How Lava Network Aims to Solve The RPC Problem For 10,000 Rollups

Transcript

welcome to another episode of the rollup Jill the CTO of lava Network Rob looking pretty Sage over there man honestly from one uh from one Sage to a Maverick it's h it's a pleasure to be here and uh it's a pleasure to have you on jaill thanks guys yeah man took a brief look at the uh presentation I'm very very excited to uh to dive into the RPC uh landscape awesome thank you for hosting me great great chaning with you guys again uh yeah so let's let's talk about Lama let's talk about RPC let's talk about modular thesis nowar everything connects uh so my name is Gil uh I'm the

co-founder and CTO of lava also one of the core contributors to the network uh my background is uh originally from cyber security over 15 years of uh experience in taking really complex systems like journals browsers uh operating systems breaking them apart understanding how they work and finding flaws in their systems uh I got into the space by doing MV so that that's what brought me in building bots on Main night we do in lava for two years now almost two years actually uh and we've built a really strong team of amazing Engineers uh so we'll talk today about you know what the model STC is what is RPC how it fits in the story of lava how

it works uh yeah and if you guys have any questions for please let me know feel free to stop me so you guys already know this but we live in a multi chain era everybody was talking about this a couple years back and we're now seeing uh we're seeing that reality come into life and we're seeing yeah from monolithic chains uh today we saw the announcement from monad uh two many many raps uh and obviously chains of Cosmos and and the likes and it is really really amazing to see uh all these new chains come and solve new problems for for users uh you know the Advent of uh da allowing for cheaper gas

uh and more transactions for users unfortunately one of the downsides of having at this current stage of the game of having so many chains is uh the fragmentation that we're seeing if it's from you know liquidity perspective uh or just like managing many many different walls from a chain abstraction perspective um and from a developer uh experience obviously building for all these chains is like super super tough for many reasons that we'll dive into uh but definitely we're seeing now that you know many more chains are coming up and why is that that is because um it's so much easier to build right so we are now basically building developing on the shoulders of giants uh we have you know

amazing execution layer medal sensus amazing projects you know um from Celestia from igen layer I da which we saw on TR man today as well crazy day uh fluent of veil you know di mention we've been seeing it clipse so really incredible projects that are building these layers um and what that allows it enables you know everybody to build faster and everybody to focus on what's really important for users and bringing their crowd uh bringing the users to actually um to focus on business stge forel which is I think Super crucial uh but there's one part that is like super super

painful uh everybody uses it you guys use it every day I talk about and uh we call this uh data access right so how do you access blockchains so quickly overview blockchains are these amazing uh disted databases so you have uh basically uh let's let's call it like an Excel file uh so if you know if it's not uh distributed or centralized just like a file on your computer uh but if you have uh the Excel Excel file on a different server on what we call Web threea note and then note syncs over peer-to-peer network with many many other note notes and those Excel files

are basically replicated and everybody agrees with the content of these Excel files excuse me for using uh Excel as as an example I think it's really simple um so now we have a this Sate and it's an amazing beautiful thing it's like a new primitive for computation right it's a new way to build things in a way that's centralized permissionless trustless so how do we read from this databases so we can't just access them on our computer because we don't have the file we don't have the Excel file so we have to talk to a server a node that has those files so every time that you are doing any basic operation with a blockchain it requires you to basically talk to these RBC nodes it's crucial

it's it's uh imperative that you have these noes uh is these RPC providers because without them you wouldn't be able to access any data so how do you write to these Exel files basically every time you send a transaction you're sending the RPC provider this RPC provider as nodes and these nodes gossip in most blockchains uh these transactions to each other and eventually there's a validator uh that puts them in a in a block U and then gossips backs to block all these notes so now we wrote to the S and then rsls updated with what we put in in in the initial vision of blockchains uh people thought that

everyone will will be running these notes right everybody will be writing notes and there's a also something called light client which we might T you on later which is a way to prove the data is correct in a simpler uh lighter way than running a note and you have a quick question uh even in the context of a like client um if a like client is a way to prove that data is correct I still have to get that data somehow and so is that is that the to access question is how do I get that data and then proving that the data is correct is is kind of another job it's another uh job for what a like client

