# How Web3 Ecosystems and Protocols evolve? - Lazar Travica | Route 3

- Channel: [ETH Belgrade Community](https://streameth.org/eth-belgrade-community)
- Date: 2023-10-07
- Duration: 30:08
- Watch: https://streameth.org/watch/yt-u8-2heU9eWc
- YouTube: https://www.youtube.com/watch?v=u8-2heU9eWc

## Transcript

thanks man [Applause] hey everybody I couldn't have hoped for a better introduction really humble thank you Dave uh so what I want to talk about today is how web3 ecosystems and protocols evolve and what lessons can we take from it and uh apply it to the multi-chain thesis the thesis the states that if we want to see Mass adoption of web 3 we need to be more welcoming to people trying to keep a certain level of sovereignty while they take their baby steps into creating their own blockchain Networks uh first of all I must say thank you to all the wonderful students we had a couple of weeks ago in patnitsa at the science research facility that's pointed towards students and high schoolers together with uh uh uh tenderly and decenter we we did that thing and I'm very happy that a lot of them attended this conference the students that we had and this presentation is really influenced by the talks that we had during those seven days with the students at betnitza so big kudos for them for asking the right questions a bit about myself so I'm a tech guy I did system engineering and GoDaddy later on I was the engineering lead at MVP workshop and uh two years ago I got the opportunity to co-found uh uh trapezius and act as a CTO where we developed a blockchain client for polygon called polygon SDK then polygon Edge and we were happy to release uh polygon supernets the app chain program that we first had touched with although this talk in essence lists all the things that pushed us forward to create create route three this is no uh this talk has no intention to Shell Route 3 but rather to give you all some insights from the industry from our experience building app chains and helping organizations adopt app chains so that many more of us can start building uh things for the next wave of adoption and not just building things for the cool new L1 but building stuff so that mass adoption can really happen I hope that you will find these insights valuable so uh the approach that we had when we started developing polygon Edge was that we treated it as as a Enterprise uh blockchain client what it meant is that we needed to cater to the needs of people who really didn't know much about web3 both from how they will utilize it how they will capture value from it and what do they need to look out for in terms of technological capabilities of these Networks this was in essence every sales call that I had with the Enterprises the one thing that these guys knew very well that they needed to pay attention to what is the scale of operation that I can have in this network that I can call my own I know what my current transactions per seconds are and I need to make sure that the network that I'm going to be deploying on can facilitate my target TPS as an engineering team what we did we put this to the top of our priorities prioritizing TPS over everything else and to be honest the Enterprises themselves they really took uh uh they didn't take anything for granted they went and started exploring the industry uh learned their ropes of web3 and uh I I must say that I was very happy to how diligent people were when they were making this kind of choices but even though we really managed to hit TPS numbers we didn't thought we could even though the Enterprises were happy and they imagined that once they have a network deployed for themselves with a decentralized validator set or not that this is this is how they would celebrate they would be ready to start deploying users will start coming and all will be good but the hard truth came immediately after and that truth was that in order to really have what you can call a fully functioning blockchain Network it's not not just the network it's it's not just the users but it's a list of huge list of services that you need to provide to your users so that they can have the same list uh uh the same ux as in rest of web3 you would need a block Explorer nobody would trust the transaction gets executed uh unless the person is an engineer can interface with the Json RPC but no you need to give your users a way to track the execution and finality of transactions if you're offering a bridge you need a nice UI around the bridge you need to have good ways of informing the users about the movement or of their funds and the security of their funds you need an RPC provider if you want other people to deploy on your network you need an oracle service for off-chain data Fiat on-ramps development tooling the list is even even longer but this these are the most common six services that people demand in their networks and that they really cannot do uh uh well and state I have a blockchain network of my own without having these things the cost of these things is humongous the business models that we see in web 3 of today is that we are creating this big uh uh tooling silos oriented towards large networks I uh I could have been treated as a uh as a really having a really highly valued startup if I just decided to be the first one to develop tooling for a very specific L1 Network which is not that lucrative you're depending on the success of that L1 Network and you cannot really easily support new ones thus the big prices for example for a block Explorer which is high in quality which is Battle tested you need about four months to integrate and you need to put aside one million dollar one million dollars just to be able to to to pay for it for one year but let's put that aside let's go back to the decentralization aspect even with uh regular things such as a bft validator of having it in your app chain Network the only way to get security through decentralization is to having a large enough validator set people are charging from anywhere from five thousand you're lucky if you get somebody