Why One Chain Can't Rule Them All: Polygon Aggregation Layer Deep Dive | Brendan Farmer

Transcript

the internet cannot run on a single computer like we're we're never going to be able to accommodate not only the the demand for Block space but also the heterogeneous requirements for things like uh decentralization and latency and execution environments like we can't uh Encompass all of those in a single chain what's up everyone welcome back to another episode of zerx research March is just around the corner and I want to make sure to give you a quick reminder to not top tick your prices of your D London tickets if you use codes 0x10 at checkout you can lock in a 10% discount on your ticket don't miss out on your chance to get ahead of the curve I'll see you in London what's up everyone welcome back to another episode of zerx

research we have an awesome interview lined up today with Brandon farmer the co-founder of polygon which is exploring the frontiers of ZK Innovation and helping to scare scale ethereum today is uh Tuesday February 6th and as a reminder before we jump into the episode you can use code0x research 10 at app. blockworks research.com for 10% off your annual subscription but anyways Brendan let's jump right into this um You Dropped a blog post yesterday actually uh and we can link to that in the show notes but you basically made two really big claims you said that no L1 or L2 can scale throughput equivalent to that of the internet and that scaling blockchains is actually scaling access to liquidity in shared State found that super interesting so I just want to peel

back a couple layers because there's a lot to unpack there can you just dive into that thesis if you would yeah sure so I think to me these claims are are kind of self-evident but I think that they they kind of are at the heart of the the scaling debates that we're having now which are between sort of the monolithic and the modular approaches and so obviously the monolithic approach says um you know we we just need to build uh more scalable chains and we can have a single state machine that can sort of scale to to accommodate all of the demand for crypto and for Block space the the modular approach says um you know we we should actually we shouldn't have a single chain we should instead have uh different layers that

sort of compose with one another and we can effectively have like a multi-chain ecosystem at the execution layer level where uh like we we can uh accommodate demand for Block Space by essentially creating more chains and um and and you know like as the demand for Block space grows then uh we can add more chains to to sort of cover that and so the claims that I made in my post were uh kind of against both of these views so against the monolithic approach I I basically argued that uh like the internet cannot run on a single computer like we're we're never going to be able to accommodate not only the the demand for

Block space but also the heterogeneous requirements for things like uh decentralization and latency and execution environments like we can't uh Encompass all of those in the single chain and uh and so then on the the modular side like if if you think about what we're really trying to do when we talk about scaling in crypto it's like scaling access to liquidity and to Shared state so if we just add like a bunch of chains and uh none of them have any liquidity then we're not really like accomplishing what we set out to to accomplish like like in some sense we we already have like there there are a lot of empty blocks that already exist on a

lot of dead chains and we still have this scalability problem uh that we're facing and so um basically that this is all to lay out like a third approach which is the aggregated approach which is like we need to figure out a way to scale access to Shared State and liquidity in a multi-chain ecosystem that's a great uh summation uh of your thesis essentially I appreciate appreciate that and it's good context for the listener but can you dive into a little bit deeper into the aggregation layer or the a layer as you guys call it for short and I guess how that architecture fixes some of those problems you just described yeah sure so if you think about like where we're at

now in the ethereum ecosystem we have a lot of rollups so we have optimistic rollups like arbitr and optimism we have ZK rollups like polygon ZK VM like Stark net like ZK sync um and if you think about what it would take so so to back up right now liquidity is fragmented across uh all these different chains um it's difficult to uh like to access State that's on one chain and not another and so if you think about like what it would take for users to trustless Bridge State or value or assets between rollups like what does that look like so if I have chain a and chain B and they're both rollups that

settle to ethereum like what does it take from for me to send a a message from chain a to chain B and to do so in a trustless manner so it requires uh chain B to wait for some sort of proof uh from chain a whether like an optimistic proof which is basically just waiting for the deration of the fraud the fraud proof uh dispute period um or a ZK proof and so so I need that and so from that perspective like we already can't accommodate uh uh optimistic RS model because users can't wait for 7 days to receive like a trustless cross train message and so but but even with ZK rollups there there's this issue

where like proof generation takes you know right now it's two to three minutes in the future it might be like 15 to 30 seconds and so you have to generate a proof and then you have to wait for that proof to be accepted and finalized on ethereum before chain B can accept uh that message from chain a and so the I guess like the the design goal of the agare is um we need a way to do this in a way that preserves trustless uh but operates at like much much lower latency and so you can think of the AG as like a decentralized service that is run by uh by stake nodes and it's in charge of uh accepting um uh like State updates and

