A Radically Simple Way to Scale your dApps - ETHGlobal Bangkok 2024

it stays with like or should I get it okay yeah I might I might have H yeah I think I think that's better I hear myself more all right so first things first uh thank you for the opportunity to be speaking here uh it's very nice and I about a radically simple way to scale your depths and what that means is a way for you to create depths that are faster that are stronger that are smarter and you can go beyond the capabilities of the think of the trading that is being done right there you can you can do amazing solutions that we're going to cover later but right what's the Doom Cas well someone another day posted on Twitter this was roughly a year ago or more uh well if etheron is really the word computer can someone explain how no one else managed to get Doom running on it and we have this guy who he did manage to cross compile Doom to risk five and run it on wasm so you can actually play it for us to do that right for this you need a stack that needs more thinking than just the traditional evm so that's why we have to do some cover of what are modular blockchains and how do they work okay they have data that's going to be used by your applications then we have the consensus layer responsible for ordering the transactions that arrive then we have my favorite one of course I'm biased here but is the execution layer and the settlement layer where everyone agrees or even an out VM right the execution layer is a very simple State machine that given a state and receives an input always produces a new state this is very simple right the problem is on ethereum on blockchains whenever we do that we name exact state it can't just be a new one it has to be that specific one so this is essentially what forms and ex so what's the kind of depths that you're building are you doing a defi solution are you building a game are you building something that needs uh intens intensive computation or you just want to to run some trading uh do you prefer running writing your apps on solidity JavaScript or python or all of them together uh do you need any specific resources do you need some specific compilers uh is introspecting low latency is always good but do you need low latency outside of it so there there's an infinite list for you to pick and the for us the ideal solution for the kind of solution that we wanted to solve was adding ethereum risk 5 and Linux to generate what is our execution layer okay so what does this bring what are the advantages that this whole mix of ethereum risk 5 and Linux brings of course you get all the benefits of the base layer with the security of the theum but then with risk 5 and Linux you get a more performant computer because before can now be used and run on the chain and this is kind of show the comparison of what evm tooling is with solidity I know open zapping is amazing it's easy to manage packages but I like to compare it as a plier it's my favorite tool it's very good you can do many things with it you can bend things with it you can pull things with it cut you can even hammer and are these kinds of tools what are tools that I'm talking about you can use open CV you can use psychic learn you can use redis SQ light and you can use any language that you want and you're familiar and it's packages and package managers you can use you can use C++ you can use go go is always trying to run from me I don't know why he doesn't like the company of python and C++ I guess B thank you rest is here okay just stay but anyway how does this happen how does the H how is it possible for us to have an execution layer that takes all of this and has the computation capacity without creating a major Bott act on on the base chain right well we're not replacing the evm we're delegating its execution somewhere else so what we do is we send a transaction to the base layer and this transaction is constantly observed by the Cartesian machine nodes okay so these nodes read those transactions as input the blockchain sorry not send send the input back but they're basically selling there and if you need to read the state fast you can just easily access the node to read it State uh there you can start to think of it as a very very minimalistic version of the how the framework behaves and of course if you have multiple nodes running instead of just one Cartesian machine you will have multiple cesan nodes and all of these nodes will be communicating executing the inputs for that specific dep and uh verifying each other to settle and this can be expended right so imagine that you need I I talked about modularity in the beginning of the presentation so imagine that you need a stronger data availability layer so you can use a veil for that for example you send a transaction to a veil or any other data availability layer what machine processes everything and then settles on the base layer but with that commitment that it was that was made for a veil in mind this way you can actually process things and process a larger chunk of data in your applications so just to look back at what we said there are two parts that are very important to this solution right the first one is the app specific rollups that that give your dab a full CPU because then you're not computed for Block space because you're delegating the processing to somewhere else you don't even have to pay gas for this processing just for the data that's being sent and the second part is the expressive general purpose execution environment this means you have an execution environment that's capable of running more that has more access to more resources you get a rootfs system you got all the libraries and all the packages that you have reused so just to have a good Insight of what you can unlock with it on these layers you unlock so many more possibilities so if we're just using the theorum layer one we are kind of bound to token sales and gentrify defy but once we start scaling the compal capacity we can start doing some gamey or even Somey decentralized AI even