#DappCon22 : Day 2 - Reshaping The World of Web3 With Infura, Eleazar "EG" Galano

Transcript

okay welcome back everybody hope you enjoyed lunch uh to kick us off this afternoon is eg gallano co-founder of infra a bit louder please and he's going to talk about reshaping the world of web3 with infra of course all right thank you hello everybody it's great to be back here so my name is eg and today I wanted to talk to you about what we're doing at inferior this is a big week for our ecosystem I always like coming back to Berlin during Berlin blockchain week dapcon East Berlin like there's so much history of ethereum here and especially

with this week being historic with the merge I thought it was good to talk about the history of inferior a little bit about the impact on ethereum the web3 ecosystem and the growth of web3 and especially dapps because this is dapcon and we should be talking about deaths so my story is I I started out my career at a small startup in California we were doing video games streaming um I was there for a few years before we got acquired by Sony I was doing some back-end engineering a lot of like devops automation type of work a lot of stuff that was like back-end infrastructure and I bounced around to a couple of other startups and in 2015 a

friend of mine at that startup said he came up to me during lunch and asked me a question about if you had a compute problem and you wanted somebody else to solve it and you wanted to then have them give that back to you and like prove the result prove that they did it correctly like how would you do it and the first thought that I had was are you leaving like where are you interviewing like this sounds like an interview question and he said no no this is like have you looked at this ethereum thing it's amazing like that it has smart contracts I'm like smart contract and you need to look at this thing and so I looked and it blew my mind and he introduced me to Joe Lubin and consensus and I ended up working

very soon after that on things within consensus met a couple of great people and we started inferior together so in 2016 we started in fira and as we were looking around the ecosystem at like what you could do it was early 2016. there really wasn't much at the time everybody was still trying to Tinker around and see what it what is it that you could do with smart contracts what what could be done with ethereum what are the ways that people want to interact with smart contracts and this is a screenshot from ether Delta this isn't like the earliest version this is the screenshot I could still find but like this is what a DEX looked like at the time right this is

what we would look at and say all right how are we going to interact with a smart contract and it was very straightforward it was a simple UI interacting with a smart contract and then here's another example some of you might remember ether role where you would send a transaction and basically bet on the outcome of a dice roll there's some drama in like the end history of ether role that you can read about but like for this we'll just focus on like the technical aspect of this was smart contract provable Fair dice roll on ethereum and here's a very simple interface that anybody can check out the code look at prove it it's all legit and then you can run that like that was at

the time that I started in the ethereum space and I think most of the people I started in fear with started in the space this is what the state of the ecosystem was maybe you could also like buy a pixel on the million ether homepage age and things like that but you have the user interacting with a simple interface and whether that was a command line interface or a web app like this it was interacting with that smart contract on the blockchain through the blockchain node connection and many of us were running our own ethereum clients or blockchain clients at the time the thing that was important to a lot of

us was the decentralized ethos of making sure that this is living on chain verifiable in um trustless and then there was also this ipfs component that people started to use as decentralized storage or decentralized data so that you can take that front-end put it on an ipfs host or server and make it so that somebody could use this standard HTTP Gateway and interface with the smart contract through a web browser or ideally in the future using the ipfs client we as inferior started to run an ipfs hosting service where you could pin that

data and it was supposed to be a Gateway and a transition path so that we could get people thinking in web 2 transitioning to web3 and that's pretty much always been the ethos of inferior is thinking web 2 transition to web3 so with ipfs you start with HTTP Gateway transition to understanding the ipfs protocol understanding decentralized data and how to retrieve it and then build the optimal and ideal solution once you've proven yes this works foreign s of adapt there's that front end and then there's the smart contract that

lives on the blockchain and what people started to do around like maybe late 2016 early 2017 for sure was they started to add this additional component that I've circled here which is adding a back-end server and the reason that they did that was for things that if you're a developer sound pretty obvious for caching for query optimization and an overall better user experience when you think about how blockchain clients store data and how they were designed they were meant to store data in a way that it's provable and verifiable and the data constructs that go into how the data is stored meant that it was really difficult to pull the data out and we've done other talks and I'm sure a lot a

lot of other people here have seen some of the talks about how different ethereum clients now store data to make it easier to pull that data out but at the time this is what people started to do and this pattern still continues today is somebody pulls that raw blockchain data indexes it so that they can optimize the queries that their dap uses but what this breaks is that ability to have interoperability with your own infrastructure like you were you're able to switch from one node to another from one API provider to another because of that interoperability when you start introducing this back-end caching indexing layer unless you're using something that that's like an open