proofs from ZK 2s in the polygon ecosystem and ensuring that those uh State updates that that are sort of the result of cross-chain transactions are consistent um and can be settled to ethereum in a way that's safe and so we do this with proof aggregation and you know we can uh dive into exactly how that works yeah I think the idea of the aggreg aggregation there is very interesting because I've always felt that shared state has been like it's one of the biggest unlocks of crypto and like blockchain over silos state which is kind of what you see a lot of in like web 2 and big tech companies and obviously one component of the aggregation layer is the polygon cdk

could you mind talking a little bit about how polygon cdk fits into the broader vision of the aggregation layer yeah sure so the polygon cdk is the polygon chain development kit um and it's just a way for anyone to spin up new chains so basically to add block space in the polygon ecosystem and uh the cdk basically admits a lot of choice and flexibility so um you could run a chain that uses a type 1 prover or a type two approver um you could right now it only supports ZK VMS but in the future it will support uh alternate VMS like miden or you know other things that we're working on and uh and so the idea is like you basically give anyone the

ability to uh to launch your chain in this ecosystem that has full sovereignty and full choice um and it exists alone but through the aggregation layer it's able to interoperate with other chains in the polygon ecosystem at super low latency and so it can access this pool of shared liquidity and shared State um and so you can think of like the polygon cdk chains is like these uh sort of this this like space of of polygon block space and the agay is the thing that stitches them together into uh you know a unified environment for liquidity and State that feels like uh using a single train I wanted to follow up there I'm

going slightly off the book here but I think there's like if you think about it right like you can have like quote unquote an infinite amount of like polygon CD chain cdk chains or I guess like other rups that lie under the aggregation there is there some point at which like too much is too much for this aggregation there or is it like quote unquote actually infinite block space I hate to use the word infinite but yeah I I sort of cringe a little bit uh inside Anytime that we use like infinite scalability or infinite block space I I I think we have a similar reaction there um but I do think that the aggregation layer has some properties that make it

really really scalable so you can sort of think of the aggregation layer as just accepting these like chain stick and proofs and that uh process like for one it it only requires a causal ordering like we we we don't need to agree on like a canonical uh total ordering of all these messages that the EG is receiving um the acceptance and processing of proofs and the aggregation of proofs is fully parallelizable and can be done with by just a single node um because what you're doing is you're doing some cryptographic work you're taking some proofs from a bunch of different chains and you're combining them into a single train but at the end of that process or you're combining them into a single

proof but at the end of that process like what you're left with is basically a certificate which is a proof uh that shows that you behaved honestly and so it it it kind of like has different properties from uh what we've seen in the past like sharding schemes and and and sort of U mechanisms that rely on cryptoeconomic security and so I do think that there's uh we need to build it and we need to figure out where the bottlenecks are in practice um but I think that there is reason to be optimistic that like this would scale largely with like the total compute available in the system to generate uh aggregated

proofs yeah that makes a lot of sense and another thing that when you mentioned scaling and the whole concept of aggregation there is what type of Novel use cases or unique do you think the aggregation there will enable I think one idea is like a distributed amm it's something that I've been like really excited about but no one's really like built it because it kind of partially seemed like the tech wasn't there but when the aggregation there came on I'm like wait this is kind of like built for uh distributed am so we love to hear like what you're looking forward to yeah I I mean that I think distributed amm are are an amazing use case and um I think just being able to trust ly permit use cases uh like Unis

swap V4 and and like different ways to Route liquidity through an ecosystem uh at super low latency low cost um I think that that's just a really interesting uh um use case I I I I do think like in some sense the way that I view it is less like enabling novel use cases and just like and enabling kind of like a bigger like combinatorial explosion of existing use cases and I think it's just really powerful to be able to um to like be an app developer and have complete control over your your chain so so your uh your execution environment uh your sequencer and you're able to like

completely control and customize the experience that you're offering to users but at the same time uh you can take advantage of shared liquidity and shared state that already exists in the ecosystem and so you don't need to boot strap uh liquidity for a Marketplace for the nfts that you're that you're issuing you can just issue those nfts um and so I I I think that we will see Noble use cases emerge uh but I think that there's like for me a lot of reason to be really excited uh just in like enabling that that simple interaction in terms of synchronous and uh asynchronous I guess interoperability or composability whatever you want to call it um what are the primary TR trade-offs between the two like what are