would do but a like client still has to get that data for it to for it to verify it so light light CLI is a is a big topic and there's many different types of light clients but in general generally you're right like think about the light client as a way to maybe get some of the data and know that it's trust worthy by talking to many peers in a lighter way and not having to se the other chain but most light clients still require a full node which is like this Excel file to talk through to get the data so what we're talking about is this core part of blockchains that no one talks about which is like how do we access the data um and blockchain

blockchain sorry for saying this but they're basically useless without without RPC they're useless just these files no one can access so so what is RPC RPC is like this this core protocol that uh almost all blockchains use to get data from from their blockchain so for it's called remote procedure call uh it's basically a communication protocol uh it allows you like to send a query to a server let's say an R provider uh it moves it to a node and it executes the query or the Excel file let's say it's sling row five to six and returns it to the user some really basic uh uh communication that you could do is like

uh you know getting a user uh balance let's say you want your balance of your wallet so you would use an RPC from the browser for example to go and ask an RPC provider what is my balance when you want to send it tou it same thing and you can see that in the center we have this RPC provider this RPC provider is someone that we are super super reliant on dependent on to execute anything on blockchain all interactions with the blockchains are made through these RPC providers this is why it's a crucial piece of of bles down so some some uh more examples of what you could do you know from deploying a

contract checking wallet your wallet balance even like making a Unis Swope trade anything that you do that interacts with the blockchain is dependent on these RPC provider dependent on making these communications uh if no one ran RPC nodes you wouldn't be able to do that again it would be stuck with a useless file that no one can access yeah these and um real quick on that Jill why um why why is there um not like a standard for rpcs where sometimes I go on you know a anchor RPC or a pocket RPC which works better than a infura RPC because that

one has a lot of traffic but is isn't it still kind of connected to the same network and it's like like why is it not just like okay we have this one standardized RPC that always is going to work what's the blocker there well you jumped the gun this is exactly what we're building right so rbcs are really really tough why because they're very difficult to scale and they're fragmented and each blockchain defines their own apis their own interfaces so it's very hard to standardize because everybody basically out of necessity like built their own solution so there's like a mish mash fragmented solutions for every chain

we're really coming to change that and to basically make it accx there for everybody we talk about that in a second yeah so when they're getting clogged Jill real quick when they're getting clogged like when there's a lot of requests these like all these calls and actions are what is being as what's happening when I click a button on you know like metamask or Rabbi these types of of calls and codes are being ran and if there's too many being run at once then it gets kind of messed up exactly and so just like you go to like say a web to you know page and everybody's going to that page if it's not properly built to scale it's going to crash right so if for example eore G balance every

time you go to your wallet CU make you see that uh you would see that on the network go from your computer to an RPC provider and then to a m and then get back to you so every single time you do it there's like hundreds of these calls being made to rco virus gotta so you know how do people actually get access to these art receipts I think that's a really interesting question so today um the majority of users and developers are using a free service using a public RPC that's either uh it's usually subsidized by um either the wallet or the actual blockchain so most of the blockchains

they subsidize access either rollups monolitic ones and app chains they will subsidize access to give users the ability to um to to actually use their blockchain because it's like the door like without the door you can't get Okay g boy so I think it's really interesting to see that the vast majority of developers are using these free services uh to actually um use the blockchain so what do blockchains do and mind you every single blockchain in this and you know going back to this multi chain era that we are in now and it's only the beginning right this is only the beginning of the number of chains number of rollups we're going to see every

single one of them has to think about these things so what are they doing right now some chains they run their own notes some chains they you know they spend millions of dollars you know Outsourcing it to vendors some chains rely on community noers and they're all doing yet because it's necessary and it's without it they would be able to give access to their users so when they are launching this chain you're a chain you're launching your chain you have to think about what's cheing these RFC notes so you know everybody knows about validators right everybody knows that when you're launching your own chin that has validators you have to boot validators but wot shoing validators is much easier

