Ethereum becoming the settlement layer of Web3 - A case study - Marek Olszewski

us about ethereum becoming the settlement layer of web 3 and present a case study of how cello became a layer two from layer one please give it up for Marik I am going to butcher your last name I'm so sorry oleski thank you thank you it's a hard last name even even in Poland um uh hey guys it's really great to be here um so today I'm going to be talking about you know ethereum becoming kind of the settlement layer of web 3 you you've probably heard of um Proto Dan sharding and dank sharding and this huge change that's happening in the ethereum ecosystem and so we're really excited to be part of that big change uh at cell and uh we wanted to share a little bit about our journey of transitioning cell from an L1 to an L2 uh and doing our part in this um kind of new direction for ethereum um and if you like this talk uh definitely come chat with me afterwards happy to answer any questions uh and then also feel free to follow me on Twitter uh my um username is just maror so you don't have to worry about my last name um at least when finding me on Twitter cool well so I was here um in the other venue uh this time last year uh at eth Warsaw uh talking about um our proposal back then to transition cello from an L1 to an L2 um so back then we had just started thinking about this big uh transition for the cello ecosystem uh we had a proposal that was out um in governance and people were discussing it and and voting on it in with onchain governance and um you know fast forward to today um we're really excited to share that uh we have a live test net um a cell L2 test net that we're calling Dango all of our test Nets are named after delicious treats from around the world Dango are these Japanese Mochi uh desserts and you can think of Dango as really the best of C really packaged in um kind of this L2 uh test net so Dango also are pretty convenient in that they have an emoji and you'll see most of my slides are very Emoji heavy uh it's another big plus for for the name Dango but again you can think of Dango is the best of cell so all of the nice things that cell offers the ability to pay for gas with tokens native even without account abstraction uh ultr green money which is this idea that every transaction on cell has a portion of its base fees go towards buying retiring carbon credits to offset the carbon uh of the chain and then all of the nice mobile features that cell offers including social connect which is a nice um way of mapping phone numbers to wallet addresses in a privacy preserving way so that you can find and send value to your friends using just your contact lists on your phones and so all of that now it's kind of being packaged in the form of L2 and Dango is kind of really the first step towards our upgrade towards upgrading cell from an L1 into an L2 and all of this again is happening kind of in the backrop of um you know ethereum becoming kind of the settlement layer of web 3 you've probably heard Justin Drake talk a lot about this you know I think this is a real an important theme right now in the broader ethereum ecosystem and you know we're proud I think we're probably uh one of the first if not the first major l1s to announce that we're making this transition and we're also uh probably going to be one of the first ones to actually do the transition uh later this year um and we really hope that cell can be a nice case study for others to follow so if you're working for an ALT L1 right now uh especially in evm compatible one you know this might be you know really interesting for you and if you're if you're not then you know I think some of these learnings may still be I think potentially um potentially interesting um so yeah let's dig in but before we do that and this is a big butt um turns out that transitioning alive L1 2 and L2 is hard uh and the only Emoji I could find for the concept of hard was this rock um and you know when we first approached this we thought hey this will be easy this is just like any other hard Fork right uh we wanted to make the seamless for developers building on top of cell so we're not launching a new chain and then telling people they have to migrate um their data themselves we want to do this as a hard Fork right so every deployed dap on top of the chain won't even really notice that the chain suddenly moves from being an L1 to an L2 but to do that in a really seamless way it turns out it ends up being a hard hard Fork so two rocks and a fork and you know you might be wondering well why is it so hard and I think one of the best uh metaphors uh for this is kind of this rebuilding of an airplane while it's in the air metaphor you know cell at cruising altitude right now um and it's got a bunch of users inside its fuselage they're happily transacting there on their phones uh and we're literally rebuilding it moving the engines from one place to another um and that is terrifying um especially if you think about just the amount of activity that's happening in inside this Fus Lodge so you might be kind of curious how much activity um one uh interesting framing um is this one that um renie here tweeted about you know if Cell were an L2 today and you