the differences I guess not tradeoffs and like in the blog post that you wrote last night or dropped last night you mentioned that you might have to have like whitelisted almost clusters of chains that like purposefully support each other in order to get that Atomic composability can you kind of dive into just the two different architectures and I guess out of the box uh for the aggregation layer like what will be supported yeah sure so so the egg layer uh will support uh basically two modes Direction the asynchronous direction where you're depending on some other chain that's sending you a message if you're train um and then the second mode is uh which will come at some point in

the future is the ability to accept Atomic bundles of transactions so this is like the synchronous mode where you have transactions that are executing at the same time across a bunch of different chains um and uh either all of those transactions execute successfully or none of them are included uh on any chain and so I think that there are challenges uh with the second approach um you basically need to coordinate uh like execution across all these chains and like before you generate a proof um if you want super low latency uh you need to be able to guarantee that like State on chain a is locked uh until uh sort of chain D is able to determine

whether it can successfully execute this transaction and so so I think that that's like that uh sort of involves a little bit more complexity but the way that I see it is the AG is sort of just concerned with providing safety so so it can ensure that there's no way for a chain to uh include an atomic bundle when uh like the other transactions in that bundle aren't executed successfully and so the other half of that uh picture is like how do we actually execute those transactions how do we guarantee liveness um for all the chains involved and I think that what the AG enables is for like novel and emergent coordin

coordination infrastructure to basically arise on top of the agare so so this could be things like shared validity sequencing or could be things like relays that fundamentally um like find the best solution for coordinating chains to basically ensure synchronous uh composability and execution and so I I I think that's sort of the way that we see it you you can do a lot with asynchronous um cross chain ad droper ability and that's where we'll start um but I'm excited to see like you know different projects and uh like different infrastructure providers in the space basically figure out the Optimal Solutions for both asynchronous and synchronous composability

okay that's really helpful and then on the asynchronous front um so basically taking a step back actually you've got the unified Bridge contract on ethereum you've got the aggregation layer and then you've got all the polygon cdk chains on top so everyone is going to at least be using ethereum for settlement but what happens in like this modular world where you know obviously we're seeing a new Celestia chain every single day or and soon to be Ian da and probably aail so what kind of trust assumptions do you introduce when you have alternative da layers but then you also have um you know chains that still rely on ethereum for da like do you there's anything there no so so the way that we've designed the system is is so

security is localized and so there there are basically like two things that can happen like you can have a data availability withholding attack um that can be launched on a specific chain and we we can't prevent this unless Every Chain is using ethereum for data availability but I I think it's a re it's reasonable to make tradeoffs for users to prioritize cost over full ethereum security for data availability and so the design goal for us is just ensuring that uh if some validium chain or some chain that uses Celestia for data availability um is attacked that that won't affect any other chains in the ecosystem and so like if you think about how the Ager works like each chain

is basically providing the like minimal amount of data that it needs to guarantee that like a data withholding attack on one chain uh can't affect the continued operation of another chain um but the the second issue that uh that also comes up is like like you said we we have a unified bridge and so all L1 assets for all chains are deposited into a single contract and this is really really good because it allows us to have fungibility uh of these assets across different chains so so I can uh take L1 eth and I can send it from like the zkv the polygon ckvm to the okx chain to the POS chain I don't have to deal with uh like a wrapped synthetic version of that

token where I have to find liquidity to swap into the native version um but the problem is is that we don't want like we want to offer full freedom and sovereignty to all chains in the ecosystem and so they should be able to pick their own execution environment but uh like let's say that someone chooses an execution environment and develops a a zero Pro for it and that prover is not sound like we can't have a situation where an unsound prover on some random chain can drain the entire Bridge uh of all funds deposited in the entire ecosystem like that would be an absolute disaster um and so what part of what the aglar does is it ensures that like chain

uh like cross chain transfers of L1 and L2 issued assets are conserved between chains so you could have a situation where some random chain deploys an unsound prover and is exploited but an attacker can only drain up to the amount of tokens that are locked or deposited in that chain um and so this is like a really really important point that uh like we we want to provide fungibility and and a good user experience but it can't come at the expense of of safety what's up everyone March is approaching fast and I want to give you another reminder not to miss out on Das London it is coming it's right around the corner and it's in March from the 18th to the 20th we have three full days of content this is your chance to bump

