Cross-Chain Communication with Sergey Gorbunov of Axelar

Transcript

welcome to the interop the show which talks about the technologies that are powering the interchain my name is Sebastian cuccio and today I'm here with Sergey gorbanov who is CEO of axlr is there okay hey great to be here thanks for joining me uh in uh you know in in our sort of you know quest to understand uh cross-chain Technologies and blockchain bridging we're are doing quite a few uh interviews with uh projects that are in this space and you know we can't have a proper Bridge series Without uh talking with axelr because they are certainly uh one of the

biggest and certainly one of the most recognizable uh actors in the space so we've got a full list of things to discuss here today we're going to be getting deep into the weeds of axillary and how it works uh but first if you enjoy this Channel please consider subscribing and hitting the Bell so you get notifications whenever we do live streams typically on Thursday and also consider subscribing uh sorry consider staking uh evmos to our Atmos validator you'll find the link for that in the show notes I also want to let you know that I'm going to be attending Cosmo verse next week in Medellin and I'm super excited about it I'm sure lots of people in the cosmos space who listen to this podcast are also going um I think you'll probably have some

people from your team there as well um and uh uh the nebular summit uh conference and anoma are teaming up to put on a side event on September 27th at 5 00 PM it is called interchange travel and we're going to be exploring all sorts of interesting topics around interesting we have three privacy security and the consensus uh we have a pretty stellar lineup of speakers uh lined up for this event so Ethan butman of Cosmos and informal systems will be speaking Susanna Adams who's Project Lead at IBC uh Zach manian who needs no introduction Yelena from informal systems Felix from

Chorus one and the list goes on Sony agrawal uh Adrian Brink and so on so if you want to attend that there are still tickets remaining and it's free and there will also be a shuttle from the cosmo verse venue and so you can get your tickets for that over at interchain Travel dot x y z putting that right here interchange travel.xyz and we'll also leave a link for that in the show notes so do attend if you're in Medellin be great to see you uh with that yeah Sergey how you doing how you doing um you're in New York for for mainnet right now that's right yep we have uh quite a few folks from the team here um have a booth

uh we're doing a side event a couple days ago so it's yeah pretty pretty fun location yeah I was talking to someone earlier about like I remember um uh you know being a moderator on a mainnet panel you know when it was still uh you know the first one when it was an online event like during covet and you know they were just kind of putting this conference together and now it's like this massive thing and uh I've actually never been to a physical one but I'd love to at some point so you you were on the test map before it became amazing yeah I was on the test net yeah I'm pretty sure yeah I remember like I was talking to the team and I'm pretty sure I didn't do the panel there but yeah um and like it was right when you know

like hoppin was taking off and like all these conferencing platforms and it's just like this you know I just get these you know memories of like entering the new world right like yeah but um I don't know about that though but you know so yeah tell me about your background and why you chose to work on on this problem of cross chain uh communication yeah I mean my background is kind of a mix of uh distributed systems and cryptography um you know I worked on things like software defined networking I studied formal cryptography at MIT in grad school doing a lot of you know protocol design system design work um and then you know afterwards I helped to uh design and take I'll Grant project to the market

um with my co-founder jorgas you know Sylvia a few other folks from MIT so that was pretty fun uh but then after launching our grand we just saw like an array of other you know beautiful ecosystems being developed and kind of realized that kind of connected them together and providing a unified seamless experience for the users to to interact with all of these applications being built in different blockchains it's going to be um you know a must to to continue increasing the adoption and simplifying the user experiences so looked at the kind of solution space understand there are a lot of problems a lot of gaps that will still need to be filled in and you know decided to focus on this problem uh kind of full-time so and it's been a pretty fun journey of the last years

and when did XLR launch actually so yeah the network went live kind of early this year around February time frame um so the network has been live uh since then you know we connected like 24 different chains uh you know everything from you know ethereum osmosis Cosmos polygon Avalanche and the list goes on we did like over I think 1.2 billion and cross chain volume to date um you know over 200 and um over 200 000 transactions I think 180 000 like wallets have been connected through this ecosystem so it's yeah pretty exciting yeah I I use it like a semi several times a week and yeah the

experience is like yeah it's pretty seamless it it feels um yeah it feels very seamless to just you know uh connect your your destination wallet and the wallet you're connecting from and and it just works and now I mean I I primarily use it with osmosis and now the integration is just so much better now that there's this uh osmosis integration built-in right right into osmosis so you don't even have to go to the satellite um UI you just in osmosis you hit the withdrawal the XLR window pops up you connect your metamask wallet or whatever wallet you want to send those tokens to and it yeah it's it's like really fantastic I think this is really the

direction in which you act the ux needs to go and and I think yeah even one point on that right like uh for for things like you know usdc um assets like you actually don't have to send them from like a metamask baller right so I think the kind of the beauty of what we have built is that you get this sort of deposit address at osmosis and then you can send um you know a deposit there from like coinbase right from binance from anywhere you have those tokens and then even though they will go through ethereum they'll go through axle they will go through you know IBC and like arrive at once it's like you're never going to be exposed to like a single step you never even have to pay gas like on a theory yeah yeah I mean yeah I think the user experience

