AION: Connecting Blockchains over a Common Fabric

[Music] good morning everyone I don't know if anybody was at the bar with us last night but thanks for making it out so early if you stayed up late I know our team had a little bit of a hard time waking up this morning but we're happy to be here My name is Matthew Spock I'm the CEO of a company called new Co we're the team behind the Aon project we're based in Toronto there's a lot of us here today so you know after the talk if you have any follow-up questions feel free to reach out to anybody in the tshirt I've got one of my co-founders our CTO in the back of the room with us and a bunch of other guys and girls from the team so please feel free to kind of ask questions we've been working in the blockchain space for a number of years now I started kind of down this path in about 2014 looking at the problems in the enterprise market and figuring out what was it about distributed technologies that would potentially change the kind of underlying infrastructure of enterprises started a team to focus on that in 2014 and eventually that turned into our company today and we came to this realization that over time we saw you know two ends of a spectrum developing we saw everything happening on the enterprise private chain side of the spectrum and everything happening on the public and on the public blockchain side of the spectrum and in many cases very different types of requirements different types of innovations but more and more starting to see you know forward-thinking companies like JPMorgan like the CME Group like a whole bunch of others realizing that there's real validity in the use of public block chains inside mainstream applications mainstream organizations so we started asking questions about how do we start bridging this gap how do we pull together all these various innovations into a single solution that could essentially scale all the applications all the requirements we have for the future of mainstream adoption one thing that we noticed along the way if you were to rewind a year ago to DEFCON 2 probably would still have been debatable how many you know long-term lasting protocols there would be at the end of all this even you go back farther you know the original Bitcoin versus aetherium debate is it gonna be one chain to rule them all you know we came to a pretty obvious adoption realization that there are multiple protocols developing in parallel there's people building customized systems with customized consensus with you know EVM functionality with fabric functionality with Korda functionality with just plain old Bitcoin functionality and at some point we were gonna realize that we were running into a very very new problem and infrastructure around fragmentation of all these technologies so we decided to build a on so we're in the process of developing this right now there's a lot more information you can find online but I'll walk you through the problems that we don't think we've solved yet as an industry that we wanted to kind of focus on and these kind of high-level be bucketing into four four general categories so the first the most obvious one to us is isolation the fact that we've scaled and built all of these various blockchain networks with no real ability for them to connect to each other without centralized intermediaries which is kind of the irony of our ecosystem is we all rely on these heavy centralized intermediaries to go from one watching to another we obviously still encounter an enormous number of problems on scalability you know the status ICO is a great example of that what happened to the etherium network on that day and obviously there's a lot of plans and development being done to solve that privacy which is mostly a requirement coming from the enterprise market around how do you know mainstream financial institutions and healthcare organizations and governments eventually trust this technology enough to manage you know your personal data to say you know how are we gonna put financial transactions and health records and all of these things onto block chains and finally although it ties relatively into scalability is this idea of performance how many transactions can we pass through the network so I'm gonna walk you through our design our thoughts there's kind of two major components that we're going to talk about one is this idea of how do we bridge block chains together and the other is a new network that we're developing called a on one that kind of sits at the root of this new interconnected blockchain ecosystem we're building so this is a on high level architecture and three kind of talking points that I want to go through so one is the design of what we call a on 1 which is kind of our new public blockchain that we're building that sits as a little bit of a router inside this network of block chains also has some really kind of core features and functionalities on a on one that I want to kind of dive into a little bit secondly is this idea of participating block chains block chains that will connect through this routing system to essentially be able to witness transactions across each other's blockchain so you know one of the the very very simple problems we wanted to set out to solve was could I write a decentralized application pulled input and logic from more than one block chain without centralizing the the applications layer essentially so today if you wanted to tap into you know the etherium network the Bitcoin network and like a private hyper ledger network you'd have to centralize a lot of that logic onto some sort of web based application so could you build a decentralized application that said well if this happens on blockchain a then trigger this on blockchain be in like an autonomous decentralized way so that's kind of the the crux of our design and this idea of bridging participating block chains into a routing system and finally I'll dive a little bit more into how does our bridge mechanism work and what we call ICTs or interchange hands actions you know what's the incentive model behind actually creating these bridges and validating transactions and data across multiple chains so let's start with a on one to two major components that were in the process of developing