State of L2s after recategorization

Transcript

[Music] All right. Uh hello everyone and uh what amazing feeling. Um when I started my crypto journey nine years ago, I uh I never thought that you know I would end up on the main stage in Khan. Uh this is this is just uh such a unbelievable outcome. Um so thank you uh everybody for coming here and I'm here uh to talk about the real state of L2 on Ethereum.

I'm Bartk. I'm a founder of L2B. Uh so u maybe let me just start by asking a simple question to the audience. Can you like raise your hand if you've never heard about Delto beat? Okay, there's one, two, three people. So, uh, so it's amazing. Thank you all so much, uh, for knowing us. Uh, for those that don't know us, L2 to beat is essentially a watchdog, uh, a community watchdog. Uh, I tend to think about this as a community for watching every single L2 uh, which is out there. And um and essentially what we're doing is something which is akin to uh traditional finance where let's say that

you want to open a new bank. Uh what do you do? You go to a regulator, you ask for a license and you would probably have to go through a kind of an audit. You know, you have to uh meet uh a lot of different regulations. You have to pay a lot of uh fees for that. Uh and not only if you wanted to be a bank, if you wanted to uh be I don't know like a pension fund or some kind of a uh someone who has some financial instruments, you go to a rating agency like Moody, right? And again uh you pay for the rating. Now uh this kind of a model uh is super common in traditional

finance but uh it's I think personally that it's actually quite impossible to pull off if everybody can launch NL2 a new chain right anybody right each one of you can set up a new uh company raise funds and launch a chain and try to attract users and then what do we do uh so we ended up at L2B essentially building an infrastructure so that we can monitor each one of these L2s and by each one I mean hundreds and maybe even thousands of them right so what you see essentially as a kind of a website with ratings uh for all the L2s the security ratings was behind the scene is that 20 odd team uh that actually monitors the

chain and each day it uh receives couple of notifications and each day it reassesses if any of the changes that were introduced to these L2s kind of break or change the security assumptions. That's a huge infrastructure that we're running and you can actually test out using our free open source tools if you want. But the whole model is kind of flipped upside down from the traditional model, right? Because it's permissionless. Anyone can launch a chain. So I'm going to talk to you uh walk you through what we've done quite recently which was a major reclassification of O2s but to fully understand why it happened I need to uh

take you on a bit of a journey a bit of a history of L2s. So let's do it very quickly. So the story uh began about seven years ago when uh a Chad called Barry White hat he published a repo uh with what we today consider a first rollup and he actually introduced that name. Uh you can today study uh this repo and it's a fullyfledged functional rollup. this rollup uh I would definitely recommend you to to look uh at uh in details but what is important uh about this rollup and I would stress it once again it was like literally the

first one that was ever published before that name the rollup name wasn't even like popular it talks about data availability guarantees right it says that it's important uh that we pass every updated leave to a smart contract so the data will always be available does the system has the same data availability guarantees as Ethereum. Uh so in my opinion data availability is one of the most uh confusing terms uh to to a lot of users and I will start you know like explaining why we uh did recoperization by by explaining to you how we actually understand the data availability. So like imagine in traditional finance that to withdraw your funds from the bank, the bank asks

you to submit a history of your transactions. That is strange, right? Like like we're not used to that. You know, when we want to withdraw our money from the bank, it's it's bank's responsibility to understand how come, you know, we have what we have and it's the bank's responsibility to prove in effect that, you know, they're not cheating, right? Like we don't need to worry about that. But here you know the situation is very very different. We don't want to trust intermediary. We actually want to disintermediate. We want to change the setup so that we don't have to trust anyone. So suddenly now it's kind of like our responsibility you know like we need to submit data to prove what we have and what if the data wasn't even published in the first

place. What if the bank just said, "Hey, you know what? Your current balance is, I don't know, 10 uh,000." And you're like, "What? No, it should be 95, right? $1,000." But the bank says, "No, I'm sorry. It's it's like 20. What do you do?" And this is just something that, you know, we we we don't think about it. And and this is not, by the way, our typical experience uh on blockchains today only because these intermediaries are trusted. you know they kind of don't cheat on us right but our goal is to build trustless systems so you need to remember that data availability it's not just about you know that that data can

be accessed uh in some time in the future you will need that data uh to withdraw your funds the problem is that data has to be published in the first place it is quite possible that a rollup operator would like claim name your balance over someone else's balance and never publish his data and then if that happens then you're essentially uh screwed right you you you really can't withdraw your funds and that's it. Uh in fact on optimistic rollups you know such a malicious operator can can steal your funds by simply not publishing data. So to kind of stress this point, it's super important. Data availability is about