why because there's inherent uh incentives right so a validator uh they stti tokens and people delegate to that and they get a commission and they make rewards from producing blocks but where is this incentive for RPC providers it just doesn't exist so it's a really really big problem in this industry um so all of these chains as they're boot sharing the validator sets they also have to boot RPC so we're talking about uh you know make it capital efficient for them it also needs to be cost effective to run these lands um it has to be scalable for these chain they

have to make it scalable for their users and uh they also have to make sure that if you're a chain you know we recently saw a few testings for example and blockchains where users were not able to to use the software they were not able to use the blockchain because the RPC so clock so from their perspective all that momentum you know is really really getting people stuck and frustrated and they're not coming back so for Developers and we see this and I'm sure you guys saw this and every time you go to claim an erir drop you get like oh RPC error we can't uh we can't do it or we

can't send transaction a new chain Just Launch a new test Just Launch you can't send transactions or your balance is not showing that this is all because of that cloud uh rpcs because you're relying on a single RPC provider and if that RPC provider is down you know it's clawed it's you know uh it didn't scale out properly then you're stuck and everyone gets gets frustrated um so more things that are problems currently with this types of setup is that when you have this RPC provider remember was in the center how much power does it have you know it sees everything that you are doing it's like this eye seeing okay you're doing defi

you're trying to trade you want to buy this nft it says everything that you do um and it says that you're actually doing it you're sending transactions so it knows it knows how to connect your wallet address with your RPC quers and then it can take the then and sell that's crazy I think I think we definitely want to build things around that in web three also yeah yeah I I think that's a really important point because in traditional Finance you see these things like payment for order flow and uh and I I think there's this common thing in even like social media and Technology applications where if it's free you're the product right you've probably heard that before and so when

most of the rpcs that are available are public rpcs and offered for free I think there's kind of there's like well why is this free why are am I the product are they profiting off of my order flow are we recreating the the traditional Financial system some of the things that we're trying to get away from in a more open free decentralized economy and so so yeah how do we solve that incentive problem this maybe is where what you're getting to I'm getting into it but I think it's a really good point right so I think that there are some uh really bad incentives today and there's some really good incentives popping up

actually so I'm super positive about the space when it comes to that narrative and I'll show you what we we built where we're building in lava that actually solve this incentive problem but just to address this one more time you were talking about uh order flow payments pay for order flow that actually already exists in web three like you have it with ethereum you have it with many other uh blockchains where basically all the transactions are routed to bots it's already happening uh and these users sometimes get EXT their their value is extracted by these Bots by different taxs so it's really crazy one more thing that's super important is

if you go back to that Excel file that I was talking about which is like the database of this blockchain how do you know this that you're not getting a stale file from two hours ago where everybody already agreed to these you know 10,000 more lines in the Excel or changes but you're still seeing that old file so the staleness problem is also a really really big problem and I think the key thing I wants people to take you know just talking about it is that this is not just a problem for one chain so as we grow as an ecosystem and the number of chain multiplies you have more and more app chains more and more rollups you have more and more of these issues uh and I think I think we have to

solve it things in the right way so Jo real quick before we jump into lava I I know you're stoked on it how much data like are are are these companies actually taking from us on our PC like like like what does infura know about me I don't want to specifically yeah I don't want to specifically maybe talk about individual providers but okay then what do they know about me they know a lot I don't know whether they col or not okay so let's say you go out you want to take a loan or you want to deposit some collateral and you're looking at different pools to do that they know what pools you're looking at let's say you want to swap

yeah but but wait but think about this not from your okay I'm interacting with the de one once what if you direct with this D over you know few weeks over a month and now they see your profile oh this guy you know he likes to invest in M he likes to uh you know to to short they see all this information but if it's not actionable yet it will be in the future and either way they can still sell your actual transactions to BX so all this is available yeah I guess what I was more so asking is like what kind of personally intrusive data are they getting they're not getting I wouldn't

say they're getting personally intrusive other than what kind of financial so in the in the you know web to World um it's more like your personal preferences uh and like what you're doing online and in web 3 is more like your financial interests like where is your minding uh what what do you like to buy um what type of tra what about like my IP address yeah of course they can see your IP address um so let's let's go one step back you're the you're on your browser you going to unisal every time you go to unisal you send a message to uh I think it's Alchemy right not and they can see obviously your IP and a lot of

