Secure Off-chain Data Availability for Rollups - John Adler

hello everyone can you guys hear me okay testing testing all right hello everyone while they get my slides while they get my slides running i'm gonna i guess introduce myself in a bit of why i'm here so hello denver my name is john adler i'm here on behalf of celestia labs uh i guess a bit about myself my background uh i'm one of the co-founders of celestia labs where we're building celestia the first modular blockchain network i'm also one of the co-founders of fuel labs which is building the fastest modular execution layer so as you can see i'm a bit of a bit quite excited about this whole modular blockchain paradigm which is a lot of what i'm going to talk about today is it working there we go okay and specifically what i'm going to talk about is secure off chain data availability for rollups using celestia let's start so what's a roll-up very brief introduction because i'm sure most of you have already had pretty vast exposure to roll ups you know they've been all the rage in the ethereum space for the past few years so what is the roll-up first we start with this layer here this is ethereum right ethereum acts as the consensus layer where it performs three tasks in this roll-up paradigm it performs data availability settlement and also it does execution what are the each of these three things so data availability is essentially making sure that the data behind some transactions has been published to the world you need this because if you don't have this data you don't actually know what's happening in the blockchain in addition to this ethereum also provides settlement so this means it'll verify proofs of what happens in a roll-up it will verify either that something invalid happened in the roll-up or it will verify that the roll-up is valid and these are fraud proofs and validity proofs respectively and the two flavors of roll-ups being optimistic and zk rollovers okay so on top of this consensus layer you have the roll-up layer and this roll-up will do execution it won't do data availability or settlement it will defer those to ethereum and we'll just perform execution and there's a variety of examples of rollups fuel arbitrarium zk sync and there's many more okay so how does a monolithic blockchain look like and this is the traditional blockchains that we've had for the past 10 12 years or so or whatever right there they've all been monolithic blockchains uh monolithic blockchain in the base consensus there at the most base layer does all of data availability settlement and execution and usually settlement and execution are tied together now what's the problem with this the problem is that if you do all these things at the base layer you're doing a bunch of work and if you have to do a bunch of work that work kind of limits the capacity of the system it limits the total throughput and this is why many blockchains are now moving towards the modular blockchain paradigm for example celestia and ethereum 2 are based on this modular paradigm where the base consensus layer only does data availability in other words it makes the transaction data available but doesn't execute it on top of this you have a settlement layer for example this could be ethereum's ethereum virtual machine and usually in the theory in the e3 model for instance these two would be tied together and then on top of this settlement there you have a completely separate execution layer uh and these can be the same roll-ups that we've seen before things like fuel orbit from zk sync many others and the reason that this model is so good is because now the base consensus layer does not have to be burdened with doing settlement and execution it can just focus on data availability and therefore it can provide a much higher capacity of data availability by specializing in this and then on the flip side execution layers and settlement layers by specializing only in settlement or only in execution can provide a much better experience much higher throughput for those things through specialization okay so how would roll-ups look in a modular blockchain stack as opposed to a monolithic blockchain stack they could look like this which is what we saw in the previous slide it could be that you have one data layer one settlement layer and then you know one roll up one execution layer running on top of this but you don't just have to have one roll up you could have multiple roll roll-ups or multiple instances even of the same roll-up there's no reason you can't have one two three ten a million roll-ups running on top of a modular blockchain stack so it could look like this alternatively you could skip a settlement layer altogether and essentially have the execution layer run directly on top of the data layer and this is what we like to call a sovereign roll-up chain this gives you the nice property that since you don't have to provide proofs to a solomon layer it means the execution layer is its own settlement layer and can therefore decide on its own rules hence sovereign the way this looks like practically is that this execution layer here in this model could actually hard fork or soft work in other words change its consensus rules change which blocks are valid or invalid and it could do this without affecting the data layer because the data layer doesn't interpret transaction that doesn't execute transactions it just orders zeros and ones without having to post proofs to the data layer but by only using the data layer for data this allows the execution layer to be sovereign okay so what is celestia and you know how will it provide going to the topic of the talk how will it provide secure data for rollups so celestia is a again it's the first modular blockchain network it spearheaded this whole modular design paradigm revolution that everyone is now pivoting towards