MODULAR INTEROPERABILITY with Jon Kol of Hyperlane

Transcript

welcome to the interrupt my guest today is John Cole he's the CEO of hyperlane it's an interoperability layer that enables developers to permissionless connect any blockchain today we're going to be discussing hyper Lane's modular approach the protocol's architecture how it compares to other interruptibility Solutions and their recent V3 upgrade I'm also dying to find out why he thinks decentralized shared sequencers are not real before we get started make sure to hit the like button hit notification Bell And subscribe to get notified when we release new episodes every week and remember that none of what we discuss here on the interop is investment advice and if you enjoy this content please consider staying with us we're live on evos Quicksilver osmosis Juno and nolus just look for interupt in the active set

my guest John Cole is coming up next right here on the [Music] interrupt [Music] John thanks for joining me today of course great to be here uh you know watch a few episodes a lot of fun so I wanted to take part cool were you at cosmoverse uh no I missed it uh oh okay I I happened to be one of those crazy people who decided to start like a crypto company when they have a young child and have a wonderful wonderful wife who uh does a lot and you know

Works still and uh with the upcoming travel for like devk and before it was too much we couldn't make it but I would have loved to have yeah it was good uh I mean I'm sure was great you've gotten you you've received some feedback on it already but uh yeah it was really great I mean this is the like the first episode we do after cosmoverse so uh maybe just like um to share some high high Lev thoughts here I I thought the I thought it was great I mean obviously like the team does an amazing job at putting on events uh and like this was no exception it was much bigger than last year uh some notable things um I'm I'm really bullish now on a I mean like I I think you know it was it was an interesting concept

that you we were sort of like looking at before uh before cosmoverse but like seeing uh all the people working on a all the research teams like blockworks research and binary and and this rmit team and all the teams that are building on a or like interested in building on a uh it just got us really excited about it so it's a it's a space that we're going to be paying a lot more attention to um the the other thing that I thought was really interesting was like there were very few VCS there and I I see that sort of as a bullish signal like there was us and lash and like a few others but you know it was very different from like a lot of other conferences where you know every third person you meet is a VC so um I think the spot yeah it's

former investor I'll tell you that's for sure I didn't know you're in the right spot yeah yeah so that was great and um yeah really just like genuinely excited about where the space is heading and to see all of the ideas of Adam 2.0 kind of starting to actually come together and people collaborating I think that that's the that's really the this the um the sentiment is that there's a real sense that people are collaborating and working together in constructive ways to uh push this Visions forward so um yeah that's my kind of highle recap on on uh on cosmoverse um but yeah glad you'll be at def connect uh we we're also hopefully we can catch up there but I'll tell you probably not going to get that

same ratio as uh you know Dev connect I suspect will be a much uh less friendly ratio in so far as the the uh you know investor to you know person kind of in the arena let's call it investor to person investor to person in the arena investors are not a oh persons in the arena okay yeah I've had uh I I've now like I I've uh people keep saying Builders and I'm like we're not construction workers you know we work in software like yeah I can't say Builder anymore it's a sensitive topic so person in the arena is my new preferred term okay cool I'll I'll hang on to that one so um yeah how did you become

interested in blockchain inter probability oh great story well so um I became really interested in it when I first got into crypto so uh um I get really into crypto around 2016 when uh a buddy we recently graduated from C where at like his budd's birth his brother's birthday and he goes like John I'm gonna put all my money into ethereum what do you think and I'm like I don't know dude I don't know what to think I barely know anything about it all I know is like you worked really hard for this money and you're going to put all of it into this thing and he's like dude all my friends are doing it so I think I should do it and we spent the weekend on on it and I read about this thing like this crazy

internet computer that like anyone can build software on and like you can't stop and you know the ethereum paper really was a lot of a lot of it was about creating financial services and it's someone who's been burned by Banks multiple times in multiple countries through like many different interactions I was like this is so cool also at the time I'm working at Morgan Stanley an investment Bank you know and I could see like I know now how some Financial Plumbing works and I'm reading this thing and I'm like this is real this is legit it's going to revolutionize everything I do I go back to work the next Monday and I'm like you guys we're going to be out of jobs not because of the algorithms because everyone was worried on the trading desk or human Traders are all worried about

like the algorithms com like that's not not human algorithms it's going to be these smart contract systems like you don't know what's coming and you know you learn more and more and more and I become a little small time minor in the next few months and start real realizing wait a second we have these internet computers but they can't talk to each other that's bananas that's not going to work I did like oh so okay so like a blockchain doesn't know anything about information that's not endemic to it okay it makes sense when you say it but like how is a world financial system going to be built on these rails uh doesn't add up and so really from very early on it was just kind of like okay I remember signing up for an internet connection like 1996 I was 6 years old and like At first my

friends who were on the grew up in Israel there were two big isps friends who were on the different ISP than mine we didn't see the same websites like we both were like connected to the local portals but by the next year everything worked together right suddenly we all had access to the same sites and it was like everything got connected so I'm I get into crypto by 2017 it's like oh this whole thing where you know there's not a lot of these blockchains now but there's obviously going to be a lot of them and they'll be able to connect like that's a next year kind of thing and then it was a next year and the next year and the next year and really the first what I would say is like the first real lling of Hope was IBC going live um at that time I'm still an

investor at Galaxy uh I really spent my first five years in crypto as an investor and you're seeing that like as amazing as IBC in the cosmos seos system at large are it felt like I thought I always think of crypto and specifically of chains as like cities or like population centers and it was pretty clear that our population centers are wide and Divergent and at best at that time Cosmos is either the you know during the Terra days Cosmos I thought was the second largest population Center sometimes it's the third some but it's it's not the first one it's not the largest one right the largest one you know being the ethereum population Center like the many Myriad evm chains

that arounded kind of like you know Suburban cities if you want to call it that they couldn't talk there was no you know outside of something like uh gravity but there was no General message passing there was no like apps that could Traverse those population centers and so that got me really uh going then came an opportunity to work on what it later became hyperland that was like two years ago cool uh yeah I mean that's uh that's an analogy that I've heard a lot right this this ology to population centers or countries and um you know Sunny talks a lot about about uh blockchains in this kind of geopolitical uh context and and