shoulders with some of the world's top Executives and have open dialogue with both attendees and speakers we're going to be focusing on a range of topics that I'll let Ren discuss for you first on the list we have Bitcoin Catalyst the Haring and spot ETF next we have a view from the buy side from investors on things like strategy portfolio allocation and more we also have a topic on Aras tokenization and stable coins which I think we can all agree are going to pay a large vot in crypto's future we'll also talk about global regulatory Frameworks like compliance best practices and the evolution of global standards that are shaping the global investment landscape we'll also have someone from the institutional front to talk about infrastructure such as Banking and payments with financial giants like visa and JP Morgan and last on the list the macro case for digital

assets so don't miss out on this Monumental event seats are limited so be sure to register today by hitting the link in the description and using the promo code 0x10 to save 10% on tickets see you in London you mentioned sort of like this scenario where there's a faulty prover right um and I know in your blog post You released last night you talked about this kind of fundamental trade-off between interop avidity and liveness is there a scenario where for example if chain ABCD have agreed to be interoperable with each other through like whatever means like relays running like food no for each other's chains that and the end state of the aggregation there when we reach sort of like the atomic bundles that a chain

could quote unquote like dos um other change or even the entire like aggregation there by submitting like in incorrect or like not valid uh messages or transactions on their like on this malicious chain is that like a possibility and what you mind diving into the trade-offs yeah yeah yeah that's a great question so so the the post sort of goes into like the different trade-offs and kind of this tension between um like liveness and low latency interoperability and so what I would say is like first uh chains have no requirement in polygon to even use the agare they can submit batches directly to ethereum they can completely

ignore the igare uh entirely um this would be a pretty bad thing for them because they would have to pay the verification cost for their p uh themselves versus having it amortized across the entire polygon ecosystem but from there there are also like different liveness trade-offs that that they can make within the agler so like the uh the the issue with liveness is basically a chain decides to accept a bundle or uh like depend on an external chain State and that chain State either is invalid or the chain doesn't submit a proof um and so there there are different like mitigation mechanisms and different trade-offs the chains can make so like chain B that's depending on chain a

could only accept a chain state from chain a Wither proof or it it could it could only accept a chain state from chain a uh if it's running a full node for chain a and it knows that it's valid and like worst case chain B could create the proof um like sort of proceeding further down those trade-offs like chain B could only accept uh a chain state from chain a if um like crypto economically secured relay uh attests that the uh state that it received from chain a is is valid and so worst case the the relay could provide a proof to the drop player um and sort of the the like most

uh I guess like risky trade-off would be the chain be accepts a a chain state from chain a just without any guarantees it just that this is like the trust me bro uh approach and uh and so in that case if chain a is submitting an invalid chain State then chain B would need to uh like it's block or wouldn't be finalized uh in the agare and it would need to go back and remove the transactions that depend on chain a and then submit a new block and so this is like I I I would say not the end of the world um that like this fault would be detected within you know a 30 second to a minute window um and so but but there

would be a rework um but I think that it's really important to note that uh this should be an extremely rare occurrence because it requires the chain a itself to be misbehaved behaving and so this can't be caused by some user like it can't be like like a random attacker can't uh can't make this happen it's fundamentally uh the fault of of the chain that provides an invalid um chain State and so like you said like there there are a lot of penalties that can be applied to that chain like if it's cryptoeconomic secure then there could be slashing um it could be kicked out of the uh out of the agir uh it could be blacklisted and so I I I I

think that fundamentally like for us the design goal is to like admit as many choices and options for chains to be able to pick the tradeoffs that make sense for them between uh liveness guarantees and low latency I wanted to follow up on something you st there I you say like that a chain could be kicked out of the a layer my understanding was that the aggregation was kind of like something permissionless is there a scenario where like for example Dow governance can like completely ban this chain from ever being inter interoperable as part of the aggregation there or like how how does that work yeah so I I I I think TBD

whether we want to like make that a governance decision and and sort of maintain a blacklist of of chains or deregister Jin um you know worst case you could have the situation where like validators would opt to ignore batch updates that are submitted to the agare from that chain just voluntarily um I I think that what you would more likely see is um other chains gossiping about uh you know a chain that had equivocated or that had provided an invalid State and they would just opt not to interoperate with um with that chain but but it would be primarily like a voluntary uh move on the part of