question around Bridges is uh is really a crucial one and it's it's a topic that has been coming up you know all the conversations I've had with you know people bringing building kind of multi-chain uh composability and and communication uh protocols um but yeah before before we talk to that about that like I want to ask my burning question which is why is axelr secure you know and um yeah what's your pitch here for like why people should use this bridge and not another bridge and why it's not going to fail like so many others yeah I mean at the very core of a security comes down through achieving it through strong decentralization

properties right we have seen it over and over in the space that the solutions that remain kind of robust are the ones that are decentralized so and the very core of it Excel network is a decentralized network powered by tendermen you know and the kind of Cosmos SDK so it has an open validator set anybody can join anybody can participate anybody can send transactions right anybody can can use it and you know a lot of other uh kind of gen zero bridging protocols they're all centralized right they have kind of two out of three multi-cigs or they have like some Federated oracles Guardians whatever whatever they call them you know and you you kind of rely then on those individual deployments for for trust and like and actually you relying on the protocol security right and

decentralization uh through the valid data set so that's one thing that's that's the very core start right if you have to have decentralization you have to have security that's how you get robustness and from the there security really is a combination of kind of good engineering practices rigorous audits security and safety features and you know I think if you look at it we have an array of all of those mechanisms we've been we've been working through kind of investing heavily in it and we'll continue doing so uh your mute what are the trust assumptions that one needs to make when using XLR and so maybe like to put this in context you

know that you know there's obviously like different types of bridging Technologies I think like on on one side on one extreme you have maybe like the super high trust uh bridge and you know you you gave an example right like a a multi-sig would be um a super high like high trust bridge where the assumption is that you trust you know these participants and then on the on the Other Extreme you maybe have something like IVC where you don't trust the there is no Bridge effectively you trust the security of the other chain because it you know there's a chain that is verifying another Chain's consensus um axelar sits somewhere in the middle what trust assumptions does one have to make um and that's like different layers of the stack let's say of the consensus

stack yeah that makes sense yeah I mean I think um okay let's say let's first of all I think like separate right like trust assumptions of the like the underlying security model right and Trust assumptions that come with implementation because um you know I think like for instance if you if you talk about IBC right uh while kind of in a pairwise world right you have the ideal security right where everything is verified through the light clients um you know if you were to implement some of this logic like let's say like in in smart contracts then your sort of trusted Computing base and the trust assumption is going to be that this logic that validates consensus implemented in a smart contract is

actually correct right which is actually a pretty big like code base that you're going to have to trust right um so acceler you know kind of like you said I think it's it's kind of a closer to the IBC model where you know from Excel to other chains that we connect it's actually similar type of like client verification right where Axel assigns messages and they're posted on the uh interconnected chains and then those interconnected chains kind of a no off the validator set of access so they just kind of verify those messages um and then on the other end when the messages are outgoing from some of the chains will connect actually validators run a consensus mechanism right that validates the messages and um

you know the majority of the validators have to co-sign and authorize every request and then those messages can be forwarded to the um to the destination chain right so so so the the difference between let's say like an IBC connection and you know an Excel connection is that um you know one-way connections are about the same you have about the same trust guarantees but the the reverse connections is because we don't want to have to parse this kind of complicated like client logic then we run a consensus right at acceler and then you then you relying the um security of the acceler validator set right to authorize the messages and uh finalize them and I guess like the other point to to mention though

um even though with things like IBC you actually get uh strong Security in a pairwise world but most of the time you know when you look at networks that will have kind of a millions of connection down the line you are probably going to be trusting some intermediate Network to process your messages anyways right um and you're going to be relying on some intermediate message because you know we can't have kind of quadratic number of connections across a million chains right so you're always going to have hubs um can you explain that a little bit yeah so let's kind of that's something I'm not actually I wasn't actually aware I hadn't considered yeah so if you take a step back right uh you know I am a strong believer in a multi-chained world where we'll have you

know thousands or you know millions of chains down the line right and then you have to understand what are the possible topologies in this multi-chain world right then and consider two extremes extreme number one everybody connects to each other directly okay yeah in that extreme you're going to end up with um you know kind of a quadratic number of connections across everybody you know divided by two um which is completely impractical to scale right like that's a astronomically High number everybody is going to have to manage their own ABC channels relayers and so on and so forth right uh so you know the internet is not connected that way no large Network can be connected in an indirect connections okay so and then you have the Other Extreme where everything is connected

through like a one Hub right and that's another extreme again like uh you have strong properties that you know every connection made to the hub uh is automatically connected with everything else but you can have you know a million chains connected to one chain right like you know this is um it's still going to be hard to manage so the reality is that probably going to end up with the hybrid model right with a mesh where you have some direct connection some hubs of Transportation so-called right and whenever you then communicate in that model your traffic will probably go through some type of a hub or network that you're going to have to trust anyways right um and so that's you know in that model

then you know you will probably will be taking assumptions that these Transportation hubs are sort of secure in the middle that process your messages you're going to be trusted it's validator set um and um um you know in that case you actually the security that you get will be pretty similar to what you get with Axel today anyways okay interesting yeah I hadn't considered that as as the scales like that um yeah the the sort of as we have more uh Sovereign independent chains that want to connect to other chains via protocols like IBC you just end up with this massive number of connections where every chain is wanting to connect to every other chain and you have like this