I think it's I think it's an interesting one it's also an apt one and feels very much like there are more connections happening between these you know these population centers if you will or like these different domains of State as I like to call them um very good I like that ter I'm going to steal it from you what's the what's the biggest thing that interoperability Solutions are competing on today like you know there's there's there's a spectrum of inoperability solutions I think there's like different um uh there's different axes to look at inoperability Solutions on so you have the the trust versus trust minimized trusted versus trust minimized right so on one end you would have like a multisig on the other end

you would have something IBC and then there are gradients in between um there's also like um this um this notion of like insured Bridges uh or bonded Bridges right where the game theory is different between each each type of bridge um you know if you if you look at hyperlane and in that context where does it sit on these axes oh that's great so let me answer the first question about like what are they competing on and I think the the axis the parameters that people think they are competing on R what we talked about it's uh it's a relation

to trust assumptions it's a relation to Economic Security ultimately it's a relation to liquidity and users and that's what I think those are different axes I think everyone thinks they're competing on what I think the competition really is from the perspective of an interoperability protocol is about the interface like whose interface is going to become the go go to interface and I'm not talking about specifically like the UI I'm talking about when you're an application developer or you're a chain like what do you do to communicate with others which interface do you plug into that and then how often do you have to switch that

interface does every time that the the state-of-the-art in crypto advances does that mean you now need to have a new interface and this is why we designed hyper in a way to be uh to really have a modular architecture and you know if you go through our docs you go through our website you'll see this term about modular security interchange security modules they come up a lot because one thing that we did start I'm getting some props uh I think for most people like so in our space of the of in operability there is a there's the product which I would consider to be the interface and then there's a big part of that product

which is uh the interface helps me communicate between chains but how do I secure those Communications right because obviously security is imperative to to this problem of interoperability of connectivity and for almost everyone that is like they just treat it as a singular thing it's tight you know it's at best it's tightly coupled it's two things but they're tightly coupled together and hyperlane like nope we decouple them those two things are separate so on one hand you have the interface this is you know think of it as like the hyperlane mailbox the thing that sends and receives messages and on the other now you're going to have a number of different security modules and you can

leverage those and use those based on the context of what your users are doing and so some of those modules are going like not some all of them are going to lie on different uh areas within those axes that you talk about so they're going some of them are going to take a much more trusted form some of them are going to take a much more trustless form for example with hyperlane you're able to secure the communications out of your chain just through like a singular key that would be the most trusted setup or you know uh recently because hyperlane is open and permissionless anyone can wrap uh their logic as a hyperlane

security module and actually we're recently seeing electron Labs a team that building ZK light clients bring their ZK light client to hyperlane as a security module so now you have these two edges of you know two edges on the spectrum of on one hand you could have a setting through Hine that is as maximally trusted right it's just a singular key and hopefully it's somewhat nice and the other is you have a zero based light client where it is completely trustless and both of those things can exist in the same system both of those things most importantly can be accessed from the same interface uh and that's really what I

think what that's what we're trying to compete on and certainly like I'm sure different people have different feelings I I want to D dive in a little bit deeper here into this modular uh concept um I I I think I'm starting to understand what it means but uh I'd like to maybe compare it or or or sort of compare it to this idea of a modular blockchain so if you could you know let's let's remind folks like what are modular blockchains and how that idea that concept overlaps with what hyperlane is doing because I think maybe in that context it it's more helpful to understand oh great framing so the I think

the the best way to explain uh modular blockchains to keep it in simple terms I'm a simpl ton myself so uh keep it easy is thinking that in a regular blockchain what some people like to call a monolithic chain I don't think that distinction is necessary but that's okay uh you get all the layers of the stack in a singular place right so you have the data availability that happens in the same place the execution and the settlement all three of those things are happening at the same place and the notion of modular chains is to start separating that so we see with you know like think of layer twos as one

version of separating you know okay so there we're taking out execution but we keep settlement and data availability somewhere else you know in a modular chain you might uh separated it even further where you have your own rollup and you're going to one place like say Celestia for data availability and you're going to another place maybe ethereum maybe you know settlement layer of choice for that and you're completely you know modularizing that stack in a sense that normally we have everything in the same place and it we're kind of like bought into it so if I'm using just ethereum manance all three of those things happen in the same place I could

now today uh and very soon with the impending launch uh Celestia will be able to do this much more easily completely separate I could have my execution happening say on my local rollup I could write my data availability to Celestia and then I could have uh ultimately settlement somewhere else uh and that's really what this concept means so now without going too deep into it where does something like hyperin and its notion of modular security coming well we now have all of these really really bright and capable teams working to advance the modular blockchain throughout the future some people want to call it the rollup

