#DappCon22 : Day 1 - The Graph goes Multi-Chain, Eva Beylin

Transcript

thank you thank you so glad to be here and so glad to hear so many people use the grass um let's get started so as you may know the graph is a default part of the stack so we are both the API layer and the indexing layer so we currently have something called a hosted service and also the decentralized network that we've been working on for over four years now but the sub graph itself is our branded API it can be composable open source you can Fork others that are existing make it your own we consider ourselves a global API so you can build a subgraph for any use

case whether it's for a dow adapt you know entertainment social it really doesn't matter any kind of data can be indexed in a subgraph and like I said you can Fork existing subgraphs so some of the most exciting subgraphs we've seen are ones you know for example the uniswap subgraph being forked and redesigned for the protocol that also may have forked uniswap or adjusted their their smart contracts so I want to get clear on the difference between the hosted service and the network so the hosted service launched over four years ago it was intended for product Market fit so we launched the subgraph standard which is the API and we really knew what the problem we were solving was which is people need and developers need a way to access Rich

data in their UI and it's very hard to do that yourself you know it really doesn't make sense for adapt team that's focused on product to also have a devops team that's focused on maintaining your server or your database and so the hosted service was the hosted alternative but you know there are centralized dependencies you know the reason we're all here in web 3 is to remove those dependencies so we've been very transparent from the get-go that the decentralized network was the goal there's no interaction with GRT on the hosted service and the hosted service already supports 39 networks so this is every evm chain you can think of test Nets Etc and now is our chance to go from the hosted service supporting all these to the actual Network supporting

these the decentralized network has over 185 node operators today so node operators in the network are indexers they run graph node they're doing the hard work serving the queries they're also the ones that earn GRT and in addition to indexers there are curators and delegators who can also earn GRT by staking on subgraphs or staking on indexers so really creating a self-sufficient economy where you don't need a centralized team like a hosted service to tell you which API is high quality you can have people in the community signal on those subgraphs stake GRT on those subgraphs and by doing so indexers the node operators know which subgraphs are highest priority so you can imagine something

like a uniswap subgraph or maybe radical some of the most used subgraphs see the most signal because they're built by some of the highly trusted teams and so indexers will then pick up those subgraphs and prioritize them the other thing about the decentralized network is that you know we already have 450 subgraphs that have already migrated but really this effort over the next year so is to migrate every single subgraph that's still on the hosted service to the network just a highlight of some of the dapps that have migrated and what's really exciting is that we are seeing a lot of these dapps you know have multi-chain subgraphs so they've deployed themselves on different chains so all of these so

far are just on ethereum so really exciting to see that traction and as we go multi-chain every dap can also continue migrating So the plan to Sunset the hosted service is Q q1 2023 by end of q1 and we're doing this in phases so by the end of Q3 all ethereum subgraphs will be or sorry you will not be able to deploy a new ethereum subgraph on the hosted service and by end of Q4 you won't be able to upgrade or make any changes and the point of this is we're creating a phased approach to ensure that developers are feeling comfortable with the network able to migrate get the support they need and also can continue serving their daps so we want to really want to make sure we're not you know leaving anyone

stranded and giving everyone plenty of time to actually migrate their subgraphs over like I said we already support 39 chains and we're adding new ones you know every month most recently we added our weave support and Cosmos Hub are we support isn't fully finished we still don't have full support for files but the goal is that really any chain like I said evm or not will be supported on the network as well so mips you might have heard of this but this is basically our next step into enabling multi-chain on the network mips is our incentivized program that we just launched called migration infrastructure providers and the goal here is that not

only do subgraph developers need time to get you know comfortable with migrating but indexers do too and we recognize that not every indexer today that runs ethereum will run every other chain so we've created this program that over the next six months or so any chain node operator regardless of which chain it is can join this program be incentivized to run an indexer and help us test and create the best network most performant Network for dapps so really excited to share that we have 50 million GRT allocated to this program we expect hundreds of node operators to participate we've already had over 500 register let's hope they actually sync their nodes we all know that can be a challenge and the program launches