World simulators and a complex a full complex metaverse so to just so you know uh I like to highlight both of these architectures because we have two tracks going on the first one is just building an app with Cartesian rollups which is giving for a total of 5,000 in bounties and the second one which is adep showcasing the Avail integration which is this case of using Ada da together with our our execution layer which is pretty nice and it really unlocks lots of possibilities so for you to have more inspiration and more idea on things that can build with ctei I'm going to handle to shahen where he will do some Showcase of the depths and show you how to build one hey hey thank youan okay guys so you just uh heard a little bit about what cazi is uh how things are on a high level like on a conceptual level let's just see some inspirational ideas that we have so your bubbles are uh ethereum blocks actually uh the amount of ethereum you put on this bubble is the size of the bubble actually it gets larger the more ethereum you put on it and the momentum is conserved here so you have to strategically think like where you want to push your bubble you can consume the smaller ones and you can get consumed by the larger ones so it's a strategic game looks simple but uh very interesting to play then we have something called cryptopolis I'm not sure if you have heard the um you have uh so here right now what we are doing is we have ported the whole economy of this game inside cesy machine so you can uh inject trc20 tokens and uh use them to have uh this economy built around this city again uh another interesting game and then we have some social applications here are two applications Comet and drawing canvas both of these applications allow you to collaborate and generate art Comet allows you to generate text based art and drawing canvas allows you to generate drawings on a collaborative canvas and everything Happening Here is fully verifiable in this inside cesy and you can verify the whole uh your brush strokes and everything inside courtesy so at the end of the day you can mint a fully verifiable canvas then if you are an AI or machine learning Enthusiast we have a lot of like proof of Concepts around that some applications using that uh there's this Biometrics classifier for you to detect some spoof some U copies of these uh fingerprint and all you can check out this particular example where you can have this model running inside cesi coupled with open CV now open CV is something very huge Library pretty much used in every image recognition out there ification and then we have Lum which actually is a Marketplace of carbon credits so if you want to build something around green five projects you could also think U around that and then for all other applications you can scan this QR code to have uh whole list of uh applications that were being built with cesi okay for the time sake I'll be a little quick because I also have to show like uh how you can build applications so if you can quickly scan this cool and then we have udmi course which is like a master class of around two hours which goes through like from the scratch to building an application also how you can handle assets inside the app chains of cesi that's going to be is there so somebody would uh be there if we are not here to answer your queries okay let me just switch it to CLI a bit oh thanks here there okay okay I hope you see the SE there it's the big screen okay but how do I use it here you have to look there it's on the extended display okay all right okay guys just a sec so uh you have this cesia CLI that you use for uh bootstrapping your applications uh creating an application so let me just uh run this simple command simple JS you write the name of your application and then you whatever language you want to pick we have python rust C C++ so it just creates a project directory for you uh this is our project directory simple JS and if you look into this direct Docker container that's where your code is going to run uh so you need Docker file for that and your actual code your entry point is within this SRC uh folder so you have this index.js file if you try to respect so this is the entry point of the payload that you're going to be getting for your application right uh you get it in a Json format uh it's going to be a hex encoded which you're going to change it to Json and then you're going to do all the functions that you need so right now it's just logging that particular payload we're going to see how you can run and uh see those inputs so this is a new example you have to build it right now it's going to take a little bit of time I have this already built example let me just show you once you build this like you have to do cesy build command and this folder that you're seeing do cartes is that the build binaries that you need so right now we're going to be skipping the build because we already have it buil we're just run it turns up the validator actually turns up the whole node so it's not just your back but the whole node which has other components as well which is going to have uh some uh servers which going to be very handy for you to let's say inspect the machine for you here are the components of this particular uh node and to send the inputs what you can do is there's this cartesi send command which you can use pretty handy command if you're testing backend applications Foundry everything default this is kind of like the application address that is default when you deploy it on the local devet and let me just do a simple hello world here and this is actually the log that we just saw in the index.js file also let me just quickly show you the Explorer if you have any questions till then keep them ready and yeah I hope I can okay it is here so as you see the input that we just sent is here and if I click on it if we check the data that's like the raw hex and puts notices reports and vouchers as per your application logic so this is like the gist of how your Builder journey is going to be thank you so much