Centric future some we want to call it the modular future I think these really are oh or even in Cosmos we talked about the AP and thesis for a really long time I think all of these really are saying the same thing with slightly different names and that goes back to Sunny's point about like different you know geopolitical areas right we might all be talking about the same thing but we all have our different languages because we come from different places um and so in that world everyone right now we're all working towards bringing that world a about and we're making it easier and easier and easier whether it's with new Frameworks uh for rollups whether it's with new businesses and tools to make launching your own rollup easier whether it's you know like

providing new data availability layers like what Celestia is doing all of these things mean that the task of running a modular chain is becoming easier and so what's the logical next step to that well they're going to become more of them if there is some actual Eon you know real motivator to do this if there's a real technological reason where this is solution is beneficial then there're going to be more of these chains but what the modular thesis kind of as it currently stands does not address is okay so you have all of these chains by definition they'll have an easy way to connect to their settlement layer but how do they connect to anything else are we just going to have

a world of just completely disperate chains everywhere um and then you say okay well we have these interoperability Solutions but the the permissioned interoperability Solutions which I would say is the vast majority of them today and I think there's really good reasons for why you would build a system in this like non-m modular much more permissioned way because you basically say hey here is the security setting it's a singular one-size fitall kind of security setting and we as the purveyors of that system we are the GateKeeper of which chains can come in and be a part of this set um of course that's quite limiting because now if you are

one of these modular chain Builders and you want to get added to that set you need to go through this form of gatekeeping so the gatekeeping has a benefit it keeps some people out but it has a negative consequence it keeps some people out right the thing that is useful about it is this thing that is negative about it um and so with hyperlanes modular design we allow anyone to create uh you know different security settings different security modules and now you can kind of be the master of your own set and so with this there are some great benefits right no gatekeeping anyone can come in but this can also be used for some uh for some

negatives for sure right some people uh might try and lure you into their chain thinking like oh well it's it's secured by hyperlane so it must be okay but actually like there is no default security setting for hyperlane so someone might pretend like this is a very nice non-malicious security setup but in reality it might be quite a malicious one and that's kind of the the other side of permissionless right like if anyone can do what they want there's good things are going to come out of bad things could come out of that too so you're talking when you're talking about this um this particular say like attack Vector uh are you are you describing a situation where the destination chain is

malicious or where the construction the modular construction of the hyperlane uh interoperability bridge if you will between these two chains is malicious in the same way that you know you can trust IBC but you you also need to trust the destination chain yes sorry so yes so that is really what I'm talking where like um especially in the modular World a lot of times we're going to be the and I guess the point of hyperline is to make it very accessible to the people who are creating their own modular chains and so in that realm we expect that certainly the first purveyor of you know of security of going into like if you're going to say be bridging into a new destination chain

using something like hyperlane chances are that it's that chain that is operating the security and one thing you can do with hyperlane is use that Chain's validator set if it's you know tenderman chain or maybe even use that chain sequencer if it's you know a rollup as the as the thing that secures the communication going out of that chain now and what are you doing in that case you're basically trusting that chain and so the same uh the same thing that ends up happening where like ultimately you have to trust that chain is going to be happening here and so if that is you know done maliciously people could get hurt now

there's another Vector with hyperlane that does not uh exist in IBC because anyone can put hyperlane anywhere someone could even do it not not on their own chain they could say create a bridge into someone else's chain that chain itself might not be malicious but they create um you know a hyperlink connection say between uh um maybe osmosis and this new like evm chain they create or you know could be like a sea level bit hyperlane is a VM agnostic so it works in many networks uh and it's not their chain so it's not that the other chain was malicious but they're like hey no we we've made this new bridge it's awesome like come use it and

uh they tell you like all the reasons why you should believe it's secure but unless you look underneath the hood you might not see that um they're doing something that's not so kosher right that maybe they're only running it where like the the communication is secured by one key because again hyper hyperline offers you a lot of flexibility and expressibility in how you secure that communication and so maybe it's a good time to get into a little bit about like the architecture and some of this will make a little bit more sense um and why it could be both a very very good thing but also opens the door for certain malicious things to happen yeah I'd like to talk about the architecture before though I want to address your hot take which was that you think decentralized Shar sequences are

not real so while we're still on the on the topic of uh modularity uh C can you uh dissect uh this statement would love to i' would love to I love first I have great appreciation for anyone who's working on this problem it's a very very difficult problem um but my contention and uh you know is that if you think about what are actions that are described by what would ultimately be a shared sequencer Network you're now talking about accepting transactions from multiple sources cating them into a single place and somehow doing it in such a way where all of those transactions are

synchronized can become Atomic and that at the same time this is somehow going to afford you scalability that is not found in any of those places individually and that I would argue that is not real what that is describing is a blockchain that accepts transactions from other blockchains and I think you know for us in the realm of interoperability we are you know we're in a similar neighborhood to uh to this problem because if you think about what hyperlane does what any you know Bonafide in oper protocol does we accept transactions from lots of chains and the one thing we're not trying to do is to or I guess some might be right but

certainly in the case of hyperlane because there is no singular hyperlane chain we don't try and synchronize these transactions at the same time we're not sending them to a singular place and collating them in such a way where they later on need to be uh you know processed automically in the some place and so I think that that gives us a view into how difficult this problem is and that ultimately if you try to say okay hey I'm going to have 10 different trains are you know I'm going to have 10 different rollups and all of them are going to be sending their transactions to me and I'm going to allow them to have Atomic interoperability right so like I I'll be able to process their

transaction such a way that they could have Atomic composability between them and I'm going to give them you know massive scalability those things don't add up to me because the first thing you need to do you know blockchains really are ultimately a lot of them their operations are about time like how long does it take the process of block and like how does this time sync with other times and so first thing you need to do is you need to sync up how often things get processed because if my rollup and your up are not are expecting things to happen at a different time well that's not going to work in this case all of us now need to U live on like a much more unified block time and ultimately that unified block time is decided by that

