ZKV presents Privacy on Polkadot - Manta Network with Shumo Chu

Transcript

uh yeah like thanks for organizing such a great event um i'm going to talk about amenta which is a privacy layer on polka dot and more um so before i get started just tell briefly introduce myself a little bit i'm a co-founder mental network it's a fun journey to like serve this crypto and blockchain world i was a phd phd graduate student at university of washington i then i went to augurin for one year uh then i did some like a zk research at ucsb as a assistant professor now i'm like a full-time inventor and this is my twitter handle um so i saw some really interesting discussion about

how to use dk in even in a broader sense right for example gid uh like verifiable computing computation option so then i happens to did a little bit work in this realm basically at my user speed time i actually wrote optimize compiler for verifiable zone launch neural network inference um so here i think um basically for folks who are interested in doing this kind of stuff um i think i have a good news and bad news i think good news is that this is definitely doable um the reason the fundamental reason here is that uh i actually can see a ken moore's law in terms of efficiency of

the zero knowledge proof systems right so if you think if you think about like 10 years ago where are we at right this is a bc tv like alessandra chiasso michael wolfish baron parno all these kind pioneers building like a super experimental uh like a zero verifiable computation scheme using zk at that time right it's way it's almost like 100 100 x slower than what we have right now i do see um this kind of moore's law of the zero launch proof computer will actually continue so that in the in the future it's definitely doable uh but in the short future so i think this kind of thing is still actually facing a pretty big performance program especially on the

approval side right but um with that said right so let me introduce manta and what is mental i think mentor trying to be the privacy layer for the entire website and uh in particular we choose to be a pokedex and kusama pershing right i think uh a lot of a lot of people like uh um uh like uh uh sota derek and uh robert have already advertised what is the value proposition being percent i would not try to be it goes into details here but just a few key points right first is that it has a very high performance consensus right especially both in terms of throughput and the

latency and secondly right so it has a pretty performance uh wilson runtime despite the awesome runtime um if you add some like zk horse functions well it makes the zk performs even better right it's already faster than almost all the runtime here and third of course uh it's a native the xcn the cross-chain communication mechanism really uh empowers a lot of powerful primitives between different parities so um i can also have an energy here right so people's the normal kind of crosstalk bridges it's really just about sending coins over another chain but if you actually dive deep into the

xcm right it's actually much much more general and much more more grand in the design so for example one analogy i always use is that xcm is kind of like rpc among different parachutes rather than just just simply uh kind of um extending different tokens here and there right so rpcs actually enables much much more uh like a powerful constructions here like one of the examples that derek just uh demonstrated right i believe in the future there are going to be more kind of uh creative usage of xcm that will power for more interesting applications and third but also very imp last but not least right

so it's have a it's a very powerful and very modular substrate uh developer framework that net you customize a lot of things right also um to to really i think for any builders right if you if you want something more than a smart contract you should you should choose substrate that's a no-brainer actually um so with that right so what is manta so our first uh product is meant to pay um we call it byot private payment basically uh it's a bring your own token private payment so uh it works roughly similar as in z cash uh it's a shared payment protocol however

um different than the cash for the first time we has a multi-multi-asset payment protocol uh byot means that uh as a privacy provider right we don't put any judgment on which kind of tokens you are using it doesn't have to be the z-class native token it could be i think right now it could be any kind of a perigen or related token within the pogba ecosystem and in the future with bridges it can be any tokens that our chain connected to so how does it work right from the user's perspective let's say if you have some public crypto tokens like dodge ethereum btc and we there is a mint process this happens in

manta um so basically we will mint your uh puppy coins to private coins we call it the shogi token inventor and first shared token you can do free transfer to another shielded address right and when you want to leave the shielded token to public token um for uh use other stuff right so you can reclaim and get back to the public crypto public crypto tokens right so uh so this is um so this is really um i think the core design here we want to have is that we want to the users have exactly the same user experience than using like uh uh using public tokens so

for example this byte underlying we are using a utxo model right in our wallet we are providing exactly the same user interface as account model so we basically handle all these kind of very nitty gritty technical details and to any user we just use uh it just uses our web interface with designer right it's pretty easy to use now let me double dive into a little bit deeper to show you how does mentor work right so basically mental is a peritoneal rental and on top of the perishing runtime we build a pallet called the mental pay palette and so instead of mental pay pallet so there is a mental pay uh like a zero multiple

circuit and this is this is actually powered by uh one of the most uh powerful zero money from system called flanka right and this circuit will generate a zkp verifier so when user starts a transaction right the user first um instead of sending public data on chen right the user will use our zkp signer called dolphin and basically uh sorry i think i'm getting kicked out uh give me a second my best i'll get that thanks uh yes

my bad forget booking the right time of the meeting rooms um so uh so basically uh instead of sending public data on chen the user actually will send um user signer to sign to sign our secret data and general zkp and the data gets sent to on-chain is the pi transfer which is a zero launch approach right so instead of kind of every uh validator do kind of the dom thing to verify the public data uh it will actually verify the narrative proofs so that your entire privacy is guaranteed um so one thing i want to say here is that kind of our value proposition is to

become a like polka dot in kusama ecosystem enabler right so i think we're going to connect to all the different chains so for example um like manta plus moonbeam we want to support the private erc erc20 and xc20 so that you will have your privacy for your xc20 assets like mental plus alcohol we can support uh private stable coins elemental plus interlay we can support private btc and as for example mental plus fallout we can support private e-commerce and many many more right i i can only need so far um so um so that's the first thing and second thing is that we are also committed to do

