Keng: Multilayer Dapps |14th Oct 2023 | Sunway University | ETHKL 2023

next topic our next topic is uh from The Tao platform and this is a decentralized ethereum equivalent bracket type 1 ZK evm now I'm not a developer so this is all a bit foreign to me but our speaker Kang from Tao Labs is going to be explaining a little bit more let's give Kang a big round round of applause all right Kang please take it away all right uh thank you for the short introduction so my name is uh Kang from Tao uh I'm doing developer relations at Tao uh full disclaimer so today's uh presentation is actually um supposed to be presented by my teammate but he couldn't come here because uh of his passport issues so I'll try my best to try to deliver this uh keynote so the title of today's presentation will be to um will be about multi-layer depths so um I guess in in today's like um L2 landscape there are so many l2s there's like um polygon Mento Tao scroll but there there is a problem right like every single layer to has to has its like um um separated individual ecosystem and what if we could um aggregate all of these um projects together um first of all it will solve like the liquidity issues and also like from from a users perspective they don't have to switch between switch between chains to um basically interact with them and from like the developers perspective you don't have to redeploy like uh smart contracts of on on every like different L2 right so with that in mind let's uh just continue um this presentation so right so in in in the in today's like uh state of like the single of each individual L L2 actually the state of the state of each individual L2 will also eventually grow too much and this leads to higher costs um because of the storage that you have to uh spend right So eventually each L2 will also grow might also grow to the size of like ethereum and gas fees will also increase um accordingly so the scaling factor of L2 is of a single L2 is sort of like limited uh by that right and um this diagram over here just shows like how um at least how Tao shares the the state between the L1 and L2 so that um um so that you know um whatever assets that you bridge from L L1 to L2 um is validated or or not right so what do you want do what do we want to achieve so we we want to achieve Atomic and synchronous transactions across CA multiple L2 chains so it doesn't feel like it's splitting up and I think that this is like the goal of many l2s and many uh super chains out there um so what does atomic and synchronous mean so Atomic means that it uh whether a transaction um succeeds or it doesn't um for example if um like if a for for a regular solidity smart contract if there is a revert statement if if your functions gets reverted every um every state change State change before that doesn't uh happen at all uh doesn't get mined at all uh and what I mean by synchronous is that um it happens in U sequence so all of the the transactions between the multiple l2s happen in in in sequence and there is no like uh weight between the two uh L2 transactions right so the first um so-called architecture that we propos to enable multi um multi-layer depths is the multiple L2 uh architecture so in this architecture uh we ask ourselves how do I transfer uh token from L2 a to uh another L2 B right so this is a simple uh bridging example Le so first of all you you burn the token on uh l2a you wait for the validity proof and then you deposit um into the the other bridge over here um which costs like a L1 cost uh like U it it cost you to the pay gas fees on L1 to actually do this bridging right then from the L from from that L1 it gets bridg it gets gets bridged to your destination chain so there's actually two transfers which is um not really efficient um and there is no uh it is not Atomic and it is also it it may be synchronous but it is definitely uh not Atomic right okay so how do we improve on this so the next um yeah the next architecture is to use a shed Bridge so in this case I think the um the general idea is to is is that users are able to just Bridge directly between um two l2s right so in this case how do how does a user transfer from one L2 to the other um first it gets burnt on the l2a same as before you wait for the validity proof and it will be actually minted directly on the l2b but this uh in this case is asynchronous you need to wait for the proof on l2a to be submitted first before actually um uh before the the token can be minted on the on l2b right and okay so to solve the asynchronous issue um yeah we we try to make the shared Bridge uh synchronous in this case so um soorry I think this should be a Shar uh sequencer yeah so a Shar sequencer will construct a super block on the L1 with a transaction on L2 a and l2b so this means like the the mint the the burn function on L2 a and the mint function on l2b will be be proposed onto an L1 block Al together and this is how we keep like uh the the atomic nature and the synchronous nature of uh this cross chain uh interaction right so just going through uh these diagrams over here um so for example if we take like a l2a to be like this uh one over here um as you can see the bridge the bridge the bridge doesn't have to go through L the L1 anymore it can just go through um uh just go um just bridge between the the l2s directly right and also on this diagram over here um the atomic cross layer transac yeah Atomic cross layer transactions but the validity rule ruls require executing all l2s and a shared sequence so I so the shared sequencer would be in this uh L1 um block over here um and yeah it will probably because because you're including like uh both um both l2s transaction in the L1 Block in the in the Shar sequencer you will probably need uh very fast uh