Unifying Ethereum with AggLayer | Justin Drake & Brendan Farmer

Welcome to Bankless, where we explore the frontier of internet money and internet finance. And today on the show, we are exploring the frontier of the unification of Ethereum. This is a continuation of our Uniting Ethereum series that we've been doing with Justin Drake. Today we have, in addition to Justin, Brendan Farmer from Polygon exploring the Polygon Ag layer. As you know, if you've been inside the Ethereum ecosystem, using layer twos is a fragmented experience. Hopping from one layer two to another is costly and takes time. Justin Drake does not think that this is the future of Ethereum, and this is not the future of Ethereum that many of us have signed up for. We just need a little bit of more technology, more innovation to refragment, recompose, unify all of Ethereum layer twos to be one unified superstructure that we can just call Ethereum.

Yeah, so David, as you mentioned, this is our third episode in a in a series, all of which included Justin Drake. The first episode, we talked about fixing fragmentation across Ethereum, and we had a zoomed out look at the landscape of Ethereum and what some of the solutions might be. Then in the second episode, we went deep on shared sequencing and we talked uh to Ben Fish from Espresso. Now, this episode is with Brendan Farmer. We're talking about Ag Layer, which is another solution. I think there's really like three technologies, I would say, or three things to know in terms of the roadmap for fixing fragmentation and unifying all of Ethereum's various layer two chains. The first is shared sequencers. Uh the second, I would say, is based roll up. So using Ethereum mainet itself as a shared sequencer. And the third is the technology we're going to talk about today, ag layers, specifically for ZK VMs. And Brendan has really pioneered the frontier here. And with these three solutions, I feel pretty confident that Ethereum is going to fix its fragmentation issues so long as we can socially coordinate, David. This is going to require the entire community to come together, stack hands, and agree that these solutions are a way to fix the problem and then adopt them. It's not a technical problem at this point, so much as it is a social coordination issue.

Bankless Nation, we are exploring the frontier of Ethereum's recomposability, recomposing all of Ethereum's different layer twos into one seamless experience. Today on the show, we are continuing this series that we've been doing with Justin Drake. Justin stepping in as a technical co host today, not just a guest being interviewed. As we talk to Brendan Farmer, co founder of Polygon. Justin, welcome back to the show. Brendan, welcome for your first time.

Thank you. Really uh excited to be here.

Justin, we've been doing this series. And maybe as you are now elevating yourself to technical co host rather than somebody that we are uh interviewing, maybe you could talk a little bit about um just prep the listener for the conversation that they're about to hear today. What are some themes? What should they be uh preparing for? How should they think about going into this episode? And also just what are your goals for this conversation that we're about to have?

Right, so um we have this problem with Ethereum, which is the fragmentation of the L2s and the rollups. Each L2 is its own individual silo,

and to a very large extent, that's due to the fact that each L2 has its own sequencer.

Now we can use ideas like chat sequencing to try and reunite Ethereum, but we need more than that. Uh, one of the things that we need, for example, is ZK as a low latency tool for settlement.

But it turns out that that's also not enough. We need a little bit more.

And basically, what Brendan is bringing to the table today is basically ideas around safety of cross rollup transactions and gas efficiency.

So if you have some sort of super intent or super transaction, meaning a complex transaction that touches multiple rollups,

you kind of have to trust the sequencer to

Add them on-chain as a bundle,

meaning to say all of them or none of them.

And what AgLayer is all about, and this aggregation fees is all about, is basically not having to trust the sequencer for this bundling and for this atomicity.

The other thing that Brendan is bringing to the table is very significant gas efficiencies that come from two places.

Place number one is this idea of having a shared deposit contract. So right now, every single rollup has its individual deposit contract, which is this native bridge between DL1 and DL2. And that means that every time you want to send assets from one rollup to another, you have to withdraw to DL1 and then deposit again. And so the fact that you have to touch DL1, even just a little bit, is kind of expensive.

And then there's this final component of uh Brendan's vision, uh, which is to safe gas on the proof verification.

So, if you have multiple rollups that share the exact same proof system, well, they you might as well aggregate the proofs

and therefore only have to verify one on chain, which is a further gas optimization.

And when you combine shared sequencing with low latency proving and low latency settlement, plus

you know all the benefits of ag layer, which includes

Gas efficiency and the safety of the bundle layer, then you get something which is really compelling and brings us,

in my opinion, to the vision that we all want, which is this united chains of Ethereum.

Justin, I'm a pretty visual person. Uh, and so uh I'd like to actually kind of try and ask you about how this ag layer uh proposal, this proposal for this ag layer, like fits into the topology of Ethereum. I often view Ethereum as kind of like this this sphere, this globe, this planet, and then we have vertical layer twos coming out of it, right? You have the vertical chains, uh scaling Ethereum, Optimism, Polygon, Arbitrum, ZK Sync. Uh in that kind of version of the world where you have like the layer twos spawning out of Ethereum, how where place like in the topology, place the Agileer for us? Like, how does it relate to the to the topology of Ethereum?