kind of quadratic uh expansion of the number of of connections and um I mean I guess in the real world that's also not how you know if you look at something like um networking you know like it like the the backbone of the internet we don't have every computer connecting to every other computer uh we do use this kind of Hub and spoke model where your computer connects to the ISP the ISP connects to some sort of um what do you call it like a um level three traffic routing Center and then those connect to each other and so we do have multiple hops until you get somewhere else and I think it makes sense to look at blockchain transactions also um as evolving in in this uh in this direction

um yeah yeah and and that's why like right like I mean the way that I think about is that we're sort of building these kind of you know airports and air traffic control system right where um kind of it's the network that's been designed to deal with you know security compromises to deal with um you know interchange communication uh and so all the protocol features of this Hub uh then kind of optimize to deal with cross chain messaging right to guarantee strong liveness robustness and you know ux properties uh and um underlying connections we actually support you know IBC right because Axel is based on a tenement change so you can connect IBC if more chains you know support ABC great will connect it but

there's a lot more than the security of cross-chain communication comes down to a lot more than what does my connection between A and B look like right um like underlying protocol can continue improving and they should continue to improve it and the verification should continue to improve it but when it comes to really Global cross-chain Communication it's really about thinking about like what are the network properties we need to provide to deliver those better guarantees at the end of the day and it's it's Way Beyond like what is an individual like pairwise connection needs to look like okay so what changes what changes XLR currently support natively yeah I mean uh I think over 23 chains have been live right uh ethereum Avalanche you know polygon uh Moonbeam

and sort of polka dot ecosystem osmosis Cosmos Hub secret Network um you know stargaze went live recently as well um the kajira is live junior is live so yeah you can go to kind of xloscan.io and you'll get a pretty good look of what's been activated okay yeah there's like I feel like the number of changes growing all the time um and yeah I want to talk about the um the um the network architecture and so the you know looking at it from maybe the the lower lower layers of the stack right and the the network itself so the validators um can you talk a little bit about what

their role is what they're actually doing so I understand there's like some multi-party computation that's going on when there are um where they're signing transactions but then individual chains are also verifying um uh transactions you know in their in their own individual chains and then relaying those to actually can you kind of describe all of these interactions from the bottom up yeah for sure I mean I think we actually have a pretty nice tag diagram let me see if I can put kind of quickly pull it out and walk through it I think it'll be easier um let me see yeah can I share my screen quickly I think you should be able to

okay let's try that [Music] there are you okay can you see this yes yeah awesome so yeah if you go to our blog actual Network blog and there is a kind of an article on the introduction to the Excel Network um so high level this is how things work right um we allow developers to build anywhere they want right meaning that you can deploy your application let's say on ethereum and you want to go cross chain so how could you go across chain we'll

expose an API essentially and an SDK that allows you to talk to what we call as a Gateway okay a Gateway you can think of it as a local kind of a router that sits that allows you to send messages to other chains okay so same as like at home you have a router you can send messages to the router what's going to happen with those messages you know infrastructure underneath it will take them and deliver it anywhere they need to go right and so in the same way sort of these gateways are kind of objects or smart contracts that you can send the messages to and then underneath it there is uh what we call the cross chain Gateway protocol which is uh the protocol that Axel is based on is uh it effectively does three functions so a validators collectively

observe these gateways and then vote on events from them right so they run a consensus protocol uh to finalize the state of those events um subsequently when they reach an agreement of what needs to happen with that event they um you know run like a monthly party protocols to collectively uh co-sign and authorize a message that then can be relayed and posted to the destination Gateway on the destination chain right so in a simple example if you're talking if you send in tokens from ethereum to Avalanche then the message to the Gateway is you know send some usdc for instance right uh that message is then finalized by the by the protocol

validators collectively sign it produce MS message that anybody can take from the Excel Network and relay to the destination chain on Avalanche post on the Gateway and from there kind of the application logic kicks into you know transfer that token to the user wallet or invoke an application call or perform a cross chain call um and uh things like that so that's that's a very high level okay that's really clear and and I'd like to maybe stop on on each of these uh each of these layers and and dive into each aspect so um at the application layer you know the application layer I think we could think about um you know the basic level will be something like axle or satellite but

it's like some system some kind of token transfer mechanism but at the most more complex layer it could be an integration and osmosis for example or directly in the application you have the ability to send tokens or make contract calls to another chain so think of it as like you're in you're in osmosis you want to withdraw tokens uh to ethereum uh that application is what you interact with and it is where you choose say I want to send these tokens over here but it could also be something more complex like hey um I want to interact with this contractor I want to interact with some application on another Chain by perhaps voting with my tokens uh or you know doing some other sort of you know more like messaging style transaction rather

than a um a a a token transfer assault transaction um the Gateway these are contracts that are deployed on each individual chain is that correct yeah so for you know um chains that have more contracts there are smart contracts for chains that don't have contracts directly let's say Cosmos some Cosmos chains right then it kind of goes through an ABC channel so it's not quite the Gateway but even IBC channels they serve an account right on chain that's created yeah okay interesting and so let's let's just look at kind of evm or smart contract chains for a moment I'd like to talk about the IBC stuff first so you have a contract over on um say on ethereum or on Avalanche or uh on Solana I suppose and