shared sequencer so I could see how a shared sequencer that is completely centralized could work much better but now if you're talking about synchronizing all these different chains and doing it in decentralized fashion I don't see how you don't run into the same problems that um a network like ethereum runs into when it tries to ensure that it takes you know many transactions millions of transactions from a maximally decentralized network of nodes it's this I view it as very much as the same issue and that's why I like to you know I don't think that this notion of decentralized ser sequencers is truly real uh I think they it's a blockchain that would take transactions from other

blockchains which is a totally fine Endeavor to pursue but but then you just run into the same problems again like that blockchain will run into the same problems of sequencing the transactions of the shared sequencer I mean like you just totally it's like a recursive nightmare uh which yeah I mean I I yeah I um I tend to agree with this idea um I I think that as we decouple and modularize different parts of the stack there are ultimately going to be some parts of the stack that will rely on uh we won't be able to decentralize fully in a way that like trust fully trust

minimize systems uh should be and it's all about like making the tradeoffs for the system that you're building uh like where do you want where do you where can you accept a higher degree of centralization or uh or in this case like where could you accept um less trust minimization at what layers of the stack and what specific areas but yeah ultimately like I don't think you can fully decentralize everything um yeah that that's what it is I think you could have in I think you have a shared sequencer that works as advertised I think you can have a decentralized but singular sequencer that works as advertised I think having a decentralized share sequencer that works

as currently advertised will be incredibly difficult and ultimately I think is uh you know trying to solve the problems of blockchains using a blockchain which maybe works I I don't know I've been in the industry for uh about six years now and I just don't think that that um slapping on a new blockchain is never really the solution yeah it might not I hope I'm wrong because if I'm wrong right like it does mean that this idyllic State can exist and like um yeah you know we were talking a little bit before we started recording I got into this like a long

time ago I just want to see like crypto Prosper like I've been doing this for a while and you know I think that my best way to improve the odds of crypto prospering of expanding the the crypto economy and even accelerating its growth is through something like working on an interoperability protocol could be totally wrong and maybe uh a decentralized St truly can exist maybe there are some breakthroughs that we just haven't seen yet that uh will happen and yeah maybe like a if it works that's better for everyone so so while we're on the tradeoff uh while we're talking about trade-offs does hyperland optimized for liveness or for safety or is that a a

function of the configuration of each sort of bridge you got it you got it there you go you this is how I know that it sank in it you would have like different configurations can land in different uh parts of that tradeoff and so two two simple examples right like we'll keep it maximally simple you could have a configuration with hyperline and certainly we've seen we've we've also helped people do this uh in testnet where transactions are basically validated after a single block on the sending chain so immediately just showed up in a block we're not waiting for finale nothing showed up in a block sign off on them and then they can be

processed as early as the next block on the destination chain and for those transactions like you know when you're dealing with uh you know faster moving block times you're talking about like a 5 six second turnaround time and end but what about forget even say what about just like I could get reared out of my uh especially if we're dealing with ethereum could easily get reared if we uh talking about dealing something like um I plenty of chains have like more than one block reor right that's not no one would get upset or like U ruffle their feathers by this this is just a

fact but you can do that and so that is something that U would be on one end of that Spectrum the other is you could actually say like no we don't process transactions for a really long time in fact we would wait um a Time designated in maybe several hours or days and a construction that we would tend to refer to as more of an optimistic Ander borrowing from the the notion of optimistic Ro it's not a oneto one but it's it's a it's a cousin um where that you functionally don't let the transaction be processed for a wide array of blocks and during that time you allow anyone the

capability to observe that transaction if a fraud proof can be constructed to demonstrate it right because ultimately like the action that is being committed On the Origin chain it was included in a block like we see it it's there we can write a proof against that if that proof can be executed then now we can veto your transaction we can block it from ever happening this I would say airs way more in the side of safety but what does it actually mean to veto that transaction you actually have to kill the channel and so you're killing the liveness for that channel and so those are I'd say like the two extremes of hey we're going one one we you know one block one block

um not so high on the safety Axis or we're going you know in many blocks but we're also giving you the option to basically like eliminate liveness in the situation because once a channel is compromised like do you even want it open like forget just one transaction going through that looks that like how do you know that there's not going to be another 10 after it so what at some point you got to kill the channel maybe you don't kill it on the but you got to kill it and then you might lose livel I mean you're not might you will lose livess for this period until like the new challenge uh channels up um so so I guess but yeah one one last thing I want to point here before we get into the

architecture it sounds like hyperlane is more of an interoperability framework than a solution it's it's more of a a framework for building interoperability solutions that cater to the needs of like different sort of trade-offs like in this case liveness or safety uh those um considerations need to be made on a per app basis uh and and it it allows you to sort of like construct your own interruptibility solution um and sort of find the right place between these two tradeoffs okay cool exactly right now it does have like some pre-made options so I like to think of it as yeah you know a box of Legos but there are some diagrams that they're like well here's how you

build an X-Wing here's how you build like a pirate ship uh but ultimately yeah there's a box box of Legos there and if you want to do some wild stuff with it you certainly can cool so let's let's talk about the architecture uh what are the different uh pieces of this this modular inability stack and yeah maybe getting into the modules and um and the the function of the those the The Interchange security modules of course so you'll need three things to basically bring hyperlane to your chain and the first one of those are we talked about it briefly earlier it's the mailbox the mailbox does exactly what it

sounds like as it's the place where mail is sent and received and think of this is the the the contract where the interface exists for sending stuff and receiving stuff so you'd need one of those on every chain uh has a function for sending things function for receiving things and it's like your first touch Point mailbox of course is open source the contracts are verified on the chains that they're on and anyone can go ahead and put it on a chain that they' like so now you need there are two more things you need you now need at least one security module and you could even have a security module where its entire