Right. So I guess I would think of it as one of the horizontal layers of the stack. So we have settlement, data availability, sequencing, and now there's this really thin actually

aggregation uh layer,

uh, which is adding to this cake. And the reason I mentioned the word cake is because there's this meme going around called chain abstraction.

And the C and the A stand for chain and abstraction.

So uh with this aggregation layer, uh is this like an intersection point for layer twos to meet before touching the Ethereum layer one?

The way that I think about it is as a small modification to the settlement layer.

So right now, the way that settlement works is that each individual rollup has a different L1 contract to do the settlement.

But if we do a little tweak to the settlement contract, we can do several things like

One is that we can make sure that the settlement is aware of these super intents or super transactions so that it will only settle if these intents are satisfied, if the bundles are respected.

Another aspect is having a single L1 contract for deposits so that you don't have to go through the L1 every time you want to move assets from chain one to chain two.

And then finally, there's like again this like small tweak to how settlement is is made, which is that instead of having one proof per rollup per slot, you have an aggregated proof

per roll up per slot or whatever it is.

Brendan, let's go ahead and and get into your brain a little bit. Can you kind of walk us through the uh the maze, the decision tree that you and the polygon team came up with to produce the ag layer? Like how did the ag layer become inspired?

Yeah, sure. So I think uh

if we go back to like a year ago, um

there there was this like question that kept being asked uh about Polygon that was like to me like a very annoying question. Um and it was like doesn't the Polygon POS chain and like the Polygon ZKVM like aren't they competitive with each other?

And for us, it was like, no, no, of course not. Like we we don't envision a world in which a single chain can like, you know, meet the world's demand for black space and can sort of like accommodate all of uh uh you know the Ethereum activity that's moving to L2s.

But

I think it was like a very reasonable question because it it it got at this fact that like liquidity and users and sort of state was being fragmented between these two chains. And so like they were, in a sense, competitive and like they they didn't share network effects, at least like in a meaningful sense. Um and

So I think we we really started to think like,

okay, um, and and there's like a there are a lot of like angles to this, uh, I think, but we we we started to to sit down and think like

if Ethereum is is really going to scale, like what do we mean by by scaling and scalability? It's like really scaling access to liquidity and to shared state.

And so this vision of having like uh this sort of constellation of fragmented rollups that live above Ethereum, like probably actually won't get us to where we want to be in terms of scaling and building infrastructure that can truly power like an internet scale Ethereum ecosystem. Um and so we started to think about like, okay, like what are the current limitations with uh rollups today? Like, so so so one obviously is um anytime you want to use the native bridge uh to get from one rollup to another, you have to touch Ethereum. And given that like Ethereum's uh block finality time is sort of like 12 to 19 minutes, uh this actually ends up being a lot of latency for users um to experience. And so

The question was like, how do we provide a unified environment that feels like a single chain, even as we uh can accommodate like many, many different chains with many different VMs and many different tokens and many different like governance uh setups and mechanisms? Um how do we like maintain uh unified access to to liquidity and state in in sort of the the L2 ecosystem? And and and so that's kind of where.

like from a sort of overarching design principle we we we ended up.

Guys, can I ask a a bunch of just dumb questions before we get uh like continue this and and go deeper? Um so so one maybe dumb question, Brendan, is is this just for polygon chains or is this for all roll-ups?

No, so so this is not like uh we we we've tried to be very intentional in like not calling it like the polygon ag layer and like branding it as like a like polygon's ag layer because like we see this as first of all like neutral

And public infrastructure for Ethereum. And we can get into like Jess and I had a really good conversation yesterday about kind of like the business decisions like that sort of drove this and why it makes a lot of sense. But

we see this as like

if the Ethereum ecosystem is to grow, and if we're going to grow the pie for all L2s, like we need infrastructure like this and like shared sequencing and and uh sort of on the front end, like chain abstraction wallets. Um but

Like fundamentally this is not something that's you know just for the polygon ecosystem or restricted to polygon chains.

Okay, so this is like you're you're trying to make this an Ethereum ag layer. And that that requires, of course, you know, we've got Justin here who's a you know um a protocol researcher, very familiar with principles of credible neutrality. That that requires maximum neutrality here. Um, Justin, I'm wondering if you could we kind of comment on this from your perspective. Is this a solution for all roll ups in Ethereum or like just Polygon? Or do you do you want to cosign on what Brandon uh said, uh Brendan said, or do you have any differences here?

Right. So I think there's definitely an intent to build a public good, um, one which is well beyond the boundaries of Polygon, one which is maximally inclusive. And as uh Brendan said,

um

There is no opinionation on the specifics of the virtual machine, of the specifics of the governance, of the token, of the branding, and all of these things.