that contract the is is it right that the validators in the axillary Network are um hold as a multi-sig key as a multi-party computation threshold key um the signing key for that contract or some sort of admin key to to interact with that contract is that correct that that's correct right like so the validators um effectively have a key that shared across them right that key you know cannot be authorized by any single validator uh but a majority of the validators when they you know sign a message that message can be posted to the Gateway and executed right or validated in some sense so so the network validates messages right and

then you have application logic that says If the message is validated by axoler do X right and so that that is sort of application layer logic that then do X could be anything from you know maintenance and a contract call and so on and so forth um so so the validators effectively are engaging in a big multi-party computation game at all times where they're receiving messages from each of the chains um um one I guess when a validator on the axillary network uh is including transactions in blocks what he's actually doing is saying these are transactions that I've read sort of in my kind of mempool uh I've ordered those transactions and um put them in this block and then as a

validator I propose all of these transactions for all of the validators together to sign a multi-party computation effectively that is executing the transaction on the Gateway chain yep that's correct right um and I think uh you know for those of you for instance that are familiar uh you know with uh kind of a an IBC model is actually pretty similar right you know in IBC model you still have kind of consensus the consensus is nothing more than a collection of validators that kind of collectively co-sign a block right and then you have IDC connections that you know verify those those blocks so it's it's pretty similar in this case that Excel validators or produce a block right or a batch of messages and you can relay that block to the destination

Gateway and and it would um and it would verify that the correct validators have signed it like it passes instead of thresholder then and then you you can execute it okay and and what's the threshold for this multi-party computation yeah so great question so we actually implemented um recently a mechanism called like quadratic voting right um that allows us to uh keep on increasing the decentralization of the network and keep on increasing the threshold um right now the threshold is like 60 of the quadratic votes uh of the validator so practical that means that you know over 2 000 validators at least have to co-sign you know every message um okay interesting and so I I think from from the point of view of like a

smart contract change to a Smart contract chain um we we've covered that what happens when uh dealing with an IBC chain is it just that axlr is is connected to IBC routes for all of these chains and is effectively minting so let's say we want to send you know usdc from ethereum to uh osmosis osmosis is a IBC chain and so therefore you're kind of minting uh a representation of that token on the axillary Network and then sending that token via IBC to osmosis right correct right so this is back to my point where kind of an interoperability

is really more than um you know what a pairwise connection looks like and so one of the core functions that Excel network does is this what we call as a translation layer right meaning that you have a messages that are arriving in certain format you want to send those messages to other chains that have a very different message format whether or not it's ABC or something else right so how do you do that and that's where kind of the actual Network takes this functionality and says well okay I can actually translate you know a message from ethereum to a message to ABC the way that I do it is I simulate that message on acceler right uh so if it's a token mint then movement it like on axillary and then you can you can route it through the kind of native protocol within those ecosystems

um you know from there so um in that case yes there's a translation layer that happens effectively between those IBC messages right and EDM messages and excellent access and sort of simulator that allows you to do the translation in the middle yeah so that this works for things that are fairly standard for like for instance sending sending tokens right so you you're gonna you're gonna create some representation of that that token uh on XLR and then send it send it either way but for things that are um I guess more generic where I guess maybe in the future when Cosmos chains have this interchange accounts um uh model you know chains might have

very unique messages and um like some ethereum dap could you know want to interact with the cosmos chain but it's like a very specific sort of message from that point on then there's just like um a direct passing of that message and because I mean because otherwise it just feels like axler would have would have to integrate like so many different message types when a lot of messages might be non-standard does that make sense yeah no it makes a lot of sense right like yeah so on our gateways we effectively expose three functions right now right you can send a token you can send a message or you can send you know a message with a token right um and I think those are actually like a really powerful combination of those

three messages so same thing in Cosmos um for now within Cosmos Wix and we extend the function to send you know a token but as you pointed out through interchain accounts one of the next uh you know Network upgrades that we'll be doing over the coming months is to extend this send message and then send message with the token function um to be able to you know translate between IBC messages and interchange accounts uh to uh you know to evm messages and yeah in that case actually will relay a generic message right and it will be executed at the destination Cosmos chain um there are certain functions that um people have asked us actually is like oh could you actually put this you know

message type or like this Logic on axle to make those things easier so we don't have to like deal with on our destination chains yeah um so I think that would be actually super powerful um you know where effectively you can have like a combination of some logic executing like an excellent to parse some message and then some logic execution at the destination chain to you know to to streamline the communication um how does um axler deal with different finality finality assumptions and you know we've talked about this also on other episodes I guess now like with ethereum being proof of stake some of the more complex finality assumptions maybe are you know around more probabilistic finality or or are being

phased out but you know I think different chains still now have different finality assumptions um how does it deal with that yeah I mean on a high level um axler implements what we call as like a finality Gadget that can be tuned in depending on the you know parameter of the network to to finalize transactions so um the validators for every change decide what the finality rule looks like for evm chains you know you have to wait for a certain number of blocks for instance right of a message though for ethereum there's like a finalization tag you can call for Avalanche you know you can wait for like one block and then you can consider your transaction as finalized and so yeah depending on evm chain that server configuration

