a Sade album. The EF/EthStaker launch party was a lot of fun and has racked up an incredible 50,000 YouTube views. The competing EthHub event is available as a podcast and on YouTube. Coin Jazeera covered the launch with their customary brilliance.

There were suggestions for the proposer of the first block to include something profound in the graffiti field (32 arbitrary bytes are available). This is what we got:

Mr F was here

Vitalik seems to approve

Image Not Showing Possible Reasons

The image file may be corrupted
The server hosting the image is unavailable
The image path is incorrect
The image format is not supported

Learn More →

According to Hudson, "Mr. F is a reference to Queen and Vitalik's autograph to them."

Deposits into the beacon chain contract continue to pile in as confidence grows. You can track them here. As of writing, 1.4 million Ether have been staked (well over 1% of all ETH), representing 31 thousand active validators and 13 thousand in the activation queue - the current wait for activation is over two weeks. Follow the Eth2 Rewards Bot on Twitter for beacon chain status updates. Nice to see that there has been only a single invalid deposit out of the almost 45 thousand made to date, and that for only 1 ETH.

Network participation rates are hovering around an extremely healthy 99%, and my Teku node is humming along nicely, keeping up perfectly, while using almost no CPU and very modest amounts of memory. This is all as predicted a while ago.

Barnabé Monnot made a characteristically thorough analysis of the first 1000 epochs (4.4 days) of the beacon chain. Tl;dr: it's working beautifully.

Slashings

The biggest drama is that there have been five distinct slashing incidents so far. Worth saying that being slashed is perhaps not quite as severe as it sounds. The immediate penalty is currently only 0.25 ETH out of your 32, plus a little more due to other penalties. The main pain point is that your Ether is now locked and inactive in the beacon chain until some kind of withdrawal mechanism is in place.

The first, validator 20075, was due to proposing conflicting blocks (equivocation). It was reported that the owner was running two validators with the same key.
The next batch of 10 all belong to Ankr's Stkr service. These were all slashed for making conflicting attestations. Ankr reports that this was due to a failure in their (home-brew) slashing protection infrastructure. It was triggered by restarting a node that was allowed to make surround votes (one of the Forbidden Acts). Clients have built-in protection to prevent this, so it's not clear quite what broke or got turned off here.
The third was a proposer equivocation by validator 18177. The two conflicting blocks at the same height have different graffiti. It's most likely that the same key was run on different validator instances, possibly in an attempt to "POAP farm".
Fourth is a proposer equivocation by validator 25645. The owner reported on EthStaker that there were two validators set up with the same key.
The fifth is an attester slashing for validator 7635. It made two attestations with the same target checkpoint.

At least three of these incidents, and possibly all of them, are the result of a bad practice repeatedly advised against: running the same validator keys on more than one validator instance. Please don't ever do this.

It is likely that People are trying to build redundant configurations. This is misguided! A little downtime (a day or two) will have an insignificant effect on your total rewards over the long term, and is preferable by far to getting slashed. In Vitalik's words, "don't try so hard". Carl B went into more detail on this on the Ethereum Foundation blog: Perfect is the enemy of the good. Raul Jordan also wrote some excellent Eth2 Slashing Prevention Tips which apply across clients generally, not only Prysm.

While on slashing, Alon Muroch of Blox Staking wrote about minimal slashing protection. This matches what we put into in Teku a while back for our local slashing protection: simple is best. Alon's article doesn't cover block proposals, but that's a trivial extension.

Testnets

The current testnet is Pyrmont, which continues to run smoothly (90% of validators are run by the client teams and the EF). There's no queue to enter right now, and we expect Pyrmont to remain the main testnet through to the end of January, at least.

Tooling

I love this community: where there is a need or an opportunity, someone steps up. This is certainly true of the growing ecosystem of tools.