chains it sounds to me like obviously ZK proofs are going to be like the secret sauce to all of this so where are we at there like how you know where did we where were we two years ago and where are we today I know you guys have some pretty exciting stuff that you've been working on so we'd love to get more takes on that yeah so uh so you know two years ago I think yeah two years ago um my uh I I had started a company or a project with uh with Daniel luo and we were acquired by polygon as part of their big push to ZK tech and and sort of transitioning from being a side chain to really building in in the L2 ecosystem and I think that people um maybe still aren't aware that like

polygon has really led the industry in the development of perform zero knowledge proof systems and so two years ago we released py 2 and and starky um and starky is still the most performant uh proof system that's that's available to people and so uh I think in the interim we've sort of been quietly working on two things so the first is planky 3 which is the followup to uh to planky 2 and starky um and we call it planky 3 but that's sort of um it's like sort of a misnomer because it's not really related to plank at all it's it's a stark Library um that supports a variety of different fields

and and and different uh variants on on proving systems and we've seen already like a really really amazing performance Improvement um in plunky 3 and we expect over the next few months to be sharing like some pretty mind-blowing uh results um and then the second thing that that we've been working on is uh like a followup to the zkm prover and so this is kind of the a more performant nextg uh version of the approver and it has like a pretty special feature which is that it supports a type one mode and so we we're kind of borrowing the uh the vitalic

framework for like classifying provs where you have type four which uh like can't generate proofs for anything evm related and then it kind of goes to type two where you can support uh existing ethereum contracts and tooling but not existing chains and so type one is is kind of like the highest level of compatibility with ethereum and it allows you to generate proofs for any existing evm chain and so we've built that and it's actually really really performant so we have been generating proofs for uh existing real ethereum L1 blocks um and we're seeing average proving cost across like a huge variety of transactions um as uh like between 2/10 and 3/10 of a Cent and actually in

the next few weeks we're going to be reducing this by 35% so it will will be like between 1/10th and 210 of a scent um and so this is like a really huge breakthrough I think for the ethereum community because it allows us to like conceive of a future in which ethereum itself is proven with zero knowledge proofs so so you could like browse to a website and that website could uh be using some or accessing some State on ethereum and it could be effectively like running a full node in the browser because it's able to instantly verify every transaction that's ever happened on ethereum with a tiny Zer knowledge proof um and so that's like kind of a mind-blowing future when you think about like crypto kind of like infiltrating

into uh different areas of our life um and the uh the second thing is it like for polygon there are a lot of evm chains that were either launched as like alt l1s or side chains like the polygon POS chain is a good example or even optimistic rooll s and having a type one prover allows us to instantly and seamlessly onboard those chains into our ecosystem and into the agay and they can immediately become ckl 2s they don't have to do anything there's no migration um we just start generating proofs and they immediately get the security guarantees uh of a zk2 um and so I think that that's just like kind of a paradigm

shift for how we think about scaling and like how we think about the maturity of of ZK Tech that's highly compelling that last part peeling back an extra layer there is this going to be like an open- Source thing like why wouldn't the optimism super chain just take this and and apply it to their own Super chain um yeah what's the so that's that's a good question so the the first uh point is yes it's released under an MIT uh Apache to license and so it's completely open source and and free for anyone to use um my response to this your second question why wouldn't optimism just take this and use it for their super train I think

like the way that we see the world is that all of this proving Tech and sort of execution environment stuff will become commoditized so it it's all free and open source software like there's no long-term Moe that exists uh or and no network effects that exist from uh developing kind of execution environment or proving Tech but there are long-term durable benefits and network effects that acrew around the agir and so I sort of get frustrated when I hear from optimistic teams um oh you know that like we're using the best tech for now and pretty soon when the when ZK Tech is ready we're just going to switch to ZK Tech and the problem with that view is like you're not

building the right thing the thing that you should be building is the AG layer because fundamentally the a layer allows for like a unified experience across all of these chains and network effects will acre to the agare like liquidity will exist in the polygon ecosystem and any chain that brings economic value to that ecosystem there will be a positive sum interaction there because like if you bring a chain that Ms a bunch of uh high value nfts or or you know creates economic value be a game like you will be able to plug in existing liquidity in the polygon ecosystem and vice versa like the polygon ecosystem will be able to interact with your assets and your users and so like fundamentally the a

layer is what we should be building and you know I I I I hope that op does uh build on on our proving tech and and and join the uh the agare I think that would be a good positive something for for everyone involved you mentioned something about sort of like evalue crew right and in terms of the modular blockchain thesis which I think we're seeing is slowly starting to play out you know people have often thought about the different layers execution consensus settlement data availability and I think a lot of people have been trying to answer the question where does like value occur in terms of like the token for each of these like layers of the stack do you foresee and I think so far like most of

