Thomas Heremans, Luke Hackett - How to run your first distributed validator

Transcript

yeah good afternoon I hope you had a good lunch next up we have Thomas and Luke from obal who are going to teach us all everything about how to run your first distributed [Applause] validator hello can you hear me properly cool so nice to meet you Thomas head of operation at OBO labs and this is Luke platform engineer also at OBO So today we're going to teach you how to run distri validators so could I just get a show of hand here who is solo staking nice nice awesome okay so I'll tailor the presentation then definitely for solo staking as this seems to be the

main audience so we'll go over what are distributed validators then Luke will do a quick demo on how to actually set it up using our Launchpad and then we'll follow up with a few nice announcements uh that we've been releasing recently okay so what are distributed validators uh exactly so why would you fly a plane sorry with only one engine you wouldn't right you want some resilience uh so distributed validators allow you to split your validators across different machines so if one of your nodes goes down you can still sit at this conference and enjoy instead of rushing back to your Airbnb and trying uh to fix it through SSH and hope you didn't have

to propose a block so as long as two3 of the nodes within your clusters are online you'll be able to perform your duties whether that's attestation whether that's proposing a block or even exiting the cluster so uh a lot of the validators uh that are Enterprise skill would use it and run all of the nodes within their cluster but for you that are Solo staking uh a very interesting use case is actually what we call Squat staking so that's a few of you teaming up and actually setting up a cluster so that you can rely on each other's uh up time and skills uh to keep validating as a team so in the end DVS are really just multisig but for validators how does OBO help with

setting up distributed validators we developed uh the most credibly neutral composable and etherum align DV implementation so as a pure middleware as you can see uh from the slide it's sitting right between the validation client and the beacon note we are not changing anything to the setup of of uh the node itself so that means you can still run your prism your Lighthouse as this and so we are not introducing any risk uh in terms of the existing clients that means that you can also make your cluster highly uh flexible and paradigmatic for adoption by just fitting in any existing clients you might already be running as part of your stack you can just add the Caron

middleware in your stack and get it to work uh within the cluster it's lean and composable Design also allows for very high performance and security as the nodes are connected uh in in direct mode from one another uh without having to broadcast messages through uh a wider gossip network uh finally credibly neutral we are not storing or actually creating the key info anywhere at any time so you're just relying on yourself for uh storing the key shares that will allow for the threshold signing so you said you're a solo Staker so what does this really mean for you so as I said if your note goes down today

uh you are at risk of missing attestations of missing uh proposals so by increasing the availability and effectiveness of your validators once you team up you can get that much higher uh peace of mind also in terms of Maintenance overhead that's peace of mind for you but also peace of mind for actually liquid saking uh protocols that you might be partnering with so I'll talk a bit later about initiatives we have with EFI and that we have with Lio uh but it's been quite difficult as you know for liquid staking protocols to actually onboard Sol stakers uh due to the risk profile of the of the different people they might want to on board so DVT is actually reducing that barrier to entry to on board many new solo sakers

so uh that's a very good Avenue to go back to the roots and level the playing field for solo stakers back into um staking at a high scale on on ethereum finally by also reducing the risk uh and the resilience of the cluster you don't need such high bonding requirements than you may also have in other systems like rocket pool so it's dvts I like to say it's really just like a risk reducing machine while increasing the resiliency of your validators all right so now look we'll walk you through how to actually set it up using OBO cool thanks uh hey everyone you might have to forgive my voice if already already had a a bit of a conference out weekend so hopefully I'll

be able to get through this all right um so yeah what we'll be going through so far is setting up your cluster configuration through our Launchpad uh this will involve generating who is in the cluster generating who you will sending the money to you through the withdrawal and fee recipient and then also just making sure that you sign on so that your wallet address or your ethereum address is actually the one that created set cluster um then we will go through the dkg process so each operator will accept the conditions of your cluster and do a distributed key generation in order to create a set of validator keys without ever storing a full key in one location uh and then I'll just go through an example of how HDC that your cluster is alive at the

very end of that uh through our monitoring where we provide uh dashboards and grafana cool so yeah this is essentially our flow from the Creator perspective um you have this process of going through all the advisories making sure that you understand what is actually involved in staking and in becoming a validator in the ethereum network uh the Creator is the one that will Define all of the information for you uh so you need to obviously like in this setup uh trust that counterparty uh you have like the cluster details which are your name of your cluster which can be free form any information you want any ask key uh and