were looking at daily active users cello would be a top two uh L2 by daily active users this is already a little out of date uh cello is processing now or C is serving I should say around 650,000 th000 daily active users uh nowadays uh so it's quite a lot of users sitting in that fuser Lodge uh if you look solely at uh daily active users who are transacting with stable coins which is a real Focus for cell we're doing a lot of mobile payments uh focus on Emerging Markets a lot of real world use cases uh if you do that then as of recently uh solo is actually a top two chain so we just overtook um BNB chain so this is BNB chain in number two and right here cell just overtook them and Tron of course is number one but we are um I looked at the data today and and actually it's uh even closer now so we're um you know on track to overtake Tron soon so pretty exciting and uh you might be wondering well what's driving all of this you know I'm not buying meme coins on SOL so I've never heard of this uh and that's because um most of this activity is happening in the payment space specifically in mini pay uh so mini pay is a really delightful really easy to use um payment product built directly into Opera Mini Opera Mini is a browser with 500 million installs 100 million monthly active users it's primarily used in Africa uh throughout the continent of Africa and and they've done a really wonderful job leveraging all of the nice features that sell offers the ability to pay for gas with tokens social connect so you can send money to phone numbers they've created this really nice wallet stable coin only that you can use basically like uh revolute or venmo but it's fully um built on crypto rails and so that's driving a lot of those this um usage um that's also created kind of a verocious demand for usdt which launched recently on solo uh token terminal was pointing out how fast us DT Supply is growing uh that's taken uh sell to be a top 13 Chain by stable coin market cap which is pretty cool we just overtook Phantom I think two weeks ago which was a big milestone for us and then blast and Stellar earlier uh in the year it there about 339 million now stable coin $339 Million worth of stable coin value on solo and when it comes to volumes people you know say that's nice but what about the actual amount of volume that's being processed this summer has been pretty pretty great for us um in June we processed almost a billion dollars worth of stable coin value so a lot of value right this is not something that we want to just willy-nilly you know upgrade and you know ye out a new um L2 we want to be really thoughtful we want to do this really carefully and so we've been really uh working incredibly hard in fact a bunch of the team is here um they've been working incredibly hard for the last year on doing this in in a way that is very thoughtful um very buttoned up and so um Dango of course is the first step towards this very thoughtful release and uh with the release of Dango you know we have kind of five learnings that we can share now from kind of our journey of uh doing this work so that's number five that's not stop you know it's the only Emoji I could find for that okay learning number one by the way um one somewhat fun game is to yell out the description of the Emoji when you see it so what's that one is it point up I think it's correct okay so learning number one is use an existing L2 stack so when we first started this you know over a year ago we said hey you know we've been building um in crypto for a while we have experienced doing complicated you know infrastructure projects uh we can build our own L2 stack um why not uh and we really consider it for a while but um the more we thought about what would be ultimately good for the broader e ethereum ecosystem the more we realize that more stack fragmentation um on the L2 side wasn't ultimately going to be that great and so that ultimately led us down the path of picking an existing stack and uh we ended up picking uh op uh and uh Ian da uh we picked them already about a year ago and then we uh did some additional further research um really an in-depth uh I would say deep dive into all the different Stacks um and then had another governance vote to reaffirm that decision earlier in the year and we think it was a good Choice uh especially if you look at um some of the throughput numbers right now um in the L2 uh space 74% of all L2 throughput now is powered by op stack um in the super chain Alan that's 57% this was in June and I think that number has probably only gone up since then so you know we we we made this decision long before this kind of Rapid uh rise of P stack happened earlier this year but certainly it's it's pretty comforting now that it's happened and so that's pretty exciting learning number two yes Embrace ethereum so uh when we made this uh announcement you know our goal was to settle on ethereum was to order our transactions uh using ethereum and have a native bridge to ethereum and to to become a true L2 and um you know we think that for any other alt L1 that's out there especially evm compatible ones which I think arguably you know in this day and age probably shouldn't exist anymore as Standalone l1s uh if