What's New in Eth2 - 1 May 2020

Ben Edgington (PegaSys, ConsenSys — but views expressed are all my own)

Edition 41 at eth2.news

Tl;dr

It's all about those testnets - all the right junk in all the right places.

Phase 0: The beacon chain

Version 0.11.2 is out, with a couple of very minor updates to the state transition and some more substantial clarifications to the networking spec. The networking spec continues to evolve as testnets get built out. Everything in this release is backward compatible with v0.11.1.

There is likely to be a v0.11.3 soon which will refactor the rewards and penalties calculations to allow for much more fine-grained testing. These calculations have been a constant source of errors in client implementations that are not always picked up by the reference tests. As a case in point, we tripped over one in Teku just today: both the buggy and the fixed versions pass the current reference tests just fine. So this planned change is very welcome.

Testnets

Topaz

ICYMI, Prysm's Topaz testnet is up. You can see Eric Conner recording an Into the Ether episode about it with some of the team (audio podcast here). It's now past 100k slots.

Very early on, Prysmatic Labs' Topaz testnet had more than one-third of its validators controlled by others (now almost a half). The significance of this is that any party controlling 2/3 of the validators effectively has control of the network; but now, no single entity has control. Some people have been setting up large numbers of validators.

If you are running validators on Topaz, Bitfly would like you to try out their graffiti wall. Looks like Attestant has been making good use of this feature.

Unfortunately, as other non-Prysm clients started trying to sync up with Topaz, it became apparent that it has a small bug in the validator rewards calculation. This is enough for consensus to fail, and prevent other clients from joining the network. Preston explains the situation. The plan is to allow Topaz to continue to run for now, and for current interop efforts to focus on the Schlesi network.

Schlesi

Schlesi is a properly multi-client testnet being built by Afri Schoeden. At some point it may even become the multi-client testnet.

The Schlesi genesis event took place on April 27th, though it's not yet intended to be long-lived and may be restarted a few times. So far, Schlesi is running (a fixed version of) Prysm and Lighthouse, with Teku just being added. It looks like Nimbus can also sync it up, so things seem to be in good shape.

You can follow progress of both of these testnets with Bitfly's explorer, for both Topaz and Schlesi. There are some nice charts. Etherscan is also tracking the Topaz testnet.

BLS standardisation

Draft version 7 of the IETF standard for hashing to elliptic curves was published this week. This is the version we are targetting for deployment of Eth2 (including the deposit contract), and is the version we hope other blockchains will adopt for broad interoperability. Sometime soon, v0.12.x of the Phase 0 spec will be released, and that version will include this BLS specification: all clients will need to update to the new version as this is a breaking change. But this is expected to be the final substantive spec change before go-live, except for addressing any issues found during testnet operation.

The Great Explainers

The ConsenSys Eth2 pages are up, and there's an Eth2 FAQ that's hopefully useful. I may have written some bits and pieces of this.

A nice, straightforward Eth2 overview article from Boxmining, that also has a guide on joining the Topaz testnet.

Jacek from the Status team with a tweet-form explainer of the components that make up an Eth2 client.

And here's a couple of great YouTubes for y'all: Vitalik talks Eth2 on The Shitcoin Dot Com Show and Danny Ryan does an ETH 2.0 Ask me Anything on Parallel Society. Danny at 2am is more coherent than I manage on a good day.

Research

Curious about the work of TXRX, one of the PegaSys Eth2 R&D teams? Joe Delong has published a team update on the latest work, including:

• The Prkl network monitor (more below…)
• Work on verifiable precompiles which could be useful for Eth2 execution environments
• Progress on fork choice tests for Eth2
• Cross-shard simulations for Eth2
• A discovery v5 simulator
• Update on the Eth1 <-> Eth2 Merge

I'm chuffed to be working with these guys

.

Coming back to Prkl: this is a network crawler for Eth2 that Jonny Rhea has used to find a bunch of bugs and issues in the spec and implementations, as detailed in the update above. It is also able to generate a bunch of informative charts about the network.

Another PegaSys R&D team is working on formally verifying the Eth2 spec using Dafny. If you are in a suitable time zone you might like to tune in to their Ethereum Engineering Group meetup presentation coming on May 13th.

At the risk of being repetitive, a third PegaSys R&D team, TeamX, is working on how Ethereum might become stateless. This is a highly desirable prerequisite to getting Eth1 onto Eth2. Many things need to change before statelessness can happen. Among them, how to meet users' needs for data via the JSON-RPC API. If nodes don't have state—that's what stateless means—how can they serve the data that Dapps require? TeamX has proposed an architecture involving Caching and Referrers.

Also on ethresear.ch:

• Alex Vlasov (in addition to his TXRX work, above) has extended his amazing run of articles on fault-tolerant clock synchronisation for Eth2: Lightweight Clock Sync Protocol for beacon chain.
• Alon Muroch is thinking about Abstract trustless pool with participant rotation - secure ways to form trustless staking pools.
• A very interesting follow-up to a post from a few weeks ago about finding efficient ways to pack aggregate attestations, from Ali Atiia.

Regular Calls

Implementers

Call #38 took place on the 23rd of April.

• Agenda
• Video
• My quick notes, and from Mamy.

API

As noted last time, there is an ongoing effort to standardise the API between the various Eth2 clients. This will mean that integration with 3rd-party tooling should be much simpler, and also, perhaps, allow for interoperability between different validators and beacon nodes.

We had a call to discuss this on April 20th:

• Agenda
• Notes from Protolambda and Mamy

Ahead of the call, Danny and Proto had prepared a comprehensive analysis of the existing state of the client APIs. The main outcome of the meeting was the formation of a small working group that is working on simplifying this and making the whole thing more RESTful. The group will be reporting back with their proposal in the next few days.

In other news

• Client team updates from Prysm, Lighthouse, and Nimbus^[1]
• Prysmatic Labs has engaged Quantstamp to audit their client.
• Infura has plans for Eth2. You can apply for "invite-only access" to their beacon node API that will be running once the multi-client testnet is up.
• Here's a simple calculator for what your Eth2 staking rewards might look like.
• And a conversation starter about how Eth2 devs might be incentivised. Personally, I'm already fully incentivised by all the love I feel from you all . But maybe I don't speak for everyone.

Ethereal

The Ethereal Summit has gone virtual and takes place next week.

Friday brings two and a half hours of Ethereum 2.0 fun starting at 9am EDT. Many of your favourite Eth2 characters will be making an appearance, including yours truly

. Check out the Agenda here - there's plenty of good stuff going on.

Registration is free! See y'all there

And finally…

With staking coming real soon now^TM, I thought I'd treat myself to a new box. It's called Metal Albert, and here it is with its namesake, Fuzzy Bert.

The staking rig is the black box on the right.

It's definitely not minimal spec—i5-9500, 16GB, 1TB NVMe, 250GB SSD—but I'm planning on running Eth1 mainnet and Goerli nodes, participating in a bunch of testnets, and maybe some other stuff like Swarm, alongside the Eth2 staking. It's going to be running for a few years, so I went for plenty of headroom. Next stop, a UPS.

Look out for Metal Albert in some graffiti near you soon!

---

Follow me on Twitter to hear when the next edition is out 🙌.

 We also have an RSS feed.

---

1. It's fair to ask why we haven't been producing regular updates for Teku. For one thing, we are the only team that does not receive any funds from the Ethereum Foundation, so we don't have to put out updates. For another, pumping out this thing every couple of weeks is plenty of work, and I can do without any more

   Tl;dr: Teku is in great shape! ↩︎