premise is just letting anyone through the door uh we like calling this like a no op security module that's but like you're going to need at least one of those and it can take many forms let's talk about a common one where it's like a validator based U uh situation where you would have validators observing the mailbox on the sending chain and once they see your transaction they see okay Seb is sending a transaction they might wait for depending on the config they might wait for finality we certainly recommend that they do in many cases and once finalized they

uh they sign off on that transaction um there are both there's both a way to do it where like they're signing where the transaction gets added to an incremental called Merkel tree and they're signing the root of that tree there's also a like per message signing uh mode there's benefits to either and they concern like different gas cost and different uh like censorship resistance guarantees and then the third piece that or sorry uh so the mud like in that validator one yes you're they're they're observing the sending chain and then there is uh a contract on the destination chain that

basically says like this these are the here are the keys of my validator so know when those are coming in these are the enrolled validators in this security module now as an application that is on the receiving end you conf you tell the mailbox when I expect messages route them through this security module and so now you could not just you could choose a single module or you could do something slightly more sophisticated where you could say wait a second what is my user doing oh is my user transferring say doing something that is very innocuous like they're transferring say $5 you know it's their first transfer

it's just $5 then maybe I'll go for that uh something that prioritizes liveness over safety and maybe prioritizes speed right like lower latency but for other transactions maybe their transaction is actually trying to drain a liquidity pool maybe their transaction is trying to make a very large uh asset transfer something or maybe maybe it's a command to actually upgrade this contract and so now give given the nature of this transaction I might say actually route that one through a different security module so as a transaction goes from the origin chain to the to the destination it is not immediately processed it first needs

to go through the security module and you as the receiving application have some routing ability you could say hey oh this is what they're doing here please send them to this module or this combination of modules or okay they're just doing this very in thing fine send them to the you know send them to the fast so you can essentially route transactions based on the nature of the transactions so or in this case like if you were trying to upgrade a contract that transaction might be sent to like a governance contract or something like that exactly right and that governance contract would then you know unlock that transactions pending a governance vote Etc that's that's really cool okay maybe so these security modules sorry sorry these security modules are are modules that

are like provided by hyper or anybody can write their own security modules anyone can write their own certainly but we don't see that happening too often right now right hyperlane is still a young protocol like we have seen a few being written by outside like the the the zkm one which we're super excited about and there's a few more like those coming but admittedly right now that's a slightly bigger lift right we don't expect you to you know writing your own Marin is kind of like going to the restaurant and they let you go into the kitchen it's like hey use the ingredients we got like the best Tak can to go ahead and make it yourself uh so that can happen I would not tell you that it's like the most common um course the more common one is we have

like we have these pre-made ones and then you decide how to use them and so you don't need to do you know kind of completely writing it from scratch it's you have a module where you designate a certain amount of you say hey I want these keys to sign off and those keys might be your validator they might be your uh you know whatever is like the key for your Dow as you mentioned in that governance case uh Keys might be a subset of your validators that you think are you know more relevant for specific transactions uh and that those are much easier to use right like in that case you're not rewriting much you're just taking something off the shelf and you're inputting certain uh parameters

into it and working that way and so that's the security then the third thing that you need to have a hyperlane integration is basically in between those chains you need to have what hyperlane terms we call a relay it's pretty similar to IBC reers um in the sense that you're not really trusting this relier with the content of your message the relayer can't meddle with the contents they are purely like a messenger um and right now senders uh when they put like when they pay for their transaction they specify the real ler they're paying for what we're moving towards and this will happen sometime in the next say six to eight months is um more of a relayer

like marketplace where instead of paying to a relayer that you pick in advance you are putting the payment ad there kind of like a bounty and then you these relay start looking a little bit more like uh you know interchange solvers because there's a payment put up by me the sender and I my transaction tells the relayer like what I want done and then they just need to go and get that done on my behalf through by send by processing that message on the other side that starts to resemble more of like an intent based system that's right yeah at that point okay I don't know that all interoperability protocols picked up on this but um I view very thin I don't I see a very very thin

distinction between an interchain intense system and an interoperability protocol um interoperability protocols right now are just might be more narrowly uh much more narrowly defined intent systems because you know the we're limiting it to like hey I want to swap this or um yeah that language just hasn't penetrated yet but like you can you know when people talk about like oh intent based architectures are like far you could actually realize one today using something like hyperlane if you build a you know you could build an application that allows you users to much more easily express an intent Beyond to say something like limit

orders um and then have just have you know hyperline reers basically fill those orders and they you know there's different language but they would actually ultimately just look like solvers so I I guess like does does this I mean cuz so there's no coordinating chain so this is very different to something like Axel where Axel the protocol sort of sits at the middle and is coordinating um is coordinating this message passing and the sending of assets um how so so so there's no Central coordinator that's doing any of

that it is purely like a chain to chain uh messaging protocol where like messages are received on chain a a realer sends messages to another chain and then a security module or a you know configuration of security modules there will relay those messages based on the types of transactions they are and and then they may have to go through additional additional um steps in order for the transaction to actually hit the chain and be executed I mean that kind of sounds a little bit like IBC to me I I don't know it's very similar I would like to say that you know hyperland doesn't exist if not for IBC you know like so what is the key difference and mean then I mean because I IC sort of

exists as a as a spec right like um it it does the IBC spec uh is it lacking in its functionality where hyperlane couldn't have just implemented the IPC spec and maybe like add some security mod like add some sort of features like the security module is the that I think really augments you know the the security of uh interoperability between chains so what yeah what's the what's the difference I guess stands is uh very complex and was kind of built you know I don't want to put words in people's mouths but I think if you look at it you see it was built with a certain chain Construction in mind uh okay and really was designed at

