zkSummit - xDai and the dawn of stablechains - Igor Barinov (POA Network)

Transcript

hi guys to talk about text day it's a they want to talk between between before now and lunch so I'll try to be fast we have like 30 minutes I would overview what is X day and why we built it and what our plans and the well you can ask question because it's quite a tiny group of people now you can ask questions whenever you have them and stop me with questions so it's better to ask questions in the middle so X is the first ever stable point with bridge dies in a TF token that's our punchline and the its its power the the biggest pop up

economy powered by Krypton history right so pop up economy means that something happened in offline and you know everyone participated and is then very used to buy food and drinks using burner wallet which is a web wallet within the browser and the that's actually what cryptocurrency is promised us from the beginning so I think about this as an electronic peer-to-peer cash system where people used to to buy food and vendors receive these cryptocurrency from from people and on ramps and off ramps using Krypton so the when we think about X die and the this this idea for

us we we think that it's kinda eerie stable to use after you tried it and this this idea was born in on it Berlin hackathon in last year and we launched it in October in 2018 with maker tau proto fire and giveth as well data so this network so when we think about this network balances of the user user balances are stable right there denominated and I native currency is die right the usage fee is stable the usage fee means what we spend on gas for for typical transactions and the supply of the network is quite small at the

moment is like thirteen thousand dollars worth of die of native token but the available the quality of native native token is massive Freitas like I'd known nineteen million dollars something died which can be bridged on this eighteen so the model is at the moment because it's a layer to say cheese it's blazing fast and we don't call this type of network like empty network we call them networks with empty capacity they undressed the coin at this term and we think that that's that's that's very important to understand that most of the infrastructure that we have now besides you know helium and sometimes Bitcoin with is with empty capacity right we have so much infrastructure

available but not enough depth to fill all this capacity and most of the infrastructure we can see now have some form of subsidized platform users here at the same so subsidized platform usage is when they first structure is saying okey guys this platform has zero gasp you're very low gas fee or no transaction fees and so forth but someone is paying for it and for for for most cases like infrastructure the people who provide the infrastructure for you paying for this right they same with X type so the consensus itself what we call proof of autonomy for now and in a form of delegated proof of taken future is environmental friendly with proven scalable and security assumptions means that this proof of autonomy

consensus works for P way network and some side chains based on the same protocol for years and secured SS for let's say hundreds of millions worth of USD for years right and scalable means that we in this case and this is why we're on this scary conference is that people can run their own side chains based on the same protocol without compromising security of offsetting of other side sheets right and in proof of work it's quite hard to do this if you want your own network based on proof of work there is a problem right because you have to move some of mining power to another chain in proof of autonomy type of

censors you just use the same mechanisms and select a group of Auditors were like that say non affiliated with each other and incentivize to provide security for the network and the security of next sidechain is less is less dependent on own on let's say on a previous idea right so the validator said this is what protects the network from from zip latex is is managed by self governed down means that validators have equal votes within this doll and this doubt can regulate itself so accords self-governed type of networks and also this dao is self upgradeable so validators can

decide on on upgradability on this as i chain or they can disable up readability if they want so the the the idea of x day that at Zeitoun of a theorem 1.0 and it's it's a it's a interface which is easy to understand for for for existing key theorem developers with what we call dynamic validator set so what data's can unborn new validators or off board but it as which let's say malicious for the network and it's governed by this exclusive self-governed down so exclusive means that validators define rules who are validators and they might accept new validators or not based on

their own Unchained decisions and there are some problems in this type of exclusive group that there are no economic mechanisms to add or remove auditors and this chain besides of these decisions of this exclusive group so the native token of sidechain is die right and the idea of having an one token on another chain as negative token it's called hard spoon it was introduced by cosmos team back in J's Christmas and Vitaly too and we think that Agata is one of the first hearts bond network from a theorem protocol so the supply of the of the x day is one-to-one from locked died on the

mainland so when the when users lock deines in smart contracts of of X dies my contracts on a theorem network the validators of the XD chain relate these transactions to X day and this token bridge is actually a part of consensus of this XD chain and the native token has breached one to one so one day is always equal to one next day well there are possible there are possible differences and balances for example someone can can lost access to their X died on the XA side and this died on the film side will be locked forever but

when the the bridge is operating it's one to one atomic swap between two evm networks so this simple illustration how this bridge is working so when user amen here Clark is died on the etherium side the group of players relay this transaction on none and 2x dice chain and for 4x die we have something like 3 or 4 multi-sig so thought so three signatures for from for Val data's are required to relay this event from chain to chain and each lock and event means new X died on the XD side and witchburn event unlocks die on the on the on the

