EthCC 2: Reto Trinkler

um which plans to integrate with polkadot but today he's talking about his new project catalysis which is a financial potential pair of chain that's going to be working on derivatives and a bunch of other stuff so yeah give it up for red oh okay thanks very much great yeah so it was great to be here so yeah as mentioned I'm one of the co-founders of catalysis next to Bruno Franco and Sophie Rademacher and so essentially catalysis is a standard framework for financial contracts so it's it's a domain chain built on substrate but let me just okay have a look at the time good okay so so like the overview of this talk is essentially catalysis the blockchain then the consensus algorithm which we have been researching which is called called Albatross the virtual machine which is called and so the kind of functionality that you can do with catalysis and finally the toolkit that that we are building so maybe to start with like so this this is incredibly research heavy but incredibly not development havoc heavy so like just yesterday like we published these three papers that the paper of catalysis the paper of Albatross in the paper of and so you can have a look at catalysis calm but yeah so the the implementation is is lacking a bit behind trait so so yeah so essentially the the the basic building blocks of catalysis is it's it's using substrate core and essentially like it blocks in the content algorithm and the run time the run time is essentially a bunch of contract templates I'm going to touch upon that in a in a moment and it's using an Oracle essentially here we're collaborating with chain link to essentially use you stare Town Crier and finally like this is all interconnected within polka dot so that obviously gives us more assets to like to derivatives on and so on so but like first I got an explanation of this I guess slogan so it's not a framework for financial contracts okay what is a financial contract essentially a financial contracts or financial contracts or mutual agreements between counterparties to exchange cash flows so it's it's really all about defining cash flows I mean if you think about an EFC 20 token contract then what but this contract really does is or fundamentally defines cash flow like how does money flow from one account to another account and so essentially like there is this this research effort which is called actors and actors is essentially like an algorithmic contract types unified standards and this standard has been developed for for close to three decades now like it's its main contributor is really brahmers and essentially like what but Billy brahmers did is essentially like he looked very close in to what finance is like like that cash flows like for for any financial product for any financial structure Roenick financial contract essentially and what he discovered is essentially that it all breaks down into into a manageable amount of cash flow patterns and the occultus cash flow patterns contract types so that's really what what I guess Octus is about and what catalysis tries to implement this like this standard which essentially discovered like there is a limited amount of cash flow patterns and it just codifies those patterns into into financial contracts and so that that's really interesting because I mean if you look at chemistry like we had something similar there before essentially like but I guess what the periodic table did for chemistry is essentially like it it took it took essentially like like alchemy like a pseudo science into proper science where we're using the elements you actually could like predict chemical reactions and you could quantify things and that's really what Octus is about like it's it's it moves finance from-from something where everyone is implementing their own standards into like a unified holistic approach to finance so that's essentially that the taxonomy of Actos it's a total of 34 contracts and with like those those contracts can really be seen as like the basic building blocks like with those with those contracts like you can combine them because they're all they do release define cash flows and you can combine them more or less arbitrarily to like create virtually any financial product you know you know you know bones loans futures you know but whatever you want essentially and so yeah so like Octus has near complete coverage like there are very exotic financial instruments which are not covered but you know for all intents and purposes its its coverage is complete so yeah so like luckily for us this standard has been like implemented in a way that is very native to to blockchain so like the way these contract types have been defined this is like with state and with state transition functions and which events which really are just transactions I guess in in in blockchain world so that's that's kind of like a lucky coincidence so it's the way it has been specified it's very native to block chains and so yeah so that's that's essentially catalysis so like we use substrate core we have our own consensus algorithm which is Albatross we have our own general-purpose vm which is and so and that kind of like on Nagas on top of Enzo black we implement that the octave standard essentially yeah so catalysis enables and simplifies the creation of financial services that are non-college to us as fast convenient and interoperable then also yeah so also very interesting I guess just published yesterday is the concept consensus algorithm which which is called albatross and that really has been like done in in collaboration between I guess my company and and dynamic foundation and its main contributors or like Bruna franca and dr. possible paschall-brown and it's essentially like a speculative pft consensus algorithm so speculative BFD consensus algorithms are a class of classical consensus algorithms that have very high performance when compared to other algorithms like ppfd and they essentially have like two modes of operation like an optimistic mode and the pessimistic mode so the optimistic mode is really like the is based on the assumption I guess this does not kind of behave they're kind of like well-behaved and and the pessimistic no this is like notes or malicious so yeah so albatross is essentially a mix between pbft and and proof of stake so it uses like pbft to create macro blocks in in the macro blocks are essentially like like the mocker of the macro blocks which don't contain any transactions they they just define a validator set that validator said this is determined based on stake and it essentially has like a random number or seed and from that seed so so once we have the pbft blocks which with with the validator set and the seed we actually use the BL s signature as a verifiable verifiable random function to determine the next seat which essentially like determines which gets to sign a micro block and so a micro block is essentially just contains the transactions contains a seat and yeah I think that's it but they're the like single sign so like it's it's performance is very close to to like a like a single sign I guess or like like a like a centralized system really and yes so that the idea really is that it's essentially like rather than never trust an attitude as it's kind of common in pbft its kind of trust but verify approach so like that the authorities of the microblog sake just kind of consign away and there are assumed to be trusted unless it's kind of proven otherwise but like this this way the speed is is I guess the performance is much better yeah so as mentioned can invalidate the selection but if