you are looking to transition to an L2 do it right uh actually settle on ethereum um and um and so we think that embracing ethereum is just really really important you I've heard from others you know we too want to become an L2 but it will only be an L2 in name only maybe there will be a native bridge but it won't actually settle on ethereum maybe ethereum will also be an L2 for our chain uh we've heard I've heard all of these different descriptions of what an L2 or a rollup means to some of these alt l1s and and how they how they want to kind of become closer to ethereum without ultimately giving up settlement and we think that that's a mistake we think that you should ultimately Embrace ethereum um the whole ethereum ecosystem has been really welcoming really excited for us as we announced this proposal last year and as we've embarked on this journey to make this transition and we think others should do it too and one interesting way um if you're an engineer uh to do this um is to really distill your protocol down to a patch Q against gu so that's kind of again all of the nice things about sell being distilled um right now Salo is its own code base and we've been working really hard to take all of what's Best in cell and create a patch Q uh that we now apply against both op gu and the optimism um code base and we think that that's a nice way to um be able to merge Upstream changes more easily and to um ultimately get closer and closer to ethereum and ultimately embrace ethereum learning number three there's a word over there that might that's right um okay and also you know there is no other Emoji where you have three kind of fingers up so this is what I had to use um migrate your chain data so this is an interesting one so Salo was launched before ethereum became a proof of stake Network um and even though it was you know e compatible uh and based on the G code base we removed all of the proof of work headers or the proof of work related fields in the headers because we didn't need them we were a proof of stake chain um but then later on people are like hey my code doesn't work because this um uh Uncle field isn't there and I need it to be there uh and so we re added it back um we also had other like Legacy transactions that we used for a while that we've now deprecated ated and this now means that c as it exists today has a lot of code that handles some of these Legacy headers and transactions and other kind of pieces of code in the code base and we can Port all of that into op stack or we can actually migrate our chain data and repackage it in a way so that it um works with uh a new code base that is more like ethereum like the op stack code basee and so this is what we're doing we're migrating our chain data and we recommend others kind of with similar uh in similar positions to consider this as well so what this looks like is you've got your um L1 blocks um being produced and then at some point in time this upgrade is going to happen this hard Fork it's going to upgrade to the new client which is an L2 client and that new L2 client will create new blocks but before that upgrade can happen we need to migrate all of the chain data for the entire history of the of the chain and repackage it in in a way where it's compatible with a new client without having to have any of that Legacy code to um support any of those Legacy fields or transaction types or or any other peculiarities and the nice thing is you can actually do the majority of this migration in parallel you don't have to stop the chain for very long because for a lot of the old blocks you can actually run all of this in parallel as the chain is executing which then allows you to have you know just a minimal amount of downtime um when the actual upgrade happens and so um optimism also did this with their Bedrock update and this is another reason why we were really excited about U picking the op stack um it really uh made this a lot easier for us to do on our end and we really recommend um other chains for are considering moving again from an ALT L1 um to an L2 uh to do the as well so that they can remove a lot of that Legacy code that they probably have in their code bases and again become closer to ethereum and you know the one interesting caveat here is that you may care about those old transactions right maybe you want to trace an old transaction using an archive node and if your new L2 archive node doesn't know how to uh reason about those transactions then you're not going to be able to to trace those transactions that's not ideal and so one solution for this is to proxy those requests from the L2 archive nodes to L1 archive nodes um and um in this way you can actually have a really nice code base uh on the L2 side and then you can have all of your legacy stuff um sitting um in these um Legacy L1 archive nodes and since the number of requests to your archive node are not going to be um mostly they're not going to be looking at these um really old transaction types you can kind of um only keep a few machines around uh to service these requests and you can kind of scale the L2 archive node with demand so it shouldn't create a lot of overhead for people running archive nodes or running RPC nodes I