Fast sequencer or global censorship-resistant consensus? Getting the best of both with EEZ

Transcript

All right, yeah. I will be opening the block about EEZ. So, I will now give you an Yeah, an overview, an introduction. We will have Jordi next who will specifically look into the real-time proving aspect of EEZ and then we will have Philip talking about the transition of NoSys chain into EEZ and finally concludes this whole section with a panel with Yeah, Frederick and Vitalik about EEZ. So, maybe going um All right, yeah. Looking back at how

over the years our NoSys perspective on on L2s and side chains and so on evolved. So, Well, soon Well, no, we are in the in the space for quite a while. So, means on main nets but of course then gas costs exploded and it was clear for um Well, we had to look elsewhere and xDAI was one of the early back then side chains. So, we started to to deploy our products in parallel on on xDAI and and and Ethereum main net. Then we Yeah, in 2021, we

in a way took over xDAI and turned it into NoSys chain and there already the question was then looming, should it be an L2? Should Should we turn it into an L2? And we first went a different route. We um followed Ethereum and we also did the merge and well, have have our own beacon chain Um, because we were quite critical of of the um, L2 roadmap um, and and the way L2s function uh, at the time. So, I gave a talk about the limits of L2 and how they are broken and uh, and why um,

or why we don't see that uh, valuable essentially for for Gnosis. In 2024, uh, I gave a talk at um, um, yeah, I took uh, Devcon at Devcon talking about native um, uh, native L2s. Uh, so, trying already to to lay out how in a way good L2s would look like and that finally now evolved into uh, what we uh, what we now call um, Easy. So, now with Easy we we have a framework where we think, well, that's that's how L2s um, should be and consequently um, this is also now um,

finally the decision to to to turn Gnosis then into an L2 under this um, Easy uh framework. So, let's again look at at the problem for a second. So, we are on Gnosis chain um, and yeah, already a couple of of applications were um, mentioned. So, we have the Gnosis uh, wallet or the Gnosis app with things like um, well, your E and Gnosis Pay. So, we need your E liquidity and we have the Circles trust graph and we have the savings um, yeah, savings token um, like S-Dai. Uh, and now um, ideally and that was that was how Ethereum used to be,

uh, how Ethereum used to be is that there are all those components, those small protocols and you can permissionlessly or you have all those Lego Lego pieces and then you can build them together as you like. Uh, and it's all works, it's beautiful, and with every Lego piece that gets added, you can build more beautiful things. But this is not how it works today. So today, those Lego pieces are fragmented on on the different chains. So just a few examples. We have for example Polygon, and there is yeah, Polymarket is on Polygon with the conditional token framework. And just a random example, it would be quite nice to say okay, I want to have a position

on on on this conditional token framework, so on a prediction market, but instead of how Polygon uses it right now, instead of using USDC as collateral, we would like to use STAI as collateral because that's interest-earning. Um but it doesn't really work because STAI is on Gnosis and a conditional token framework is on or that specific instance is on Polygon, and it just is quite complicated to to go back and forth. Another example is in Gnosis app, we would like to support better privacy, and there are of course privacy protocols out there. For example, privacy pools.

Well, but it's deployed on Arbitrum. So kind of going back and forth to Arbitrum is again difficult. Another example, on Celo, there is this self protocol, pretty cool protocol where you can take your passport and yeah, kind of scan it with your phone and make proofs that in your passport whatever, you have a specific nationality or you have a specific age, and that would be quite nice to combine this with the Circle's Trust Graph. But again, it's such a pain to to do it because they are not on the same same chain.

Last example, your e-liquidity wouldn't be nice to kind of um share this liquidity and have overall then stronger Euro liquidity against ether and um uh and uh and and dollar stable coin. But again, it's fragmented. So, there's some liquidity here, some liquidity here, some liquidity on Ethereum, and um it's all a mess. Again, why is it a mess? Because ultimately those layer twos um or those those other chains have essentially very slow communication with Ethereum.

Uh and that is kind of the seven 7-day settlement. For for Arbitrum, um or for yeah, for um it it essentially takes 7 days to to Ethereum. So, if you would want to go the proper path of using the canonical bridges, it would take you to kind of go back and forth from gnosis or I mean to to just implement the things I just described, the communication between the chains would take yeah, many days, and it's completely impossible to build an app. So, you would need to fall back on third-party bridges, but then again, you have a fragmentation, they might only support a subset of those things. So, it's overall um a total mess, and Ethereum therefore, and I think that's a big reason why um yeah, why Ethereum is in the position it

