[Music] and is also another co-founder of 1 in and he's going to talk about the technical architecture that's powering this new crosschain world. So, uh to give you an overview of all the things that you need to know about and how things work under the hood, please welcome Anton.
Welcome Anton.
Hi Kik. Thank you.
Okay. Uh hi everyone. I'm here to reveal like a high level uh architecture of uh crosschain stuff in one inch. So yeah actually we use three different layers. The bottom layer we call it intense layer.
Right now it's one limit order protocol. This layer is able to uh use user funds by their digital signatures. Next layer is built on top of bottom layer is the price discovery layer. It's what we call 1 in fusion. Uh it's basically adds Dutch options.
And then there is one more layer on top. We call it 1in fusion plus but that's actually uh atomic swaps layer. We have estros there implemented. So yeah we will start from limit order protocol. So that's uh not usual diagram that's diagram of value flows.
You could see how different different assets are being transferred like rights for those assets are being transferred among entities on this slide. And here you can see also transaction that resolver or how we call taker they can do transactions to fill orders while signers or makers how we call them in this protocol they only only release um uh order with digital signature we call it sign it order here. So yeah those who want to create orders they create them off offchain and those who going to fill them bring them onchain and yeah that's how it works. Uh moreover limit order protocol have uh some features which developers could use to achieve something uh outstanding. For example it have predicates.
So there is a basically predicate is a static call which check arbitrary things and it can return true or false meaning it would let order to be filled or not. So for example, if you would just integrate Oracle there, you could make oracle which works like a stop-loss for example, it could check price on some oracle and uh then you will be able to sell only if price will go down. You you can't place order right now because it will be filled immediately. But yeah, that's like one of the examples what could be built using predicates. There are also interactions.
So basically limit order protocol uh handles two transfer froms transfer from maker to taker some funds and then some other funds from taker to maker and uh this protocol is also notifying uh you if you need to do some pre-interaction mid interaction and post interaction those are interactions between those two transfer fromms and one more cool thing is dynamic prices. This is actually a thing which lets you write arbitrary smart contract with arbitrary logic for forming a price. This means that limit order is not only working with a fixed prices. It also works with dynamic prices. And this means that you could build like incorporate arbitrary logic there.
This could be different auctions, exponential auctions, oracles, stop-loss orders, like you could came up to anything here. And yeah, actually limit order protocol have huge variety of flags. Using some of them, you could achieve like interesting things for you. Like for example, if you want this order to be as cheap as possible, you could enable one of two options. It could uh you could allow only single field orders or you could allow only like full field orders.
If one of these options is enabled, this means that you will be uh spending 15k less gas on invalidation. It's using like bit invalidation in this case for example. And there are many other options. I recommend you to check it out. That's how it's work for fusion actually.
Fusion is uh uh limit is Dutch auctions on top of limit order. So you could see orders each have its own auction and resolvers are participating in these auctions. Since this is Dutch auctions how they work they part participate passively. So they just wait while price is being improved over time and they compete with each other. That's why when exchange rate is affordable for them already like little bit profitable they can't wait more because someone else can steal this order.
So yeah that's how they compete and we think that that's actually one of the greatest mechanism for price discovery. Uh yeah that's how usually one of the auctions look like. This is example from onchain. This was a almost 13 million swap from USDT to rapid ETH. And here you can see chart how price was uh going on this auction.
And you can see multiple fields happened in like less than one minute and this would this gave uh 200,000 surplus comparing to market price if you would try to exchange it with market order. So yeah that that's like one of the examples and also visualization what is Dutch auction this is a thing where price is moving from uh from good price for ser maker to worser price but it's getting improved over time for takers resolvers uh yeah what is fusion plus fusion plus is built on top of 1 in fusion so basically it's like next layer we have limit order protocol We have one infusion with Dutch auctions set up for price discovery and we have this extra layer which is like atomic swaps, escros, HTLC, escros all all these things. HTLC means hasht time locked contracts. So I have an interesting scheme how it works. I tried to make it easy to understand.
So I hope you will be able to easily understand it. So what's happening here? Some arrows have numbers. That's basically numbers of transactions or like interactions. Other arrows they are showing how funds are being uh transferred among different entities.
How rights to those funds are being transferred. So here we can see that user originally uh creates two things. It creates signet order including hash of the uh secret key and they also create this secret key which they do not reveal until something could happen in the future and then when resolver is uh happy with this Dutch auction price happening with this order they are bringing this order to limit order protocol on source chain you can see it's arrow number one and then funds are being locked inside the source escrow which is locked by timer and by uh secret key. Then uh call number two happens. Resolver uh interacting with limit order protocol on destination chain and it's locking destination funds destination assets on destination escrow and then user uh resolver is waiting for user to reveal uh the secret key.
So you could see it's arrow number three here. So when user is happy about destination escrow parameters and finality most of these things on EVM are being checked by one hash computation because those escrow they're basically uh create two proxies which are deriving their addresses from stalk which is hash of all the possible parameters of the of this atomic swap exchange and yeah what's happening next when user is happy about finality and parameters of destination chain escrow they are still owning their funds via not revealing this secret. So once they decide to reveal it it would unlock both of the escros on both chains because they are locked with the same secret. If user would decide not to reveal the secret for some reason then destination estro can be cancelled first. Its timer is faster and resolver funds are safe.