data publishing and you need, you know, a trusted group of parties that would attest the data was published and when you publish data to Ethereum, it will be your Ethereum validators. Uh so that would be the ideal solution and then later you need to retrieve data, right? And most teams seem to focus on later. they kind of forget about you know the left part of this diagram. So very quickly uh the community realized that publishing data on Ethereum is kind of expensive and we need to find a cheaper ways uh to publish data and I think it was Stark uh uh Starkware that uh has proposed this concept that we today call

data availability committee the DAC it's it's like a group of attesters if you like right it's like uh essentially it's a multisig that would attest that the data was published by the operator. Uh and if it was published then the chance that you will be able to retrieve the data later on for your withdrawals of course you know uh it's it's uh it's uh the guarantees are are much higher and notice you know the original proposals I wanted to stress one thing that data valid committee should be made up of reputable entities trusted by the committee and the market and they are compensated for these services. So it was understood at the time that it is

very important job and this committee should be uh is essentially a trust function and users should have certain guarantees. Uh so the funding members of the first stark x uh data availability committee where consensus infura never mind uh few other companies uh like a network of unknown companies. This is how this concept started. Uh then uh and notice by the way the year right we're talking 2020. Uh so five years ago and and then uh two years later um arbitum announced a construction for optimistic rollups which is essentially kind of an equivalent uh if you like uh they called these types of chains antitrust chains.

You will see a lot of L2s following this model on the right especially the orbit chains and they follow it because it's much cheaper than publishing data on Ethereum. And again in the proposal of arbitrum you can read uh you can read actually here um as an example there might be 20 committee members and an assumption is that at least two of them are honest. This is a very weak assumption, right? So imagine, you know, 20 reputable companies attesting that the data was published, you would need to corrupt most of them so that the operator can actually steal your money.

So I think as a community we kind of agree that you know as long as the the data availability committee is made up of reputable entities, as long as uh this number is relatively big, you know, these constructions can be good enough for certain use cases. Um for other use cases it's much better to publish data on Ethereum to get the full security of Ethereum and there was a lovely discussion on Twitter how these constructions should be uh called uh Vitalik was like insisting that it should be called plasma but at the end of the day you know these constructions today are called validiums. Uh validiums is a term that was suggested at that time by by by

the founder of starkware elements or so. Uh so so we've got validiums uh stark was like ZK uh company behind uh ZK rollup. So today we're using the world validiums for uh ZK uh constructions and optimiums for optimistic constructions. Uh and these constructions don't publish data uh to Ethereum. And then uh quickly later Celestia was launched. Notice after four years of research very complicated infrastructure availability service and uh layer as one of the main services that they launched. Uh they

launched DA. What about proof systems? Again if we go seven years back everything was proven. the malicious prover who does not have the the keys the private keys to the leaves they cannot change a leaf uh only the person who controls the private key can. So in this in this original constructions it was impossible for the malicious operator to steal the funds because they had to p uh produce a zk proof and sooner uh very soon uh the first system from was launched uh over the mainet. it was diversify. Uh then optimism has launched their first

version and again in the first version they talk about uh fraud proof system. This is the core purpose of a fraud proof system is to defend users funds. Uh ZKVM uh was launched by ZK sync soon after. So that was more or less the setup when the uh rollup centric uh road map has been announced. Uh this is how it all started and this is when Vitalik said hey things are looking really great. Uh it looks like rollups are the future and that uh will be our future as Ethereum. Now let's fast forward and see where we really are uh today and why a lot of projects have realized that well

if DAX are cheaper than rollups then maybe having just one member on the data committee will be even cheaper. Why not? let's do that. Some of them went even further. They completely confused data publishing from data availability and they realized you know what maybe publishing data to AWS is super safe. You know everybody knows AWS right you know if we publish data to AWS then everybody would be able to download the data from a AWS but they kind of forgot to mention what would happen if they never published this data. So and other projects started to post data to Celestia or Avail or ADA but they kind of failed to notify Ethereum that the

data was actually posted there. They didn't use what we call uh an DA oracle or a DA bridge. So in fact uh Ethereum doesn't know that the data is posted to Celestia. What about proof systems? uh turns out that proof systems are kind of easy for simple use cases but if you wanted to prove the full EVM execution uh problems arise. So there are a lot of problems that were uh discussed for fraud proofs. There are a lot of potential problems and complexity discussed for uh ZK proofs like uh certain um vulnerabilities that were

