Scaling Blockchains: The State of Layer 2 Solutions - TOKEN2049 Singapore 2023

thank you my name is Joyce Yang and I'm the founder of global Queen research we're our Venture Dell and I'll have all the panelists here introduce themselves and give a little bit color about what they're building why don't you start e sure uh hi everyone my name is I'm the co-founder of scroll so scroll is building a general proper skating solution for ethereum based on the honor proof we are building a bicode level compatible with the KVM so so for you then developer you know we are just like ethereum but much cheaper faster with a higher throughput and we believe that the future will be cryptographic so we are building you know the most translates you know they are not approved technology to remove trust to math hello my name is Alex I'm co-founder of matter Labs which is the company behind GK Singh protocol zika Singh is a protocol on a mission to scale ethereum public permissionless blockchains in a way that can reach Mass adoption while not compromising on the core fundamental values that makes blockchains valuable to us and the technology we're building is the ZK stack it's a ZK VM and it's currently launched on ethereum as uh ziki sync era for the past couple of months we've been the most actively used L2 on ethereum from all the l2s with over 30 million transactions per month and accounting for almost half of the daily active users so it's a yeah very interesting Forefront of innovation in the ethereum space hi everyone I'm Ed Felton I'm co-founder and chief scientist at off-chain Labs we're the company that developed arbitrum arbitrim provides a pragmatic scaling solution for uh for ethereum there are two public arbitrary chains arbitrum one uh which is the flagship chain and arbitrum Nova which provides lower transaction costs through use of external data availability and then we also support arbitrum orbit which is a program that allows people to build their own chains with customization hey everyone I am Sandeep I am one of the co-founders of polygon and uh polygon is a you know a solution layer to Solution on the mission of uh scaling ethereum and uh you know we have very deep uh conviction on the ZK technology and we are building ZK based uh scaling solutions for uh ethereum and uh we also like you know have our Flagship product polygon POS chain which is now also upgrading into a ZKA validium and arguably it is the most uh you know one of the most adopted chains in the ecosystem with large amount of uh you know developers across D5 nfts gaming and Enterprises they're using we recently launched polygon cdk and made it fully open source and it is available for anyone and everyone to launch their fully ready ZK Roll-Ups which have permission less proofs fully tested audited and a mature ZK approver for them to launch their own dedicated uh ZK uh you know chains and we are also you know building an interoperability layer with which all of these chains can interact with each other thank you all for that introduction so why don't we just quickly give folks example here of an application building on top of your project right I know um scroll hasn't launched yet but maybe an example of what you imagine it would look like if it were building it on top of you yeah um currently we are in the like uh it's our last test net and we will launch our manual very soon so what I'm expecting is that because I just last month went on like African trip to understand you know different use cases or blockchain because people still doubt you know whether blockchain is really useful whether it's just a token game like you know for for rich people or like you know it can really serve for some real world you could take you know for example in some places there are you know like currency inflation and people need to access to some stablecoin and also like you know their financial infrastructure is not so stable and you need to do global payment and also build a better credit score system really for people so I'm really expecting like you know our different you know local communities like we have very strong community in turkey and in some other places I went to Africa and I I do want to see like some local Builders can build something for people to solve some real life real world use cases and as far as you know our contest Nets like you know there is large you know uni swap you know already deployed and like you know after minute I hope like all the ECM applications will deploy because it's so easy to deploy everything works it's so seamless and then I expect more like you know use cases really for example you know real world not just you know some game yep Alex uh anzike sync we are since it's a generic scaling solution we see all the kinds of applications that people Deploy on ethereum uh I can particularly uh say that we see a very strong activity in in the areas of defy gaming and nfts um what's interesting that's being built on ZK sync that that we don't see in many other places is um wallets using native account abstraction which as you can think natively supports uh so things were you know like wallets that support native biometric authentication signing transactions from your phone and playing fees and any tokens or using guest list transactions uh so because we believe that the ux has to match or even exceed the convenience and and the comfort of the webto Technologies to reach the stage of message so arbitrim is a is a general purpose scaling solution like like all of these we