L2s in 2024: Surge, Adoption & Security | Nicolas Liochon - Consensys

desk put your hands together wow what an audience thank you all for coming um I'm I'm super excited um to be here um Rachel thank you so much for the introduction um we don't have that much time it looks like we've got only like 20ish minutes so how about we start by um talking a little bit about what Linea is for those of you in the audience who aren't you know fully aware of what Linea is and how it connects to this L2 ecosystem and by the way my name is Sam kler I'm an editor at coindesk um and yeah I'll be moderating today excited to be here so hello everybody uh I'm Nicola Leos and Linea is a z evm so what is a ZK evm it's a layer two that works at a bu cut level exactly like e main net uh and that does prove the correctness of the execution and the problem that it solves is it makes transactions much cheaper basically 40 times cheaper than mainnet today and it's going to decrease again and it may makes this price as well more stable and with much less fluctuations due to activity because it scales so there's a difference of course between ZK EVMS and optimistic Roll-Ups can you talk a little bit about you know the landscape like where you fit in um you know with I I think there's something like 20 or 30 different l2s that have come up over the past year or so um let's talk a little bit about that yeah so at the very beginning what we add was Z rollups it's basically back to 2018 the idea was in order to scale ethereum we need to do the calculation of chain and to prove on chain that it's correct but back then it was uh a very Advanced cryptography we did not really know what was possible and what would be the impact on performances the complexity of the proof and so on so what was done at the beginning was what was called specialized rollups just for payments and and the idea that we could execute the bu code instructions of the eum virtual machine that was defined much before without any knowledge of what was the proof system Advantage was nearly seen as impossible so for this reason many people tried different branches so you had some zs that were specialized then some people said yeah actually the zero knowledge technology is too complicated we're going to play a different game we're going to have fraud proof Mo that was the optimistic cups so you've got one branch Z One branch Z specialized another Branch optimistic hups then people said okay we're going to try to have generic execution but not with the evm it's too complicated so we will have specialized virtual machine and that would be yet another branch and then people said yeah but actually there are different ways of doing virtual machine so let's try other branches and other branches and two years ago uh it became clear that actually doing the evm with proof was actually possible that's created in a way the last branch which was ZK the proof plus the evm eum virtual machine and that's a technical decomposition but there's actually another one which is uh I would say business and functional people said okay uh some people are building technical layers optimistic ziz and all those layers but actually we need to build an ecosystem so people said okay I'm going to reuse one of the technical layers optimistic to build one ecosystem specialized sometimes for gaming for defi uh and so on and so you have multiple technical layers multiple ecosystems and you do a big Matrix of all this and you have uh what we have today so there's a few different things that I want to cover today um you know obviously I want to talk a little bit about Linea I also want to talk about l2s generally the state of ethereum so first off um you know zooming out a little bit ethereum is about to undergo a really big update called Proto Dan sharting which is this thing that is supposed to help make transactions even cheaper um for for rollups um like your own and maybe you can talk to me a little bit about what that upgrade is and what impact you expect it to have on services like your like your own yeah so uh the idea ofup and Z rollup is it as secure as ethereum and what we mean by this is if a rollup disappears completely you can still recreate it just from what you have on ethereum it means that everything that is happening on the layer two on the rollup is first proven but as well the data is written on ethereum so anyone can recreate the full state so when you do this and you do this with a lot of transactions the issue that you have is you need to write an incredible amount of data so we have some optimizations around compressing the data around removing the useless data there's many things that you can do but at the end of the day you're writing some a lot of data on eum layer one so in the past we were kind of using the generic uh uh layer one uh system like evm allows you to write data so you write data that's fine but it's the idea of the foundation was let's have something specialized just for data that will be used basically mostly by for rollups and the nice thing about this is because it's specialized it's obviously much less expensive but uh the second point that is quite important on the long term is something that can scale easily so when they're going to start in a few weeks uh they will start with a reasonable amount of uh capacity but it's something that can change basically multiply by 10 20 50 without any uh major difficulty it's basically changing a setting so thanks to this we've got the capacity to multiply uh immediately the capacity of a roller by something like five and then by 250 uh without having to spend too much brand power so that's what this EIP is providing so um I was at a dinner the other night and a lot of folks were talking about l2s the proliferation of so many of them and what it actually says about ethereum as a protoc call and both um you know the successes it's had in in cultivating this ecosystem on top of itself but also maybe the failures that it's had as an ecosystem the fact that it needs so many chains built on top of it in order to fulfill its ultimate goal and I wonder like what your response is to to this idea that ethereum in order to