decentralized network that some of those now exist for indexing this becomes a proprietary point that means your dap is no longer really adapt and at least in the form that we were talking about back then the other thing that people started to do was they started to not run their own blockchain clients they started to use inferior many applications so if you've used metamask if you've used a lot of the popular applications out there that if you're not running your own node you're connecting through inferior and what we did was we we said that we wanted to help the the user adoption of ethereum we knew that from the beginning

we were starting a centralized service but we said our goal as inferior was not to solve a decentralized problem initially because the decentralized solution at the time already existed it was run your own node like that was the decentralized solution at the time and we said we're going to try to make it as easy as possible for people to get that data themselves through our API and see what the trade-off is is anyone actually going to find Value in this and I remember going to conferences and our team going to conferences and talking about what we were doing and people would ask those questions like but isn't this against the ethos isn't this against decentralization and we said no

because it's interoperable what we want to build is something that's interoperable with running your own node so you can switch to us and switch back we're just trying to make it easier for people to get started because if they wanted to try ether role or they wanted to try making their own version that was better and there was that extra barrier to entry additional friction they might just give up and say I'm going to do something else maybe at the time they were equally interested in ml or AI or something else and we said let's make it as easy as possible for the the brightest and smartest people to try experimenting in this space and then if they find Value we'll find a way to decentralize in the future so that that was the choice that we made

we said we wanted to bring more users it was a conscious choice that we made and we said yes we're going to be centralized we're going to optimize for bringing users into the space bringing developers into the space and a lot of the applications out there started to make a similar choice they said we want more users to be able to interact with this maybe they were driven by just like wanting users maybe they were driven by like the value that was created for themselves and and their team by bringing more users but people started optimizing for adoption so you can look I'm now talking about that like 2017 era people were optimizing for the adoption of make it as simple as possible to use their

application do you do the heavy lifting so that for us that meant run the blockchain nodes run the blockchain clients so that you provide this API that people interact with and the thing about it was the centralized versus decentralized Choice some people were aware of the choices that they were making other people were just putting out a fire and doing whatever they could to like solve a specific problem at hand and I'm talking about like the Ico craze in 2017 we had engineering teams reaching out to us saying we were running our own nodes for this and our Ico is like in the middle of the sale the sales going for like two more hours can you help us and we would say

sure like um we would get waking up like woken up in the middle of the night by teams saying can can you help us run our infrastructure and then we would open up our endpoint at the time we we didn't have a paid service like we do today it was just fully like a a public good that we provided to the ecosystem but when you look at what happened during that phase this is um this is something I pulled from the Reddit cryptocurrency subreddit where it shows about 6.5 or 6.3 billion dollars that were raised in 2017 through icos so look at all that value that flowed into the ecosystem For Better or Worse into certain projects

but it flowed into the ecosystem and imagine what this would have looked like if you said hold on before you can actually participate in this ecosystem before you can like put ten dollars into this crowd sale you have to go to this website understand what a note is understand how to download it install it wait a couple of days to potentially a week for it to sync some people might do that but there's there's no denying that that extra friction hurts um hurts adoption and will put a limit on the adoption of the technology so it's understandable that people especially maybe if they were they were funded by investors had to make decisions that were best for scaling for

user adoption rewinding a little bit the choices that we made when we started in fear in 2016 were yes we agree we're going to be centralized when we start this thing out but we're going to progressively decentralizing the future we'll just keep trying to solve the problems as they're put in front of us and then we're going to find a way to decentralize as things go on the two key functions that we provide as inferior are read and write it's very simple it's reading and interacting with the blockchain data and it's actually propagating transactions to the network that that's like the core function of inferior and it's also served by the

ethereum node so many users that we've had over the years not just use us but they'll also use their own infrastructure or they'll potentially use another API provider node providers now are popular because what we do is not run a bunch of nodes behind the load balancer like they're fantastic clients out there Aragon nevermind go ethereum basic I don't want to leave any specific ones out the parity client um is uh or was amazing and turned into the open ethereum client there's there's a lot of things that were done there to try to optimize within the client itself but within computer science and like the