etherium side right there is a aggregation of this affirmations of validators on x decide and we joyce benefits of player to consensus for example bridge smart contracts are white listed on XD chain side and it's free for validators to aggregate signatures and also this aggregated signature can be related back to to the main net by any user so this way we solve so called exit problem when users afraid that they will not be able to relay their fans from Sochi because while the eaters will stop to aggregate the signatures so they don't have problems with aggregating signatures

because they run out of money in their accounts and if they run run out of money on their accounts on a video my net then user can relieve this aggregated signature from mainittu from x date to main net so that's that's a matter of of token bridge there are multiple modes of our bridging solution the the one mode which is which is used specifically for its diets lock/unlock of EFC 20 token on one chain and mint and burn of a negative token on on a sidechain so this this model involves integration of the bridge into the consensus of this 18 and this is the first type of heart spoon sidechain

integrated with interoperability solution between two networks right at the moment the the it's a market cap of this network is quite small and it's a thirteen thousand dollars and that's whatever people relate from network to network but on the other hand we can see that bridge was used by 875 unique users and it's quite exciting that you can see real usage of of the bridge between two networks and the the market cap of those networks shows actually what what is the value bridge for for this specific application of platform usage and well in future we plan to have a heart spoon

sizing with double token and one token will be native and we will keep it as a as as die for for platform usage and staking token will be a separate token which will be not visible for for for regular users and we think that this paradigm will be adapted by other projects and this way one year you know look into all dramas around level one solutions for example with a theorem and how they compete with the theorem the one of the problem is okay for you can deploy your application on on our chain and this application can benefit from you know

speed and empty capacity and the price of execution but but also you have to make transition from one ecosystem to another ecosystem with a change like is that it's it's not the same because we think about this as like extension for existing pavilion 1.0 and the native token of the sidechain is not competing at all right die is not competing with Hume about our token which is if well our first network was pure a core and we try to to make this to solve the problem of scalability by by having like different consensus with Dow which manage validator said and quite succeed

in this because it's working and the the model is adopted by multiple chains and what is different between the the POA core and the X die which is a secretariat second generation of the same idea is that the the platform usage is non-volatile at all right when you think about the theorem we have fought our token price about our guest price so it's two-dimensional volatility with X dye the biggest price is subsidized by by by validators like in almost any level 1 or level two solutions nowadays a subsidized guest price and also the token is non-volatile the decentralization assumption of this type of consensus are you know discussable we call this a minimum viable

decentralization is what is good enough for for the for the experiments that we have now and we can see that this type of a minimal viable decentralization can power like largest pop up economy like happened in Denver with thousands of people used cryptocurrency at one place and tens of vendors used to collect the scrip the currency and off ramp from X die to die from died USD well within one event right but also it's we know that and we can claim that this solution is it's a minima with decentralized at the sense that it's

it's not cross you know the command service business and maintenance mission regulation because validators are not are not able to to modify balances of users on this network right so the model it can be scaled in and scaled out so that's a beauty of off-site cheese for example and vertical scalability of Auditors can beef up their nose if they see more traffic they can subsidize essential depths and we already have this with bridge operation so breaches whitelisted on XD sites about theatres are it's free for them to aggregate signatures on the side chain so some of course can be subsidized and we plan to go to decrease price on elliptic curve and also implement changes to EVM from you know

if in one point ice road map so the the R&D team can optimize protocol which is which is also interesting when you have more than one R&D team working on the same concept and here we know multiple companies working from the same type of protocol and which are willing to continue to work on theorem 1.0 so the same can be applied with parallel side chains and this model is interesting for what what we see now that the the model of shared security layer like we have across most pocket that theorem 2.0 is interesting from from from many perspective but if you think about users which are interested in their own side chains with their own experiments some

of them are interested in having their own ecosystem and they don't want to have shared security and they want to have their own experiments for example cool protocol which launched on the network like X die so they used their own CLN token which is easy 20 token to heart spoon they are talking from a theorem to the even 1.0 sidechain and they built their own bridge their own modification of the bridge for multiple FC Twente to multiple ESC 20 which is not we not a use kiss for us but the use keys for them because they want to build something like decks on their side chain right so they used bridges they used no EVM compatible network they used a model of XD where the bridge is integrated into consensus