achieve its ultimate ends has kind of relied on a lot of ecosystems like your own that um you know and we'll talk about this have varying we're going to talk about security a little bit but we're we're still in the process of building these so a lot of the traffic that you know is now going on to ethereum is going through a lot of these services that don't have the core kind of um you know use case I guess value ads that ethereum seems to to promote itself as having yeah so it's interesting because I think that the view changing with time and we don't know yet what we will have at the end uh so when we started the need was we need to scale and we don't know to scale the evm so let's try different things and now we're at a point want and and the idea that is behind this is okay as any way we can't do the evm we can go into crazy experimentations uh let's try a totally different virtual machine let's try specializ up codes let's try everything uh what appears now is that actually the evm we know how to scale the evm and a side effect of this is it puts back eum in the in the loop because what linear is doing it's executing evm instructions exactly like uh the main net so it means that linear implements the evm specification defined by the eam foundation so eam is back in the center of the world uh a foundation basically decides a part of the future of linear because we're implementing this specification uh and there are other people implementing this specification as well which means as well basically more power for theam foundation because it's not a single team that decides everything and that can say no I don't want to do it there are multiple teams doing it they're all uh kind of competing in a very positive way to have the best possible implementation the more aligned and the fastest uh which leads us back to the situation when we have theum layer one client when you have one specification by the foundation and four clients five clients it's back to the to the story and it's The Next Step very likely that we expect at Cina is actually rollups will be secured by multiple different provs because as there are different implementation you can use multiple of them to increase the security of the system and once it will be kind of um validated enough relable enough seen as rable enough it's very likely that it will move back to the layer one and the layer one will be using this technology it's a future in maybe 5 to 10 years I don't know uh depending on how fast we are working but it's a very likely future so what trust assumptions are there built into Linea right now we're going to talk a little bit about security and the stuff that you're you know in in the process of building but there's 355 or something if I checked right before this million dollars you know in this system who are those depositors trusting with their funds what pieces of your system yeah so fundamentally you trust that uh eum layer one works like if eum layer one disappears uh everything is lost as simple as this but in terms of consensus that's the first layer the second layer is you trust a little bit that the proof system works and the proof system is quite complicated because when you say that you trust the proof system you need to trust the mathematical definition of it that's one thing you need to trust part of the implementation the implementation that is going to verify the proof not everything but the part that verifies the proof but you need as well to trust that we did a good job in what is called the arithmetization of the execution which is 600 pages of specification so something quite big so and I say you need to trust this a little because given the complexity we know that uh it's not going to be perfect day one yeah so we have another layer of security which is the number of people that can send a proof is limited so we call us having multiple layers of security making the system uh a single failure won't make you lose funds that's that's how it works so that's called training wield that's how we probally call that and so and that's a generic mechanism and there a last trust assumption is as the system is evolving we can upgrade the system so here you have a security Cil validating that the upgrade is actually valid so you have this trust in the security Cil which is which is at the end of the day basically conses that conses is not going to be a very bad upgrade uh without any warning so there are some delays and so some precautions but that's a basic assumption so there's um a tool out there that you know some of you in the audience might have used L2 beat which is where you can get a lot of this information on your own ecosystem and and others and I'm curious so you're you're currently like a stage zero as are most of your competitors um basically all of your direct competitors are quote unquote stage zero rollups which refers to the kinds of training wheels that we're talking about that you still have I wonder if you can like you sitting here in 2024 you know think about or help us understand when will you be a stage one or stage stage two these aren't perfect or like you know scientific measures exactly but when will you kind of start taking these training wheels off to make it to those latter stages yeah so actually what we're working on right now is adding training wheels to add more security to the system so a lot of the things that we're doing as let's have over type of verification over type of of checks in order to make the stem more secure uh the training wheel the biggest training wheel and that's uh the big biggest pain point for everybody is really the fact that we can upgrade the smart contract given the changes in the technology it will take a while before we remove this let's be let's be very clear uh so we really on an idea that we're going to add layers of security uh for at least three years and every new layer of security that I can add I will add it and then we will start to say okay it's stabilized we can remove as layers but to me the ultimate head game is really what I was explaining about having multiple provs because we know that the technology is complicated we know that it changes