that premium has gone to the settlement layer do you foresee a world in which like the aggregation layer becomes like an additional layer of like the modular modular blockchain stack and that actually becoming the value of C layer or like the dominant area of Val cre yeah I think you know to give our perspective we are less concerned with like where like value acrel at this stage of of development like like I I think right now everyone every project in the space should be concerned with finding product Market fit and like getting people to use their ecosystem and so thinking about like where we can extract rent or where we can sort of like divert value from from users to our

token is not like I I think it's a completely fair question and it's like a good question to be asking but it's not something that we feel like we have to completely answer now so with that being said I I I do think that there is room for value acral at the agler um like the the way that we think about it is uh like if we grow the polygon ecosystem polygon is going to be operating some public chains in that ecosystem and so like any sort of positive sum or like growing the pie uh mechanism is good for polygon like it's much better than if we tried to have an ecosystem where we run all the chains and we sort of like you know are trying to extract as much value as possible um there's also I think a lot

of room elsewhere in that stack like by running coordination mechanisms by running sort of shared validity sequencers that enable Atomic uh transactions and synchronous composability I I think those are areas where there's the possibility to like for Value to a crew in a way that's not kind of rent seeking or um or extractive um I'm not sure that the AG will become like the point of value like value acral or sort of value extraction I I think that the settlement layer is will will still sort of dominate but um I think that the EG layer will be a really important point where Network effects uh begin to acrw and and where like um it

becomes really obvious that you should be in the polygon ecosystem system because it's just a no-brainer you you get to preserve sovereignty and freedom for your chain um but you get to plug into uh shared liquidity and shared State and offer a better experience to your users makes a lot of sense to me too growing the pie over uh just trying to extract value I see the the logic there but I do want to ask uh why do you think the ZK EVMS on Main net today have failed to see significant traction when compared to arbitr and optimism do you think it's incen do you think it's it's uh maybe you just didn't have that 10x cost reduction which maybe now you do with the uh Subs Cent uh cost for transactions but yeah I would just love to get your take there yeah so I think

it's a lot of things so I think the biggest culprit is that uh arbitr and op had a really long head start and so they've been able to like build a market position that I think is difficult to displace and like the market right now is sort of like a down cycle I I don't think it's demanding super super high throughput and scalability I think it's demanding like good execution and access to liquidity and incentives for a small group of users um and so I I I think that's part of it I I I do think that there's also a lack of understanding that uh of like the true costs of an optimistic rollup so we're we're going to be like writing a post on this but um

it turns out that optimistic Roll-Ups are actually a lot more expensive uh than ZK rollups if you look at the actual costs that are born by users and so if you look like like we measure kind of capital inefficiency as uh as a as like sort of correlated to um the or as directly correlated to um the fees that users pay to third party Bridges to exit the chain because B basically optimistic rollups like impose this cost on users where if they want to exit via the native Bridge they need to lock their funds for seven days and so users that don't want to do that are forced to pay uh fees to third party Bridges and so if you look at the aggregate fees that have

been paid by users say on arbitrum um they're like tens of millions of dollars and if you compare that to the cost of ZK proving every single arbitrum transaction in history that's like less than a million dollars and so I I think that there's this huge cost discrepancy that's not really being felt by users because a lot of the usage is incentivized and and their air drops and sort of there are all these like distorting forces that that are affecting Behavior but I think if you actually pull back and and see like what the structural uh advantages of ZK rollups versus optimistic rollups are there's a really really clear argument um that ZK rollups are are superior and and like we

we just feel that the tech is here and and this year we'll we'll actually see um some shifts in in user Behavior I think one of the points you made in your blog post last night was that this aggregation layer differs from a shared sequencer um and you know share sequencers especially for optimistic rollups introduced for example like additional latency additional trust assumptions and I think my question here was that the aggregation layer unifies liquidity and sht state across all of the different chains right and I can understand how it removes the need for a Shir sequencers in order to provide the same like trust assumptions for like Atomic interoperability or like

asynchronous or like synchronous composability but what would a Shar sequencer on the aggregation layer look like because I think maybe not now but eventually people will start picking up on for example like Crossing meev and even with the aggregation layer like perhaps like Searchers or some other entity would want some concept of ordering transactions across like these different Ro uh so yeah could you run run through like what that potentially would look like yeah sure so I think to to to hit on your your first point first um it it's like a real real structural disadvantage for optimistic rups um to have to rely on on like a shared