ZK proofs very optimized uh ZK proofs or even some like optimism to to to to scale this solution so yeah this also um is together with the idea of like a multi- prooof system so uh perhaps we could have some blocks that are proven using using validity proofs and like maybe uh 100 blocks in between use optimistic proof so like every 100 blocks you have like a a checkpoint using uh validity proofs yeah for this uh infrastructure okay so another way is to actually scrap the entire bridging idea all together and use a L1 call and uh or L1 delegate call so this idea is um basically you don't have you you won't developers won't deploy any smart contracts on the L2 um but yeah all of the contract will will live on L1 actually so like your your Unis swap openc everything on on L1 but use l2s to submit calls or delegate calls that will be pushed onto L1 so so your l2s are basically just commun communicating like the the state change that will be reflected onto L1 and I think this will solve the problem of um um having to deploy contracts on every single uh chain right so I think the diagram here um kind of um shows give gives an idea of uh how this might work so on the L2 you there will be a smart contract to just um uh execute a call um or or rather this L1 call which will indicate like the the address of the L1 smart contract and also the call data so like I don't know like a swap or or whatever and and then this L1 smart contract will be included um the transaction this whole L2 block this um request for State change will be included um onto uh L1 block like a regular rollup and and this will actually um indicate your state changes on the on the L1 chain so yeah so I think this would be uh uh probably a pretty good solution um uh it's it will be cheap and convenient to yeah access the latest L1 data but I don't think anyone has uh done this yet so it's pretty much experimental and if yeah and if anyone could do this during the the hackathon like test out this idea that would be great so basically using any L2 to call uh L1 contract yeah so okay so yeah so like all other l2s like Tao also has like also um wants to have like a a super chain and we call it the the Tao Singularity so with the same Concepts that we have in mind uh it could be the the delegate call or the the shed sequencer we want to have uh we want to parallelize the deployment of l2s to scale uh ethereum yeah so I I I pretty sure this diagram is um uh common across um most L2 super chains yeah so yeah I think that's the end of the presentation it's pretty short um yeah and these are like further readings and yeah research that contributed to this short keynote thank you thank you very much really appreciate it as we understand your colleague could not make it unfortunately so it's nice of you to be able to step in and still give the presentation really really appreciate it we could you open it to the floor does anybody have any questions uh about Tao about the presentation I'm not a developer so all of this is a little bit okay we have one yes we got some developers in the audience just please introduce your name and then also your quick question that you want to ask thank you hey I'm alas thanks for the presentation I just have a question about the bridge from L2 to L2 I understood the share sequencing just can you briefly re explain how um L2 to L2 is is um is uh sorry trust minimize like trust minimize or trustless Okay so this requires like this smart contract called the the bus uh bus smart contract so it actually um um I don't have the code right here but this bus contract basically um indicates for every transaction on on uh on like a single L2 it you have to modify your smart contract to um inherit this this bus so that it can um transfer data to the destination chain yeah but um yeah maybe I can explain uh uh more in detail I think if you see the the yeah the the sample code maybe you probably understand okay and sorry the question was about the shared Bridge oh the shared Bridge Okay um so like how how do you yeah so part of the shed Bridge um it actually uh yeah it actually adapts this um bus contract but uh yeah it's not here so or if you have the code later you can show it to him perhaps after yeah yeah yeah sorry about that might be no problem no problem at all perhaps later some things are a little bit complicated you can't fully put into words sometimes you need that smart contract next to you but no worries uh any last question one more two more I like the term Tao Singularity it sounds nice uh okay let's have one more question up here all right please introduce your name and the question you uh so thanks for presentation so I'm martinet I have a question on the uh super like I guess the synchronous uh with the sequencer so you have a super sequencer right there uh my question is if if we're waiting for one of the l2s and that l2s is like let's say crashing could I'm considering this case and how do you handle those cases like because if waiting for all of them then that kind of becomes like a Dos vector uh if that's a requirement so I'm just wondering uh what's the solution in this case um so like the example would be um if you submitted like a a bridging a bridge um transaction so if you wanted to Mint a token on l2b and you already burnt um token on L l2a and what if like one of the chain crashes H okay so um I think in the in the shared sequencer um you include both transactions in in a single uh block on on L1 so um if either um if either transaction doesn't happen on either chain I uh the block will be just like reverted on on L1 yeah okay hope that answer the question if not you can always find king and ask more questions after this let's give him another warm Round of Applause thank you very much King