it's a layer that only does data availability it does not do settlement hence the little cross here it only does data availability and this is different than i think probably every other chain that is kind of trying to be modular but is actually semi-modular where they still do settlement and data in the same layer here celestia only does data what this means is it can be highly specialized for this one task and can provide many orders of magnitude more data throughput than you can get out of a monolithic blockchain okay so what's the celestium which is kind of the we can finally get to the topic of this talk a celestium is essentially taking the best of both worlds from celestia and ethereum and using them together synergistically where celestia can provide highly scalable data availability throughput and ethereum can serve as a settlement layer and a liquidity layer because as we all know ethereum currently has the strongest the liquidity mode you know is the most liquidity out of any other blockchain especially like programmable money blockchains or smart contract blockchains you know it has the most by far so this allows applications on ethereum to leverage what celestia has to offer okay so what does this last dm look like in practice you have a execution layer just like you had before which you can think of like a rollup right so you could have one or two or however many you want and then rather than you that using ethereum for both settlement and data availability his execution there uses celestia for data availability and uses ethereum for settlement in other words it'll post its roll-up block data to celestia where celeste the celestia chain will guarantee through its consensus protocol that this data is available and users can verify that this data isn't available through the consent through the celestia protocol which used this as novel technique called data availability sampling that allows any node in the world to verify that data has been published without fully downloading that data i don't think any other blockchain does this or even close to this other blockchains especially monolithic blockchains require you to fully download data in order to verify it's available which is part of the reason why monolithic blockchains aren't very scalable in this case celestia through his data availability sampling allows anyone in the world to verify that data is available without fully downloading it and then this celestium this roll-up essentially would use ethereum for settlement which means it would post fraud proofs or validity proofs to ethereum to adjudicate to handle deposits and withdrawals of tokens okay so how is this accomplished it's accomplished through this new technology that we're developing this is all public and open source the entire celestia stack is public and open source now this isn't like some you know secret thing that we're hiding it's all it's all out there this quantum gravity bridge technology works as follows that you have your roll up here listed as an l2 operator you have your roll up and it does two things one is that it will produce some sort of proof or some sort of new claim or hegestation on what happens in the roll-up in other words the effects of the transactions and then simultaneously to this it will also post the actual transactions that cause that effect right and again these two things don't have to be placed in the same layer you can use different layers for each of them you can use a separate data layer and a separate settlement layer and this is what this does so the transaction data which you can see on the right here gets posted to celestia and the celestia val data set will attest to the availability of this data and if you do not trust the celestia validator set as i said before anyone can run a node to verify that the data is available or not and it can verify this without downloading all the data thanks to data availability sampling then the this attestation this these signatures over the fact that the data has been published to celestia are relayed to ethereum through our quantum gravity bridge contract in the form of a data availability attestation and then downstream roll-up contracts can use this bridge contract to query for the existence and the existence of a valid attestation essentially it's like a life client relay just a this is a fancy like client relay so in this scheme rollups will post their proofs to ethereum and post their data to celestia and the bridge contract is essentially a drop-in replacement for a data availability mechanism which traditionally roll-ups co will either post their data directly to ethereum so it can replace that or some roll-ups have have used a technique called volidium where you actually leave the data off chain but you use something called the data availability committee to attest to the availability of the data now you might say well why don't i just use a data availability committee why would i go through all this hassle of using an external consensus layer instead of just seven guys seven large corporations signing off that data is available the issue with a data availability committee and what we're trying to avoid is that a committee has no penalty for lying so if the committee lies and they attest that data is available when it's not they don't actually get any penalty there's no penalty whatsoever they just get away scot-free the only penalty if you want to call that is the reputation gets hurt that's not exactly very robust in this case if this celestia valve data set lies about data being available when it's not they will get slashed this is a slashable offense in the celestial consensus rules which means potentially billions of dollars is getting burned through the slashing as a very strong cryptoeconomic guarantee the second reason that celestia the celestium and quantum