they used black Scout which is our open source glycine Explorer and some other tools but they launched the sidechain with their own let's say security model right they don't want they don't need to use existing security token and there was security I mean token which provide security to to the consensus and this is interesting for for many projects which already had their own in token offerings because for them it's quite hard to convince their token holders that they have to use another token and yeah that's that's what people say so can be different application specific side chains that that's also desirable bye-bye projects right the ecosystem

effects diet started in October as a community project and now we see that it's listed on state of dabs like anyone deploy it next I can deploy their own depth there are multiple wards which are supporting or is died right now the the most exciting one is burn reward because it's it's within the browser you don't need to download it it's open source world created by Austin Griffith there was also a special version for for for each damages birthday and recently Austin Griffith published an instruction how you can set up the event like a general event so a construction how to bootstrap this type of economy for it for your local event also the this model adapted by by purchased world and they

also integrate grass release which can allow the apps to use meta transactions on this chain which is a convenient way to send tokens because users don't need to pay gas for at all without having gas va0 right because if we have gas V is zero then we might have some denial of service attacks and with meta transactions we don't we less likely we have this type of denial of service attacks it's also it's supported by the export of all status and nifty world and now it's supported also on trust walls at the part of trust core so there are multiple zero note solutions which are

developing and GDI which is a winner of each Denver hackathon and the edge tech is planning to implement the token model all next time and it works with customer P seen as anything 1.0 compatible network and also it's interesting that the participation the validator said is open and wallets are most interested party in in participating and in validator said so there are there are applications from Portis from alpha world from my crypt and from Berner world to give all details of this network that's like for new applicants [Music]

with for existing for dieter's which we have now now there also a black technology company Block C deployed their analytics and you can see the spike in the middle it's even relevant but also you can see that after the spike people continue to use burner wallet and the the most interesting here that it's quite viral idea to send you know X died and just experience this solution I'm I'm happy that other projects which are working and on different stack let's say with payment channels or copying the burner world interface and ideas so recently launched dye cart is carpet interface of burner

wallet and that could that the idea adapted and I think that for us as a filmmaker system it's great to have electronic peer-to-peer care system within the video maker system and not you know within somewhere else let's say lightning well we support black scene Explorer for forex diets a full featured Explorer it's a with all features of all most features of either scam it's open source so anyone can contribute they the the Japs are migrating to 2x time because as I said it's it's interesting for two types of use cases for one is a peer-to-peer payment use case and second

is convenient to use platform for smartphone checks with empty capacity and also with matrix which lets say mobile developers used to collect about the users like a lifetime value or price of a price of acquisition of a user so this type of matrix it's it's it's easier to to collect and the easier to predict on a platform with stable platform usage so that's why we think that this model of having execution token of the platform as a stable token will be used by by other cheese and we'll be happy if this will be the case for for new platforms all right so about the future of X died so we next week we

are going to publish our new digital proof of state consensus which start with the typo it's also implemented as a form of down so it supports two token model 1 token will be a native token with the ERC xx native bridge and another will be is taking token with FC 22 you see 20 bridge so it will be two layered consensus honeybadger 50 on DFT layer and post down which is our proof of stake directly proof of stake decentralized autonomous organization for 0 resistance and reward distribution and the upgrade to honey badger will bring censorship resistant I think it's

the only one censorship resistant protocol which is available as a library at the moment so it's asynchronous which doesn't have problems of synchronous protocols like pbft and some others at it has on chain randomness and instant finality we are very proud to to implement this protocol and we recently had official review of this protocol by GPO Masson from caddesi security and yeah we think that this protocol is ready for use in production and the exit ID POS system will be based on honey badgers underlying EFT protocol

and post ow which is our zebra control and the reward distribution algorithm implemented entirely in smartphone charts and this model is also interesting because if you look into implementation of DPOs and different systems it's it's quite hard to for for for for most of developers to fork it and launch their own protocol based on existing delegated proof of state with different settings right especially if you if if you want to modify some format say of additional reward for for for validator sort of delegate us and you

want to let's say you want to take fees from the bridge and distribute it to validate or you want to produce an additional staking token if you think about this type of modifications to existing did I get a proof of stake it's quite hard right it's it's usually on on low level it's distributed across files you have to fork the decline you have to collaborate on your own fork and we saw examples for for egos for steam at and you know it's it's hard right this post our consensus is implemented in smart contracts so the good thing that will be audited as a whole thing and also formally verified and it's it's it's it's much easier to to modify cell eg code then modify you

know under rank code of the of the protocol and the protocol can can can move forward and your Depot system will be implemented in smart contracts so you can you can forward with the protocol or you can stay at some point and just wait for something right and that's something that I expected to have any theorem for four years and we didn't have it and now we see that it's transition to a new protocol so I think this will give some opportunities for developers who are staying confident 1.0 to get this idea of a dedicated proof of stake on a theorem which brings us from