September 20th so really excited to see that over the next six months we're going to see hundreds of node operators often are likely new node operators ones that currently aren't indexers participate and really what this means is node operators you know if you're running a gnosis node or you're running an optimism node you can now also run an indexer you know this wasn't really possible over the last two years as we gave our indexers in the network time to be self-sufficient on ethereum working at any kinks but now this means that if you're running a node for that chain you can also support the DAP ecosystem whereas maybe previously you're running the node for mining purposes or whatever else it might be you know really indexers are the backbone to supporting dapps so those daps don't need to run the nodes themselves and really excited to share that gnosis

chain is the first chain that's being supported on the network let's hear Round of Applause for gnosis so exciting and the reason we started with gnosis was firstly we think that gnosis is incredibly aligned with the graph ecosystem the values of decentralization evm based very easy and user-friendly for nodes to to be run and we also think that the ecosystem in the gnosis chain um you know world is is really full of dap developers so we want to make sure that any dapps that are on gnosis chain can have the opportunity to migrate immediately and again if you have you know adopts across many chains as the new chains flow out you'll be able to migrate that subgraph to so the process for our mips program will

be a loop per chain basically Our intention is that indexers can set up their infrastructure prepare the chain so sync their node we're also in talks with all of the chain foundations to make sure that we can get snapshots or whatever is needed to make it as easy as possible for for you guys to run those nodes start doing some stress testing on Gurley we recently migrated our testnet too girly beginning the actual migration to the network so that might be testing some of the subgraphs queries testing the Gateway on our test net and then eventually monitoring quality of service on the network at that point when we have you know high quality of service we have enough indexers serving then we'll start migrating production dapps to the network so it's really important to us that dapps aren't migrating to the

network and don't feel again like they're stranded if the indexers aren't ready themselves and we take that very seriously we think that you know again the graph is kind of the solution to make sure your dap is fully decentralized so we want to make sure there's enough node operators that can provide you that quality of service we also have what we call I don't know if I'm going to butcher this name digstra which is our L2 test net so we are uh deploying uh on arbitrum to make sure we can have scalability in the network this is really important for indexers as well as subgraph developers to make sure costs are low and so indexers that choose to help out and participate in that testing Forum can do so and then every indexer will be able to earn rewards based on their

participation so really encourage you if you run a gnosis node to start syncing programs start September 20th and then new chains will be rolled out every few weeks we expect four to six weeks per chain but really that's dependent on you know how node operators behave you know what challenges they see what you know how long migration takes Etc so really the goals of this mips program are to test the network so you know like I said testing the Gateway L2 all these critical components that we've only seen with ethereum so that's one chain and as you add more chains it increases the complexity for the indexes themselves expanding multi-chain so of course supporting all the dapps that currently use the graph on the network itself and rethink this is really important because

you know ethereum is you know obviously the first and maybe the largest ecosystem but again we know that many dapps have deployed across chains we want to make sure you're not having half your subgraph on you know the hosted service and half on the network but having them all migrate at the same time and then of course migrating your subgraph over and you know the end goal is that all daps across chains should be fully decentralized you know we shouldn't be building for one chain this is a multi-chain Universe um we believe in being uh chain agnostic although we are built on ethereum we want to make sure everyone is supported in web 3.

like I say we have 185 indexers already on the network we've got nearly 10 000 delegators which is incredible delegators are folks that have GRT they don't want to do all the work of setting up an indexer but they still want to participate and so they delegate their GRT to the indexers and what that does for indexers is it actually increases their capacity to serve the network and so there's a multiplier on their ability to serve and allocate to subgraphs and that's also based on their delegation so for example if you are a GRT holder and you don't want to run an indexer the best thing you can do is delegate and you can also earn rewards from the network curators like I said are the folks that signal on the subgraph so this is very important because you don't

know which subgraph is high quality you know maybe with touch to the wind but otherwise the point is that indexers follow curators to know which subgraphs are most important um and what's really exciting about the curator role is that the intent is for the subgraph developer to be the first curator so every subgraph has a bonding curve so obviously the earlier that you're able to Signal the better it is for you and the intent is that the subgraph developer the one building it or maintaining it is able to curate instantly and so as the subgraph gains more usage the bonding curve is Works in their favor and they can earn more rewards over time and lastly we have 466 subgraphs already migrated very exciting but not enough so we strongly believe that multi-chain

is not only for the dapps that are building on that chain but in itself creating multi-chain capabilities really enables web 3 to be what it is meaning you don't want to have your infrastructure split across different infrastructures you don't want to have an inconsistent experience for your developers and so we think that once you actually have multi-chain in a fully decentralized way you can have completely new use cases that maybe weren't possible before because the ux or you know difficulty of doing that was was just too high oh uh oh let me just go back so top daps are already powered by the graph Network this is again just ethereum today but anything from

websites that are a bit more static you know like a web 3 index or a snapshot to nfts and Dows Juice Box style actually is especially my favorite because if you all remember Constitution Dao and I don't know if anyone participated um but at the time of constitution Dao's Peak they were actually running their own server I think they might have even been hosting their own subgraph and they couldn't handle the load that they were actually forced to migrate to the decentralized network in the middle of the auction because their website kept going down so that's a really great example of how the network itself can be more resilient than any centralized service because you don't have to maintain it yourself there's no central point of failure and you know just like ethereum where you expect at least one node operator to always you know serve