and I'll talk a little bit about where we're at in our development timeline and roadmap we're having a pretty major release at the beginning of 2018 where you'll see some of this implemented in a in a live test Network but for now you know two major characteristics of the and when watching that we that were building one is a completely new virtual machine focused on higher performance and scalability parallel transaction processing inspired by kind of characteristics of the java virtual machine to pull in you know some of this enterprise scale that large stream large mainstream organizations have just relied on for a long time so the idea being that we you know will still be able to support languages like solidity we're developing a new smart contracting language a more creative name to be determined it won't be called the Aeon language and obviously everything would just get executing them on what we call the AVM so that's that's one characteristic in the meantime what we've done while we're researching and implementing this we've also made some modifications to the EVM so our first release is going to have a modified version of the EVM again optimized for performance so I'll talk a little bit more about that in a second probably one of the longer-term and more ambitious parts of our project is the new consensus algorithm we're designing and if you can follow this this design it's a little bit tricky to understand but you can dive into this a little bit more deeply in our white paper we're building is a delegated hybrid consensus system that essentially gives us the characteristic of instant finality on this connecting network so when we realized hey we want it to route transactions across many block chains the first problem we realize we needed to solve was that we couldn't have at every layer of routing the need to wait for block confirmations because then this routing mechanism would become completely inefficient if every step along the way I needed to wait six to ten blocks for a transaction to be accepted and then transmitted to the next network so we had to design a effective incentive mechanism in a consensus model that was instantly final so you've seen this in some networks models like delegated proof of steak with a novelty here that we're building called proof of intelligence so if you follow this diagram with me you'll notice here this is intended to be a brain and in a dollar sign so we're building essentially a consensus mechanism that has three major actors on this on this blockchain as participating in consensus one is the validator himself who participates in the block creation in fact kind of good timing that this follows the Omni presentation I BFT this would not be too dissimilar at the core of the network to something like the IB of T consensus that amy has been working on and then behind this validator is the idea that people can support validators by either staking tokens into the system or doing something we call solving running our proof of intelligence algorithm to essentially prove computation of the network the long-term goal here is to create deep neural network capabilities on top of the blockchain so that you know when we get to the long term potential list technology we could be running artificially intelligent applications on top of blockchain at scale so we're starting to train these neural network nodes to essentially prove a problem to the network as an alternative to staking tokens so we're gonna have this kind of three actor three types of validators that participate in the consensus process a lot more to dive in on the on here but I won't go too deep but you know there's also a mechanism of how do you become a validator how do you nominate yourself how do you get removed from validation if you're seemed to have been malicious or acting dishonestly in the network so we've kind of laid this out we'll be we'll be doing a lot more kind of diving in and explaining this in some subsequent papers but you can read about this online if you follow the project so this is a on one at the core of this concept is also the fact that the main network has another interesting functionality and it's this idea of being able to bridge participating block chains into what we call our connecting network one high-level philosophy when we set out was that we're gonna build a design of a network that allowed you to connect other block chains but we also had this philosophy of all chains being equal inside the system meaning I could build a connecting network but that somebody else could also connect to me and become a connecting network to somebody else so instead of this being kind of a two-tiered system with one network and others surrounding it we see this kind of evolving more is like a very complex spiderweb with everybody becoming participating networks to other connectors so that at some point you can route transactions from one blockchain to another recognize what's going on and act on as a result of those transactions so one of the the pieces of research that we did going into this was the fact that you know one blockchain connecting to another was not all not all that novel of a problem to solve there's been a lot of interesting research projects big BTC relay is a good example of people that set out to essentially solve this concept of one-to-one blockchain connections how do you trigger a transaction on the etherium network as a result of the transaction on the Bitcoin network that's what BTC really set out to sell so this one this one blockchain the one blockchain problem essentially being able to prove that transactions validly occurred on another blockchain it was relatively solvable what was not solvable and our hypothesis is that we're gonna have tens of thousands maybe hundreds of thousands of block chains at some point and then not so distant future some that are built privately some that are built publicly some that are built and backed by governments and some not and at the end of all this you know the concept of the next generation of the Internet is off people like to refer to this technology is not going to come from a whole bunch of isolated networks but of one kind of common ecosystem connected block chains so essentially we've