information about where you're coming from you're coming from Unis swap what page you're coming from from information like that what tokens what tokens you're looking at so if I have like if I have a social profile connected to a wallet in and on that wallet they can this these RPC providers can see my connections and my IP address and yeah my FS yeah not great they can see our fs and they can see all that most of this these things they can surely say um they can see it's because again blockchains are useless without it why because it's

the pipe that connects you to all the data so when you're asking for the data you're as what data are you asking for what's the context of that question so once you have all that information you can really build a lot of stuff around that and uh IP addresses are geographically or oriented and unless you're using a VPN and so if I am trying to claim an airdrop or trade a meme coin and for some reason uh I'm in I'm in a jurisdiction where it's not legal or allowed to to claim an airdrop

or trade a mecoin who is doing that censorship is that the front end maybe if it let's say the front end says it's okay they don't care would the RPC censor my ability to claim that airdrop or or you know illegally trade a mean coin they could it's possible they could I don't know if they do it I know that in the past there were uh disc talks about rpc's blocking tornado cash from example so it is definitely something that they are capable of but not to the full extent that you would think there probably are are ways around in the end for the user but yes you can definitely censor and

block users uh based on what you're doing on the block by just lying you can lie to them I could tell you your balance the chillion Heat or I could tell you it's zero and you have no way of knowing true I'm just telling you it is what it is and you take it as a fact so I think it's really important um that we solve uh for these issues freshness accuracy censorship and and this was part of the reason that Satoshi found it so important to run our own full node is that is that because then we couldn't you couldn't really fudge the data that you were giving like to someone who was getting it from their own server on their own computer if you run your note there there's it's much

much much harder to trick you why because you are verifying State you're verifying that Excel file and it's on your computer that if the file is on my computer I can just tell you you know oh this is what's inside the spal war you're good but so we' we've been building this in a way that we been building love in a way that solves these issues and allows you to trust the data you see while not having the file locally because as we get more and more blockchains get more and more files and no one's no one wants to host it and it's becoming so difficult it's expensive you know requires a lot of bandwidth of storage and far as develops very it's you know

every time there's an upgrade you have to update your executable so maybe you can do it with one blockchain but what if you have 10 20 it doesn't scale so each one of these providers that join our network is basically a responsible for doing this that you know they might be very good at ethereum but they're not doing yeah so or they might be very good so but they're not doing ethereum so this allows the network to bring in many many participants or each optimizing is and is an expert in his domain so I think that's a really good point to think about so as we were talking about these modular layers right with consensus there bu one's talking about talk about

settlement execution right what's missing what's missing is the the the layer uh for for modular blockchain data access and this is what we've been building it l it's stading a bit more so when we were designing this from scratch two years ago we were trying to think what are the key qualities uh that an access layer that is universal for everyone agnostic needs to be needs to have and it has to do this flawlessly because remember this is the layer that connects everybody to all the amazing already existing beautiful Technologies of everyone's big building but this is the Gateway this is

the the door so it has to be Flawless has to uh work perfectly so one of the critical Parts is scalability it has to really handle the search of love when you there's an nfty uh you know draw I want to be able to to not have an RPC error you know when I'm claiming an air drop I want to be able to to transact and don't be frustrated with the with the system because if the system is frustrating I'm just going to say you know I want to go somewhere else I has to be cost efficient right so in lava uh the service is actually scored and the rewards adjust Bas still score so if you're a bad provider you get less

rewards you get less users if you're really good providers you get more rewards and uh more uh users it also have to be cost efficient for the the chain remember chains are usually you know when they start it pump a lot it there's a lot of traffic so it's obviously it should be more expensive to round but maybe a few months later you know stabilized somewhere the average um and in that scenario they don't have to keep paying what they were paying you know huge amounts of money to operate the the public RPC for everybody so they can lower their cost and make it more like a market where you know the the man um there's like a a balance uh it also has to be low