see a lot of D5 applications arbitrim is arguably the leader in D5 with more total value on chain than all the other l2s combined as well as high uh the highest trading volume of l2s so in the defy space there's a lot going on we have a really large and robust ecosystem um the gaming is one of the is one of the real growth areas and um to that end we have been working on or we just recently released the stylus upgrade which lets people write smart contracts in in regular programming languages like rust C or C plus plus compile them and run them in a fully composable way on the same chain with evm applications that not only uh helps to onboard a lot more developers and to be able to use Legacy code like libraries but also gives a 10 to 15x improvement over evm and performance so we think that's going to be a big boost to the gaming ecosystem please yeah I mean at polygon uh you know I can go on and on about the use cases because they're like just like probably with many accounts like probably more than 50 000 applications different different kinds of but uh you know as we see uh we keep seeing a lot of Niche uh web3 use cases like we are still uh you know like the in the ethereum ecosystem the third largest defy ecosystem after ethereum and obviously arbitrum but interesting part is that this difa ecosystem is uh you know hasn't has any kind of incentive events for last two years and it's it seems to have an organic uh sticky effect to it it's uh you know doing well then on nft side also uh you know off late polygon is like not rank two only after ethereum mainnet uh and then on the gaming side like uh you know generally uh you know polygon is considered to be uh you know a pretty big leader like we uh in the market roughly like uh you know 1990 did 95 percent of the games that you see especially after uh like immutable X which is which was a very big player in the gaming space roughly controlled 30 40 of the markets uh they are moving over their infrastructure completely to the polygon build chains so on the gaming it's really really strong uh you know in 2022 itself I think Delphi digital had published a report that around two billion dollars were funded into on-chain you know decentralized games and out of that around 1.5 billion was directly invested into the uh you know polygon and immutable uh based games and on top of that like around six seven hundred million was ethereum which also nobody is going to build on ethereum the games actually most of the games are going to be built on the layer two so you know we assume that large number of games would also come there and then finally the brands and Enterprises I think uh you know many of the people in the audience would know that you know their polygon you know has been doing a lot uh you know very very good on some of the largest institutions in the world from from you know ready to the Starbucks to the Disney's and Instagram to the world everybody like you know was integrating now uh you know and and and experimenting and recently I think there are some leaked uh you know uh Twitter uh tweets from people that Singapore in Singapore grab has launched uh you know uh NF this uh proper like web3 integration and this is not like you know run-of-the-mill it looks like because with the some of the things that we have seen these are proper like fit like almost like stable coin payment use cases not stable coin but payment use case uh in the form of nfts and nfts are kind of notes users can pay uh using the grab wallet and this is huge because grab has around 180 million users uh in in the Singapore and uh basically like 35 to 50 million daily monthly active users and uh you know around 15 million transactions a day in if even if a small percentage of those transactions start happening with these nft kind of bills basically you have ten dollar five dollar two three dollar bills and you can pay you know it could it could bring in a lot of uh you know adoption and polygon believes that you know with opening up this Enterprise space we open up the pi for all the you know amazing protocols like uh you know that we are building in the ethereum space and uh you know this hopefully we'll be able to grow the pie for the entire industry on the same thank you for that there's about currently 10 billion tvl across L2 scaling Solutions and I imagine that probably to accelerate very quickly given all your recent developments how do you guys think about each mainnet versus the scaling Solutions as in dominant liquidity hub um so because all the layer tools they are kind of liquidity are locked on some smart Contracting layer one so definitely like you know the layer one liquidity is like way larger than like you know because I need to add up all the layer tools and all the D5 protocols on each family one so it's still larger I think for layer 2 you know what we can talk about is that eventually how many transactions will be executed on different layer twos instead of and eventually ECM will just become a settlement layer instead of you know like sending different D5 transactions moving apps here around it will only be verified some proofs and I believe that in the future there might be like over 95 of the transactions will be settled down on some Layer Two and layer 2 will become the entry point for ethereum because the next period of users definitely can't you know enter ECM through a layer one because it's not so scalable and for layer 2 you want to become a really a great entry point for the next billion of users and