parameter that gets deployed um with um you know with the finality Rule and then you know your latency and everything else is then sort of a function of that um of that finality let me actually share this very quickly because I think this is pretty powerful as well so if you go to you know acceler scan which is our um you know explore right you can actually see different how different finalities of different chains affect latency of the Cross chain messages so it's like Excel scan like General message passing and so you can see like from Avalanche you know to Moonbeam transactions take you know on average like a little bit over a minute a minute and a half so from binance more chain it's like two minutes right then you

know kind of interesting difference from there we actually expose like very rigorous statistics around it um you can see kind of average time execution from every different chain so ethereum is the slowest uh and you know the transition to proof of stake did not help that um I think it actually really it's even slower now right on average because you're you're waiting for like 64 to um uh to 96 blocks until an um kind of two epochs have passed and so in reality the latency from ethereum has actually increased and okay interesting I hadn't considered that um so does does axelr also or well does it or will it also provide a callback so

you know if you have a application calling a message and that application wants to respond so in a in a sort of interchange queries interchange accounts uh fashion as it's been uh proposed would you be able to have a return back from the application yeah we already have Return of the application we actually have applications that are built in with it so let me yeah take a step back so I think what we're building right like is really kind of the you can think of it as sort of the you know UDP or TCP IPS or player right and then from there you can build all kinds of applications on top of it right and I think like callbacks is a type of an application layer protocol that can be built on top

of excellent right so it's not I mean that we have to provide anything for that um any application could instantiate it you know have examples for how to do callbacks and uh you know you can have non-st execution you can have sequence execution um so Excel Network by itself guarantees kind of an ordered delivery of messages right and you can on top of that build various application layer properties to get ordering of messages to get like callbacks to get you know encrypted messages and so on and so forth and so those are all like application layer protocols and like we gave a bunch of examples how to instantiate those okay interesting um yeah so I want to come back a little

bit to the security aspect and so what are where are the where are the security risks I guess like every you know there's there's liquidity being held um in axillar and you know if you go to axelr scan like you've got all the liquidity pools there is that liquidity on the axillary chain or is it being held I mean no I guess it's being held in the individual contracts on each uh chain right that's right yeah yeah and so you know I I guess you know looking at you know something like you know the Nomad hack recently which was like a smart contract bug um and there's been like tons of smart contract bugs um

yeah you know I think one of the security risks there are that uh people should consider is that like their asset the the liquidity of the of the of the pools on each individual chains like are at risk of being say hacked or something like that I'm sure you guys are doing like tons and tons and tons of auditing to make sure this doesn't happen but but it is a risk um what are some of the steps you're taking to mitigate you know potentially uh you know worst case scenario style effects yeah that's a great question so I think security right like is really a function of um you know kind of a three different things I think you have to start with the core design that's secure right so kind of uh decentralization is like the

best primitive we have to guarantee security and diversification of the validators from there you have to have kind of robust you know engineering practices and it follows by kind of operations audits bug boundaries and like safety mechanisms like you said to deal with worst case scenarios right so recently actually uh wrote A Blog on security as well um you can check out on our on our website um and when it comes to you know the latter um a you know I think we're doing like reverse audience right so if you look at it like we've done like over 27 audits um of our Network you know everything from the contracts to the back end um be uh we are um

um have functionalities that allow us for instance to kind of uh um suspend a traffic from a malicious chain right so as an example when when Tara was was going down uh we we suspend the traffic from from Terra and that uh you know prevented any kind of you know arbitrary mint um you know of those uh functionalities of of those tokens like on other chains and then finally on our gateways um they have things like rate limits right so our Gateway actually processes up to a certain bound of messages um that are token related uh per given interval and then when that limit is reached you know you have to wait until a reset um of that limit and it kind of depends on the asset and you know the volume and

things like that so in the worst case if there is like a bug on the actual Network um then uh what can be drained sir from from the gateways Gap um and uh yeah we also have you know bug Bounty program for over 2 million that's running right um and uh you know through uh immunify and you can kind of take a look at that and uh you know that's another kind of a piece of process that that we're putting in place um and and have you guys like as part of like the security um so your security practices do you have like internally do you have like a process or do you have I guess like I don't know some sort of a war plan for

like if uh if a hack does occur you know we've done all these audits we have these rent limits we have this we have like um have you done like dry runs for like how to mobilize the team how to contact you know like so that you're you're ready if this this this has happened at some point yeah I mean lots of work right like been put around those things I don't think I want to you know talk about like specifics of it but uh you know absolutely right like I think you know if you look at our demon like background I think building secure systems what we've been doing for the last you know kind of a 10 years right it's not an easy thing and I think this is a kind of one of the hardest problems to build uh securely in the space right now but uh

and you know I think we take it like seriously throughout the whole stack and like you said like operations and is a part of it and we even coach our validators uh even validators have things like key rotations that they have to do on the network right so I think this is that people don't realize but like you want to make sure that the keys are actually refreshed and updated um that protect the networks and so validators are required to do that as well yeah I mean another aspect I want to talk about here about security is the um is the is the the validator network itself uh currently I wasn't able to get an idea of this but like the acceler token isn't um isn't on public markets yet but um I'm sure does it have a price already