from the state of minimal viable decentralization to the true decentralization because validator set will not be exclusive anymore right with delegated proof of stake which is also interesting in this model that staking token can be any talking so there is no vendor lock on someone else token distributed in the in the way if you look into some especially recently launched protocols of auditors of these DPI's protocols you can see that many validators names are like Capitol or VC you know like this most fun that validators and and that that's because of the distribution of the token this

type of protocol can be devised we'll buy with with any staking token for example one can launch this post out DPOs with tether which is most stable stable token or with die which is the most decentralized table token as a staking token and there is like less problems like with nothing could take attacker than some others and and mint their own staking token on the side chain and in some period of time let's say in a year they can switch from the like one token model is taken for two token models is taken or switch from one staking protocol to another so I believe that delegated proof of stake should be a shiny new ID which will replace coal

based networks for for the most use cases and we need to make this dedicated proof of stake concept you know working with without like previous problems people claimed you know in US and some other chains about centralization yeah that's that's basically it lets anyone want some free x-day now generates an I have yeah that's that's actually it's using committed transactions and no this one is not this one is a yeah this one if you want if anyone want to scan QR code and when it's die for free no one come on okay you get oh it's just it

just need to open it you any questions so far I will continue to generate - so you can yet well now it's 1x data so usually when I start with 1x I like one or two people now will be no not yet well we have oh yeah private key in your local storage so it's birth it depends on in which mode you are if you're in private more than yes if you are like normal mode and but but you can go to advance and copy your private key and it's burner world and yeah if you burn it it's fine

so anyone did you get one nice day okay 5x day oh that idea that we that you it's it's a world you can you can use and and forget about it later on so be it's a there is no as you saw there it's not like password creation or seed freeze or nothing like this so it's designed for for the the the fastest boarding anyone got this five day five well if you claim that someone else clintus it's a yeah it's unfortunately okay 10x day anyone there aren't a nice day

now okay so between 50x died now I invite you to go to form that view that network there is a post our subform will post our our paper there but for now you can you can you can read the beautiful description of a honey badger beef team how it works and it's a well address here in the red shirt implemented honey badger mostly with with our team and them this is one of the most complex consensus algorithm out writing and implementation and the it's complex

because it it contains sub protocols and and also we added some sub protocols on top of this complex protocol itself right for example to have dynamic validator set and distributed key generation and these in the in this beautiful writing you can link and you can get sense of this protocol and in you know human readable format so I invite you to read you know all these three parts of how honey badger is working so now let's get up let's give away 15 and one two three are you ready okay yeah 50x die so if you you know my time is also the last last

thing if it's in claiming that someone else claims it yeah if in the within the burner world you have you have exchange right and and in here you can you can offer am from X died today and from die to e to offer em from X die to die you're using bridge built by POA Tim token bridge and with three or four validators required to delay this event it says maked up ratify POA and giveth and when you offer them from die to Eve you're using Union swap right so basically the sex day to eat is without

a custodian between you and and I well but when you offer I'm from die too if you have to you have to have some gas so you have to use some well there's a gas station built in where you can sell some of your ex die and get some day oh yeah that's that's like when we think about bridges between networks and we usually think about like one one hope right Oh let's move this from one network to another in this use case we'll have like three hoops and on on 8th January we had four hopes because there was additional easy Tryon token built on on top of effects day for convenience of vendors and yeah that's that's the future future

is interconnected and multi hopes and noncustodial with different settings of decentralization but designed to work for use cases what I like about text day once again that they say K system is not competing with the film its complementary thing and it's scalable so it scalable both vertically and horizontally it's open to use with still with exclusive about data set like when anyone can apply but still exclusive and it will be migrated to more open and censorship resistant consensus algorithm and that's our plan for the for the for this year thank you guys so any questions oh yeah

well attacks are possible right on all on almost any system and and here you can tune the the number of auditors and candidates and so in our implementation of well it's so it's a long story but yeah we we basically have every epoch we have a fixed number of non validator selected for this epic and and up to two thousand well candidates which can be selected and we use unchanged randomness for now for our test methods is run Dow with consensus fault reporting's and but for in honey badger will have function random so we will not need to use and down and out of 2001 s21 deters are selected into the into the

epic right and they have in our design setting they have equality word distribution so within like this twenty each each blog they were distributed equally between each validator right so no matter how big the stake is but delegate us can start to delegate on this like selected validators and the award is separated within the award of in each individual validated between the validator and the delegate errs and we have some protocols have like top saturation point like after you delegate some amount the award of validators is not going down right we have bottom saturation