so if we can have a system when a single rollup is uh secured not by a single proof but by multiple proofs we have here a perfect solution and that's the end game and the end game requires to have multiple uh zkm implementing the specification perfectly which going to take two years and then we need to have this type of uh gentleman agreement that sharing the implementation it would take much less last less than a year and we'll be ready to play so we've talked about this before in a previous panel I think it was last year but in terms of how you communicate these constraints and and and these you know security limitations these training wheels to users how do you um a as someone who's building this technology think about you know how users should be considering the the safety of their Investments relative to locking them up on ethereum Direct directly what would you tell a family member or a friend about your your system or or others um in in this orbit um should I put all of my eth you know into Linea like is it a safe you know way to to to secure my funds yeah so the security is very Global so it's not only like you you put all your money on eum itself uh and how are you going to manage this and I think it's really a key topic uh globally for the ecosystem because what we're seeing with with rollups it's basically in the past we are saying okay eum cannot be used anymore because there are so many usage already that the transactions are too expensive we can't put more more use case there we're at the maximum and we're removing this limitation now the transactions are much cheaper so you can speak to much more people and again it's going to scale so you can have more use cases and and the point is is now we've got this it means that we've got new users and there as new users uh they may not understand how eum Works they they may not understand how wallet Works uh there's a lot of complexity that they don't know and we think that on top of rollup itself there's a Global Security approach to have because the data that we have from metamask is when someone enters the ecosystem and loses his phones in less than 6 months basically he won't come back said okay it's too complicated I'm leaving forever so we really need to be sure that uh it does not happen basically so people who are new to the system can do things that are not that will not put them at risk and I think it's really a major responsibility uh for us to do this if we want to build a long-term ecosystem so that's how we do that so so we are increasingly I think seeing this trend on ethereum that we've you know see it really started I think on Cosmos this idea of app chains so blockchains that pop up built on top of some base chain or that can draw security from some other you know system that are specific for one particular use case whereas Linea um and you know the the you know scroll polygon and a lot of these optimistic rollups um they're general purpose um the ones that we're talking about when we're talking about like 32 L2 chains or something how do you see your model this general purpose model fitting in in a world where it's easier and faster to design ZK circuits that are more specified to particular use cases um yeah uh so that's link I mean it's interesting to look at successful applications and a successful application that we have today is for example farcaster and you know there's a debate abch layer free layer four layer five is farcaster a layer fre I don't know but it's a successful application and why because basically they've been able to say okay some parts are very important and needs to be secured but the eons so it goes onto the Layer Two And there are some parts that are very complicated to do in a decentralized way it's still a decentralized problem you need to distribute all the images all the cast all the data everywhere in a secure way maybe you need to be able to prove that you're actually Distributing it and you're not lying and providing some data to some people and some other data to some others and you need to do that and I think it's very interesting to look at it from a what do you need to to do do you say oh I've got a rollup so everything is a rollup well no that's not the right way to do things the right way is to say okay I've got a problem where I need the blockchain to secure this part and then I've got other problems that can be quite complicated in terms of cryptography and distributed systems and they need to be sorted out so the way we see things is uh we should have a single level of security which is the maximum the story that you can play with options all over the time saying yeah I don't need this security it's just too complicated it's confusing and you're just going to get it wrong to me just an excuse to put the responsibility on the user saying oh yes uh you were told that it was not really secured it's your fault no it should be secured properly without playing games and then you need specific tools to build the other parts of your application and you should not twice to force a rollup when it's not a rollup so what we have in linear uh in many cases what you want to do you want to have some Advanced cryptography and for example to calculate options some results and to put them on chain and we have a library to do this that is actually totally independent from linear that is actually used by I would say nearly most many blockchains layer TOS all over the place called this library is called knar and people can use it to build Advanced uh features uh independently of a rollup uh but they will basically at the end push the result of this calculation on theup and uh it's a key different in approach I would say online that we say applic something that seems a little bit obvious in my opinion but application should really look at their needs and not try to use a blockchain for everything but just use the blockchain for what they need awesome I I think we're at time um but thank you so much um everybody for joining thank you for for you know um giving me the opportunity to talk about Linea and l2s and ethereum and yeah enjoy the rest of the conference thank you very much thank you very