like do you have an idea for like how much money is backing the um the validator network I don't have idea yeah because the token is not rolled out but so it's been sort of you know delegated to some validators but you know it's going live uh I'm gonna next week and um you know I think from there kind of the market will will dictate it but I think one thing to note on that um I think while everybody surf you know talks about like the price of excellent things like that I think with things like you know interchange security right the security across different networks uh can actually be shared right where like actual Network and borrow security from other networks that it connects it can land security to other networks that may be smaller and things like that and so I think we're

going to start seeing more and more of that where like the value that we transfer kind of actually is directly related to the security of some of the networks that we transfer it from and you know we can borrow the security from them through this interchange security properties so yeah so like if you know if the car if you know the cosmos Hub after it enables interchange security say has like you know two or three hundred percent the security of the average um the average Cosmos chain or I don't know like let's say osmosis starts doing shared security uh you know you would you would consider doing share Security in order to increase the security of of Axel art and and sort of boost its security profile relative to those other

chains yeah for sure right like I think uh for chains that have a lot of traffic right or have a lot of uh you know benefit from axler they're all going to be interconnected you know one one will borrow security one will execute it in like different different combinations so um you know like osmosis for instance as well like I think there's gonna have to be some type of a shared security and cross-axle and Osmos is given how much traffic will get there um so I think it's it like the whole ecosystem is going to be more uh working as one when it comes to you know Economic Security as opposed to these diverse uh individual um you know Security based on like one token or one asset

yeah I think that's one of the most interesting things that I'm anxious to see play out over the next couple years is um this this idea of like meshed security across different ecosystems and different networks and certainly interchange security is like one step in that direction but as other chains also enable shared security yeah it's gonna be interesting to see how like those Network Dynamics play out um and and also sort of in that cross train world right um yeah I want to talk about user experience a little bit uh because I know that's something you're also kind of interested in and passionate about but um what's the what's the opportunity here for um for Mev because the the validators are in effect ordering transactions uh

as they assume come in in their main pool um you know are there Mev opportunities on axelr and have you seen any of that activity already um yeah I mean I think when it comes to cross chain you know MAV kind of gets to you know the next level in some sense right you can have Med you know sort of an axillary right and like relaying mechanisms and things like that I think inevitably MAV needs to be solved like at the application layer right um through things like you know commit and real uh protocols and will actually you know work and then kind of have worked on um you know some some designs there um we haven't seen it yet or at least not that I know um but you know yeah probably start

seeing it at some point and then we'll have to we'll have to solve it but not at the network layer I think this thing will have to dissolve at the edges um so so you are you of the approach of you of the opinion that like axelr shouldn't um you know should let me be like the market sort of you know play out or should axelr take a role in preventing nav from happening what's your are you Pro muv or or against that maybe um I think it needs to be at the application decision right I think um so at axle Network you know we can solve it uh if we need to and we actually can't solve it very efficiently because we have this you know kind of

multi-party computation so you can actually efficiently deploy like a multi-party kind of encrypt and decrypt uh protocols and Excel networks yeah uh but I don't think that will I guess even if we solve an acceler I don't think that's enough to solve Mev for applications because you're still going to have MAV you know and relayers you're still going to have Mev at the you know um at the destination chain and things like that right so it's just like we can solve it but if not enough that's why I think in a world with multiple chains with potential in multiple hubs you have to solve it at the edges um so so solvent interaction is just not going to solve the problem at the core okay yeah so let's talk uh let's talk about a

little bit of the the application stuff so what are what are people building on Axel like we've seen you know we've talked about uh osmosis already and obviously satellite has has been you know kind of the flagship application uh what other cool things are people building with XMR yeah I mean I would say like all the applications that you have seen in the space that have any type of a product Market fit I'm now being redesigned in a cross chain native way right so things like cross chain swaps you know cross chain borrowing London protocols you know nft across chain nft so like mindao launch and uh cross chain nft collection you know with axler um you know you have gaming applications and things like that so all of them are

starting to think about what does my cross chain NATO experience look like right how do I redesign my application we have like over 100 projects like across pretty much all of the categories that are that are building like in I guess you know as as you know Bridges start to become commonplace um what what is the ideal future look like you know so like right now you know if you if you want to move tokens from osmosis to another chain like you're still interacting with this kind of interface where you have to specify where you want to send it and stuff like do you think

that there's a future where you just go on an application like openc you know you've got a wallet with tokens it may be in ethereum it may be like an ethereum wallet or something that's compatible with openc it may not be and you just get to pay with whatever tokens but in the background you know the the those protocols that are executing transactions across chain making swap Etc and moving those tokens to the wallet of the you know to the destination address and then like providing you with your nft for instance see like is this the Holy Grail or like how how do you think we'll get there and yeah yeah I mean yeah I think the Holy Grail is that as a user