latency so no one wants to wait you know you want to see the price right now you want to make a swap right now you want to you want to interact with whatever you're interacting right now you want to get your account balance right now you don't want to be wait you know 5 Seconds to see it uh has to be agnostic in modular because there are so many Changs has to make something that actually enables people and anyone to basically build and bring this new chains quickly to love uh has to be accurate right we talked about this the data itself chel lighted has to be fresh you have to make sure you're getting the latest Exel far it's what's right now they financials right now not financials

today and I think really important is the centralization it's not just one person it's not just one provider it's many providers distributed across the glow so that you know if one of them is down because of surge or whatever or they got censored or they got their servers were disconnected from Power whatever you can still talk to many others there's always an alternative just like in traditional blockchains when we're building a blockchain we're making sure that uh you know it's can you can suffer up to let's say 33% faulty or malicious participants it's the same for RPC we have to make

sure that if most participants are um giving good service the service is going to keep running so how does it work from the rollup and the provider T it's very simple a new rollup wants to launch let's say it's even compatible it takes our what we call spec which will'll show you later demo it's basically this Json file it's like a schema and it defines the different types of apis so if it's evm compatible you just inherit from evm and you're good to go just change the chain in uh change a few parameters that are unique to your change your chain you deploy it on chain and you deploy your token let's say you have roll token you deploy it in

what we call public RPC pool and you say uh every month distribute 1,000 on this R token to the providers now the providers are like wow I can get rewards from this just like R they do so they come to Lava they they running they start running your software your node and start giving service to your users they take laava tokens they stake them on your spec on your rollup and that basically gives you the economic certainty that you're getting fresh and Anor data because if they lie they can always get slashed just like validates is that the only um is

slashing the only uh mechanism of enforcement no you can do jailing as well jailing is where you just you get kicked off and your your funds are locked and you are now basically can't serve users so you can get paid so this is the so boss who's the jail boss lava is a decentralized network so there's actually a protocol is that's built there's an algorithm that is built based on the service that people are given and the ratings of that service that automatically uh JLS or slashes so how do deps connect to to La

so we're talking about providers right so now we have this huge list of providers why because 10,000 or 1,000 rollup token you know that's pretty good incentive for everyone to join start anyone was good with running downs from you know any regular person or that knows how to you know run stuff to actual like super bespoke providers on our Network that are they have like data centers or they have um their profession they can come run servers and give you the service so now you can access get this list of them it's called pairing list so um you ask lava can I get a bearing list for all these user all these for a DI menion let's say this

examp you get this list provided one two 3 four five six and uh you basically can talk to them direct remember we said it was decentralized so in the tradition model you just talk to one sir in our model model in the laa Network model if you use R sdj you are basically now directly peer-to-peer talking to any one of these providers so there's no middleman there's no um there's no way to uh if even if you want sales you can always fall back to the pafic how does lava choose which RPC provider is going to service my request okay amazing question so basically on chain providers take lava

tokens on the roll up chicken uh this what allows them to give you service on BL met for examp so um there's a few ways that we are now building this list when I say we it's basically the onchain algorithms the network so think about uh how much tokens you st and how many tokens just like a validators like a validator so on delegated to you so if you have let's say 10,000 lava State and delegated your chances are higher of getting selected to the for the list um this only only at one like a lottery it's not like a lottery because it's uh weight it's

weight stick okay but like if I have 50% of the Stak I have like 50% of the lottery tickets yeah yeah so yeah if you look at like that 100% so it's a random about it's it's weighted function this is one aspect the second aspect is also uh quality of service so as users get and start using your service they keep creating it continuously eventually it gets on chain uh and all of these ratings are summed up and providers will provide better service over time will get selected more frequently uh the third one is a more practical way of uh filtering it's like

what you need let's say you need somebody only in Europe so everybody's in the US who don't you know when you ask the algorithm it's still going to give you somebody it's going to give you some so these are the uh the different types of um pairing selection algorithms that we have you get the list and then you can talk to the providers gotcha so um maybe a silly question but um what if I want access to Lava but I don't care to have any stake or opportunity cost or Price exposure to the lava token um you know how do I overcome this this bottleneck to get access to the network yeah so to get access to network you