Huge shoutout to Beaconcha.in for the amazing mobile app for monitoring your validators. I am using it, and it is beautiful.
Scripted installation of a beacon node and validators. (Pyrmont only, and Lighhouse only, so far.)
CoinCashew added a section to their fabulous client installation guides to cover using Google Cloud's free service to monitor uptime on your node. There's also a video walkthrough.
While we're on Google, here's an article from Nansen that I previously missed about Ethereum 2.0 ETL in Google BigQuery.
BLS Horcrux from StakeWise uses Shamir's secret sharing to trustlessly create BLS threshold signatures.
My colleague Adrian Sutton gave an example of how we are using the excellent chaind tool (Medalla data challenge gold medallist from Jim McDonald of Attestant) to monitor the performance of our own nodes. (Adrian found a bug in the queries today related to endianess, but it'll give you a flavour.)
Protolambda has a test generator for client APIs.
Here are the beginnings of a project to create connectors to monitor Eth2 node statuses. Seeking contributions.

Staking services

Some lists of staking services are emerging:

Here's a fairly comprehensive list from stakefish
A list from StakingEther
Beaconcha.in's list

The EthStaker community is putting together the Ethereum due diligence committee to grade staking services. You can see past discussions on their playlist.

Reminder that your staking service could appear here!

Roadmap

Vitalik updated his take on the Eth2 roadmap, with the addition of progress bars.

The main change in emphasis, as we've noted previously, is the moving forward of the Eth1/Eth2 merge, which we'll now be working on in parallel with sharding and light clients. The leading candidate for the merge is the executable beacon chain, currently being prototyped by Mikhail Kalinin.

Vitalik also proposed some protocol tweaks that we might make in short- to medium-term. One of the big challenges for clients is processing the epoch transition in a timely way: the first blocks of epochs often get orphaned on Pyrmont, and attestations are often incorrect around the start of epochs. This doesn't affect the beacon chain itself much just yet, but it would be good to ease this soon - that's part of the goal of the proposals.

In the even shorter term, an additional type of withdrawal mechanism for stakes is under discussion. Currently, when you make a deposit, you commit to a private BLS key that will be used in future to withdraw your stake from the beacon chain by some mechanism that hasn't yet been designed. This is fine for individuals, but is a problem for trustless staking pools who need to manage withdrawals via a smart contract. Now that the Eth1/Eth2 merge is much closer, it's possible to design such a mechanism and give some certainty about how it might work. We reviewed some of the proposals in the last edition. Dmitry Shmatko added to the discussion.

This has all gained momentum now that genesis is behind us. Lakshman Sankar wrote an article on supporting staking pools, and appeared at EthStaker community call #12 to discuss it. Since then, a pull request to the specs has appeared, and Superphiz wrote some commentary. None of this is technically difficult. It doesn't even require a fork of the beacon chain. It's just good to agree a clear commitment to a mechanism.

Articles and Media

Regular Calls

Implementers

Call #53 took place on the 3rd of December.

Agenda
Video - A lot of Spanish people were getting quite upset in the chatbox. Something about their YouTube channel having been hijacked by an Ethereum scam of some sort.
My quick notes, and from Mamy

No big news. We had a bit of discussion about the future of testnets. Tl;dr - we're keeping Pyrmont through end of January.

Networking

We had another of our occasional networking calls on the 10th - the first for three months. According to my records, it was #7, but Danny has it as #6.

Lots of good discussion, but you really had to be there for the detail.

Upcoming events

Two on Monday from EthStaker

December 14 @ 7am CDT / 13:00 UTC, EthStaker will hold a call with Nimbus. YouTube
December 14 @ 5pm CDT / 23:00 UTC the EthStaker team will meet with Rocket Pool founder and developer Dave Rugendyke YouTube. David did a special session for ConsenSys last week, and it was brilliant - I highly recommend this one.

And finally…

This newsletter was brought to you with the help of 1984 Sade. I listened to the entire Diamond Life album five times straight through while writing. I'm now feeling quite mellow. Hope you all are feeling chilled as well.