Scaling Wars: The Race to Speed Up On-Chain Infrastructure l KBW2024

amazing hi everyone I'm Diana from 1kx we're a crypto investment firm um and I'm delighted to be joined on stage by our guly four of the most exciting projects in the space right now to discuss uh scaling Wars I'm not totally sure about that choice of wording um but we'll Begin by um some introductions so if you could please uh introduce yourselves in your projects for sure hey guys I'm Smokey I'm one of the co-founders of bar chain bear chain is an evm identical L1 powered by proof of liquidity which allows us to turn liquidity into security and use it to power the applications building on the network such that the chain rewards its users and applications hey everyone my name is NC part of the founding team here at Mega meie is the first realtime blockchain by settling on ethereum where able to achieve a level of performance that is really a zero to one for applications so 100 ktps 1 millisecond block times hi my name is keone Han co-founder of monad labs we're building monad a br performant evm equivalent layer one that's introducing parallel execution a high performance database and a bunch of other improvements from the ground up to help the evm scale hi everyone I'm Sydney I'm the product lead at Eclipse we are building ethereum's first svm L2 um the easiest way to think about it is we are building salana on eth so let's start by uh stepping back a bit and thinking now it's been nine years that we've been trying to scale ethereum um so what's the argument for scaling evm chains maybe ke we could start with you yeah sure thing so I think the first thing to keep in mind is that there are incredible Network effects from the evm um there's so many libraries that are built out so many applications and if you think about the core premise of blockchain one of the biggest arguments for blockchain is atomic composability so being able to compose on top of a rich landscape of applications that have built um Lego blocks like building blocks um for defi is incredibly powerful and at the same time it's ethereum is really Limited in throughput right now for a variety of reasons um but it's really necessary to rebuild execution stack from the ground up to massively lower fees within the L1 and ultimately serve many many more users any other takes on on why why we're here yeah so I mean I would highlight I think there is a difference between scaling ethereum and scaling DVM I think the value of scaling ethereum is that ethereum is the most decentralized element layer in the world so focusing on uum and scaling uum is different than scaling DVM but I definitely agree with every Keon has said around how valuable the evm is and the network effects around the evm I think it's also very fair to say that the evm is perhaps you know to date the most battle tested you know piece of infrastructure that we have right um bear chain itself is evm identical but we're not an eth L2 we're not you know eth in that manner but I do think that when it comes to the degree of liquidity that degree of developers and the degree of users they're already um you know active and engaging in an evm adjacent ecosystem uh it gives you a pretty rich base to to start compounding from yeah and I can quickly add on to that um when we were first designing our architecture we were we were very intent on kind of selecting the best pieces of the modular stack um that's why we end up going with you know the salon virtual machine as the execution environment um celesia for da as it's scalable and verifiable and ultimately we decided that ethereum was the best fit for settlement because of you know a lot of the reasons that was mentioned before you know it's highly decentralized it's Lindy um and our validating bridge is on ethereum and we're able to inherent a lot of the the benefits of um L1 security as a result I'd actually love to dive in there more and with the salana ecosystem being obviously popular and somewhere where we can go for Speed um to understand how you think about competing with salana if you do um and maybe Sydney we can start with you and I'd love to hear more about like is it um is it competitive and and how do you view that yeah um so we actually get this question quite a lot as you can imagine you know why not just use salana and I think for us we fundamentally have different design goals um as in L2 we deeply value verifiability whereas salana it's less of a priority as toly had stated in his you know endgame presentation or keynote at uh breakpoint last year um so that's something that we care a lot and as an L2 there are certain advantages uh that we that that we uniquely have right so the first one being um as an L2 we can we don't have the same burden of consensus as the L1 and we can fully so focus on or solely focus excuse me um on the optimizing of uh the block producer so that way we can really up the harder requirements for our sequencers while keeping re-executing costs relatively low um and that way we're able to get a lot more performance out of it um and I think there is a world where Eclipse will be cheaper and faster than salana however I do think that we serve different user base at the end of the day um we are more so targeted towards the ethereum users I really care about eth as kind of the canonical asset and I think we're actually complimentary as in eclipse and salana in a lot of ways in that you know we obviously benefit from a lot of the research that's being done on the salon front um and we're looking to contribute research back as well um and we're working we're actively working on open sourcing many parts of our stack yeah I can go next I