uh to the multichain plus cross chamber right i think xcm is really really nice it brought us all the connectivities to this uh parity and relation asset at the same time uh we were enabled like decentralized bridges to connect to other other chains for example ethereum avalanche bs bsc polygon you can name it right so it's great that folks like nomad are developing like truly decentralized bridges so that uh we also like hide cheap gatsby bridges so that we can uh we can use them actually to connect a broader uh token world right so i think um um so actually instead of just just

showing you here right us because we are the actually the first project um to actually run a zk snark on substrate and uh we in last from the last december uh we actually started the um test dolphin test net uh instead of actually uh just uh showing a slide here um i think actually better uh i can show you the uh runtime demo so can you guys see my screen um alright so this is um uh this is a um this is the dolphin test

net right so it's running a substrate runtime so for example now you have a test polka dot right so now you can see this is in the public realm and so to convert it to the private polka dots right it's very easy um so first you need to install the dolphin signer i already installed here right you can see uh this is this is also the latest version um so you just say hey i mean first i will switch to a private right you can see now i have like a uh four thousand and one hundred and seven private polka dots right so then let's say if i want to mint more

um i just click 2000 right then um i click to to private and then deposit right so this is what prototype uh prog.js and i use the protojet to sign the transactions um is being processed and you can see you can see the resulting uh subscan support for dolphin right so to to give you a better idea right so since so this is actually

real-time on-chain data right since our deployment of the test net right so we already almost passed 20k transactions uh which by the way for to give you a comparison right so this is actually more than like a one years of tornado cash private transactions and it kind of really demonstrates uh that basically um because we are have a very user user-friendly privacy project we're getting actually more people adopt on privacy right so to show you some like unchained data right so uh let me show you one one example of

the private transactions so for example let's look at this private transaction right so you can see right so the core part is that all this kind of unchained data doesn't leak kind of any of your information right the only thing you had is a piece of zero laundry proof which exactly is uh like a 192 byte of data that can you can think of kind of like encryption computation that doesn't leak any of the doesn't leak the source the source uh the destination the amount and and and so on right um so that's um that's a short uh demonstration of that dolphin uh welcome to try that and

uh get some test tokens from our discord right so that's um so that's um that's where we're at right now so then i will i would like to share what we're actually moving forward right so i think we're going to have a pretty uh big improvement of the next version of the manta which is mental pay 2.0 test net and and uh this will be uh this will be launched by the end of the february or like early early march so what's the difference the first difference is that we actually will enable reusable addresses right so this is a big big feature bump from the previous uh test9v1 essentially it will be even more user friendly

and secondly we're going to enable view keys for selective disclosure also for regulation compliance so for example uh if you want to put your tax returns right so you can just um use your viewing keys and generate the report right so uh so it will be easy also it will be uh like a it it will be like a action regarded by xero under proof right so i think this is a uh this is basically the kind of privacy model where we're trying to promote right so uh privacy doesn't mean that we you don't you don't want regulation complaints at all um the last but the least the pre pretty exciting thing is that uh we're going to have the 10x performance boost on the

approver right so so basically our in-house zkp team works very hard to shave all the uh to shave all the constraints so in the new survival dementor will only be have well only having like a one tenth of z cash disabling uh constraint comps uh which is actually really awesome technical achievement and to translate to actual performance right the prover side will have much much shorter latency than previous um so last but not least i would try i would like to advertise the underlying zkp system that powers manta right so basically uh this is a joint um like a this is a joint effort with many many folks for example folks from uh

poli-chan uh folks from elio and also folks from uh all the different like cardano all the different places right we kind of want to create a public good zkp system that benefits for everyone right so there are a few design highlights i think first is that it's a universal snark which is a plunk right so it's uh doesn't have to you you don't have to do triceps per second right and secondly it actually uh uses our clothes curves right for example bn like dis 381 and bis377 right so i think uh upward curves uh like jeff perfect mentioned all really prove it uh provide a really powerful primitives for you to build zk systems

so uh in addition right so it's compared with kind of like a first generation clunk it also supports advanced features for example some of the features has has been already implemented for example customized circuits uh or like customized gates right this is kind of our secret source to be kind of like a b to z cache performance and also like lookup tables and various other techniques and there are going to be more added in the pipeline for example cost based optimization also like a star configuration basically you use the faster ridge solenoid encoding to to make it start also makes the prover prover speed much much faster right i think um with that um i'm basically happy to take any

questions thank you um i think there are a couple of questions in the chat which you just go over quickly whilst anna and yamo come up for the panel right um so i don't know if you can read it on this one here it says why is the sign a download required this is from i think that was from the demo oh great that's a great question i think in a long time we want to get rid of designer but in the short term uh it's really because there was an improver speed is not there right so um and native speed is about five to six times faster than five to six faster times faster than the

wilson so that it cannot be a full web front end so we actually did some benchmarks right so if we are using the watson it's about two minutes latency i think for some people it's okay but for us i feel like we basically feel like uh like we shouldn't wait ask the user to wait that long for sending a shielding transaction and i think uh like i just said because we're keep improving the zkp performance i think for sure that we're going to remove the signer one day so that you have just like a full wasn't