charging you 5 000 for a validator per month all all the way to fifteen thousand dollars per month of course even if you have POS while your token worth is worth nothing on the market you need to have a way to pay the validators in Fiat this amounts if you really want to have a secure network you need to have Byzantine fault tolerance most commonly the the the consensus mechanism we developed for Edge ibft 2.0 needs a minimum of four validators to be able to uh tolerate any kind of failure both Byzantine and crash and this amounts to 20 to 60k per month for startups this is unimaginable you need funding just to be started with your app chain for Enterprises who are not saying look I know I will have this in production for the next three years but they're budgeting for projects that they need to carefully examine what kind of Revenue benefit can it bring to their organization these projects get drowned before they even show value because you're spending 60k a month just on infrastructure compare this to a live uh life of a web 2 startup somebody who's building everything on AWS gcp Azure they're being thrown Cloud credits at simply because everybody thinks that they will be a high paying customer in the future they are taking a bet that this startup will in the end become a huge lifetime value customer for AWS for example in web3 we just don't have that there's a huge gap in what startups and Enterprises face between starting uh with their needed infrastructure and Tooling in web 2 versus web free if you take a look at the cost of decentralizing that's just shy of half a million dollars per year so that you can have the essential infrastructure to really call yourself a web3 network that's not good but I want to answer why is it like that how we came to this point and what lessons can we learn from ethereum I'd argue that it's all because of how web3 protocols evolve and the slow process that they had towards it uh I think the best thing to do is to turn back and look at the most recent and in my opinion the biggest uh event in evolution of web3 networks which was ethereum's transition from proof of work to proof of stake uh it's one thing I'd like to make clear right off the bat is that uh I don't want to even tackle uh the Energy Efficiency uh some might argue uh green washing aspect of all of it no I think that's a good counter argument to everybody attacking the web3 industry with web3 we've got this nice thing that we can correctly estimate how much power per year does a network consume nobody asks the question how much does this Bank consume how much does this entertainment service consume and I don't think it's a good uh uh good argument to be even getting into because proof of stake is not an effort to improve Energy Efficiency it's excellent that it did that but if we just stop there we are simply giving the counter argument what I would argue instead is that movement Pro from proof of work as the instrument for securing the network to proof of stake is simply an effort not to depend on large external companies which are producing graphics cards such as AMD and Nvidia in essence the sole output of the the product of these companies is what made ethereum secure of course it was the thousands of Miners and Mining farms and everything but if we cut off the supply for from AMD and Nvidia essentially ethereum would be rendered insecure because the hashing power would be dropping dramatically with the movement to proof of stake some might argue okay but we're not we are still at risks of decentral of centralized liquid staking uh derivatives that are securing the network but are still we are all being scared by it no uh Lido isn't Nvidia Lido was born out of this industry lydo has uh uh incentives to keep themselves in check not to become a risk for the network that really let them grow and let them be born into success I'd say that we must keep that in mind when we're talking about the risks of proof of stake as it is something that's coming from the industry is uh carrying the web free spirit forward and we cannot we cannot even compare those two especially with the rise of uh of the large language models cheb GPT open Ai and everything if we were to say that crypto was just a small nuisance for these uh uh big Hardware manufacturers is becoming even a smaller dot on their radar because suddenly you have this whole new use case that you're going to manufacture gpus for and if we start thinking about it what guarantee do we have that gpus could even be uh uh could even possibly secure uh networks through proof of work of tomorrow because they're driven by other motives and web3 wasn't the reason people started building graphics cards so what how I really love to to explain proof of stake and to to talk about it uh there's a entire family of proof-of-state consensus mechanisms called uroboros which uh is a ancient symbol for a snake eating its tail which is a good illustration of how if we have the results of block production in ethereum which is uh basically ethers being redistributed in the network if we have the result of block production being the same thing that securing the network that's a powerful enough thing that we can call this network completely self-sufficient and completely self-reliable and that's the point of the entire movement from proof of work to proof of stake luckily we got Energy Efficiency as an added bonus so with ethereum biting its own tail securing it uh securing Itself by its own uh products we see how the evolution of web3 protocols must take necessary steps and must wait a long time until it can achieve complete self-sufficiency and health proof of stake wasn't a new thing whoever took a look at ethereum's roadmap almost eight years ago you can then see uh vitalik's thoughts on proof of stake you can see that people are starting to think about it but people were also very aware that that change cannot happen overnight not not simply because of technical reasons but because of socioeconomic ones I want to I want to uh take a look then forward we're