Turning ZKP Potential into Real-World Solutions | Mary Tr, Orochi Network | ETHTaipei 2024

hello hello hello hi guys I guess I am a few girl talking about zkp today uh so my name is Mary I am a co-founder of oroi uh we've been building oroi for almost three years and thank you organizer for having us today talking about ran pasta a new coding uh folding conditional folding scheme paper and my um problem solving is cor with the database all right my topic I try to pick the topic because I want you guys to like observe the way we actually turning from the paper into application all right we have uh looking back a bit about the current state of uh zkp in general so apart like initially zkp is used to build up a verifiable computation but then many people try to uh you know scaling up the ZK uh the blockchain because blockchain develop quite rapidly crazily so we need a solution to scale up and people working on many topic to uh you know actually try to scale blockchain and or uh proving system so some uh some people working on Z love and one of the biggest problem of Z love is proving times and the parallelization so in order to let the calop uh actually being practical we need to solving the um verification times and also the pro size and also someone working on verification time and B side the thing with this is um it need to be optimiz because if we want to being practical we need to optimize the verification time and the bsite too and then someone working on very verifiable computation to build up the ZK VMS or ZK CPU but um ZK VMS or ZK CBU itself is considered like kind of a zanic circuit and it causes a lot of difficult for for development and the security as well and then um I mean like in in the market for now people trying to build uh ZK VM ZK uh CPU and then for my team we choose another approach by weing uh we soling the data correct next let me moving that I guess everyone here quite hungry right I'm just Landing 30 minutes ago and from my one of the observation of my team from the uh zkp State out of the market we developing ran pasta it's a new folding conditional folding scheme for the first time you can proving Ram program in parallel efficiently and also our conditional folding scheme in Ran pasta allowing you to folding non complete by n three it's more efficient than folding the complete B the tree also um you can see that people try to solving the scalability of blockchain by man Way by vertical way and then uh many paper and so research had develop for but then if you uh I think it's already D to the upper threshold of scaling by vertical so our team we thinking to build up and to scaling the proving system by horizontals by horizontal scaling side so um with ram and pasta with our conditional folding scheme actually we can allow you to folding with non complete bory we can allow you to proving ramram in parallel and this is a first time ever you can Prov in a parallelizable verifiable ram program also with our conditional folding scheme we allow you to build up the IVC which mean at the same time you have IVC you have PBR RAM and also uh from those approach and from all of the characteristic of ramen pasta we think that it's going to be the very ideal solution for us in terms of solving another problem of verifiable computation is about data coric and uh let me so next slide so for what happen with data correctness because uh normally the comput know of data is just like garbage in garbage out and it's going to be waste of time at it for if you just like solving like keep solving the uh inaccurate input in the beginning so that's why we solving the uh data cic by our ZK database and first time ever we proving we providing the provable data for web three all right uh what do you think about on gen data availability this is a big question right and um in order to answer this question we looking back to the transformation of onchain data and the vable data so what happened with onchain data normally at the moment still people a lot of people using uh blockchain to storage data but lochain is a synchron system it cannot soling a synchron event uh like that and also at the moment people still like fitting the data from the smart contract to the and also uh it's kind of familiar concept with our we three people but for traditional people for web to people oroo is still somehow unfamiliar so we need another approach for people to actually interact with the um web three database which we call provable data and SEC datab more efficient and easier with vable data we actually use blockchain to store the commitment only the commitment and we fitting data directly to Smart contract with a zkp instead of fitting everything on blockchain and of course the data can be fit with a zkp see the concept and we also have a comparison to see the um advantage and disadvantage of onch data and improvable data so we can see that with unch data the pros and the cons of brch data very obviously so it's more expensive because the data have to be store on chain and also it's slowly because it's require the corpor transaction to get the uh the the data necessary the necessary data and so we cannot like eliminate the third party trust why with the probable data it's cheaper because the data store option is just commit with the ZK approver and also it's faster because it just require one single transaction to fit the data and also we completely eliminate the third party trust in vable data of course uh going with the the pros also we have the cons compared to the onent data data on data for now somehow it's more familiar because the market already get educated uh somehow it's more familiar is it like like this familiar to uh the concept of on data but for rable data we might it sometime to actually educate people how necessary of rable data all right so before that um let me oh no no not this all right uh I'm going to talk a bit more about the basic concept of ZK database uh which mean we build the first ever verifiable data pipeline what it does and for zik simplify we need we just query and update the data fitting to from data and with zkp from the app to the smart contract that directly and with the uh fitting is with the kind of a commitment we can use some commitment scheme like K to uh stor the on data and make sure that no one can kind of like insur or changing the uh data processing all right with uh this see we actually having some Trac 3,000 install of the database meaning that actually it's very necessary and make it fit for for people so inviting everyone if you are also want to exploring and having the same approach as uh solving the data corrent next you can come and talk to us and having using this uh we allowing user actually like friendly only for developer um managing the databas and solving everything regarding rable data more easier okay so what's improvable data um I think V can come up with a concept of proing proof of everything so in ZK database we have the concept of proof of everything of uh regarding data right the provable data meaning we do the sampling data sampling correctly to processing the data correctly and to the lock up R and transform the data correctly to also otable in storage and this is the another C of basic concept of verifiable dat pipeline so the information in the real world information is the raw data and uh we are going to make sure all of the process of the data is verified by the way we like uh for example with the verifiable sampling that the data is going it's going to be transformed to the construct data by uh verifiable sampling and then from the construct data we're going to make sure sure that no one can CH in church or like having any in Che into the data processing and at the end everything will be uh verifiable data P like this way uh all right we talking about application of Z database sorry um this is some application of Z datab actually we are providing provable data because I believe that uh in the near future when we about data it's not just data or raw data anymore it's about provable data and we can providing this kind approvable data for anyk for thef by the way we Pro providing very friendly dashboard and uh we also becoming aiming to become the data availability layer for any blockchains any smart contract platform all right uh also another like application dkml verifiable private computation and of course um our main purpose is we we want to use the ZK database for high performance thef and on on onchain games require like real time interactive uh right what's our next step I think I talk a lot about appli and our release product and actually for from this year we planning to to to bring Z kabay as a service for everyone to use it and for now you can totally visit our ori. Network and download it and using that with our uh Z we actually can delegate in proving so the server side proving on high performance computer without sacrifice the decentralization one of the very interesting characteristic of uh Z database all right again as a Serv for SEC database and we providing provable data for multiple platform uh more about our teams actually my team with 20 people uh my on the team have four people we are working on reman pasta uh the new folding uh conditional folding scheme which allow you to um folding nonbinary data all right and then we going to have uh we also contribute for a lot of UR work like Microsoft Nova you can Che it here and you can visit our grub to uh interact with us and you know if you if you are looking for some play to uh talk with to chat and also to contribute regarding uh what we have been doing here we are welcome all right uh yeah our slogan is just secure scaling web three and this is our Twitter you can follow any question please [Applause] so we have five minutes qer time does anybody have question for Mary or for Orochi Network okay thank you Mary