for us like to become that entry point and become like you know part of the the ecosystem we want to stick to what ECM has all the philosophies or ECM or the values of vcm remain open remain Community Driven remain incredibly neutral and become that trustworthy layer like you know people need to people think of maybe School eventually as just an extension of exam it's not just technology technically you are scaling exam but you are part of like your extension of issue and which you know because you share the same value and you become you are building only for example building one thing which is the most trusted worthy entry point for every users so that's our mission and we believe that you know that world instead of you know focusing on some like you know being a technical provider which builds a lot of fancy features and making our system more and more complicated interconnected we just build one system super simple to use you don't just worry about anything and 100 times cheaper so you're saying just to clarify you're saying that eventually L2 will become the settlement later and also an entry point for users to come in so layer 2 will become the execution layer and layer one will become the settlement layer for firefighting the proof and the users enter you know like ECM space through a layer 2 and then we want to build a layer two that stick to all the ecm's values not playing any kind of aggressive marketing not silly but for some research focus on solving the fundamental problem for example build a system that is really usable that works in practice and that is trustworthy yep yep Alex dots I fully agree it's been in the ethereum's roadmap for a couple of years now vitalik has set up this vision for ethereum to embrace the roll-up Centric approach where ethereum becomes a small place to settle things roll ups and ldus in general are becoming the actual place where all things happen you can think of it as in the analogy of like ethereum is an island with where you can only build small houses right like it's it's fundamentally Limited in space now roll ups are something like skyscrappers so you can take Singapore's analogy right like they allow you to have the much more floors much more activity happening on the same surface but uh that's probably still not enough to accommodate arbitrary number of people you know a city of a fixed size can only grow so far even even with super high skyscrapers so at some point you will need to transcend that and go grow Beyond uh what Roll-Ups are capable to doing go Limitless and this brings you to things like validiums and volitions that are more like reclaim of the land on the sea where you can extend indefinitely further Beyond this island yeah if we look at where the liquidity is today on ethereum there are hundreds of billions of dollars of liquidity some fraction of that a small fraction of that has been bridged to various l2s but on the l2s it tends there tends to be more active more activity relative to the amount of tvl so people who are just holding um value are more likely to have it on ethereum people who are actively using it are more likely to be on L2 which makes sense because uh trading activity is cheaper lowered gas fees on L2 but there's been another interesting thing that has been happening that we've seen on arbitrim which is that in addition to having the largest amount of bridged value from arbitrum or sorry from ethereum on arbitrum we've seen a growth of uh of uh tokens being launched directly on arbitrum So currently there's more there are more arbitrary more value in arbitrum Native tokens on arbitrum than there is in bridged uh in bridged eth or or or tokens so I think that's an increasing Trend that we're seeing that the cost of minting and managing a a token is also cheaper nl2 and we're seeing increasingly activity happening there yeah Cindy yeah I mean I in a very simple term I feel that the future is is like you know how we say that the businesses have B2B model versus b2c model sure so for ethereum ethereum initially or till today is like more like use u2c model user to chain model and now we are seeing ethereum moving into this chain to chain model like that means in in next two three years four years you would see ethereum to be a you know the base settlement layer which is providing security and the the you know settlement guarantees and you know uh uh you know security features to these chains but then I see the user activity on ethereum you know on the downtrend only because users would want to interact on the on the layer two and especially now with Advanced Technologies like you know ZK Roll-Ups and optimistic roll ups uh you know which have like with like what arbitrum is doing with permissionless uh you know fraud proofs or uh in with ZK rollers permission there's Melody proofs and all that users will feel more and more confident about out having the value existing on these uh you know fully ethereum secured chains and ethereum becomes the source of security and settlement so it becomes like a chain to chain uh business model instead of a user to chain involved that's helpful and thank you for all your data points I think uh in this current market what we're seeing right now is that majority of the roll-up costs are going towards the deuterium right for data for um settlement as you said so what do you think will happen to you guys and what are you guys planning for to reduce that cost right we're talking about 80 to 90 of the transaction costs uh going to ethereum right