talking about ethereum going in the way of uh L2 scalability uh being being the base layer for validating all of these proofs whether they be validity proofs or fraud proofs but what I truly believe is that uh even after the L2 scale and help us move the problems somewhere else we will truly have a multi-chain many multi-chain ecosystems even vitalik agrees with this this is a Reddit post where he talked about being optimistic about multi-chain future but the multi-chain future coming back to the start of my talk is currently riddled with with really really huge costs the bft validator costing 20 to 40K a month amounting to half a million dollars a year is something that's very clear in how it prevents both Enterprises and startups from entering the scene one size shouldn't fit all is another problem like uh if if we take a look at some of the coordination points of people building the ethereum network such as The ethereum Magicians forum and if you scroll through it you will see people having lots of different conversations with a majority of them would never go into Mainline ethereum protocol majority of them will never see the light of day because a fraction of those are simply people trying to make ethereum better for themselves but ethereum is a public good ethereum should be good for everybody and should support as many use cases as possible if go ethereum's issues list on on GitHub uh 240 issues at uh at the moment of this screenshot if their issues list has the structure that we had at with issues in polygon Edge it's mostly people wanting to put in changes that make sense for their use case even a private deployment of go ethereum or something else where in reality this is something way bigger than themselves and uh I'm telling you this because it's very it's necessary to understand in the app chain thesis that if you want to have your own networks you cannot just simply go redeploying other networks clients but you need something that you're able to customize for your own use case the next thing that I I see is a great Pitfall to adopting the multi-chain thesis is that people need to start valuing security over decentralization I've seen projects shoot themselves in the foot attaining high costs attaining huge governance and coordination complexity from day one simply because they wanted to say look I'm decentralized nobody even asked the question do we have anything to secure at worth securing at this moment so that we need to decentralize it from day one and this gets especially ugly if we're talking about on-chain governance and governance of the validator set this is the state of the app chain thesis as I see it people uh the majority of people who really want to go Maxi on everything web3 think that you need a huge validator set from day one but in reality you would be paying a huge amount of funds Fiat funds to people who are going to secure a Network that has nothing worth securing it the key takeaway from this is that decentralization should be done only because of security not because of decentralization itself not so that you can go ahead and state my network is super decentralized I have this many validators no you should know and you should remember why you're doing it in the first place and why was it uh uh why we needed decentralization in the first place another point of this argument is of course some might argue but look uh there are Services right now readily available infrastructure tooling services that I can go subscribe very affordably spin up as many validators for my own app chain that I need but the the answer is quite simple to why this is not a good idea we're talking about decentralizing in a way that you have autonomous validators in the validator set if one company uh one blockchain as a service company manages the entire validator set what's the difference from having a single validator Network there is a difference but not in security there's a difference in performance where one validator network simply because of network messaging complexity is much more performant than a hundred validators managed by a single company that's why we need to really be aware of why we're decentralizing and how we are approaching the community that's going to help us do it the second uh argument towards uh being uh hyped up about the novel uh uh novel Technologies being made in this space like roll-up Technologies which are amazing which are the only way to really scale performance that we know of today be people think that from day one they need to start their app chains with these kind of Technologies these kind of Technologies where again if we're comparing web 2 and web 3 there's a there's this huge even a very very old joke of today about startups who from day one start using kubernetes like it's not a very smart thing to do it's going to raise your cost it's gonna raise your Manpower costs and things are gonna get very weird very fast when you want to be quick and iterate what this means in the web 3 context is that a lot of people are starting to build Roll-Ups even when they have no users they're opting in to having huge infrastructure costs these are really machines that need better and special more specialized Hardware than typical bft validators and it's not always a good thing to to think that you're going to have scale at day one because the costs of having that capacity made possible for Roll-Ups May prevent you from even attaining the user base that you are wishing so start using roll ups when you reach adoption and need to scale don't use them simply because a novel Tech breakthroughs and because you want your pitch deck to look better in front of the VCS but uh with all these things there's a bigger problem and that problem is that uh cross chain communication is inherently very complex and very insecure we're talking about two disconnected State machines two disconnected pieces of uh uh uh two disconnected networks that somehow need to trustlessly relay messages with one another the bigger problem is that I lied to you a couple of minutes ago