really don't need Too much exposure all you have to do is buy a subscription on chain with laa so you have to buy laa one time you can buy it for a whole year for example and then you don't really get exposure is this uh similar to chain link in the way that uh uh an application or or roll up might uh pay for Oracle services and the link token it's similar yes um I think that link is more continuous whereas in lava you can buy it's like a sub it's like it's like a SAS where you can come on chain and just buy a package uh you can buy it for one month for three months for six months for a year so it's really

up to you cool so do I as a user as a chief Dej trying to fight for my life on chain and farm airdrops from Africa do I need to um pay lava to use your high quality services or is there some sort of like subsidies or like other opportunities for uh you know just your average user yeah so as we were saying everybody today uses public RPC remember all the users almost everybody so who's paying for that roll up token in this scenario so roll up token in this scenario pays the providers on your have so you pay nothing they you just telling you the service for free for free they they but

in this scenario at least you know that there's no like ulterior motives for providing you to SLE service well what is a lava over lava the LOL uh red circle that okay so remember I told you you have to ask for the pairing list mhm how do you ask for the pairing list who do you ask the pairing l so have and the pairing list is the uh like the the ordered pair of provider and app no is just the list of it's just the list of providers for one app of the list of all the providers for one app yes and this list changes every

every few blocks every eatbook you call it every 15 minutes this this allows uh opportunities for smaller providers to also get into the list so it's not fully dominated by got it okay so the providers are running the rpcs from rollup to user providing access to the Excel sheet um but you need access to lava's Excel sheet in order to have access to all these providers that are going to give you access to the Excel to the to the state and the and the balances so so you run lava over lava to get the providers for any given app yeah we most to like clever names in laa so

LOL so this is our solution for this centralized problem right because if you're just asking our server please give us the parist then it's like oh so what's the point you're just trusting One S instead of that we have what we call La over la which is anybody who stes to provide lav over lava is a list of providers that you can load from anywhere from your computer from from from a website uh from your snapshot of the blockchain take a list very simple fact and you can just talk to these providers always get the pay list and it uses the same protocol so everything is trusted uh and accountable for I guess quick quick question going

to the to the uh like the economics once more is it uh necessary to pay in the lava token if I'm just a user or a rollup or could I could I for instance like taking a step back if I have usdc could I commit that to the network and then in the back end the network uses that to buy lava and and purchase the program yeah so the rollup itself can pay whatever token they want for the providers so they can put whatever like cup token they can put lava token they can put whatever token that is reachable to a cost chain with ibcn um for the user they I mean I'm assuming that someone that build like really cool

interfaces to basically just Swap and buy the subscription usdc or whatever token okay okay and then eventually it has to end up in lava to uh to uh follow some slashing rules so just to clarify when I say user the only users that will actually pay are really high throughput apps that need anything above the publicy the publicy gets you wherever you want to go but it's not it's not something Unis swap would use because Unis swap uses a bajillion quests and it's not fair for the providers to serve a blian quest whatever rollup token uh is placing the O so someone like udis SW you know go on chain and buy the higher

tier subscription and this and this payment will go to the providers as well okay and then is it only it's only the providers that could get Slack right it's not Unis swap or applications or rule ups that would get slashed correct slashing is so a provider stakes in order to give you Service as a user why does it stake because staking is a way to ensure that they have economic uh you have an economic certainty that they're not lying because if they're lying there's a penal got it okay okay so some really cool features that I I'm going to Glimpse

over let me know if you want to dig in deeper one is the provider optimiz so as I was saying you get this list of providers yes they're rated by quality of service and they're also rated by stake so you know how much you trust them but maybe you're in somewhere that is not really uh traditional maybe you're in a in aition continent but where there's no good providers so from this list uh the provider Optimizer this is all done completely the back it checks it basically probes it provides one by one or uh and checks who's the fastest it also checks who is the latest data um so it guarantees to give you from this list the best provider for you

not personally um the other feature that we have is constant availability so we're talking about how uh this you know the key qualities of you know what's important for building a universal access layer right uh so what if a block if lava blockchain holds I mean we've seen that happen in other chains right all of the RPC and all the other chains is going to stop working that's Unthinkable right we can't this is something we cannot have so when we we built lava we built it with this feature basically it allows provider providers to see that lava stop working fed and they can now in the background