is in, where it's I mean, um where it's certainly not been as successful in adoption of um of of protocols and of of uh of dapps, uh it's certainly not as successful as we would have wished um it to be. And yeah, I I very much blame this uh level of of of of of fragmentation to it. So, what is um where's easy um going towards to? So, easy will essentially be a set of chains um that are much more tighter connected to Ethereum um that essentially do what we call real-time Yeah, settlement. And

real-time in terms of Ethereum means every 12 seconds because every 12 seconds there is a new Ethereum block. So, if you settle every I mean, if you settle your chain every 12 seconds, you are essentially kind of in real-time with Ethereum. And now, things changed dramatically. Now, essentially, all those all those change chains that choose to kind of join this economic zone, they can do Yeah, they can do instant transactions back and forth. And kind of technical technical term would be they have synchronous composability. So, in the first step, those are two chains Well, well, Nosys

will move into this EEZ, and we are building a project that we call Rollup One, kind of a the first Yeah, proof of concept or or I mean, more than a proof of concept, but kind of the first instantiation of a EEZ chain. So, those are the very first chains that will join this EEZ, but of course, the potential hope is or the path should be that any chain that wants to join that Yeah, essentially, those network effects can join the EEZ, and then it overall becomes this much stronger place where Yeah, innovation can happen again, and just more positive network effects

um happen. So, again, today, it's quite complicated to Yeah, launch a new chain. If you launch a new chain, there still so many pieces you first need to get on the chain to make it attractive for developers to build on this chain because you need different stable coins, you need different on-ramps, you need price oracles, you need There's a long list of things that are basically before you even can start building interesting stuff. And in this new easy world, you could launch a new chain that is easy enabled and by default you already have everything available. All the tokens, all the protocols

that are that are available on Ethereum, you by default can already plug into them and you can really add to it instead of first rebuilding what was already done. So first you benefit from the network effects, but then in the next step you also contribute to the network effects. So every again right now even if a chain overcomes this hurdle and then become successful, then it becomes successful on its own and it's essentially still its own yeah, its own ecosystem and its own cosmos and it's not really contributing to this overall thing.

Again, what this would change with um the easy. So one more one more perspective on this. So the promise was made quite often. So I think I'm I'm certainly not the first one telling you, okay, this is how we fix inter operability. So there have been attempt I mean many attempts to say, yeah, we are doing shared sequencing and we are doing this zone where we will have this inter operability. I think the big difference is that yeah, we put Ethereum at the center. So if you if you build a new protocol that says, okay, within this protocol you will have shared sequencing and you will have the synchronous composability and

so on, then of course that is only useful if if many projects join that that zone. So if you are the first one then well, congrats, you have synchronous composability with yourself and then well, okay, maybe you can convince someone else to join, but then you are still a small subset. This is again different with easy because the first chain that already joins easy will have this yeah, synchronous composability with Ethereum and that's immediately already a huge huge value add. So this it basically doesn't have this bootstrapping problem of first requiring enough chains to join before it's useful. Easy is useful from the very beginning because Ethereum is

part of it. A bit technical slide so I mean there are many concepts and I just wanted to yeah, quickly quickly say so there is the idea of based rollups. So essentially rollups where they they don't have um they don't have their own sequencing um and there is the concept of native rollups that I introduced at at Devcon. So essentially a rollup that doesn't have its own token that basically just I mean this is is as neutral as possible and kind of is is basically just an extension of Ethereum.

And then there is easy which essentially means you have this yeah, real-time settlement on this synchronous composability. And you can combine all of those things and that is what rollup one does but you can also still and this is what noses is going to do you can also still have sequential sequencer with with which means faster block times and other things. So yeah. So one more question I sometimes get is should we even have L2s anymore?

Aren't we in a place that L1 is scaling and that basically just everything should live on L1? And I would say there are four arguments why we should still have L2s. And one number one is simply still costs. So those are the costs I took from today. So, swap on Ethereum costs yeah, plus minus 15 cents. Borrowing the same. And those are based costs of $10 or something like that for for a swap on Ethereum. So, 15 cents is great. There are still many many use cases that require significantly lower fees. So, our Gnosis Pay card where you do a I don't know. I

mean, you can you can buy a coffee for for for $2 or $5. And that potentially even a swap because you would might want to swap your interest earning tokens then into the token you pay with. So, basically such a transaction cannot for this specific use case can certainly not cost one cent. Definitely needs to be below one cent. And 15 cents is completely completely impossible to build that. So, that alone means many use cases are not feasible on on Ethereum L1 just because of costs. The second is of course speed.