developer space there's a limit to what you can do with a single node and then it's inevitable that you start trying to do things to scale the traffic first you do things horizontally and you say I'm not going to run one I'm going to run two and then 20 and then you put a load balancer in front of it and it's a very straightforward scaling solution but the problem with blockchain Clans is each of these things have their own database and that your requests get routed in between them and they're all slightly out of sync and so centralized providers like inferior end up building custom microservices and proprietary code to do that indexing into traditional databases to be able to detect when a request needs to be routed one way or another so that people get a

consistent user experience and that's not something that is Trivial to reproduce that was something that solved the read problem for getting access to the blockchain data but then on the other side of that is the right problem so transaction fees this one's showing 2018 started to go up in um because of additional traffic on the network and icos and things during that phase but then you go for another couple of years to 2020.

and that 2018 Peak is just like a little little tiny Hill next to this massive mountain of transaction feed growth that happened in a defy summer uh I mentioned that because uh going back when uh when we were in this phase of 2018 this was just talking about ethereum like there were other networks that are starting to come out there and um were kind of in their infancy but in fear I didn't choose to start supporting those it was a little too early for us to really decide how we were going to approach that but once we got to 2020 it was pretty clear that this was a problem that was hurting adoption and so what we said we were going to do was go

multi-chain we said that as the developer tool that many people in the ecosystem used we needed to support the chains that developers wanted like it didn't matter like if we favored one chain or another it didn't matter what the technical merits of one were over another if developers had an interest in deploying their applications to a to a new chain we should be able to offer that as an option to them and so now you can look at this list of um chains that we now support I can't even remember the number in my head it used to be he's I used to be able to say oh it's two now it's four like we'll have to count those boxes um but there are many evm compatible and now non-evm compatible change that we

support on inferior and over or over 40 percent now nearly half of our users don't just use ethereum or don't just uh they'll also be using a layer two or another like ebm compatible network and about a fifth of our users use multiple layer twos so let's go back to this previous diagram of what the DAP used to look like this was very simple you can see how you could make this pure and decentralized you can say I'm a user I'm going to put my UI on ipfs and I'm going to run my own blockchain node but how does that work now when it's a

multi-chain Dap were you aware that now that's a lot more difficult and was it a conscious choice that you made when you were developing your depth so if you now built this Gap that interfaces with polygon arbitrum optimism several other chains it works it works great your users are getting extremely low fees and a lot of optionality in their choices but now if they said and you might say my my UI is open source you can run it yourself if you want to this is now the requirement that you have to be able to make that uh match what this initial intent of that pure dap was from back in like 2016. is somebody here willing to run an optimism arbitrum Etc like node

like five six different blockchain nodes like clients are not like clients it's it's going to be a challenge right and it's something that hurts adoption and usability so that's something that I wanted to highlight was that when we're optimizing for user adoption we said we're making it simple we're doing the heavy lifting and we're making those intentional choices about whether we're centralized or decentralized when when you're building your desks whether you're building them at one of the hackathons this coming week or just as part of your product all of the things that we should be doing as an ecosystem is making sure that the choices that we make are

well-intentioned and deliberate the DAP complexity is a inherent centralization as daps get more complex as we introduce the concept of multiple protocols and bridges and other things Mev Mev protection it it introduces this centralization to the DAP design so for inferior's evolution we want to be more intentional about how we approach this we had said in the beginning that we're going to be centralized focused on solving a specific problem and increase the adoption of the ecosystem we feel like we've gotten to that point we've gotten

to the point where adoption in the ecosystem is now sustainable and not just by us there are many other players in the ecosystem that provide value to developers and so at East Berlin this coming Friday we're going to be talking about the next evolution of inferior which is how we're going to continue to approach on that of this Progressive decentralization problem starting with some things like something as simple as being able to sign in with your wallet like metamask is it is a might seem like a an obvious one but it's a significant change if you just include that type of design into your application I also wanted to invite you to come and

celebrate the merge with us like I said it's a big week for all of us in the ecosystem so we're we're having a merge party this Thursday sounds like there's a lot of merge parties this Thursday but it's also our our chief researcher Pat McCrory's birthday if anyone was able to uh check out his Workshop this past Sunday Patrick's awesome you get this it's a two-for-one special with uh Patty's birthday so thank you very much everybody [Applause] cheers we've got two mics in the room just raise your hand if not since we went a bit over time

um I'm just gonna thank you again hope you have a great great party on Thursday [Applause]