got three different types of or two different types of block chains that we're looking at connecting so one obviously is a legacy block chains block chains that exist today the first bridge implementation we're building is to the etherium network I'll talk a little bit more about how we're doing that at the token level but also just a generic bridge about how do you recognize state across multiple block chains so this is connecting legacy block chains we're doing some research and some experimentation of how do we connect Bitcoin the most obvious block chains to connect are the ones that have like a second layer some our contract ability to be able to create that logic and a second layer above the network so Bitcoin obviously doesn't have that natively design in the protocol but we do have some ideas of how we could do that and then secondly is this idea of a on compatible block chains so we hear a lot about new token projects launching of scaling requirements at each of these individual projects so if any one of these projects has enough scale that on its own could cause you know it could be too large for any single blockchain to share with others at the computational level you know we're starting to design this concept of building your own blockchain for your own application so if you can build your own blockchain for your own application and you can have that blockchain tied into an ecosystem of other block chains you start to just take on your own scaling requirements rather than sharing scaling requirements are there with other applications on the same network so we'd see this kind of fragmentation of industry of black chains connecting as more and more people launch their own block chains with kind of pre-built functionality with pre-built modular components whether they're coming from hyper ledger and the fabric protocol whether they're coming from modular versions of the etherium protocol or building something from scratch so the idea is that as as other people build their own block chains they all become what we call a incompatible and they can synchronize into this this communication protocol that allows all these block chains to sync to each other so let me talk about the bridge for a few minutes a few things to understand about how this works there's a there's an economic model that we're building essentially to incentivize people to connect block chains to one another so at the core of all of this inside the a on one block chain is a system of registries that we use to create bridges track bridges and incentivize participants on these bridges so these registry contracts are essentially smart contracts on the connecting blockchain that act as the identifier of all these bridges that are being built inside these contracts there's essentially two major functionalities one is the fact that the bridge needs to be created by a staking of tokens after that staking of tokens has been put into the contract other nodes can also participate in that or other token holders can participate in that bridge by staking into the same contract a set of rules around the economics of that bridge would be published as part of that contract so it's up to the Creator the bridge to say my bridge is gonna have this fee structure and it's gonna be this decentralized and it's you know that whole mechanism is free market opportunity for there the the bridge creator to decide so we we envision that at some point there's going to be a competitive landscape of bridges connecting block chains and there's gonna be faster bridges and slower bridges and you know expensive bridges and cheap bridges that connect watch ants one to the other inside this registry contract all of the nodes that participate in stake they're tokens will essentially become bridge validators these bridge validators have the role of witnessing and validating transactions that we call interchange transactions so they sit commonly on both block chains that they're bridging they can witness transactions on one lightweight consensus we're testing honeybadger BFT on the bridge it'll be a single round voting mechanism on the bridge to say hey did we all witness the same thing if yes then we transmit that message to the to the connecting network and the connecting network and then reroute that to the blockchain and it's intended for the the other part of the bridge registry that's kind of critical is the fact that we use the bridge registry to keep track of the route hash and the state of the blockchain it's connecting so at any point in time independently somebody receiving a transaction could go back to the registry contract and prove through that route hash that the transaction they're looking at is actually included in the state of that blockchain so the the the registry will keep track of all these route hashes of all the block chains that connects to the interchange transaction is the format of transaction that we're designing essentially to have a common language across block chains of what a transaction format should look like so very very simply if you think about a single blockchain today all you really need to identify is who is it coming from and who's it going to now we add a second layer of who's it coming from on what network is it coming from and who's it going to and on what network is it going to so we have to create this idea of identifying block chains so we can effectively route these transactions and the the idea of a ictv which is the interchange transaction fee is essentially a layer above the transaction for you to already pay inside the blockchain that you're operating on you pay a second layer fee to cross bridges throughout the network so if you needed your transaction to be witnessed on another blockchain you'd pay that fee so over time what you'll see is you'll see transactions being able to skip from one participating blockchain through a connecting network into another participating blockchain paying fees and crossing bridges along the way and incentivizing these actors to essentially act as the validators on these on these bridges so there well we're also publishing an economic paper in the next few months that will dive into you know the economic model behind this a little bit more the token