Aggregated Blockchains: A New Thesis | Brendan Farmer - Polygon

[Music] a new theis hey everyone thanks for coming to my talk um I do have slides that's that's my first slide what oh oh there we go um I I will say I I just want to clarify um before I start I made my own slides and so what you're seeing is not low effort it's uh it's just low Talent um but so the uh the thing that I'd like to talk to to everyone about is something that we've been really focused on lately at polygon and it's what we call the aggregated blockchain thesis and you can think of it as uh a response to the fact that crypto like our when we talk about scaling in crypto we're sort of stuck between the monolithic and the modular approaches to scaling and so uh if you think about what the monolithic uh view says it says that when we talk about scaling and scalability in crypto what we really mean is scaling access to liquidity scaling access to Shared State um and so the idea that we can scale by by adding a bunch of different chains uh doesn't fix this problem if it means that we are fragmenting liquidity or fragmenting state and so the modular view the response is um you know we no ju just like the internet cannot run on a single server we can't run an internet scale crypto ecosystem on a single chain and moreover uh different applications have sort of heterogeneous requirements for security and centralization and latency in ux and so even the idea that we would put all of those applications on a single chain uh doesn't make sense in in in a certain a certain sense um but the problem with the the modular approach is as we can see on ethereum uh we have a bunch of l2s so so we're we're scaling horizontally by adding uh uh more and more l2s um but we're running into this issue where liquidity State and users are fragmented across all of these chains um so you have uh like amm pools and liquidity on arbitrum on optimism on polygon zvm and it's actually really difficult to share liquidity in State uh in a trustless way between those chains so I'd like to claim that when we talk about unified liquidity uh this this Vision requires two things so it requires that assets be fungible across chains so if I I have eth on polygon zkm I want to do something on polygon POS like I want to swap to usdc I should be able to take my eth and move it to polygon POS and get eth back I shouldn't get a wrapped synthetic version of that token uh and the second requirement is that we should have super low latency for cross-chain interactions both asynchronous and synchronous so if we think about the current state we we don't have either of those things right if we want to to uh to bridge trustless between two l2s like let's say we want to go from uh polygon zkm to arbitrum or to ZK sync I need to take my eth on polygon zkm I create a transaction to withdraw that eth I have to wait for a proof to be generated I need to wait for that proof to be verified on L1 for that block to be confirmed and then for my deposit to be credited to arbitr or zync so all in that takes like 40 minutes and so that's an unacceptably long uh time and delay for good ux and for unified liquidity and so in some sense the polygon's aggregation layer the aggregation thesis is a response to fixing this problem so it's uh the goal is to provide safety for asset fungibility across chains and for super low latency um interaction whether asynchronous or synchronous uh between chains so we I I didn't create this GIF this was the very talented polygon marketing team but we we can see that uh and we'll get into this but the introduction of the aggregation layer allows us to move assets between chains in the polygon ecosystem uh in a way that is fungible so we can take like all of the L1 native assets and L2 native assets are locked in a unified bridge and we can take eth from uh polygon zvm and move it and get eth on polygon zkp or on AAR or the okx chain um and crucially we can do that in a super low latency way still safely um and without uh a footprint on L1 so so we're not limited or bottlenecked by the capacity of L1 to do this cross chain interaction uh and so the the first element of the aggregation layer is providing safety for this unified bridge and so I don't know how technical the audience is this might get a little bit technical but um but the goal with polygon is to create an ecosystem that uh gives chains sovereignty so uh you should be able to run your own execution environment you should be able to run zkm type 1 or type two miden uh uh custom zkv Ms with uh something like sp1 um but the problem is is that uh we need some sort of guarantee that if any uh ZK VM is unsound then it shouldn't be able to drain the bridge of all funds locked on that bridge and so the first thing that the aggregation layer does is it provides safety for an ecosystem that has heterogeneous execution environments um so even if uh you know the the custom zvm that runs uh like chain number 1,000 is unsound or is compromised um it can't drain the unified Bridge of the funds that are deposited uh the second thing that the agation layer does is it provides safety for a synchronous uh composability across chains so basically what happens is uh if I'm chain a or if I'm on chain a and I want to move funds to chain B uh I can chain a can send a message asynchronously to chain B and chain B can commit to the state of chain a that resulted in that message and so the aggregation layer gives a guarantee a cryptographic guarantee that uh chain B cannot be finalized uh on ethereum if the state that it depends on of chain a conflicts so th this is a really important point because if you recall like what's required for trustless interaction between two chains we need a proof to be generated but we also need the state corresponding with that proof to be finalized on ethereum otherwise the chain could equivocate so this is a way of ensuring that uh chains cannot equivocate and must be consistent uh within this ecosystem similarly um for synchronous composability uh a user can submit a bundle of transactions that execute across a large number of chains and uh some mechanism whether a shared sequencer or Relay can execute these transactions across chains and the Ager uh can enforce that these transactions cannot be unbundled so this is like a really really important safety guarantee for synchronous composability across chains because if we don't have this guarantee then we could have a bundle being partially executed uh in the ecosystem so so you could have um like a mint transaction being executed on chain B but uh not have a corresponding burn on chain a and so the bridge could become under collateralized and so uh I I guess to sum up there are three safety guarantees that the aggregation layer provides the first one is safety for asset fungibility it guarantees that even if aover or zkv somewhere in the ecosystem is unsound it can't compromise uh the shared bridge and drain the ecosystem of funds so in some sense like soundness