a time where if someone said my blockchain design is going to be the most dominant one you would believe them like that was a reasonable statement to make four five six plus years ago uh today where we see a number of designs certainly the evm might be the most most dominant whether it's good or bad right but there are good reasons why we still haven't seen um you know that there are not many reliable IBC implementations for evm chains is not coincidence it's not for lack of trying it's not that there's not enough smart people trying to there's many smart people working on this uh it is that in my view that the IBC spec is

quite complex and was designed with tenderman chains or you know something that looks like a tenderman chain uh in mind right a chain that has um light line verification a chain that does have fast finality but for us we wanted to focus on extensibility I wanted to see something that ultimately could let anyone connect any chain now IBC lets anyone easily connect a tender chain but I'll give you one example right like the um the the the call backs that are necessitated in IBC are very useful in the context of two tender chains talking to each

other but you have to so basically like you have to send you know like a a a call back is being sent from the destination back to the origin to kind of let them like Hey we're all good the thing happened um if you're dealing with like slower finality and more expensive chains like ethereum now you may that interaction much more honorous and so implementing the IBC spec as it is is something where would have limited or at least in our view would have very much limited the extensibility of so then how are the callbacks how callbacks treated in hyper I also wanted to ask you about that you know they're not required is they're not

required yeah okay so you could send assets to another chain and not know whether the assets were Min yeah so the way that you know is handled through a much more like uh call it like less rigid process where you could just check like the hyperlane Explorer let's say and see oh well my think arrived um but observing the observing what happened on the destination we think achieves that same um achieves that same benefit in majority of cases um and so you're sort of relying on the application to check uhre and giving them the tools to make that much easier right okay rather than the chain actually getting a

confirmation that exactly yeah because so I think this is interesting yeah I think this is really interesting because I like I I I was thinking about the web right and how like interrup ability happened at in in in the web and you were talking earlier about how um you know like your email uh Communications were you know uh hindered back in the '90s and then it all kind of like fixed itself but I I think inability and in the web really happened if you if you take a step back and look at computers right so you have like Unix computers you have Windows computers that you know there might be like other um types of os's out there and interoperability between these

systems happened at the application there so like we're talking right now through HTTP and you could be on a Mac I could be on a Linux computer any other computer as long as like we're in HTTP and we're in a browser we can interoperate because that application layer in propability works it's not we're not like talking at the kernel layer we're talking at the application layer through a protocol which is called tcpip right and it seems like what you're describing here is you're relying on the application layer to ensure that that finality happens rather than like the protocol actually enabling the improvability that's yeah that's uh I think that's an apt description uh we do

leave more to the application L because I think the if you want to be more extensible what's at the other end of that is being more rigid and if we want it to be more rigid like IBC is already there like like uh but with the expressed goal of kind of maximizing the extensibility and I think this has been proven out as now we've seen and the coolest thing is that these are completely separate teams this is not our team has been built for kmm kmos DEC uh for for move the the movm and for uh even it's even being built for

cardano's environment and then as well as a number of smaller more bespoke VMS by teams like Caden and teams like Masa that have again more unique constructions and they're implementing hyperin for their own environment and if we opted for something again that's much more does all of the all of that within the protocol very unlikely for that that that could have happened interesting um so I want to take a step back and just go back to what we were talking about earlier and um the role like who are the you we talked about validators we talked about relays can we just expand on the role of those two like uh parts of the of the

the the transaction those two participants in the transaction and are there other participants that are facilitating transaction and what are their roles of course so we'll start with the relayer because that's the easier one the relayer needs to be involved in every hyperlane action between uh one or more chains and it is you know it's it's the Hermes of the system it's the messenger it uh a re anyone can run a relayer it's open sourced it it's a fairly simple operation to run and what a relayer does is it observes the mailboxes of the chains you want to connect you know you can have a single relayer that's observing so even like 15 different chains and as it

observes these different mailboxes it sees hey Seb is sending something it looks like he's got a little treat for me in the form of a fee I like those fees that help me run my business and then I check wait a minute okay so in in their message Seb says that it's going like to this place this application and that these are the security modules being used is there are there valid signatures tied to this they're supposed to be at this location you know maybe they're on ipfs maybe they're on like some publicly available like uh Cloud destination I find them everything looks good I initiate the processing so that's really all the relayer does it is involved again as I said in every hyperlane

transaction um because because anyone can run one you could even Envision a situation where users are doing it right like U I remember seeing sometime in the last two years like a really cool script that allowed users to like inject their transaction or sorry to relay their own like IBC uh transactions that is ultimately possible with hyperlane as well full disclosure not not happening right now no one's I guess someone tried to do that in the hack on but you know it wasn't uh we called production grade even though they put a valiant effort uh now onwards to something like the validator uh something like so just just one one second like on the relayer um

the relayer uh is um so the re relay is remunerated for this work obviously but it doesn't have to be but you hope that it if it's not why would they do it so so just to make this comparison to IBC once again that the difference with IBC because I he does have realators is that uh there is no light client verification so like chain B is not checking chain a and chain a is not checking chain B right yep that is uh okay that is correct and then so when the relayer sends that the relayer is not tied to the security mechanism in any form or fashion the relayer is purely a messenger whereas uh in the IBC case

right like the relayer again I think I wouldn't say that it's tied to the security function in a way that it performs it but it relays something about it back to whereas yeah in in hyperlane the security module which can be expressed through the the validator agent is where that uh happens so there's a world where you where the relayer relays something to the destination and and then it gets blocked by the security module where basically like you know getting relayed is not a guarantee of anything other than getting relayed so a transaction can still not arrive at its ultimate destination based on the logic uh where it gets ran