think salana is awesome and when I wonder why why do I think this is because there's experiences on salana which are unique applications that there are only possible on salana I I think that's awesome I think that's cool when we look at Mega our focus is something similar right how can we facilitate a level of performance that can actually scale to you know hopefully millions of users fully on chain right and for us I I don't necessarily know if that's competitive with salana I think ouris is there's a level of performance which is fundamentally Impossible on layer one and it is possible on Layer Two giving this given the security Frameworks uh that you have as a layer too right especially if you settle on ethereum so yeah I mean our focus is on realtime apps I think that all the applications which were built on meie are applications that couldn't exist on salana so we're not necessarily competitive in that manner but I am just a big fan of like all of the Innovation being done on the layer one uh you know like scaling road map we just don't necessarily think it's sufficient for extremely high perf fuse cases so with scaling we always have to talk about tradeoffs um and I'd love to get your takes on what's worth sacrificing in the name of speed um or is there how how are you addressing that question uh Smokey should we start with you yeah for sure you know I think that when we look at this panel we're probably the the slowest of the in that be chain I think is very much focused on um you know providing a solid user experience and providing a product Centric experience uh but from I think from a different perspective right so to answer your question uh directly Diana I think that one should have um you know few trade-offs when it comes to speed but very carefully chosen ones um I think decentralization is probably the the most copout and real answer in that case and that as you have a wider and more distributed Valor Network it simply becomes a little bit harder to make that all happen but I also do believe that there's a happy medium that can exist there so when we look at something like bear chain um which will be able to support l2s rollups Etc um you know while we might have a 1second block time ourself you know we can actually think about how that could support many other things like Mega E in the future um and in our case you know as a team that uses Comet bft consensus we'll still have a validator set of 256 and hopefully growing validators but our Focus isn't necessarily on I'd say um improving the user experience the product experience from a speed point of view I think it's actually improving the user experience from a a cold start problem point of view um because I think it's very easy for us to forget that you know we can build Ivory castles and we can build like Towers in the sky about the most interesting technology but if we don't build products that actually users want to get their hands on and access then it's all kind of useless and moot um and the whole idea of of what we're doing at proof of liquidity is not necessarily optimizing for Speed but optimizing for actually driving fundamental economic value um to applications that are building on the network such that they don't just have a tech advantage in terms of evm identicality by building on be chain they also have a of bootstrapping their own liquidity bootstrapping their own user base and making it easier to get off the ground and really get that activation energy which tends to be very difficult as one goes from zero to one building in a relatively Niche landscape like like web 3 right now right so um I think the decentralization this the the easiest answer there but it's definitely not a perfect one by any means I think that we're all in the space because we value decentralization we value the implications of decentralization such as censorship resistance self- custody um and we value the fact that there doesn't need to be a centralized Authority that is administering all balances and has complete control over um the state of the world instead there's shared Global state with hundreds or thousands of nodes running um synchronously like keeping up with each other and maintaining this this shared Global State um the implication the implications ultimately are quite far-reaching but just to mention one specific one like if you're a small business owner you could take payments directly on the blockchain just by running a node and validating the all of the transactions that are going by and um knowing the entire Global state so in order to do that we actually need better technology that can help um make this network much more computationally dense and support many many many more users I think that's the fundamental job for the L1 and um you were talking about salana earlier so I think salana has definitely done a good job of showing um all of us here in this room what's possible and it's on us to then take up that mantle and um help carry the evm forward given the network effects yeah I think we've made some pretty interesting design decision design decisions so I can Chim them in I I think aerum is the most decentralized layer one and I think that's very powerful for us our FIS is okay if we have this extremely decentralized settlement layer why don't we Leverage The Fear the layer to security framework to build the most performant execution environment right and those have tradeoffs very very obviously so we centralize block production we have maximally decentralized block validation but we still are able to centralize block production to achieve this very very powerful level of performance and I think the reason we do that is because we're