validity sequencer or you know the requirement to run a full node for every uh chain that uh you know you're the chain that you're sequencing is interacting with like that's a real barrier and a real bottleneck for scalability and so I think like structurally in my like my view is that this will prevent optimistic ecosystems from actually scaling because we we know that like if you have to run a full node for you know 10,000 or a million chains that's just not a practical like we're sort of back at the Modo ethic uh view of the world um and so uh to to get to your uh your question like would there

ever be a shared sequencer that runs on top of the aggregation layer um that basically handles me and and like or or at least like redistributes me um I think that there could be but I think it's important to note that like the the main benefit of the agare is that it doesn't impose like a coordination mechanism on every chain and so you could have chains where there is a lot of cross chain am like like let's say you have a defi chain and then you have a bunch of validum that are batain transactions to be executed on that defi chain um there will be a ton of cross chain M that might be extracted uh in this setup but uh I think that while you could imagine having a shared sequencer

for in that case for that cluster of chains um you wouldn't need to have that same infrastructure uh like managing coordination between those chains and like the some random gaming chain or uh like like there are cases where you don't need synchronous composability between sets of chains and so fundamentally like a huge advantage of the AG layer is that it doesn't impose a single coordination mechanism for every single chain in the ecosystem chains can organically decide like how they want to coordinate with one another and can kind of pick the uh the optimal tradeoff for for their use case and and you know

their uh their user needs from the user experience perspective in the end state of this like do you envision like one wallet and you're just doing all these different activities using different chains and you really don't even notice is that kind of like what you envision yeah I mean to me that's kind of the Holy Grail I do think it's like easy like like like we're we're building infrastructure at polygon right and so so it's easy for us to say like yeah we have this amazing infrastr structur design and it provides safety and you know we can have these coordination mechanisms but like in order for this to work we really need a lot of uh of help and investment on the wallet side and so I I I think like to your point that's exactly the end state

that we need and and there's a lot of work that's required on the uiux front to actually get there so what do you expect for the aggregation layer in V1 that's launching I believe sometime this month and then as well as for v2 and timeline on that yeah so uh so for V1 V V1 is really um it sort of less the aggregation layer and more the unified Bridge so it's what we call the lxo bridge and so you'll be able to have fungibility across chains but you won't have this super fast low latency interoperability at first and so that's what we want to provide in B2 is the ability to do proof aggregation and the ability to allow chains to safely interoperate um even at lower latency

than than ethereum than proof generation so so so V2 is like you know we we might not get there immediately at V2 but like it will be about accelerating or you know lowering latency from uh a couple minutes to you know less than 30 seconds less than 10 seconds eventually we we should be get able to get to to you know one to two seconds or or even below and then polygon proof of stake I know this probably isn't you know maybe what you work on the most considering ing you were previously at miror but um I guess when is that slated to kind of become a polygon cdk chain that's actually ZK based yeah so so our proving system uh optimizations have have L like

a a big motivation for for doing those has been to allow for the POS chain to to seamlessly upgrade uh to a zkl 2 um I will say I I I don't work on this team which I I think is a good thing for me but um it's it's really complicated to think about uh like migrating the bridge of the POS chain to uh basically convert over to the lxy bridge and I I know that the team that's working on this has a really good um and safe plan to kind of manage risk and and uh and handle this migration in a secure way um but I think that that's going to be uh like the

major source of complexity in the upgrade because it's a train that's running and has a ton of activity and you know we we just have to be really careful in how we manage uh that upgrade I had a question about um sort of coordination infrastructure right um going back to the point that we had on this fundamental trade-off between latency and livess I think a few questions there the first being what do you think the ideal coordination in infrastructure looks like is it like a relay that every polycon cdk chain like Ops into that's crypto economically secured by like the P token or what do

you foresee like the end state of the coordination infrastructure looking like are there like maybe like two big like groups of chains like everyone's just using the same thing or is there like 10 different groups depending on whether like you're like a gaming group like a high lat a low latency def5 group what do you foresee that end State looking like the future will be heterogeneous so so you will have different chains that have wildly different requirements for liveness and low latency and synchrony um and they will take different approaches to coordination um but I I do wonder if the the the dominant coordination mechanism looks like a relay or it looks like um sort of like Builders so uh basically