would say um learning number four I haven't seen this Emoji before what is this this is Photoshop because there is no emoji that represents four um but what is the learning number four the fourth learning is really around rethinking where your token lives right so if you're an L1 and you're transitioning to an L2 you have the choice to also deploy that token on ethereum and this is exactly what we're um looking to do and and we're looking to do this for two reasons one it makes us a lot more compatible with kind of the super chain which is really nice uh but two it also allows us to encode some properties uh around the solell token um that can now live um you know immutably on ethereum so for example um cell has a a$1 billion Supply cap we can actually enforce that now in ethereum and the way that you launch a token for a live L1 on ethereum is you at the time of the upgrade you can create the token and then you you put the entire Supply into the native Bridge um and that way you know the entire Supply is in the native bridge and um uh folks can move their tokens from the L2 into the L1 if they want to because it's already living in that native bridge and so this is pretty exciting okay learning number five five five I don't think it's high five stop I don't think it's stopped I think it's Palm I think face palm it's not face palm um okay the fifth learning is to do multiple upgrades right we talked about how much value there's on the line here we talked about um um you know the the risks and the stakes um Stakes are high you want to do this carefully you going to do lots of upgrades um and you know that's why our road map looks a little bit like this we've already done this uh Dango upgrade so Dango is an actual Fork of our developer chain Alpha Horus we forked it and then we upgraded it um next up we're going to be upgrading the alpha hores um developer test net itself without forking it that should happen either later this month or early next month and then we have another test net and then finally main net that we're hoping to upgrade by the end of the year um and you know we're doing all of these and we've done countless more countless more I think ja is in the audience and he's probably done I don't know how many maybe hundreds at this point um and we're working really hard again to do this in in as thoughtful uh of a manner in a way that doesn't put any of those funds at risk and certainly doesn't impact any of those um many many users that we talked about before okay so if you're curious and you want to learn more um you might want to check out the salad Dango um block Scout instance big big uh kudos to the block scout team for launching this uh in a very timely manner um you can find it at cell- Dango dobl scout.com uh you can also experiment with uh the testnet native Bridge which you can find at Super Bridge so you can go to test nets. superb. apppp and you can start building and one of the nice things again about this whole way that we're doing this upgrade is that anything that you do on tapac Cella today uh will work seamlessly tomorrow after the L2 upgrade later this year and so if you're interested in the hackathon um today and this weekend we have some bounties definitely definitely check us out come talk to us we have a booth we're here to help um yeah and let's build something exciting together we have time for two questions does anybody have a question for Maric that was an amazing presentation who here would like to ask him a question let me look around the room any questions yes please let's run a mic to him yeah I have a question why did you choose optimism instead of arbitrum for example yeah really great question um so we looked at optimism we looked at arbitrum we also looked at ziki syn in in polygon ziki ebm um actually all of them went through our um uh forum and went through kind of our community governance to try to entice our our community to kind of go and vote for them and ultimately the community was like these are all really exciting but we don't know how to evaluate them so then they ask clabs which is where I work uh to evaluate all of these different stacks and so um our team worked really diligently to look into all of these uh and they're all really in exciting they all have lots of pluses and you know some minuses um I think um couple of things that really resonated with us with OP stack was the the minimum um diff Philosophy for GU having maintained the GU fork for a while we felt the pain of merging Upstream changes and we really like the design of separating kind of gu from the rest of the uh L2 stack so that was one thing that was really nice certainly all of the work that they had done on their Bedrock upgrade uh everything around chain data migration everything around um the um archive node proxying was really really interesting for us as well um and um I would say the license was also particularly compelling I think um um you know you mentioned arbitrum arbitrum has I think since changed maybe they're licensing a little bit but not still not quite as um liberal as as op stack I think op stack should get a lot of credit for having such a liberal license um I think it really is appealing to a lot of Builders thank you