we are building another protocol that we think will be useful to a lot of projects that start off with the idea of creating an ear c20 token on the on the etherion blockchain and eventually want to natively move that token into its own blockchain coin so the first bridge that we're gonna be releasing at the beginning of next year is a token bridge to essentially allow for a burn and mint functionality native to the ERC 20 contract so if you can prove to the ERC 20 contract that you've burnt your token then on the other side of the bridge on this new native blockchain a coin will be minted for you and kind of more interesting than that is the fact that we want to make this a bi-directional bridge so that again if I prove to the protocol that I'm connecting to that I've burnt a coin then I can actually ream into the ERC 20 token so we want to be able to create this fluid concept of tokens being able to transfer from one blockchain to another without changing the underlying supply of that of that currency so that at some point you can see token economies that are scaling across multiple block chains rather than single block chains so this is something we're gonna be building and testing for our own token but as a protocol that we think will be pretty useful and available for other people to apply to their projects as they go from ERC 22 native coin at some point in the evolution of their technologies so other kind of potential applications we're looking at obviously there's a lot of talk these days about decentralized exchanges the only ones we've really seen function to date are ones that operate on top of a single network so the ERC 20 exchanges because they all operate in the same logical environment of aetherium but these centralized exchanges across block chains is not something that we've been able to figure out yet and that's because we don't have a decentralized home at the middle of all these block chains we just have these centralized intermediaries so decentralized exchanges across currencies and tokens but like a real decentralized exchange it allows us to connect Bitcoin to like going to ether to any other blockchain that has a native currency interesting to you know our enterprise clients where we started testing this is this idea of hybrid public and private blockchain applications where you can start writing some of your logic and keeping some of your data on private chains and if some of your logic and some of your data operating on public chains and still be able to build these into a single application so a simple example we often use when we're talking to our enterprise clients is you can imagine having a health record managed on a sink on a private network of hospitals an insurance policy managed on top of the etherium blockchain and a payment being triggered in some sort of banking Network an application built there would essentially require that I'd be able to trigger actions on all three of these block chains pretty autonomously you know I go to the hospital my medical record is updated my insurance policy is verified and a payment has made all of this being built in a decentralized app that sits on the Aon network connecting to all these block chains so that's kind of this idea of decentralized applications across multiple block chains we also see this as a really interesting solution to scale aetherium you know as we see lots of applications that are running into scaling problems as they go through something like an IC o---- we're really all the etherium network needs to know is the final state so if you can run all of your transactions on another blockchain and and just simply record the final state at the end of your sale at the end of your your application back to the etherium network you can use this as kind of an on ramp and off ramp for aetherium applications as you're developing them and then the dream is Unchained artificial intelligence this is the idea behind what we're doing in proof of intelligence so that you can start to create you know machine learning functionalities that can be called by smart contracts on the chain without having to take advantage of centralized AI systems that are being developed today so roadmap this is our first release coming up in 2018 what you can expect to see there as well have a we'll have a live network operating likely in January but you know we'll see engineering roadmaps tend to be a little flexible modified EVM big change being moving from 256 bits 228 bits in the ATM design we've modified the echo hash mining to essentially increase ASIC resistance by increasing memory usage increasing computational complexity this eventually will get swapped out with proof of intelligence which is being developed kind of in parallel but needs a little bit more time still running as a proof of work system before we get to that hybrid consensus we're gonna be open sourcing our java api which is like an enterprise api we've used for a lot of our large-scale enterprise clients focusing on transaction scale and then finally we'll be releasing the ERC to a on bridge protocol and here's a look at our roadmap over the next kind of year and a half to two years both on the research layer what the papers will be releasing and in the releases of all of the functionalities we expect everything on our white paper to be released by early 2019 but with live iteration so the network kind of going throughout that time so that's oh and then just quick update on progress for those of you following the project everything's been is in development right now with the exception of the a on virtual machine which is still kind of in its research phase some of these things are you know 2 month roadmap some of these things are a year and a half long roadmap so we've got kind of pieces of the team that are working on all these pieces of the project in parallel and we're constantly giving updates out so you know follow us online would be happy to get you guys involved in learning more about what we're doing this is who's here at Def Con so we've got I think 12 at Def Con most of them I think they're all in the room so feel free to come find us if you have questions and you can follow us at an entire network thanks a lot guys [Applause] [Music]