Ben Edgington (PegaSys, ConsenSys — but views expressed are all my own)Edition 25. Archive. Trouble viewing? Load original.
Ben Edgington (PegaSys, ConsenSys — but views expressed are all my own)
I've been hiding within an Ethereum 2.0 bubble for well over a year now—the intensity of the Eth2 effort kind of demands it. So it was wonderful to surface for a while at EthBerlin and take a look around at the wider landscape. My takeaway: the level of activity across the Ethereum ecosystem is astonishing 🤯 Ethereum today, with all the constraints and limitations we know about, is wonderfully alive and creative and vigorous! Just imagine how things are going to be when we finally unleash the full power of Eth2 🚀
This edition's top pick: take the time to listen to Danny Ryan's Into The Ether podcast. A great overview of where we've been, where we are, and where we're going. (Direct link to MP3.)
Oh, and read to the bottom to find out how to help me win a T-shirt 😆
Version 0.8.3 is out. This is the selected spec release target for client interop work going on in September.
There are no substantive changes to the core specification, only updates to tests, some clarifications, and an update to the networking spec.
Protolambda has created a very useful summary of the various optimisations that clients have made in implementing the specification.
Since the Phase 0 and Phase 1 specifications are no longer such moving targets, it's now the season of close scrutiny and evaluation. We've already seen the results of Runtime Verification's audit of the deposit contract. I am really delighted to see more of this happening: the more eyes on all this, the better.
This week, people have been looking at the Ph0 fork choice rule and the Ph1 proof of custody construction.
Ryuya Nakamura (a Casper CBC researcher) described a flip-flop attack on LMD GHOST. This is by no means fatal, but under some circumstances an attacker could prevent the beacon chain from finalising blocks for a few hours. As I understand it, the attacker (who needs to control a significant fraction of the network, but less than one third) can "save up" attestations during a network outage and then use those attestations later to prevent the network agreeing to finalise blocks for a period of time after the outage.
A possible defence may be to modify the rule to FMD (fresh message driven) GHOST: attestations would be valid only for a limited time, so couldn't be saved up by an attacker.
Meanwhile, Dmitry Khovratovich was commissioned to do an audit of the proposed mechanism for generating proofs of custody using the Legendre PRF. For an excellent introduction and background to the proof of custody problem, see the first page of his proposed fix document.
In his audit document, Khovratovich finds several issues with the originally proposed construction, and a possibly improved attack on the Legendre PRF itself. He proposes a new protocol, still using the Legendre PRF, and there is a PR to incorporate this into the Phase 1 specification.
The Ethereum Foundation has funded some bounties for attacks on the Legendre PRF, and even for the most interesting paper on it.
Here are some open issues in the specifications repo that caught my eye recently.
Call #24 took place on the 29th of August.
Lots of good updates. I won't repeat everything here, my notes cover things fairly succinctly. A couple of things — links to the results of the Protocol Labs' EthBerlin bounties (some great work emerged!) and planning for the rapidly approaching Interop retreat (see below).
Some topics from ethresear.ch.
Something that is increasingly gaining attention is how to maintain validator privacy/anonymity. This is important for a couple of reasons. First, since validators need to keep their private keys always available for signing block proposals and attestations, we can expect that they might become targets for attack. Second, if block proposers are known too far in advance, then it is easier to target them for bribery or denial-of-service attack.
The second issue could be addressed by using "secret single leader election". In each slot a validator can discover that it is the sole proposer for that slot, and can prove later it acted correctly as block proposer, but it cannot be worked out ahead of time. My PegaSys colleagues proposed such a scheme a few months ago. Justin Drake has managed to improve the efficiency of the scheme such that it may now be practical to implement: Low-overhead secret single-leader election.
With respect to the first issue, the PegaSys team is back with, Anonymity: a ZKP to remove the mapping ip address / wallet’s public key of a validator, and Mikerah is also working on Privacy-Preserving Casper FFG using Traceable Ring Signatures.
On another topic, there is an interesting discussion about differing Ether issuance rates on the Eth1 chain and the Eth2 beacon chain, and possible consequences of this: Where will reward money on Beacon chain come from?
Highlights from two weeks of Gitter channel.
Way back in May this year, during the NYC Eth2 meetup, Joseph Delong proposed holding an Interoperability "lock-in". The plan being for all the client teams to try plugging their implementations together, seeing what works, and fixing what doesn't. Joe Lubin kindly offered to fund most of the expenses for this.
For the past three months, all the teams have been working diligently towards this Interop retreat: implementing and passing the common reference tests, collaborating on networking specifications, implementing tooling to quickly get test nets up and running. The spec freeze in June has been a huge help, giving us all something stable to target.
Well, the time is upon us! In a week's time, forty-five of us will be gathering in a lakeside cabin in remote Ontario, including all the client teams, some EF people, some Whiteblock people, the Phase 2 Quilt team, and the EF Ewasm team. Antoine Toulme, CTO of Whiteblock, is the appointed technical director, ensuring that we make orderly progress, and resolve our differences effectively. And perhaps, even, amicably.
I won't be publishing another edition of What's New in Eth2 until it's all over, but watch this space for a full report in three weeks or so. And, meanwhile, look out for plenty of exciting memes on Twitter 😱
Here's a teaser for what's to come: Nimbus talks to Lighthouse.
Wish us luck!
![[Twitter]](img/twitter.svg){kind=small}
![[Peepeth]](img/peepeth.svg){kind=small}
Follow me on Twitter and/or Peepeth to hear when the next edition is out 🙌.
![[RSS]](img/rss.svg){kind=small}
We also have an RSS feed.