issues are quarantined uh to specific chains where the prover is unsound and then the second two are safety for synchronous and asynchronous composability so the way that we think about the AG layer is it provides this Foundation of safety that lets uh other coordination mechanisms like shared sequencers or relays or block Builders uh sort of be built and exist on top so uh you could have a shared sequencer like espresso or noit um that's actually executing synchronous transactions across a large number of chains and in this way we're able to have a multi-chain horizontally scalable ecosystem with thousands or tens of thousands of chains but for uh subsets of those chains they can operate uh synchronously and and and and we can give sort of the uh the feeling of using a single chain even as we're in a horizontally scalable ecosystem and so so it's sort of like the Best of Both Worlds where uh some chains don't need to interoperate like like if you're a game or uh like you know a social application you probably don't need synchronous composability with with uh like Unis swap or some defi primitive and so you don't have to pay for it and so we can have like a heterogeneous execution environment sequencer setup um and uh we can sort of allow chains to make trade-offs that make sense for them so uh and again not to get too deep or too technical but um you you might look at this uh at this guarantee or how the system has been described and say uh okay what if uh chain a does equivocate and so uh it submits uh one state to chain a or to chain B and another state to the a layer and so chain B cannot be finalized uh to ethereum there there's like a 10 to 15c liveness delay where an equivocation is detected and chain a might be slashed or something and so I I I think it's a really important point to say that uh if we are uh allowing cross chain interop at lower latency than ethereum finality we have this intractable problem where uh we have to choose between either providing safety so providing this guarantee that chain states can never be inconsistent that we'll never have uh a situation where the bridge is under collateralized or where users lose funds um but we could have delays in in liveness if if chains misbehave or shared sequencers misbehave um but uh fundamentally we have to choose between that view or uh the view where we'll never have liveness faults but we also won't have safety because we we'll just trust a shared sequencer or the Ager or different chains um but uh in the event that those uh those entities misbehave then uh they can steal funds that are deposited on other chains um and so I there's sort of more to say about this and and you know you can read our blog posts but uh it's really important to um to emphasize that this is sort of the best that we can do if we are providing lower than ethereum latency finality um and and safety so uh in terms of comparisons um I think it's really important to emphasize that uh this view this sort of vision of uh safety and asset fundability and low latency interrupt is something that is only possible with a ZK ecosystem we we can't do this with an optimistic rollup and you you you can sort of easily see why because uh the challenge period That's imposed by optimistic rollups right now it's seven days for arbitr and optimism um this makes it so uh like the latency for a cross chain interaction is actually even worse than ethereum finality because we need to wait for the challenge period to expire um and so fundamentally like being able to have very fast finality with with ZK proofs on ethereum is what allows us to to offer safety guarantees and and to allow uh low latency interrupt um further uh for ZK rollups like I I think that there's a misperception that when we talk about aggregation and the aggregation layer uh we're just talking about proof aggregation and it's really important to emphasize that this is not the case because you you can imagine uh two ZK rollups that are trying to interoperate and uh you know you have chain a and it produces two states like A1 and A2 and both of these states are valid and and so both are secured with zero knowledge proof but uh chain a equivocates so it sends A1 to B and A2 to the proof agation service then uh B is depending on a state of a that never happened and so we're not providing a consistency guarantee that allows us to ensure that funds that are on B are safe uh so I guess I have six minutes for this slide so maybe I can slow down and and spend some time on this one but uh so the the vision for polygon uh under the agair is one in which uh we have an ecosystem of thousands or tens of thousands or hundreds of thousands of chains uh but fundamentally These Chains uh retain freedom and sovereignty to basically operate how they want so they can use whatever token for staking for operating the chain whatever token for gas they can use custom uh execution environments and zkm um and so they're they're not giving up anything and I think that there's a real contrast here between proposals for other sort of rollup ecosystems where um you know there there might be uh Revenue share requirements that are enforced by uh close Source licenses um there might be restrictions on the types of chains and the types of execution environments that you can run um and so the the vision with polygon is not to be rent seeking and not to be extractive like our our view and and our bet is that um like if you have an environment that's horizontally scalable and where liquidity is is shared across a large number of chains then it's purely a positive sum game for uh chains to join this ecosystem because if you think about how uh things work for l2s right now uh you sort of have to play games like you you have to uh launch your token or uh airdrop or or do liquidity mining incentives to bootstrap an ecosystem that will allow you to provide liquidity on your chain and this is sort of like a completely different approach because the liquidity already exists in the polygon ecosystem and it's permissionless to access it can be accessed by any chain um and so as a builder you can take uh you you can focus on what's actually important and and what what your source of Leverage is so so if you're building a a game or a def5 primitive or an nft mint you can uh focus on that and you can access the liquidity that already exists in the polygon ecosystem and so you you don't need to worry about bootstrapping uh an nft Marketplace on your L2 in order to launch an nft mint you only need to oh okay uh you only need to uh focus on building the best def5 primitive the most Capital efficient defi primitive or the coolest nft Mint or the coolest game uh and you can take advantage of an existing uh pool of shared liquidity an existing user base um and and other Primitives and applications that can help your users um so yeah thanks everyone um you can uh you can learn more about this on the polygon blog on my blog on Twitter but yeah really appreciate everyone taking the time and