Ethereum has 12 second block times. And again, for many use cases you want to have faster speed than 12 seconds. The next one is still number three is still what I would call distribution. So, we saw this concept that many let's say centralized exchanges want to have their own L2. And I think it kind of makes sense. So, they can basically I mean, in a way it's one way to think of it is to say it's a it's a franchise model. So, they can they can kind of take what Ethereum uh essentially built and have their own version and in that they can have their own try to use their existing distribution channels to get a user base

on it and they can take a cut essentially from from the fees or they can have other business models. But, it but it's still and again, this changes with easy is is it still feeds back in heavily into Ethereum and and because as soon as you do then transactions with Ethereum back and forth, yeah, well, you you strengthen Ethereum and again, it strengthens the network effects. And the argument number four for for L2s is not neutrality or in a way it's the opposite of neutrality, being opinionated. So, I I think Ethereum and we all I think we all agree about that Ethereum should be the place that is yeah, absolutely neutral. But, being

neutral comes with a cost and I mean, the obvious one is in the case of hacks. So if there are hacks and unfortunately, there are still way too many hacks, um it comes with a big cost of being neutral. So, it comes So, basically saying, "No, we are not we are not hard forking. We are we are not censoring. We are not" I mean, we're basically letting the attacker go away with the hacker go going away with um yeah, with with with stealing money essentially is a big cost and or is is a big trade-off. And again, I think it's correct that Ethereum well, stands here for this or is this anchor piece of of credible neutrality even at the cost of essentially letting

hackers get away with with hacks, but I think there's definitely room for L2s to have different opinions and to say, for example, on on our L2 and and so, what so what I'm suggesting for gnosis to say, "Well, we are opinionated in this case and we say if there's a large hack, we are in a way, we are not neutral and we are trying to then have a process in place to to revert that if if in any way possible." So, now let's look a little bit in in the technicals. The next talks from Jordi and Philip will will go more into it.

But, just to give you an intuition of rough intuition of how it technique technically works. There are two major components that that are required to um uh to become essentially an easy chain. And the one the first means to to your state. So, you cannot uh basically as Ethereum I mean is basically need to within the time span of 12 seconds of a new Ethereum block, you need to commit to Ethereum what your new new state is going to be.

Um and and you also need to kind of commit in such a way that you Yeah, you cannot go back from it. So, if you want to do this um trust in a trustless way, then you need something like ZK proofs. So, you you basically uh make a proof that this is your new state and you also prove that it's correct, but you can of course also use uh methods that um assume some level of trust. So, it's sufficient if if whatever. I mean, with with TEEs or uh with even a multi-sig. As as long as there's a firm commitment, this is our new state and you have this commitment fast enough, that's kind of the real real time component of it. That's the component number one you need.

The num component number two you need, you need to accept that Ethereum is the ultimate source of truth. Uh and that means if if this commitment if you made this commitment on Ethereum, then it's binding um and usually Ethereum goes only forward. So, basically you make this commitment, then it is what it is, and then kind of in the next block you make a new commitment. But, once in a while it happens that Ethereum reorgs. Um so, once in a while it happens that you commit something to Ethereum, and then Ethereum kind of forgets about this block.

Um so, it happens roughly three times a day that an Ethereum block was created um, but then a new Ethereum on top of this one, but it's instead building on the previous one, what we call a reorg. And that happens roughly just three times a day. If it happens, you need to kind of accept that or you are layer two needs to accept that. So you committed already to the state, but this was forgotten. So potentially you need to reorg as well and you need to yeah, essentially you need to follow the reorgs um, of Ethereum. And generally speaking, that's quite quite doable, but it is something to keep in mind.

So again, um, one two all right, two three slides and it technically works. So let's say we have um, Ethereum block 100, block um, yeah, 101 and then 102 and those are anticipated blocks at this time. So we see our we are here in time, we we we see this block and we anticipate that that those blocks are coming and essentially thanks to the beacon chain the 12 seconds in the future. So essentially we have to do a bit of time traveling. So we are here in our L2 and we basically already here. So let's say 6 seconds before this block actually happens on Ethereum. 6 seconds before we

have to already peek into the future and say, "Okay, we are doing a block here and it will have transactions that in a way only happen in 6 seconds in the future and it and they will happen in this Ethereum block and we need to anticipate already what they do. So again, we have here for example a user trying to make a trade um, on Ethereum. So it's starting on the layer two, sending some tokens to Ethereum, doing a trade, and getting some other tokens back. And we need to anticipate or in in in in 6 seconds and already kind of continue our chain with the anticipated result. Um and kind of

then essentially um uh commits to this and then see if if that uh transaction that that settlement transaction of essentially the settlement of our L2 state actually works the way we anticipated it. And if it does, great. Then we are uh then we are um in this state. So, things uh can get complicated here, but so just roughly uh an overlook. So, of course, it might be that um that um that this doesn't happen because because you are uh because we we don't or from the L2 perspective, you're not fully

controlling Ethereum. So, you might say, "Okay, I want this transaction to be in the next Ethereum block and I expect it to have this result." But things can change. Someone might get a transaction before you in and then your transaction does not have the expected result. So, there are tricks you can do to essentially uh kind of test or or or um deploy or or um yeah, ex- try to execute multiple versions in parallel and they basically so say, "Here's Here's a settlement, but if that doesn't work, then I can use this one. And if that doesn't work, then I can use this one."