uh keep serving users and keep them accounted for when the lava comes back they can send all those relays on chain and basically uh get paid for the working down while lava was fire so we made sure that even if lava holds you can still use it and we've actually had holds on testnet because testnet we sometimes deploy uh very interesting patches check different scenarios and we tested the the feature to work and it worked amazing um one more think is a native res sck so I was saying we have a lava validator which is this traditional cosos validator which basically builds blocks uh and then we hand votes in the network and then we have providers which

take tokens right yeah and then they can give you service in RPC so uh when you stick to a when you delegate tokens to a validator you can also delegate the same tokens basically restake them to a provider when you delegate to a provider you can restake that on a validator we've done it so that first you have better Capital efficiency um and also uh it uh balances the network the voting in the network between the validators and providers it's very important to keep the NDS that they don't try to uh maneuver on one or another it also improves the uh basically allows to reset the the token

which is I think a really interesting way to promote better optimization of these uh the quality of service across the network um yeah I think these are some super uh cool features um Let me show me what a spec is so spec or specification remember uh when the rollup is coming to deploy he's coming to deploy his uh his chain of lava provide this access layer to his blockchain he builds this uh what we call specification spec inside the spec it describes what are the apis that are needed to perform uh this this RPC the service that the users are expecting to

get so they are a very simple file that uh anybody almost anybody uh a little technical could build and inside of it there's some two cool features I want to talk about one is add-ons so if it's a bit technical let me know or feel free to ask question guys so um add-ons are way so in ethereum for example uh there's the usual apis that I showed you eth underscore Gap balance if you remember but there's also thebug underscore trace for example uh these apis are way more computationally intensive to perform so not all of the providers want to give them because they might be too expensive for them but

they're not optimized for them so addons allow you to put that aside and only let some of the providers choose to service that for an additional payment extension are uh are something like an archive node so remember that Excel file so most nodes only keep the sucel file they only keep let's say the the latest 10,000 lines but uh an Rive node is an Excel file that is the entire history of the lob so it's like a huge huge huge huge file so it's very expensive to keep it so what are extensions allow you to do

and this is very unique to RPC this is something very new and I think super cool it lets you basically tell the consumer what which provider to talk to so if you have archive support you get paid let's say five times more for this query and the query will go to you if you don't have it then the user will send uh and in if the user asking for a you know one of the newer lines will go to the regular so it allows you like this Mar reality of request so what are some options and which you can use all these are available right now by the way Jill real quick um I'm not sure if this specific functionality is what

allows the extensibility of lava or something else but my silly question is how does lava scale to thousand chains and do it where other kind of heavier antiquated Services cannot because of the SDK like like why how so SDK is one element that allows for that because it allows for direct communication with providers whereas all the providers today they rely on the central point where all traffic goes through two different NOS lava is more like um this example where you see this uh red circle or the mention R for for

example instead of going through this one Circle you go directly to the orange provider so you let's saying you have 10 of things or 100 of things so you don't go through the central point so it makes it more scalable and more uh it doesn't have a single point of failure because there are many points to D many providers the other way in which it scales a thousand chins is using this spec whereas you know provide whereas uh traditional providers they have to go and they have to bu infrastructure defs development to bring in a new chain in laa it you can do it in 10 minutes it just build this back you deploy it and that providers join these providers

might have access there might be validators on the Chain they start giving service and then the incentive program is just supposed to work such that as soon as the specs are deployed there is enough incentives for um providers to get active yeah if there's not enough incentive they're not going to come so where as a chain you have to make sure that there's enough incentives if your chain is really heavy let's say you're Salon you need a lot like a pasty machine to run up then your accept should be higher if your chain is super likewe maybe you can put a lower uh a lower uh incentive so it's really like a uh a market for these types of providers and it should be in just join uh and at the end of the month we tell you the

network tell you tell how much teach provider get and the quality of service and you get a reward so one thing to think about lava it's like a distribution Network for your notes so we have the server kit it's the goang implementation of the of lava uh protocol so basically uh it allows you to it talks to the providers for you we have the lava SJ which is the same thing only written in typescript so you can run it from the browser let's say from the UN SW app for example and we have the Gateway which is more like a traditional familiar RPC provider where you just get the URL you're getting some