found in the proof system that could affect every single ZK chain that's using them. Uh in a way when people today think about these proof systems they they see this complicated maths and all sorts of potential crypto economic problems and a lot of projects came with this ingenious idea and I'm being sarcastic right now of course that having no proof system at all is cheaper than having a proof system and also it is potentially even more secure because you know these proof systems are complicated and by the way if you and instantly upgrade rollups, then maybe the security of your system is the same as having no proof system, right? And

suddenly we got an avalanche of uh systems with very very simple uh data availability guarantees like very very weak uh some uh projects like literally posting to uh to to one server and have data committee consisting of just one entity. So I'm not really even sure why they call it committee. um and they don't have a proof system. So to sum up in most of the tools uh permissions operators can simply steal your funds. And now notice like this is not how we started right when we started you know most effort was actually launched uh was put into trying to make sure that the the users funds are secure

uh proof systems are sound and data is actually published and it costed money today seems like it's more about user acquisition TVL and whatnot and this is all at the expense of security. So I think we got to a stage where we kind of decided as a community that we need to change that and as L2B you know as a as a as a watchdog that is run by the community and for the community. We did a survey we asked you guys like what do we need to do because you know it seems like the whole rollup evolution is going to a very strange place where no one seems to care about security anymore.

You told us to be more stringent. You told us to actually push these projects further to actually uh make sure that they uh are made up to the standards that we expected them to be built uh five six years ago and we announced that uh change uh back in Bangkok uh in autumn and we gave projects six months to adjust half a year to adjust and I'm here happy to say that you know This reclassification, recategorization has been uh um introduced by L2B. Uh and projects that do not meet these uh basic security and decentralization criterias

like they don't have a proof system or they don't post data to anything that resembles data available committee and not to mention posting to Ethereum. We have put them from uh rollups category or validiums category to the others category and we did that change uh two weeks ago more or less and look what happened. Uh we ended up uh originally we had like 120 rollups or so we ended up with only 24. So these are the systems that truly post data to Ethereum and they have uh they have uh what we would call a solid proof system. Uh with validium category it's even worse. Today we only have three such systems. Meaning

that everybody else, every single system that used to be called Validium turns out that these systems are actually posting data to a very small data availability committee or they're using Celestia or a veale organ with no infrastructure to inform Ethereum that that's the case making data withholding attacks very very easy and very cheap if operator is malicious. Everybody else was kind of relegated, if you like, uh relegated to a category called the others with our hope that they will soon uh enhance uh their security infrastructure and soon we will be happy to announce that uh they are back into

the rollup and validium category meaning that they actually uh take security and decentralization seriously. uh and many projects approached us and they are working heavily on that. Uh we have uh Hubie from L2B here on stage. He'll be giving out uh awards like Khan awards to to every single project that actually survived recategorization. And to sur survive recategorization all you need to do and granted you know it's difficult you need to do two things you need to uh post data to data availability committee that consists uh uh must have at least five independent

public attesters. Um or of course you you can post that of Ethereum if you want or if you're using all the solutions like Celestia you need to use a bridge like blobstream uh blobstream Celestia has built such a bridge very very few people uh and projects are actually using this bridge uh you need to be able to inform Ethereum that the data was actually published uh to Celestia or aail or a DA uh or any other frankly uh LDA uh solution and also it's not enough to declare that you have a proof system, you need to have a running proof system. You need to have a proof system that actually works, that is sound, that can prove every single transaction. And if you're using

optimistic roll-ups, you need to allow ideally anyone to challenge uh the state or uh if you have a a white list, this white list needs again to have five independent entities that can validate uh state roots. cannot restrict uh validators to just one unknown entity because it's kind of useless. It can be your friend that you know can easily collude with you. Uh that's not what we want. So this is uh recogorization. Uh talk to us if you're hesitant. Uh this is all done as I said before uh by the community and for the community. And because we are a public good uh company that is funded by by directly or

indirectly by all of you, you know, I want to uh extend my my uh biggest thanks to all of you. I mean, if not the community, uh we wouldn't exist. I think we are serving a very uh important uh function here. and with hopes of uh of the uh the usage of Ethereum in in the real world by by the the the the real financial institutions uh this function uh will be even more important in my opinion. Thank you very much.

[Music]