So, you can then basically say, "Okay, there are a bunch of transactions that are unaffected regardless of this and I can always confirm them. There are some transactions that uh depend on whether they get into um into Ethereum." Uh yeah, the whole thing gets then complicated once you have um multiple easy chains and you have this those those blocks that's that sync with the Ethereum and you might have some blocks in between and you might have but yeah, too much too much for this talk but but just as a kind of um in in in yeah, in depth things are of course a bit more complicated but Okay so

last two slides I believe. Um so number one is just again step back who or what is required to make easy happen and the good the good answer here is no one. So we can do it right now. We don't need any hard forks. We don't need any buy-in from from someone. It can just we can just permissionlessly build it and we will build it. Of course, there are um a few entities and specifically builders and validators um validators who ultimately uh control the settlement of Ethereum blocks that if they work together with easy it can make the whole thing I kind of tried to describe in the previous slides easier because uh it becomes more predictable that your um L2 chain can actually do the settlement it

expects to do. Um so one one last one is uh I already hinted at it. Yes, Gnosis will um will become an easy chain but in parallel we are doing uh what we call roll up one um and just a few yeah, comparisons. So we will um uh roll up one will be fully based so just using uh Ethereum timestamps uh or a block time 12 seconds. Um it will have ETH as as um as um as gas token but yeah, on Gnosis we will keep extra. I use 2 seconds block time so um yeah.

Finally, very last slide. Um Very rough timeline. Uh we hope that by the end of the year um we have uh yeah, Gnosis with limited easy functionality so with with uh with the ability to do synchronous calls, but yeah, not I mean you you can do very complicated synchronous calls that go back and forth, and maybe we'll start with with calls that only go in one direction or only have one return value. Um we will still try to bring concrete tangible advantages to noses earlier, and in parallel we are developing this roll up one or started as a test net on Ethereum, and it can as roll up zero, and it can turn over time um into into a roll

up one. So, bottom line is in summer hopefully we have something that's already live on Ethereum. By the end of the year we have already um functionality to to noses for for noses chain. Okay, um that was my overview and or introduction to easy. Thank you. >> Thank you. >> I'm short of time. I might want to >> take questions. You can take questions. >> I'll happy to answer questions.

I think I see uh maybe we start >> Great talk. How can everything be solved and there's not something coming up which breaks the whole thing? >> Um pretty confident because yeah, we have we have in a way yeah, create proof of concepts end to end. Um so, we have tried everything I described here. Um So, yes, it boils down then to what level of reliability um can we get to? But here I would say um the whole Ethereum block block production process has already shown that you can go to a high level of um of reliability despite all this complexity. So, if you just look at how

what the complex flow is of building an Ethereum block right now with relayers and bundles and bundle merging and and builders, and and I mean there's a whole I mean you can I mean it's a really complicated process, and it works extremely reliable. So, I would say those two things to me are proof that we can do it. >> We've time for maybe one more. If we're quick. >> I think there's Do you want to get your mic? No.

>> What is the What is the general incentive? >> Yeah, yeah, yeah. Right. So, so the question is um do we need to lobby people to join? So, the good thing is um I think that in a way we we I mean we want people to join, but we also don't care that much because I mean essentially as as soon as Gnosis joins, we already get huge advantage for Gnosis because we are fully compatible with Ethereum. And if others want to have this example advantage as well, then um they should join. So, we don't need others to join to get an advantage out of it. That's that's point number one, but then maybe you're still kind of thinking about when does it make sense to join? And then I would say, "Yeah, I

admit kind of for the large ones, and that's maybe right now it's probably Base and Arbitrum are are probably the chains that have um yeah, sufficiently created their own ecosystems. For them it's maybe slightly less interesting, or at least um because they are in a way already strong enough. But for all other chains that uh struggle with having enough liquidity, having enough stuff, for them it will be a big um a big benefit to say, 'Okay, finally I don't need to yeah, whatever, spend all this money on trying to create incentives for liquidity, on on paying paying then essentially uh use Ethereum.' So, long story short, short time I would say it's most interesting for chains that haven't

reached escape velocity, but definitely I would say long term, um, it should be interesting, uh, for everyone to to kind of benefit from the network effects.