I'm curious to hear your thoughts on this um um so for us like you know because every layer to need to settle down ECM and currently like you know what we can do best is that you know we do our best to to move the bottleneck for you know our throughput and uh like you know to to to move that bottleneck become you know Essence bandwidth bottleneck like their da like you know bandwidth limitation and that's what we can do best right now and because before before like tank sharding like even not before because 44 can only make some of your data become cheaper and also some one controversial opinion is that you know even after I believe that when people you really think you know it's after 544 you can become like 10 times cheaper or even more but but after our analysis the reality might be because you know like the first launch of it for four might be just very a few blobs and the people maybe different layer tools will compete for this data blob so even they have different you know like metering for this for this data space it might still be expensive for a while it might be like two times cheaper three times cheaper but not not that significant so only if if I moved to dang Charli and it's really increased you know the parallel data chart and like you know you you really increase the limitation that for for low up throughput so that's why like you know what we are focusing on is that you know we do the best we can before everything comes and then like you know we help ECM to build the future so we are not just building for ourselves but we aim to kind of in the future delegate more engineering effort to help ECM to build and for the whole ecosystem to grow together I think that's that's very important for us Alex um so the uh yeah in order to understand how to lower the cost you need to understand what causes the bottleneck we have different technological components of the uh that make the price of a transaction it's the execution these um data storage data availability or publication and then the um the bottleneck of the sequencer like how many transactions can you just sequence and put into the block and if you look at what's the most expensive what's the most scarce component today it's clearly by far data availability right not the storage not the execution not the sequencing but the uh data availability which you have to pay to ethereum to broadcast the transaction through the ethereum network so that everyone can reconstruct the state so what we can remain a trustless permissionless rollout and this cost is uh indeed not going to improve significantly with ap 4844 with product sharding and that which is hopefully coming this year but if you look at the actual data throughput the bandwidth uh what the AP 4544 is bringing it's roughly a half of what ethereum is already capable today so it's like it's not really yes it's exclusively reserved for Roll-Ups right so initially it will drop the price a little bit but then we're quickly gonna saturate it and it's and we're going to be back up to the prices very hard today so um and uh even if we wait a couple of years and ethereum implements gradual sharding I think it's always going to be a slower increase in capacity than the increase in demand so data availability is the central point of our efforts what we do with ZK sync are two things I spoke about them in the previous talk one is we're using State divs instead of call data like we're publishing not the inputs of transactions but the outputs which is a very interesting approach which allows you to compress data if you do like 100 or 1000 transactions that all right in the same storage slot for this transaction type you only have to pay like a single data slot at the end of the block and so you can take examples as oracles and and Arbitrage transactions and so on like uh which will become massively cheaper and the second approach is just extending the data availability into Evolution into something that goes off chain but still seamlessly interoperates with your roll up accounts that fully inherit security from ethereum and this is what we're building with ZK Porter I I guess I'm a bit more optimistic about this than than Alex um I think several things first I think it is true it is true that currently for uh Roll-Ups the cost of posting data on ethereum is is the dominant uh part of the cost um and our team has worked really hard to try to minimize that cost by using the best data compression and timing when we post data for when ethereum uh gas prices are lower and so on but there is a limit to how far that goes right I think the fact I think currently on ethereum uh call data which is what we use is artificially expensive because the original ethereum design was optimized for a different world and with ethereum's pivot to the roll-up Centric roadmap there's a new emphasis on data availability as one of the core functions of ethereum and you see the eip4844 that is coming in a few months as a first step in that direction but I think over time the cost of data on ethereum will go down so it's not too much higher than the cost of General commercial data storage um and I think people will still want to be able to do things more cheaply but I think the advantage of this sort of external data availability Solutions will be less as ethereum kind of matures into a mode where data availability is one of its major uh is is one of the most important parts of that protocol so I'm optimistic long run that these costs are going to come down in the meantime we