with vitalik's post I intentionally cropped it to hide the fact that the one reason that vitalik is scared about the multi-chain future is because we don't see the future of really uh secure cross Bridges across chain Bridges happening uh soon I'd argue that they are happening and those Solutions lie in the role of Technologies but we are not seeing them as bridging Solutions we are seeing them simply as scaling Solutions to illustrate this and how different tab chain networks in their ecosystems approach security and approach uh cross chain messaging I want to show you for example the tvl graph of novel novel App G network completely completely just for illustration purposes I don't even know the coin that I took a screenshot of for this so especially in its nascent days the majority of the networks tvl if it has bridging enabled with other cross chain ecosystems is made up out of wrapped tokens that have come from other ecosystems and brought in value because the net Network's native currency doesn't have that much inherent value yet they're depending on liquidity to come from ethereum from other l1's from other l2s from other app chains in the ecosystem so that's all nice and dandy you have your network is responsible for securing a large portion of funds that came from other ecosystems which have value then you take a look at how many POS app chain ecosystems exist exist now and uh how how these uh uh how these ecosystems try to secure the network is by saying okay you will take my ecosystems token that has nothing to do with your network because your network token worth nothing these total value staked is usually much less volatile in value than the value of the total value locked people can bridge in and out all the time every day but rarely in this nascent networks do you see validators taking unstaking more uh more delegated stake being pumped into the security pool of the network and this is this is a an illustration of that where you have the tvl going up and down but the total value staked uh tends to tends to be constant does anybody does it can anybody does anybody have a hint of why the network is vulnerable in this exact points on the graph the answer is very clear if you are talking about the stake being the amount of money making you liable for properly securing the network these are the times these Peaks above the the black line which signifies the value at stake these are the times when you're being a fool for properly securing the network because if you play it against the rules if you took those stake tokens and exited on ethereum claiming they were your own you would have made a profit even with the stake that you are going to lose and that's a huge problem we need to make consensus mechanisms that do not count on anybody trying to be fear or playing fair but being properly punished when they do and this is no punishment you can see why before staking third party tokens the solution to all of this that I see is that before trying to secure your network through third-party tokens before uh taking a plunge into somebody else's ecosystem in an economies you really must try to build your own because if you have an own economy and you're trying to uh depend on another ecosystems token for securing it you're gonna have a bad time and you're gonna have a those Peaks that go over the line where validators are being foolished for being honest so to recap everything that we talked about I want to show you the recipe as I see it for a successful app chain first you need predictable and affordable costs you as an organization cannot make that big of a decision without knowing how much it will cost you you need the ability to make choices not every size fits all and I very well know that decentralizing because of security is the way to go you do not need to decentralize from day one especially if you're not sure you're even validated your business model and reach the product Market fit with your app chain fourth you want to scale because of lack of throughput you want to scale because there are these users that cannot find their turn in your blocks to put their transactions not because you've heard of these new technology that solve the problem for ethereum but you're not ethereum most probably and last you want to launch with proper financial incentives so that you can really keep the network secure and help build cross-chain applications and let users move liquidity through different ecosystems so predictable and affordable costs by facilitating the creation of fair market values we need the marketplace for validators we do not want to be in the hands of people who will have a five thousand percent markup on the validation costs we need the ability to make choices we need to favor the lifetime value of our app chains instead of getting a quick value because of uh we are accruing value for a larger ecosystem that we have nothing to do with we need to decentralize because of security by supporting decentralization but not forcing it we shouldn't Force anything upon anyone especially when we're talking about budgets and early startups that have not yet reached their product Market fits nor Enterprises that are still experimenting we need to scale because of lack of throughput not because of tech novelty not because of our pitch decks we but we need to have a clear path and support so that tomorrow we can adopt these Technologies if we succeed enough that there's such a need that we cannot scale otherwise and we need to launch with proper financial incentives by having the customer success as the one and only incentive what we're building with Route 3 is uh we're trying to bring this power of ethereum that came through the snake eating itself that came through self-sufficiency taking careful small affordable steps along the way to having a successful Network that can scale in the future like ethereum does we're creating Route 3 don't want to chill but feel free to approach and ask any questions that you might have thank you any questions [Applause] no questions your honor amazing