having some asset you get to interact with you know any application any chain with one click from your wallet right like that that's really the Holy Grail without having to think about how do I Bridge how do I move my tokens like how do I go from one network to another how do I pay multiple gas fees and so on and so forth and I think with us most as you know we're starting to get into that experience right that I mentioned that you know can you go to osmosis it will give you like a deposit address right for assets that you want to transfer like usdc you can go to you know coinbase or binance you can withdraw directly to this address um never have to touch it either you never have to pay gas on ethereum never have to pay you know gas on axle never

have to bridge anything and like you'll you'll asset will just arrive in this application so we're really going to connect an application to application directly through all the infrastructure underneath it and the users can kind of go between them so that's what I think uh is going to happen more and more and that's certainly the world that I want to see right where people can interact with an application from multiple networks through multiple wallets without having to think about how to do it really one click experience what we're after and but the competition is fierce right like you know you guys are you guys are one of many protocols that um are are building interoperability Solutions um we are automated but we're the only ones that serve decentralized and secure

and can scale to you know many to many connectivity right um so yeah well I think some people would would yeah I I hear you I think some people would also say that uh they are decentralized and can offer like you know many I mean I I certainly think you guys are like in the in the top contenders to you know like yeah I can encourage people to look at the report like life I put a report right comparing you know security and other properties of different uh you know Solutions in the in the market uh kind of um you know it's pretty pretty clear there I think who is actually decentralized who put out this report um li-fi li-fi okay I'll have to put that in the

show notes okay cool I haven't seen that thing um yeah cool and um one other question I had about yeah around like user experience how do you think we saw the fungibility the fungibility problem of bridge tokens um yeah I mean I I mean I I think the users don't have to think about bridging tokens right I think the next generation of applications will um what I call be cross-chain native where maybe the tokens are bridged on the back end but they can be you know kind of a swapped right for for the right representation or maybe the tokens are not even bridge in the back end and you just kind of sending messages to composing applications right so as an example if you've taken a cross chain

loan you can do it by sending tokens to the destination chain and taking a loan against this asset or you can send you know you can lock your token on the source chain and just send a message to the destination chain that says you know I lock my tokens in a contract on a source chain could you please authorize a loan to me on this destination chain right in that case you actually don't move the tokens You're simply moving a message that allows to execute a particular sort of user action right on the destination chain um and so I think we're going to live more and more on that world where uh for some vocations it makes sense to move tokens but for a lot of applications it makes sense only to send a message to to execute a user cross-chain activity okay so I guess in this loan example

um let's maybe try to construct an example here so let's say we had a lending protocol on um chain B you have some native asset on chain uh a let's say usdc you want to borrow some tokens over on chain B but that chain B pool doesn't support usdc directly or natively uh on chain a you would lock your tokens send a message to chain B say hey I've locked these tokens there's my collateral um I can borrow against it but you're never actually moving the tokens over to chain B you're just kind of locking them as collateral somewhere else set that's what it looks like yeah okay interesting

yeah it's an interesting model yeah and so then yeah therefore you yeah you don't need a rich asset you don't need to bridge acid right like I think the way that I like to think about it you can move assets to the program or you can move program to the assets right and like the the model of sending you know messages is really a model of where you can write your program as a message right you can move it to the asset execute an action and then return the result back to another application um you know as a sequence of contract calls or you know user user activities and so kind of traditionally by making users move their tokens we kind of turned users into sort of human routers what I call right with General message passing we can automate all of those like user

activities and write them down and like abstract them and send programs to be executed across chains so then effectively you have okay this is cool uh so you could have you could effectively have some sort of a lending pool that has collateral that is represented as claims to lock tokens on like tons of different chains wow okay I never thought of that that's okay that's kind of that's really interesting so so I think that that's why I said right like there's going to be a paradigm shift of like what it means to build a cross chain native application right where it's it's not about like which wrapped assets you're gonna list

that's not what it's about it's about architecting your application in a way that can talk to other applications other chains and send messages and kind of benefit from all of the liquidity of the whole ecosystem without having to fragment it and like have users move the tokens back and forth so this is what I call is like cross chain native deployments which is what we're working on with you know a lot of the builders on top of Excellence hmm super cool um yeah one one like kind of side question here when you move let's say you move usdc like native usdc from ethereum uh to osmosis right now you're you're getting you're getting a rap token

um but osmosis will support uscc at some point at least that's what Sunny is claimed uh well in in the future then will you just get native usdc over there like will axlr be making that conversion or like kind of having a liquidity pool there that just gives you the native one or will you still get this kind of rap thing yeah you can definitely you know just do like a swap and get the native we actually have a project uh kind of squid um building on top of axle Network that that does that right so it uses exceller usdc as a routing asset uh to send messages across different chains with that axle wrapped usdc but then on destination chains it gets like swapped for the native usdc and like the message is executed about the user now users never have to take those actions right

because they can they again like code it as a message and you just get like the usdc of the deepest liquid hmm interesting and I guess so right now like usdc is is probably like the asset that gets circulated the most I'm just looking at the total lock value and it's something like 60 million usdc uh uh or 68 million and then the next asset after is like eight million dollars worth of wrapped East so uscc does serve as sort of a routing asset um in in different ways right yeah that's right okay interesting and do you think that will continue to be the case and if that's the case like what kind of

maybe regulatory liabilities that you fall under especially now with this um all this uh talk about regulatory stable coin regulatory uh action well I mean actual Network itself is sort of decentralized network right so in the same way as you know you have a kind of ethereum that produces blocks to transfer messages so in the same way axiller you know produces blocks that transfer messages across chains right and uh you know then the network itself and kind of the properties of it um yeah kind of preserve the decentralization and security aspect of the of the underlying consensus I just noticed something here so it's