do offer the option of external data availability for people who are willing to make the trade-off of not having the strongest guarantee of availability which you get from ethereum but but people are willing to trade off making a just small additional trust assumption in order to get lower costs I know that all of us are probably exploring those sorts of approaches as well yeah thank you yeah I mean we from our side we try to take like a 10-year 15-year you know outlook on this and uh you know I agree uh with with the fellow panelists here that um you know in the future even when ethereum brings in uh you know Proto uh dank sharding or I mean uh you know this sharded version of uh data availability and we have a huge amount of data availability is that still enough for the whole world to bring to come to web3 I am not sure about that so that's why like I think that I have a more use case specific view uh on that and which which also very aligns with the what Ed was saying is that you know the for the people for the use cases who are willing to take the trade-offs they can still have like off-chain data availability Solutions in very it is different ways for example gaming and you know social use cases and all that you don't need ethereum's data availability guarantees whereas for defy where you know you would want uh that those kind of guarantees and then only the pro the the chains which are actually which want to have huge amount of liquidity and the D5 applications they will fight for or bid for those those kind of data availability costs and actually users are would also be transacting on a much bigger uh you know kind of transaction value so they would be willing to pay uh that kind of fees having said that like you know polygon is like is known for its ZK promise we also believe that because of the ZK validity proofs the the role of data availability in terms of providing the the security of course it's still there but it kind of gets reduced to a very large extent because if you have a roll up uh you know data will validity proof ZK validity proof then as a sequencer you can do this kind of long range or like you this kind of like data withholding Ransom attack which is kind of which is possible for sure but it's kind of looks very very like rare and it will be possible in very few kind of scenarios so that's why like I believe more in the use case specific uh you know kind of security with with ZK proofs you have fair enough security all the validiums also will have fair enough security but if you want D5 level of security then go for the ethereum data availability guarantees thank you for that so it sounds like we have the group who thinks it's depending application specific and user specific and then um these two gentlemen here thinks more that it's more often more optimistically that eventually we will just um get to the good and safe stage um I love to delve into decentralization right this is a topic that's been kind of pressuring I think a lot of the layer two projects and from the overall crypto Twitter and the crypto industry how are you guys thinking about this I'm particularly interested in which one do you decentralize first verifiers sequencer what do you guys think of where are you in the road map here maybe you could start um sure um I think for us like because we are building a liquid up solution so they are like you know multiple parties inside mainly like you know there are two parts you need to decentralize one a sequencer um so sequencer is generating your block and then there's a proverb proverb is basically taking a Blog and a general proof attached to this block saying that you know all the transactions in this block is is valid and that is amazing proof of Unchained so for us like you know because all the robots are in a stage where you are not you know really delivering what a layer 2 should deliver like you know you haven't removed your multi thick you have multiple training wheels like you are in a transition stage even like on Monday they are you know Alpha magnet beta magnet they are different testing stage and uh too early to decentralize everything is that it is a very dangerous signal for me because you know like even if you decide so before you think about designation you need to think about like why you want to design a sequencer and why we want to design approver and for example like some reason for design sequencer is that you want to avoid like in a real-time censorship resistant reset resistancy and there are some other like you know white single point failure but there might be some other approaches like you know which people are you know proposing like you know using some other crypto you know graphic parameters like maybe you can encrypt your transactions so that you know you still can't you know like since a certain type of transactions and there are some other way like you can send your transaction layer one and the enforce Layer Two to execute that transaction so like before like you know we talk about decentralized we need to think about like you know what's a what's a what's a goal you want to achieve and then like for us we believe that you know design a signals are first at least right now like you know we have multiple proposals for that we already have some kind of proof of concept implementation for how we decentralize this but it's in early stage we want to discuss with the community about how we decentralize and what's the philosophy and what's it going to achieve and instead like we might want