of the benefits of decentralization but not all of them uh this is an SDK uh example so we built this tab called all blocks and what it does it uses lava SDK to directly communicate with the pro all the providers in L so in this example if you see on the right you can see all of the probing and all of the relays that the SDK is doing to many many different providers and in the example I think we're going to click on one is relay and you can see this one is served by some somebody called palmer. you get the latest Spock and if you click on a different relay see if we do it here then you you will get a different provider uh so we

talking basically to many many providers at the same time it's one thing to think about all the all the chains that are going to be live and deployed in the multi-chain future and then your demo just reminded me all those chains are also going to have test Nets and so uh it just expands this RPC scalability issue even even further and Defence and test nuts and multi rep yeah yeah I mean it's been super exciting to build this project you know it's it's such a it's like the underbelly of of

web 3 it's like the aillen Seal of web 3 right no one talks about it no one talks about how it's super centralized and how chains are now Boot surfing RPC but everybody feels it when you go to claim an airdrop when you're trying out a new test set when a new chain launches when nft drops everybody feels it so you know we've been building this and uh we're super excited I'm I've been super excited to share this with you guys there's uh the old adage of don't trust verify and I think right now in the current landscape this is very much a step in the right direction because Alchemy and and furo without naming you know the rest of the centralized

providers like they are very centralized it's very much a trust that your data is accurate and secure uh there's no verification and then this you know there's obviously more decentralization um and so we're trusting a decentralized network instead of a centralized network is there an element of verification basically how do you continue going in the direction of of more verification less trust uh for this network I think that's a really great question and you there's multiple ways to approach this problem and when we were thinking about this two years we went with what we call the optimistic approach so just like rollups when you when you talk to work

like use RBC to S transaction to rollup and then it tells you oh confirm you made the trade you made the swap you transferred the funds then your friend looks at the transaction is like okay I can give you the pizza web um so now there's a there's a an element of trust why because again rollups most rops today are optimistic in nature that means that they post the updates from the da or or let's say on U on E for example and then there's a dese period of seven days and so like op stack still hasn't for example fully implemented uh fraud so how do you know like that you know it's actually not going to be

reverted or that it actually happened correctly it's very difficult to do uh but this deep dispute mechanism allows you to slash a bad sequencer right that's a core idea so as you go up in speed you can't verify you just can't if you start verifying it's too slow it's way too slow so what we've built is a similar approach to uh optimistic blups uh so for example you get a response it tells you your balance is 200 every once in a while you roll the US and you ask another provider the same question what's my balance and then you compare the two results because blocks chance the Excel

file is deterministic everybody agrees in the same state uh for the same block if there's any uh mismatching the hashes of the response you take this what we call conflict and you upload on chain on chain we have a voting mechanism so all the providers uh that are chosen by the jury reexecute the query and if the majority says that July that the one is the providers the majority wins and the the minority might get slashed and the majority might get a reward St for days yeah and this uh deterministic property of of state is also the reason that uh like data availability sampling Works in Celestia because as you as you

sample the likelihood that you find any kind of conflict decreases exponentially uh as you get one commitment two commitments that all all agreed to one another it's very it is increasingly unlikely that you find any kind of conflict exactly yeah so the the more you check so it's it's like a mechanism to deter bad behavior so this Ro of the dice if it catches you you could lose everything you worth so that it doesn't in the economically doesn't make sense for you to to give B results awesome well uh Jill it was an absolute pleasure having you on I feel like I learned quite a bit about this this piece of the modular

stack that is not uh talked about very much and uh I really do appreciate you shedding some light on it thank you it's been a pleasure to be here and uh let's hope we talk about it a bit more and let people know that you know they still need access to their data they need to access definely that's what we're building definitely uh definitely and I know that it's the uh you guys have a a staggered set of launches up here so I hope that those go all smooth I know that the community is pretty excited about the points program and things do you have anything on on that yeah you know feel free to to come join us join our community we have a huge Community bring your chain Be A Champion or push the the different specs on chain

help us uh bring the word awesome man thank you for coming out thank you guys so much