gravity bridge approach is superior to a vanilla data availability committee is that with a data availability committee users cannot verify that the data is published they have to completely trust the committee or they have to fully download the data which kind of defeats the whole purpose of some external person the testing to the fact that data is available if you have to fully download it yourself with celestia users can actually use data availability sampling to verify the integrity of the bridge before using it and even during using it so at any point they know if the bridge is compromised and then they can enforce the slashing conditions on celestia with the data availability committee you don't actually know that the committee has been compromised they've compromised you but you don't actually know it so this is strictly better than a data availability committee in those two axis okay so now you know okay this is more secure than a data availability committee but why would anyone use this scheme over just using ethereum for data availability right well it's largely the same reason why people use data availability committees and why all these royal projects are exploring alternative data availability options because of this chart sorry if you post data to ethereum you're looking at something like a swap being two dollars on a good day with low transaction fees this was taken this morning or maybe last night oh this was taken this morning on l2fs.com made by david mihal i think so even on a good day with not too bad transaction fees a swap on a roll-up that poses data to ethereum will be two dollars if you use an external data availability solution like celestia that cost essentially drops down to nothing it's just a fraction of a penny so this is why roll-ups have been exploring off-chain data availability solutions this whole time because these costs are simply not sustainable they're going to get worse over time as ethereum gets more congested because there's such a strong demand for ethereum and block space and that demand will only increase ethereum gains more adoption you know we need a solution and posting all the data through ethereum even if you try to do some fancy compression stuff you're still going to get numbers like this and it's only going to get worse and celestia through celestium through the quantum gravity bridge offers something that's essentially as secure as ethereum but that provides you the low cost of a data availability committee okay so if you'd like to find out more here's our website celestia.org our github is the celestia org handle again all of our code is open public and open source free and open source nothing is done behind closed doors and if you'd like to follow us on twitter it's the celestia org handle and also as a disclaimer there's no token sale going on so if you see any of that that's a scam now do we have a few minutes left i don't know if questions are allowed but if they are then we have a few minutes for any questions can you quickly explain like how a fraud proof or uh you know contesting a fraud proof would look like in the celestial world yeah so in a celestium well first of all let's take a step back and you know consider a fraud proof in an optimistic roll-up right in an optimistic roll-up the fraud proof is posted to ethereum and adjudicated and resolved on ethereum in a celestion exactly the same thing happens the roll-up will still post its fraud proofs to be executed on ethereum the only thing it does differently than a vanilla roll up is that it'll post the transaction data of the roll-up to celestia but the fraud proofs for an optimistic roll-up and the validity proofs for zk roll up are still verified on ethereum exactly as they are today so can you also query data like which was posted into some celestia or is just like it just well like does just doesn't have a little check and then it just discusses data okay so i think the question was can you query the data that's in the celestion okay so uh the way it's currently constructed is that every single byte in a celestia block is attested to via the data availability attestation and the quantum gravity bridge so you can act you actually get access to every single byte that's in the celestia block through a merkle root and then you can later on provide merkel proofs against that merkle root how is like celestiums like data availability more cheaper than you know using it over like l ones like what makes it more cheaper in general so the question was why is uh why would a celestian be cheaper than using any alternative l1 right the reason it will be cheaper is because celestia is specifically designed for data availability which means it can provide more data availability throughput than any other blockchain in the world more capacity means cheaper transactions can you elaborate on why why that happens like what oh uh well it happens for two reasons uh the first reason that celestia provides more data throughput than any other chain in the world is because it doesn't do execution and settlement by not doing those things that essentially frees up consensus nodes to just spend all their resources on bandwidth just all in data availability therefore they can do more of that because they don't waste resources doing execution that's the first reason the second reason compounding this is that celestia uses data availability sampling so that nodes can verify that data is available without fully downloading it this allows you to have trust minimize like clients and it allows you to have nodes that contribute to the security of the network without being full nodes without fully downloading everything and the combination of these two properties is what allows it to have a much higher throughput okay no more questions sorry but you can find me after i'll be around thank you so much