to decentralize approval first so there are multiple reasons for that one is that you know prove identity influence you know the block generation so otherwise you are producing some legit block even if your circuit has some bug the prover can only do some stateless proof generation for for certain blocks so it lower your your risk for you know like being hacked even if your system has embarked and the second thing is that people are really bullish on like some some Direction like you know making they can even become time faster which is using Asic using customized Hardware using even more technology to make proof generation become faster so we want to encourage people building better and better proving technology by opening this door to the community so they can start earlier because they know like there is a huge opportunity to make fourth generation faster and then to to enter our ecosystem so that's our goal and also we have published our paper talking about how you use GPU AC and fpga to make fluoride faster um the first academic publication and so I think that's that's very important to start early and then after one year you you deliver those kind of Asics and then you can make you know your DK proof system become even more more efficient and also avoid some single point filler in your proof system but also trying to avoid like you know the fast improval of the wing and I hope some you know Philosophy for that so our roadmap like firstly we are building a really simple to use transparency system and it works and it works very well and then later decent has approver to avoid single point filler to incentivize the community to build better you know Hardware um and then like to make this further more performant and then later move to more like a sequencer designation um and like you know the philosophy inside yeah thank you I think he did a really good job going through different components or aspects of decentralization so I'm going to do the same and and offer our perspective so we'll start with the validation I think we can take a moment to celebrate a decentralization of validation at least on ZK Roll-Ups today because every single transaction that happens on any Ezekiel rollup is enforced in validity is verified by every single full note on ethereum so we're already more decentralized than any other scaling solution the second component is the sequencer like who decides what transactions go in the uh in the blog this is a massively important component like from the perspective of the mission and for physically sync everything we do is measured against the north star of the mission which we outlined in the ZK credit Manifesto so um you need to enforce censorship resistance and incredible neutrality of the transactions and dealing with Mev in a credible neutral way and you also need to make sure that the system actually remains resilient to network outages to like like you don't rely on a single entity on a single validator for the availability we see it with the Roll-Ups and Roll-Ups the sequencer sometimes goes down which is unacceptable for blockchains right so like just for this from this perspective alone we have to decentralize the sequencer as soon as we possibly can and GK sync is working very hard on uh with high priority on the consensus based decentralization of the L2 the third component is the prover and the prover is important but I would argue that if you have decentralized sequencer and you have permissionless proof generation and and then the the the leader election in the sequencer basically places the responsibility to generate the proof on the leader for the current block you already halfway there and if the leader fails then it just like goes to the next leader and so on so like if the sequencer is decentralized your approver is kind of already half decentralized so when we think about this conversation with approver it's really about decentralizing the proof generation to an extent where the prover is actually distributed across like it's not done in one single cloud or like in in multiple different Cloud providers it's actually widely spread across different small machines a lot of like you know just like mining is done potentially on the uh consumer hardware and on gpus and Mining machines on gaming machines and so on so for the we made like it's really important far goal for us to decentralize sequencer to this extent this is why the approver the oh sorry the the pro technology we're using is already optimized for Consumer grade hardware and the new proof system we recently introduced called boojam is provable on our gpus with under 16 gigabytes of RAM the the very last part is the decentralization of the governance and I I think that that's ultimately the most important thing to reach credible neutrality when no single party can decide on the future of the protocol and the way it evolves and this is something where um we will announce plans but this is something we think a lot about education yeah thank you and please sure so in in designing arbitrum and building it along the way we we've had a decentralization roadmap in mind and we've been executing that as we go along and and being careful not to introduce things during the development process that would get in the way of decentralization now I think it's important as Alex said to recognize that decentralization is really about reliability it's really about making sure that the system can operate the way it's supposed to operate and it is intertwined with Notions of