point so after a certain amount the reward of all day that is not going down between so in the worst case scenario about that as well yet but say thirty percent out of five percent of their rewards each validator we get like 25 date of each father to get five percent reward and within this reward they distribute between old alligators and the the validator I understand that it's quite messy right now what I'm explaining and the that that's how the the reward distribution is working but like what if someone will purchase all staking tokens and then we'll get all validator seeds right it's

it's possible but let's say this like let's say we have malicious actors which are trying to buy more validators seeds but also non malicious actors can can buy validator seeds and we have tortoise of magnitude more candidate seats than validator seeds and this randomness random selection will decrease chances of malicious actors if non malicious actors will put more more stakes to be candidates so it's a game and it depends how market will react I think that once such a small change like X diet would be hard to to get malicious actors right because no one cares about this in a

small type of of network for now and it's it's a problem for for money change how to bootstrap how to bootstrap there is taking token here the good thing that you don't need to boost trips taking token because you can use existing staking token but give reward to the stickers in different token which will be as taking talk in the future and give some supplementary reward for example from exit fee or entrance fee from the bridges so exit fee from the bridge is a natural way right yeah you have money like you have ax die now you don't need AG that you won't die because that you can convert the food right or two drinks so when you exit now a fee is zero but eventually this will be non zero I don't

know 0.2% something and for you it's fine to pay zero point two for fifty dollars because you know it still yeah it's it's a good deal right to get 50 die and after pay 0.2 to convert it today which is basically USD so that that's the exit fee will go to validators of the sidechain and this will be collected on the bridge automatically so that's incentive model for validate also the small GPS system while the staking token is still you know brewing and the enterprise's yeah that's interesting to see I don't know yet I hope to see more attacks on the of the algorithm so please

yeah in few days validity is our paper and we'll be happy if someone will put some critique on it yeah any questions about anything else yeah white guess price is fixed but it's more like psychological decision right you can subsidize gas price you can say yeah gas prices is zero like any net transaction can be can be accepted and you know malicious actor what can write

a script quite easier and just feel all blocks with transactions and will not allow other people to put their transaction to cheat so it's called denial of service but not just able to judge denial just consent transactions into blogs access if we have price of one one Shannon right here one gig away which is mentioned this price can be it's fixed for now but they can be increased you nobody's asking on the same key we're under attack what are we going to do the natural ways to make you no grass option like any freedom so the second idea which is implemented in some protocol is to have some like priority queue for transactions and some metering for for accounts let's say if you have an account with specific

balance you can send hundred transactions per day without being throttled right and this can be implemented well and it simple is implemented within some modification of a theorem protocol without any changes in consensus right it's just a way how auditors are agreed on my own including transactions so for now who have this gas prices is a fixed and subsidized and it's only subsidized for the peacetime right on wartime wartime Co can do different things right and here there is no there is no CEO but this time let's say wartime Dow you can think okay guys we are under attack

what should we do and they have multiple things to meeting here this attacks right increasing gas price is one priority queue and the throttling of transactions per account is second third is spinning more and more change like X day if you look into World of Warcraft realms they have hundreds of same realms right where they up to 30,000 people can play on one server so what they have more service people can move from from shard to shell basically right and them with some form of bridge where you can move all your stuff but not your friends between charts so this just price fixed to put this okay this is a baseline

where we start if everything is fine right if we're under attack then we can we can think about methods to prevent this attacks and as I said increasing gas prices one priority queues with which is you know actually started in Bitcoin right this priority queuing of transactions within a small esv and third this throttling of a transaction on on cues within the one of the ball daters mempool yeah but that if you put something zero and it's naturally people can start you know dos 0 you said 0 well here you have to pay small but this

price and also increased price from 1 to 2 from 1 to 5 psychologically easier for validates at an increased price from 3 to something right you promised free now it's one way why but that's just a game theoretical aspect of selecting parameter Y 20 violators you know yeah like what what what does a good number of our data's in a network what is a good sunny what's a good number of fathers yeah waiting for distribution yeah but you know some people refer to him about 20 21 right when I'm thinking about 20 there is a

paper decentralization Bitcoin a theorem by a mean bone in there is some further some claims about 20 volt data's if you have a forum system 21 day there's a good enough and comparable decentralized as you know current state of the theorem up to 2017 so like it you saw some numbers it's it's act as baseline and after you think okay what's how how can we optimize it lunchtime thank you