through the security modules that can happen and is expected Behavior got it okay and so so we understand the realators now um so the validators play a role in uh in specific security configurations so specific security modules they don't always play a role but they can play a role yes um I think that's a good way to but like they don't yeah so they no no decided role in uh security they're purely a messenger and then all of the security handling is related to like what they they are transferring something

from uh like from one chain to another that is completely secured by by The Chosen security modules and if it were to be you know altered in any way the signature that there would be passing along would not match the uh the signature from the security module and so something would be a miss and that message would just kind of fail by default because what happens on the destination chain is that relayer it initiates a function in the mailbox called process and what process does is it looks to the contents of the message and it says oh okay well I'm supposed to go to here's what I'm supposed to do I'm supposed to go through these security

modules okay cool let's run the logic there and the logic might say oh well is there a signature from this validator is there a you know or if it's like the optimistic model we talk has it been enough time H you know like okay yes has it been enough time check has there been a fraud proof submitted no check okay move on to your destination if you fail any of these like that's it you never get past the mailbox and it is totally intended or expected behavior for lots of transactions to never make it past the mailbox got it okay very cool and so there comes in like the role of the security modules and like they can yeah

they can take many forms we have some presets like U one that's more of a proof of authority one where you designate which signers U we're about to finalize an economic security one where it's that same concept of the proof of authority except now those signers also have value at stake this enables um a really cool form of reaking as well because it can accept any asset uh as given uh and then yeah like I said people can kind of bring their own you know they can bring their own home cooking there and so it opened the door for a really really Lial team and electron bringing their ZK light client as a security module so in the case of

uh rollups so if you take a a roll app or roll up where the rollup doesn't have validators itself it is relying on you know uh other layers of security and the validator sort of sit at the bottom layer how uh how would the that particular security module work there who would be signing um we like doing so let's say if you went with like the the easiest setting the authority based one you're now as the initiator of this you're you have to designate like who is going to be signing the messages that go out of my rollup and so you could lump it in with your sequencer

that is one setting uh at that point I would personally I think that is the preferred setting because as a rollup user to me that just says I'm receiving the same trust model as using this rollup just any in any other setting um so that's one possibility the other is validators that are either related or unrelated so you can like kind of bootstrap a Consortium small as you may want of uh validators they might even be putting value at risk like they might be staking you know atom they might be staking your um your whatever is your native currency on that rollup because

obviously not all rollups choose to use the same uh gas currency okay and uh so yeah which brings me to my next point which is um gas I know you guys have some uh uh like a gas module how does that work yeah so one thing that would be really unpleasant is if gas wasn't handled by the protocol if gas was left on to the user and you do all everything we talked about but you would also need to specify where you own the gas tokens on that destination chain that's kind of shitty that doesn't accomplish almost anything

of what we wanted to talk about and so we have to have a way for hyperlane to deal with destination gas for you as the user because again right like if you were initiating a transaction from m uh and you're ultimately executing it on arbitrum like if you needed to hold e you know on arbitrum to do that that's really unpleasant that's really not great so uh the way that hyperlane handles it is through the relayer primarily and that's really the reason why they're collecting a fee in most cases so this is not a committ role but almost everyone that runs a real layer

does this again because otherwise they're the service they're Prov providing is very very thin and so the relayer will hold the balance of the gas tokens of the chains that it's connecting and think of them kind of like an inventory manager so uh it charges you on the destination chain in the local currency but it charges you uh an amount that basically covers his expenses for processing your transaction on that destination so going back to that osmosis to arbitrum example you're paying an osmo on osmosis but you're paying the full amount that includes the tally for what the reer is going to have to pay on your behalf once

you get to arbiture okay that makes sense so uh you guys just recently released uh hyperlane V3 uh with some uh some notable improvements and um for developers can you talk a little bit about um yeah what's what's new in hyperlane V3 and yeah so with hyperline V3 the the the there's a couple new things but the most important one and it might not be most important for everyone uh on the core team here but it's the most important to me certainly because this is been like a thing that's been bothering me since uh since day one is that uh now there are two

functions that you needed to hit to to uh to send a message with hyperline you needed to initiate a function called dispatch and you also needed to initiate a separate function which is like the the the interchain gas function so you said hey I'm sending this thing and in that same transaction you said by the way I'm paying for that thing over there and there was a good like technical reason for why to do that well the protocol was early it prevented like too early of acific Etc that didn't mean we liked it I certainly did not like it uh it felt kind of janky and so that is now lumped in there is uh a dispatch

that basically allows you to pay at the moment of dispatching without the need for a second call again this might seem small but uh you know when you are like us and we work very closely with the developers that build on hyperlane we feel their pain and you know doing this so many times and seeing so many people like ah dang I forgot to do that which is like I understand because it's kind of like you go to the store and you go through one counter uh and they say hey I'm buying this thing and they're like cool you're good and then you oh no you pay over there um and so that is something that always bother me but that is not you know the most exciting to me but certainly not the most exciting uh uh the the bigger Innovation is the addition of what we call Hooks and so

hooks are an easier way for people to bring into their transaction security mechanisms that are not um you know that are not uh like hyperlane native let's call it so imagine you are transacting with optimism maybe maybe going into optimism you would just rather use the native Bridge so hooks let you do that much more easily they let you kind of uh route the transaction through mechanisms that have nothing at all to do with hyperlane and that kind of expands on this point that I made earlier earlier on of it's about for us it's about being that interface and so now with hooks

it's even more likely and more easy to use hyperline it's a singular interface and uh route your transactions through whatever mode you find ideal for the task at hand so who who's building on hyperlane and are you able to subtract the the volume of transactions happening in the say hyperlane ecosystem if you want to say call it that right the the ecosystem of that are using yes I no this is uh this is something I I am very unhappy about so um yes on chains that we know about and track no where we don't and so uh right