So yeah, that's how it looks like and that's what I uh tried to show. But the thing is that all those s source blockchain and destination blockchain you could see like limit order protocol like escrow there. Yeah, that's how it looks for some chains or maybe most of the chains. But some chains like Bitcoin for example, they do not have any like smart contract functionality. their their nature is a bit different.
They built with UTXO where outputs are programs themsel which you have to satisfy to be able to spend. So yeah, the architecture of underlying blockchains could vary, but the thing is that they are actually I think almost every chain is capable to have this HTLC functionality for atomic swaps. Since this idea of atomic swaps, it's initially appeared on Bitcoin talk forum in 2013. It's like 12 years ago. uh there were only bitcoin like chains at this moment and yeah if I think if it's if something is possible to build on bitcoin you probably will be able to run it on any possible chain yeah I'm even not talking about smart contract chains yeah so yeah we believe that actually atomic swap scheme is the only scheme which is self custody for the users uh and that's like architect aritectural step to make this thing uh like we we want this thing to succeed and be there like forever.
If something is built on technologies on architectural level uh with these things inside like atomic swaps for self- custody yeah it should work finally. So yeah, if you're yeah because working on custody solutions is something I still think it's like not very long-term vision because yeah there are like too many risks here. I mean the more uh things we would have built permissionless the better future would be because we always can build like custody things it's easier. So yeah, let me check from my presentation. Maybe there's something more to add here or we can switch to questions.
Oh, I think that's it. Anton, uh, thank you so much for for that. I think the the part that uh is interesting right now is that uh the questions are actually being taken over and hijacked by Sergey on the live stream because he's now just talking to the to the hackers and they're going back and forth on a couple of these. Uh but maybe I'll give a couple minutes to anybody on the live stream um to ask any questions to Anton, especially on the architecture. Uh the one obvious question we got is that there's so much incredible technical detail there.
if we can get a link to your slides, uh we can share that with the attendees. That's like the first question we got asked. So, um Pomic, I'll uh ask that question. I'll get you the link and then I'll get that um posted in the live stream chat.
Um
yeah, for now I I sent it to you, but yeah, we need to make make it as a link. Yeah,
we will get that slide up. Uh the other piece I guess my question to you is um as you've kind of looked into this and and obviously Bitcoin is like a perfect example for this because you kind of have most parts of what you would like in a smart contract there but with some limitations. What have you found like is is kind of working and not working when you talk to people building on top of Bitcoin and um
and they want to integrate DeFi from from EVM like what are their wishes? What other complaints like how do you think that's going to change?
Oh yeah. So yeah actually functionality for computations for variety of computations is very limited on bitcoin chains and bitcoin like chains but I think some of bitcoin like chains have more uh op codes which are working actually in bitcoin a lot of op codes are disabled for security reasons even if we do not understand those reasons it's actually not very easy to find the reasoning behind some uh decisions in the Bitcoin protocol and yeah the thing is that what I saw from the market what people tried to build this is like a different layer two approaches for bitcoin and other things but there's a huge difference between layer 2 built for EVM like chains and for bitcoin like chains because on EVM layer 2 chains are being governed by layer 1 chain actually by smart contract cont being deployed to layer 1 chain. So every layer two have a special smart contract deployed to layer one and this contract is actually guaranteeing the safety of this layer two. This guarantees that block producer of layer 2 will not steal everyone's money and all computations should be correct and signature should be correct. But that's what's not possible on Bitcoin.
So how layer 2 on Bitcoin work? they actually use bitcoin for data availability and for um ordering sequencing of their transactions. So people just put EVM like transactions into bitcoin as a plain data and this data is just being stored and order it in the right way. So on other side they just extract all transactions they can find in this blockchain they put it in their own chain as a separate blockchain and they can run it. So they're using bitcoin as data availability here and yeah I'm not sure this is something what can be called layer one because whatever logic is happening in their chain is not being covered by bitcoin in any way.
Uh yeah, I afraid that Bitcoin network with allowed op codes is only being able to work for like few specific use cases like multi. Okay. They built lightning network. It's payment channels. Uh yeah uh thing.
Yeah, we could have payment channels. We could have like atomic swaps. Yeah. I mean, but I would not expect variety of applications for Bitcoin, like tens of different apps. Yeah.
No, no way.
If the only tool you have is uh HTLC, um then you're uh kind of limited to everything you just said because uh that's kind of the the max you can do.
Um yeah.
Uh so maybe my my last question here and um and uh we can sort of get back to the rest of the hackathon for the hackers is um are you looking forward to people building some specific projects? Is there something on your wish list that uh people should have to hack on? And I say that with like a warning to everybody who's going to hear the answer. It's going to be a super technical thing Anton's going to play. So you could say the
actually when I originally saw the list of different uh blockchains of different blockchain ecosystems which have like different nature. I mean you have to learn different skills to develop software for those different platforms because they have like absolutely different programming languages, virtual machines.
Yeah,
I was actually amazed by the number of the projects because I usually work within EVM ecosystem and I am like slightly touching some other ecosystems. Uh but yeah, I was amazed it's like 15 to 20 different blockchains. Uh that's huge and I don't know who can cover this. Only hackers on hackathon can cover this. No, no any way for like any single company or a few companies to build this.
Yeah, I think we we need cumulative efforts of uh all the hackers.
Amazing. Well, and thank you so much for uh that amazing presentation. We shared the slides already with the with the hackers and uh I guess we'll see you in uh in two weeks when you get to go look at all the amazing submissions that come out of this.
Yeah. Thank you, Karthik. [Music]
Automatic transcript — names and jargon may be misspelled.