our random function based and BLS signature which I guess is very elegant and yes or like a short comparison yes albatraoz SP EFTPOS hybrid like compared to like I guess to known contentious algorithm on each one each one is our ground and the other is Ouroboros so our ground is is more like pbft like and Ora Boris is more like us like and so yeah comparisons comparisons I guess to our ground art it's kind of uses pbft and to the Ouroboros really it's it's POS but the main difference especially to to Ouroboros is that it doesn't have this notion of time slots which which really like in the optimistic case makes it faster than Ouroboros then the the second novel concept a gas is stable machine which is called and so other like the paper has been published and so yes and I supposed to be like a general-purpose virtual machine and so essentially the idea of answer is to pack the app logic into the state rather than into the state transition function so I guess like Bitcoin is is almost trivial example where like the app logic is in a state transition function rather than in a state whereas yeah so the idea is like the entire the Taira approach ik is is in a state if you have like a general-purpose VM then you can update and modify this state much simpler essentially and yeah so some some basic functionality I guess of catalysis really so like collateralized loan so you just have like like essentially it's just to contract types of actors which is an annuity contract and a collateral contract so like the example for an example here would be honest ones to sell or always once you put up like a house it's kind of tokenized and once who as collateral and once again alone Bob like merchants trading is quite easily done and they're like like short trading is essentially very similar to two merchants trading as well yeah same it futures ratings of futures trading same as merchants rating requires an Oracle that's essentially like done by chain-link similar for options and they are really like any combination thereof then like a bit to the toolkit so like the toolkit that we're building around catalysis is kind of like like this image wallet which is really a custodian solution we are building an exchange which is called Agora trade and finally cut the losses yeah so like the image world is really like this concept of you have a QR code in the QR code is a seat using the BIP 44 standards like we derive private keys and that works for pretty much any network so like with that file it can hold pretty much any cryptocurrency so like you can you can imagine this file as like any hiriam account which you know with one account they can hold all of these different ear c20 token with an image wallet you know just can't hold out see Stephen Pearcy 20 token but it can hold Bitcoin litecoin and pretty much any network pretty much any coins of any network and so we use we use that concept in our non-custodial exchange so non-custodial really means that in this dis exchange like we separated the matching engine with the settlement engine so the matching engine is kind of central but the settlement engine is decentralized so like you never have to give up custody over your funds and that obviously like works very well with with the image wallet because if you want to trade for example a theorem against Bitcoin and the settlement is done client-side and noncustodial this means that like like a normal user of this exchange would need to have like a Bitcoin wallet and in etherium wallet if he wants to trade the third I said then he needs to have a third wallet and so this really doesn't scale but it with image quality is essentially just one file so one file for any network so like yet the idea of us and it's it's really user friendly so by the way we you can use it so if you go to a go to the trade you can register and and you I can allow service catalysis which really like is really all of those derivatives essentially and and Andy I'd like to make everything as user friendly as possible so like all of all of these three tools that are like kind of like the building on top of each other so for example you can list like a future contract on Agora like a catalysis future contracts on Agora and like you you hold all of your your assets within the image world okay so that was essentially many thanks if you want to learn more more about catalysis like all of the papers are published like since yesterday catalysis not cool thanks hi could you explain a little bit more about this exchange so you can an example exchanging heater and a Bitcoin and it's no wallet so how do you do how do you do the swap is it an atomic swap or yeah exactly so so right now we're using an atomic swap in particular hash terminal contract but like once once polka dot is ready like we can update the settlement engine as well so that we can use like in the chain communication for for the settlement process sure but are you using the core features of the polka dot or you're implementing it yourself sorry you are implementing this exchange yourself or you are using polka dot internally so yeah we're implementing this ourselves right now we're using hash table contracts but yet the idea is like one sprocket that is ready we're gonna upgrade so you'll get will get rid of your implementation and adapt yeah I mean like a large a large part of the implementation we can keep we just need to like update the settlement engine really okay and the second thing and what's your Advent advantage of building your own virtual machine why you don't use it was more anything like pork oddities doing so it makes like so like if you if you want to do an upgrade or an update like it makes everything much more fine great so like so substrate as it is right now it's essentially like if you want to like update your runtime you essentially swap modules and like having everything in such a virtual machine like is like essentially allows allows you like to make very fine-grained changes and essentially within a transaction really okay but is this virtual machine based on and stinkbert one machine or you are yeah pretty much which one bottom I guess okay can you explain how the actors contracts are are going to work with cataratas catalyst OC they're gonna be building into the protocol or if they're there will be like smart contracts yeah so there are essentially objects in in and so in the virtual machine and really like like objects or like yes essentially they're the like contract templates I guess and that's that's really just like an instance of such a contract and that's like you you can more or less look at them as small contracts so you said you use your own consensus and it's optimistic consensus so if I am a bad actor and I try to act nicely when the somebody is validating me and I cooperate with for example many of of peers how is it going to prevent me to do bad things yes sir like there is an entire section on cannot tax in the papers yeah it's probably best to just have a look in there but that tax I mean that is like a range of attacks but the other the paper is very detailed so so how does a native support for Oracle's work or how does it look like sorry how does native support for Oracle's it was mentioned under yes essentially like we're using or like we so this is all the research right very little very little implementation but yet so the idea is essentially to use like chain like to like for a data source essentially but yeah like this is so in yeah in particularly the Oracle it's that's very little done actually on it all right thank you so much right oh I think it's great