able to effectively settle on etherium as opposed to another layer one which might not be as decentralized I think that that is the complimentary architecture of Mega E with etherium and it allows us to take such extreme tradeoffs in the name of performance because we settle on Eve I think uh all of us you know land somewhere on the spectrum of decentralization versus performance we will have we have taken a different design approach when it comes to that um for us I think we're somewhat similar in in kind of what uh namk has said in that you know as an L2 because we've outsourced kind of the centralization and inheriting some of the uh security properties from the L1 we're able to push performance a lot further on the L2 um there's obviously trade-offs with that in that we have a smaller block producer set um however we are mitigating this in a couple ways the first one being we're inheriting um eventual uh censorship resistance from the L1 through uh Force inclusion um number two we are enforcing safety via uh fraud proofs and by keeping re-execution costs relatively low and lastly um we are starting with a small decentralized sequencer set which will enable a larger degree of liveness and sens sensorship resistance as compared to a sing single sequencer setup I have a question for um I guess both eclipse and mega eth which is you know how do you guys reconcile the fact that there's this 7-Day fraud window um if you have really high performance that means in those seven days that's a lot of transactions to go by like how do you how do you think about that yeah I think you make a great point it's almost like a suffer from success situation if you have so much activity I think the real answer answer is it's a it's a waiting game for us to get to a point where we can migrate away from optimistic architectures and towards zero knowledge Solutions which aren 7day periods yep same here we're planning on using risk zero for our ZK fraud proofs um but eventually OB we want to go full ZK and really want to shorten that window to whatever is is possible and best for the user and followup question on that for for Mega eth even if you're using a high performance client does that not get rolled into uh other evm rollups or how do you think about sustaining your remote as a high performance chain yeah I think that um I think I understand your question uh let me know if I don't I think big picture there is we're very excited to see architectures like Mega Eve pick up because in the world where you have multiple 100 ktps execution environments being filled is a world where everyone is using crypto and that's like a directional win I I think before we get to that stage though we are very big fans of you know let's let's try and have as much in one execution environment as possible right I think that the ethereum road map while we are big fans of you know the the effective like security framework of layer 2s we don't necessarily think l1's can scale to the extent that is needed I think that maybe people have gone a bit too far with like let's have a cadrillion rollups let's have like app trains left and right and then you lose like Atomic composability which someone on this panel mentioned so I think our feis is at least in the short term let's have uh effectively one Mega rollup right and that's where we want to see as much performance done we want to continue to optimize our product to get to a point where we can have more than 100,000 transactions per second keep working down that path but who knows maybe in the future other people will be able to build their own Mega E FL products who knows and keeping on the topic of Moes so in a world of chain abstraction how do you think about uh having a mo as a chain and and I'd love to to hear from each of you about what you're building and and offering for the communities and and what's drawing them to to each of you maybe if Sydney we can start with you yeah this is something that I'm actually very passionate about um I'm a firm believer that user friction should not be anyone's Moe um so in that regard eclipse is very happy to be you know both the destination chain and the origin chain however I think chain abstraction only makes sense in uh a world of you know infinite chains which fundamentally goes against our thesis um we believe that in a world where we're able to support app specific sequencing uh local fee markets high performance and value redistribution uh we're able to kind of support the mass majority of us cases and maybe there's less reasons for apps to become app chains and in general um I think there is something to be said about Atomic composability as as it was mentioned before um but that's how we think about it I think that you know one thing that we've seen in the space is that um layer ones if um you know if there's the right mentality around community and there's the right level of respect for um just the incredible power of community then um it becomes something that it the community basically becomes um an entity that can carry the mission forward and spread the word of decentralization and drive users to the applications that are built uh within the layer one and so for me personally I'm incredibly grateful to um the incredibly passionate monad community that um probably everyone here has heard about just how awesome the manad community is it's really just the result of individual community members that are contributing at an incredibly high level and really passionate about the space and really creative and um now are supporting ambitious builders that um are trying to bring decentralized apps to life so I think that you know when you ask about the Moe I