those blocks same here you know you always expect at least one indexer to want to serve your queries and if for example your queries are too complex or there might be some reason that it's not as attractive the index it can also set their query prices so everything is very Dynamic and we expect that four dapps themselves this is exciting because you aren't paying maybe one giant subscription fee that you don't know how that's bundled up or you don't know how those costs are attributed you get to pay per query here so you'll have a fairly consistent price you know per month but it'll be based on the queries that you are actually experiencing not you know some some larger tier so the best thing that you can do is begin migrating your subgraph if you haven't already done so it's actually

quite simple you know we are improving documentation and we have a lot of support Engineers but otherwise you know it's it's free for all you can do it yourself firstly you publish the sub graph on the network itself so if you already have a subgraph on the hosted service this is very simple it's just a toggle in the studio then you signal GRT on the subgraph so like I said subgraph developers are really the ones that can benefit the most from signaling early create an API key copy that URL and then migrate your endpoint and it's very easy and the great thing again about the network because it's open permissionless and the subgraphs are open source as well is that once you migrate your subgraph anyone else can benefit from it so you actually as a developer don't even need

to be the one to migrate it if someone else wants to migrate your subgraph they can too but again the goal is that you know the benefit is huge here for everyone we as the graph Foundation also have a lot of migration grants so if you are a Dap or a developer that needs help migrating your subgraph we're giving out grants to help you bootstrap that experience so bootstrap your initial signal some of your initial costs and to actually walk you through that in our White Glove service so I don't know if anyone heard about subgraphdal we launched this sort of subtly but the community has launched the second Dao in the graph ecosystem we have the advocate style already which is a community Dow focused on community grants and the subgraph Dao is focused on all things migration and subgraph

support so really the goal was how do we create an environment where you can in a decentralized way support the community so it isn't the graph Foundation or one of the core developers that has to help out so firstly migration support this can be anything from taking calls to literally helping you migrate demoing sharing screen pair programming whatever you need subgraph QA and testing so one of the things we've seen in scaling subgraph development is that you know sometimes people are brand new they don't know all the best practices they build a subgraph that has issues or maybe it doesn't sync as efficiently as it should they come back to us saying what's wrong with you guys why isn't your subgraph working hey you know you actually could have built it maybe a little more efficiently so we want to make sure that the subgraph Dao

creates a community of subgraph developers that can actually do quality assurance of each other's subgraphs and we've already started this process of interviewing folks who feel that they're senior enough to take on that role and actually be mentors for others subgraph standards so this is really exciting for me personally because what we've seen a lot of is people forking subgraphs or creating you know similar versions but they all follow very different standards and that makes it very challenging for dap developers to then build you know really Dynamic apps it you know it takes more time to understand what type of subgraph you're using how to integrate it and so misari one of our core subgraph developers is actually building standards across all protocols they started with D5 lending

and so they're trying to create a standard schema that all defy lending protocols could follow and that way it makes it much easier for a D5 app to be able to actually integrate and we'll see more of this but the goal here is to have work in groups really figure out what are the standards that daps need per category so in defy nfts Dows Etc migration grants and bounties so like I said we have a lot of migration grants but we also are hoping the subgraph Dow can do what we're calling migration bounties for facilitating the migration um so this is you know if you know of a subgraph that needs to be migrated maybe that team doesn't have the capacity but you want to help them you can earn a grant or a bounty for actually migrating the subgraph over so again we're really

scaling in a decentralized way to make sure that there isn't one individual who's an expert and the goal here is that those folks that help you migrate can then also continue helping you you know maybe you hire them eventually maybe you have an ongoing Grant but you know you don't have to take on all that work again because we are in an open source ecosystem and lastly subgraph grants themselves so every quarter we fund probably around a hundred thousand dollars worth of Grants to actual subgraph Developers for building their subgraphs maintaining them creating new things sometimes it's tooling and the subgraph dial will be taking those over long term so really our goal is to remove the foundation core devs from this process as much as possible to make it as easy for you as as possible too

and coming soon so in addition to us expanding multi-chain we've got a lot of work going on that can help those on multi-chain but you know really anybody building on the graph so number one is sub streams this is work coming out of streaming fast one of our core developers which enables parallel data processing so we're hoping for up to 100x Greater efficiency in data streaming and the difference here is that the data model is actually quite different whereas the subgraph follows a schema the goal with the substream is that you can just Define you know maybe it's all uniswap token Pairs and stream that data and eventually you can actually have the subgraph consume that substream as well so the goal is to actually open up different use cases here we know a lot of folks actually use subgraphs for their routing algorithms