like super nodes that create uh like bundles and and uh and you know propose uh blocks to a bunch of different chains at the same time enable synchrony in that way or whether it will be like a true shared validity sequencer where you just have the same entity or you know decentralized mechanism that's sequencing a bunch of chains um I think that there's like a lot of room and opportunity for people to build this infrastructure um and I think that like the design space is very open and so I'm I'm more just excited to see uh which direction it goes and and what people decide to build I had a follow-up question on that um I think when the PO token

announcement first came out there was this idea of sort of this I'm going to Butch out the term but like multi-utility token like you can like restake it to do like different things like sequencing uh proving what do you think the P the token itself fits into like this whole like aggregation there if it has a role to play yeah so so the way that we see it is that um stakers will uh stake the PO token to participate in in the iare so to be able to accept um batches and and proofs from chains and then aggregate those proofs um I think that you could definitely imagine scenarios in which um you know restak poll is uh is also used to crypto economically

secure relays and different infrastructure I I think that's a very plausible uh view of the future um you you could imagine a situation in which like restak eth is also used to secure the agir and and and I I I just think that there's a lot of freedom um to experiment with different designs and uh but you know I I think Paul will be will be really core to to uh the ecosystem and and to the to the uh design but it's not going to infringe on individual chain sovereignty so so chains are free to use whatever gas token they want whatever staking token they want they can sequence uh however they want and so there's no like like like there's no

penalty that polygon is extracting from chains um I know that for uh optimism and for arbitrum there are like Revenue require sharing requirements and like there's an extractive mechanism for them to redistribute sequencer fees um back to arbitr and back to optimism um and we just don't think that that's a long-term viable model and so for us it's it's much more important to um to to focus on Purely win-win positive sum uh relationships with chains it feels to me like a big piece of the puzzle is going to be really just garnering developer mind share and getting some of this stuff built and establishing those Network effects that

you just you know talked about a little bit earlier so what's the game plan there just build the best developer tooling you possibly can and provide support or is there like going to be an incent program of some sorts for them like just any light you can share on that strategy would be cool yeah so so all the above um I think that uh the nice thing about polygon is that we now have uh a really really good technical team and like a foundation to to build like really Innovative and and kind of best-in-class products um and so that's combined with like a lot of strength obviously in BT and so there are a lot of chains that have already agreed and

committed to to sign on to the agir or to join the agair and they're going to be bringing like a ton of capital and and TBL and uh resources and so I I think the idea is that um you know there there will be incentives and there will be you know BD mechanisms that kind of bootstrap this ecosystem but once the ecosystem gets to uh like reaches escape velocity in terms of the available liquidity and and the interesting apps developers that are building um I think that that then it becomes a lot easier and it becomes sort of the no-brainer because if I'm a chain um like I'm not giving up anything by joining the agler like like there's no there's no penalty

there's no like restrictions or or requirements that are imposed on me I just get to to uh access shared liquidity insurance State and so we think that that's like a really easy sell for uh for chain Developers awesome well said well ren do you have any more questions or is this a good spot to wrap it I think this is a perfect spot to RIT awesome Brandon thank you so much for coming on we'll have to bring you back here in like couple months after uh the agler actually launches and we can get some updates from you but seriously thanks for taking an hour out of your day man yeah thanks for having me really appreciated it this was a great conversation really looking forward to the future of the aggregation there and really like how seamless the user EXP is

going to be and I think it's going to be a big unlock as crypto's still scratching their heads over what interoperability especially for ethereum is looking like right now thank you you yeah I I yeah I'm excited for the uh for the future and it'll be great do you have any uh last uh I guess things you want to tell the listeners you know maybe where they can follow you on Twitter yeah sure I'm ATB farmer um I also have a Mir blog that I'm going to be publishing more posts on um and then you can always check out the polygon blog uh you can always reach out to to anyone from polygon we'd be happy to to chat with you um but yeah stay tuned for for the launch of the egg layer for more announcements um and uh

and Innovative product launches coming soon so yeah we're excited awesome sounds good I recommend everyone check out Brennan's replies by the way there's a bunch of gold in there some good conversations and threads but all right we'll talk to you later Brandon take it easy yeah thanks guys hey everyone thanks for watching today today's zero X research episode I wanted to take a second and remind you about an upcoming 2024 digital asset Summit in London this March seats are limited so hit the link in the description and use the promo code 0x10 to save 10% on tickets see you in [Music] London