now like if you go to the hyperlane Explorer it tracks chains where there are deployments that uh the US and the core team have played a role in and these are like and what's being tracked is the movement between those uh between those mailboxes those are Megan mailboxes that we know of and are aware if you put hyperlane on your own chain tomorrow and you then you use it to communicate with it uh we're probably not going to know about it unless you tell us and so that we don't there's some like Petry stuff that we can do but uh isn't done yet kind of

in the you know in the free willing Spirit of crypto and it is unfortunate I would say uh but yeah we don't get a full picture so um I mean it's kind of feature and a bug right I mean it's yeah exactly now it's something that you know we're working uh to make easier for people who are like growing the Hy so we are creating a registry where people can more easily add like hey I just fored hyperlane and I put it on Aptos or hey I'm just you know I'm putting it here putting it there um because certainly by now there is many deployments of hyperlane that have nothing to do with the core team as there are of the ones that we've played

a role with and that's certainly the goal right like I want to you know a year from now that the vast vast majority of them have nothing to do with um with the coure team I think that is ultimately like what it means to have a decentralized protocol um but certainly from a day-to-day perspective it's like oh we can't see this as easily so to your other question of like who's building with it it's primarily I like to talk about you know kind of we we have a lot of themes that we draw from uh from Star Wars our name actually is you know hyperlane in the the Star Wars Universe actually didn't know this until one of our colleagues proposed it hyper is like

the easiest path the safest and fastest path between like Star systems it's what they use to travel between Star systems like it's the warp of of of Star Wars of Star Wars show yeah exactly it's like the warp drive path and um so you know I'm he's more of a Star Wars guy I'm more of a treky my dad uh pet was he was a massive fan and so we we we have like a little bit of both where our we've created like someone's going to build a competing solution and they're going to call it warp drive well we beat him to it we beat him to it so we have uh we have War right I know where like the easiest way to bridge assets with Hyper to bring them to any new chain is through what we call War Prout so that

was like stick sticking some Star Trek in there suck it uh but so there's with our expansion in t OS um through the the kmom and ultimately the cosmos SDK module you know will uh likely be live on chains like osmosis and now we're working very closely with Neutron which we're super excited about and all of these can be used now as a hub to reach uh other rollups we've been working very closely uh with the Celestia team on a number of uh number of initiatives we want to play play a big role in how uh rollups that use Celestia Forda are able to

continuously do that easily do that and securely do that um beyond that again we are solely F pretty much solely focused on the uh rollup or modular uh ecosystem or as you want to call it the app chain ecosystem I think all those uh words kind of combined into the singular thing and uh that has been most of our activity but that said we still have a number and a growing number of just what we call uh interchain applications or just think of this like Just straightup Apps not just app chains that are incorporating hyperlane most recently um

uh a for its new governance initiative where they are transferring governance between chains elected to use hyperlane alongside uh chain link and and layer zero which we think is really cool uh and then there's a number of smaller apps like Dam finance and Yama and then you know super form that do different like defi things some of them are like interchain stable coins some of them are like an interchain uh lending protocol uh that are using hyperlane between a more um you know let's call it established set of chains but through hyperlane they have the ability to expand to any chain they want and no Reliance on us so one of those teams Dam Finance they wanted to uh have their

application be available specifically have their stable coin be available on a chain that no of the interoperability protocols were covering in in Asar a chain that is quite Big in Japan and they they brought hypine there by themselves right they uh and that is something that we think is like really cool really powerful and it's kind of the it's the thesis at work very cool uh so how can developers uh start using hyperlane on their chain and um yeah what would you like uh to call upon our audience to do great great question so what I'll say is if you are an app developer building

uh you know you're not interested in creating your own chain still reach out either join our Discord or follow us on Twitter if you want to have some interchange capabilities come to us we can show you how to get started the docks are obviously a great place but uh don't be shy we again as I mentioned before go uh there's a whole community of kind of hyperlane builders that is very happy to help uh new people coming along and you could find us all in the Discord and so certainly do that if you are building your own chain if you wanted to start testing around with hyperlane you can follow the deploy guide if you know that's something that you want to do but you're not that's not something you actually want to to deal

with we are now uh offering access to hyper where we will uh our company here will perform will basically run hyperlane for you as a service uh so even if if if you're like hey I want this but I don't know if I can run myself don't be shy just come we'll do it for you um that's becoming an increasingly more popular option as we help people get off the ground and the cool thing about it there's no lock in so because of hyperlink architecture the ultimate ownership and uh responsibility be transferred over and so you could get started with uh someone else doing it for you and then quickly as you like you know get your SE legs from under you you

can just uh say hey you know what we're good we learned how to do this let us take over now yeah which uh was one other question I wanted to ask what's the business model how how does hyperlane make money uh so there's two there's the business model that hyperlane and the protocol um might adopt and that I think looks like a quite simple one where you have a you give chain a chance to opt in to a specified set of security modules where they're they'll also be sharing some of the fees with some like uh hyperlane down that's one option or sorry that's one that's for the

protocol itself beyond that there's like well we formed a company we employ people and for us uh we'll be looking to continue to provide uh services on the Rel layers and services on setup um which just look very much like a kind of boring software business but you know boring software businesses are are acceptable sometimes boring software businesses make the most money so sometimes they do yeah yeah great uh John thanks so much for coming on uh I have like so many more questions here but we're wunning pretty long um it's been really fascinating and uh yeah I really look forward to seeing more come out of hyperlane and I mean I think just

generally like inability Protocols are great for for the ecosystem as a whole and we want to see uh more you know easier liquidity flows between different domains of State different ecosystems and um really glad to see hylane uh performing that function interesting and model novel design so thank you well thank you so much it was great to be here