reliability and and sort of liveness and progress that you're going for I think it's also important Direct recognize that we need to we need to think realistically when we're thinking about decentralization if we say that something can be done by any user then we'd better make sure that it's the case that an actual user with the actual laptop they own will be able to do that function otherwise it's really just an on paper sort of right to participate so what you really care about is making sure that you have resilience and making sure that you have decentralization of power in the system and that is both Technical and economic so as far as the progress of of arbitrim Technology on that decentralization roadmap we're getting close to the end of that road map to get to the decentralization goal the making the validator role fully permissionless for that we have a code that has been published with the security audit and it's in the pipeline toward being Deployable with the approval of the arbitrim Dao on the on the main chains and then the remaining piece is sequence or decentralization and we just announced a couple days ago a research partnership with espresso systems to a research and then develop a decentralized version of the arbitrum sequencer so we can see the end of that um decentralization roadmap we have executed most of the steps and and we're pretty close yeah from our end also like we have been very lucky at polygon that we have uh actually three different ZK teams which are you know progressing in different directions so you have polygon ZK VM which is like a single sequences roll up that we have launched and then we have polygon 0 who is building like you know who's breaking the boundaries and building new paradigms into into how you prove chains and you know they we are already very far ahead in the in the process of uh you know creating ZK proofs where you can prove the entire decentralized change itself which essentially means the decentralized validator chain you can directly you know prove them so all of these things are under process under under very high progress from our side uh and uh so you know we believe that once these kind of probability is available you can easily you know you can easily decentralize the uh you know sequencer or validator or you know these combination of these roles but I think like we again on this one we have a very use case specific view on this for example like if you are running a public defy chain like you know and and which has let's say these uh lending borrowing markets which needs to be have which needs to be having liquidations and all that there you need to ensure that there is like credible neutrality in terms of uh you know achieving liquidating transactions let's say you know I mean if there is a large D5 chain and then there is a liquidating transaction coming in in a very high volatility environment and let's say some of the Insiders like till till now we have great very responsible projects which are doing all of these things so we don't expect that to happen but then and we trust these teams to not do it but in future if some other D5 chains which doesn't have these trustworthy uh you know people sequences running the chain and there is a such kind of scenario which I am talking like there's a liquidating transaction and that and the party running the sequencer itself has a big loan which is a which will let's say can get liquidated they can essentially censor the transaction so apart from the liveliness which uh you know both Alex and Ed said you know in terms of the liveliness guarantees of the chain it's or the reliability of this uh this this network uh you know decentralization is important but for D5 change specifically you need decentralization for these other purposes like instant censorship resistance also to have the fully functioning uh you know and and vibrant permission less D5 ecosystem over there so for that kind of use case you need you know decentralization uh absolutely but then let's say again you have you have an application and you that application has a social applic social uh uh you know let's say dab built on top of it and all that there you might want to rely on a gaming application you might want to rely more on uh like a centralized sequencer which is providing a good enough service and and uh you know you don't need to decentralize but you want to go with a higher performance for the better ux or the users so we have a very user you know user experience and uh use case focused view on this particular particular thing but as I said that you know because we being uh very lucky in having three different ZK teams you know uh one of our teams is already you know pretty far ahead in that and we'll be able to have decent life sequences uh and decent live streams very easily for the people uh or for the builders and also we will add these functionalities into our polygon cdk so you can actually we give an analogy of having a Subway sandwich like you if you are launching a chain you can choose what you want to do you want to see centralized sequences decentralized sequencer what kind of data availability you want to choose and you can have your own uh chain as per according to your requirements this is super helpful thank you everyone's already plotting given how great these answers are so thank you so much we're out of time really appreciate everyone coming on stage and giving their perspectives so we look forward to hearing more from each of you and your projects thank you