or non-ui use cases so not just querying the subgraph for their dap and we hope that substreams can actually make that much more feasible L2 scaling so like I said we're incorporating arbitrum in the network this is part of the mips program as well so if you're very interested in testing this please join shell proofs also incredibly exciting we announced this at graph day and shell proofs are a new efficient snark that are being built by our snark Force it's two two to three core Dev teams coming together working on a snark and the goal is to use this four verifiable queries in indexing you know so the graph right now migrating to the decentralized network we still have some manual components where arbitration

occurs you know with a group of people but the goal is that all arbitration or governance can occur in a verifiable way so over the next few years you'll see verifiable queries verifiable indexing coming out and this also has impacts on eip44s where you can use the graph as the verifiable solution for querying historical data instead of using a full node and lastly Gateway optimization so we know that the network is only as decentralized as the actual stack of the network we've done a lot of work to make the Gateway more efficient so to make it actually more efficient in querying also the the processing of the transactions and the goal is to keep adding those features and expanding to more gateways we really hope that you know eventually

everyone can run a Gateway we're on the path to like clients but for now just getting more out there and making sure everyone feels good about how they can access their data in a decentralized network is most important and that's it we're going multi-chain we're so excited about it we already support chains on the hosted service but you know the true vision of supporting on the network is coming true this is cool thank you so much so so I think we have uh time for maybe a couple questions but let's make it snappy let's do it uh is this the first one right sorry

so sorry for the gentleman in the front I took the Mac first in going multi-chain is just about indexing or do you also plan to move your contracts like the bonding curve and all that to a layer two because like interacting with the the graph contract on mainnet is one in my opinion one of the biggest barrier to people adopting decentralized Network yeah great question so going multi-chain is supporting new chains on the network but like I mentioned we are expanding to layer two I mean arbitrum is is the network that we're expanding to so part of the mips program will include testing because this is brand new again you know there's obviously been a lot of hacks

and whatnot in the past with with Layer Two Bridges and so we want to take the opportunity to make sure that's tested well before it goes in production thank you and there was a question right here yeah actually this was also like the question that I wanted to ask but I have another one uh so uh do I understand correctly that also like uh the the indexers for the um for the other chains that you that are going to index like not only ethereum like what do you expect them to be like do you think there will be like those who are running the notes like for different chains or like it will be for example let's say people are running like the archive notes for the blockchain itself and they will also be running uh like notes for the for the uh for index index hours of

the car of the graph yeah so that would be the goal you know um because we've been on ethereum only the last two years all of our index most of our indexers are ethereum-centric and so the goal is right now as we roll out a new chain that the node operators who are the experts so if you're running gnosis chain right now please join us um to actually run graph note as well um there's maybe a bit of a learning curve you know you have to deploy graphene you have to work out some Kinks but you guys are the experts and we also think that the ones that are running the nodes for that chain are actually most incentivized to continue supporting that ecosystem too so you know really it's not for the graph it's for the dapps in that ecosystem so we want to make sure

that all gnosis daps have enough indexers that are supporting them but really anyone who wants to run a noses chain especially because it's so easy can can do that okay so basically yeah sorry so basically it's like it's going to be more you think vertical than horizontal so like indexers will be the ones who also run in the notes of the blockchain itself and not like the let's say the indexer who is running like the indexing notes or for like different chains like ethereum losses and all that of course like it can be both ways but like what is what do you have like any statistics on that like what is the expected way yeah great question because we're just starting out we don't know the stats there I expect that for some like gnosis we will see overlap because they're evm based it's fairly easy I

expect maybe for something like near we won't see as many current ethereum indexes running near just because it's not their their bread and butter but the cool thing is that all indexers get to choose so you really get to choose whichever chain you want to run and then also you know depends on like the the cost in economics because you as an indexer are earning from the network a lot of what we're seeing is indexers choosing which chain they want to run based on the economics that they foresee so for example polygon notoriously hard to run more expensive however it's one of the most used chains and so we could see you know the the indexers who are willing to maybe have a larger setup will do that but really I'm excited to see what happens so yes I hope to show next you know data maybe next stop con

yeah thank you super cool last question is there a Dap node dap app that you I can run for an indexer or something like that dude I would love graph node that would that would be cool yeah super easy what are we waiting for I don't know I don't know guys it's gonna do it thank you so much thank you thank you for being here