I'm discovering this total value locked page on uh on axillary scan it's quite interesting so you've got this slide uh over here so it says 8 million to evm 82 million to Cosmos is that total value logged or is this transfers um yeah so my guess I I don't actually know I'll have to double check with the team but my guess is that two EDM meaning that uh it has been minted on those you know evm chains from other ecosystems right um and then two Cosmos is probably from evm from all of the evm to Cosmos what

has been okay so a lot of the traffic from evm to Cosmos is that is that what we're seeing I think okay this is an interesting indicator this is this is interesting to see that you know we actually have like quite a bit of liquidity quite a bit more like 10 times more liquidity going or volume going from other ecosystems to the cosmos ecosystem and rather than the other way around yeah um an interesting metric uh I mean it makes sense right like because I think in Cosmos you know like every time a new chain is at it like it needs to bootstrap itself right then it needs to have liquidity in so you know I think once we make a connection like that liquidity kind of floats flows in like you have to like it is a bootstrapping effectively tool for a lot of the chains

that are that are launching within Cosmos yeah all right well let's um last topic here yeah the you know the the token is going live uh I believe on the 27th uh that's right we have people asking questions here in the chat like on which exchange will it be launched on uh I don't know if you have answers to that yet but um yeah talk about you know so the road map going forward as the token launches and how that's going to affect you know the balloter is set and all these things yeah I mean as a token that goes live right like the the token on the network is uh sort of a utility token right that you can use to to stake with validators that provide security and pay kind of gas for processing cross chain transactions so I think one one

interesting property actually of excel token is that the the token is used every time you go in Cross chain without you even realizing it right so every time you send it like a cross chain call from like ethereum to osmosis user pays gas like in the token that they transfer let's say usdc but um that fee then is taken to be paid gas fee like on acceler on relayers right on on osmosis kind of destination um and same thing for for the evm chain so um yeah kind of in the token music as a as a primary mechanism to um you know to secure the network and protect it against kind of DDOS um you know civil attacks and things like that and um yeah like like in real terms

what's what's happening on the 27th in terms of the because currently the validator set is I guess all the investors uh in XLR right or most mostly not really no no I I actually most of it is community so we ran like a test net you know for for the last year a lot of like Community validators you know everybody from like Cosmo station right to you know chain nodes figment cores um they've been running notes participating in test Network and you know when they um when they felt comfortable enough they they uh upgraded to the main Network and then you know some tokens were delegated to them to run validators on the mainland and so yeah as of 27 you know because the token is going to be widely circulated then anybody can delegate to any validator

that they choose um you know and you know then that will adjust its surf top performing validators based on this permissionless set um set from there okay and then at that point also the validator set will expand because currently I think it's limited to 50 and you have plans to expand the values that probably through governance or something like that yeah it's a you know it's a parameter right in the network um so I think we'll gradually keep an increase in it as we uh as we get more people as we get more confidence and like you know stability of the network and things like that okay well at interop we're certainly uh interested in running a validator and we've been looking into it and uh certainly will um we'll launch a validator whenever it's whenever it's possible so if you're

watching this and uh you have uh in the future and have XLR tokens and want to stake them with us uh the link uh to to that will be in the show notes um thank you future delegators um yeah and then so then so the axillary token will be minted on the on the main chain and then there's this raft XLR that will exist on the on all the different um chains that you connect to as a representation so do you plan on having these tokens also like you know have sort of like liquidity pools on those tokens um with access wrapped axillary pairs in order to um facilitate like um the flow of those tokens in and out yeah so kind of yeah back to to a point

right like I think this is going to be the first project that launches with like chain agnostic token in some sense right so you can have you know it's representation like on ethereum like an axler and other evm or chains on like Cosmos chains right you can move back and forth across them and it's sort of mobile right um yeah when it comes to you know like liquidity and other things uh you know it is going to be public and permissionless so I'm pretty sure there's going to be like Community programs that will that will uh you know continue making use of it and like various applications and dexes and things like that but we'll have to kind of wait and see yeah for sure um well thanks so much Sergey where can people go to uh learn more about XLR uh maybe

um is there any information about the token launch that that you've written about or posted and and where can people go to learn about building um with XLR I think the best place to start is just go to our website right axel.network uh from there we actually link to you know various uh tutorials of how you can participate in governance right or stake on Excel network uh through the token um you know kind of a Bridget all those functions are outlined there there's a blog post you know you should subscribe to our socials like at axella core is the Twitter handle sign up for Discord there's a pretty vibrant developer Community there and yeah get engaged awesome well thanks uh thanks for coming

on and enjoy the rest of uh of uh mainnet and uh yeah we'll um we'll definitely uh follow up I think we'll have to follow up in some time um when uh after the token launches awesome well thanks for having me it was it was fun thanks