then below that is where you decide how many operators you will have and how like their addresses uh so you have like a handy button to say use my address if the if the Creator wants to be part of that cluster but the creator doesn't have to be part of that cluster also uh you can type in here ens names um oh sorry I'm doing something to the side there you can type in either ens names or just straight ethereum addresses uh and then afterwards you generate what is called an enr which is essentially the operator's unique uh like ID for that cluster uh and you also are able to pick here how many validators you want you can run one you can run thousands um in one given

cluster depending on your uh capacity and your setup uh so after you do that you go to the withdraw configuration stage where you can either use just straight up addresses or you can do something a bit more complicated with splitter contracts uh through our SDK which we'll talk about a bit later on uh so yeah once all of these details are confirmed you sign this cluster configuration and it'll generate a invite link for the other members to use which we'll see uh in a second there's yeah multiple signature requests here that don't charge you anything on chain so yeah now that you have an invite link I'll just show you briefly from an operator's perspective

for people see uh so you can see already who's accepted and who you're waiting for in the Launchpad flow uh you just need to get started read a similar set of advisories um sorry quite similar to the Launchpad that you would have for the standard ethereum uh validator and deposit setup just understanding your group risk that it's still Alpha and Now Beta software uh and yeah that you need to be careful running it uh this is where you see that you want to trust and verify the addresses there that you know who you're dealing with uh and that you know where the withdrawal and fees are actually going to and you generate your own unique

ethereum name record uh in order to verify your ID uh locally when you're doing the dkg process so yeah once you confirm all of those details uh you'll get back to a operator status page where you will wait for the rest of the operators to complete this flow uh then like the DK part of this that comes afterwards uh we provide in our documentation um like the commands that you will be running uh you will need Docker installed um pretty basically uh but you can also build from Source um we have yeah Caron is completely open source uh you can build it and run it locally if you want to do that um so

yeah at this stage uh what we give you is a link with the dkg command and some simple information about the cluster we store the cluster definition in our API for ease but if you again want to like fetch it from somewhere else you can store it locally and then distribute it uh without using any of our services um so yeah once the dkg process starts running we wait for other peers to connect uh in this situation there's four peers so three other peers that we're waiting for uh they communicate through a relay this relay again very configur durble you can run it yourself or use one of our public relays we also have four I think Partners right now

that are running public relays so if you don't want to use our infrastructure you can use someone else's and it just yeah again leads to that like credibly neutral approach that we're trying to take where in any situation you will not have to use anything that's hosted or provided by us um apart from the software um so yeah once all of the peers are connected you have the dkg process starting which can generate any number of keys it will take slightly longer the more validated Keys you have and then it will push out a link that will lead back to our uh Launchpad which I'll click on in a few seconds that will be where you can see the current status of all of the validators in your cluster so yeah I click on this link here

hopefully in a second .5 speed and and I'll finish up with a cluster all of the operators accepted a waiting deposit you can verify all of the information one validator pending activation can't claim any eth yet can't exit because it's not deposited and you will have output as part of that dkg deposit data that's identical to any other ethereum validator that you can either deposit through our lunch pad or through the ethereum lunch pad so yeah that's very briefly a demo of how to set up a cluster once you do that you will have all the configuration files needed you run your normal e validat stack so your consensus layer your execution layer

your validator and you run Caron in the middle of that uh and just to give an example of like a cluster fully up and running what it looks like is this will be our status page that or overview page that you can see uh if you download cdbn our repository uh we have these um Caren overview graffan dashboards in them and you can see right now simply for our main net cluster this is the first TV that went live in December of last year uh we're still running it we have three other nodes connected one one that cam game node is actually mine uh running on a d node and you can see the active validators that we have right now and simple information that will help debug any issues with your cluster um so yeah this is all publicly available

online online you can look around our documentation see see if you can figure out how to run in the future and yeah hope you all give it a try thanks cool amazing uh thanks Luke so where do we go from here so you've seen how to run a DV which like initiatives can you really take part on so I'm going to share a few with you so the first one is actually we're super excited to announce our mayet open Beta uh so we announced that two days ago as taking Summit and that that will go live on the 1st of December so what is part of main net open Beta so as Luke just showed now the DV Launchpad is open for anyone uh whatever your profile is to run a DV up to one validator at first if you're

interested in running more than one valid just talk just talk to us the second thing that we're releasing with the open Beta is our OB SDK so today you show uh you saw how to create a cluster via the DV Launchpad but you can also create and access cluster information through our SDK so that opens the rout for any protocol uh like we're doing with FL and eifi for example to integrate uh with us for pragmatic um cluster creation programmatic cluster creation finally we're also excited to announce our obal split so validator rewards need to go somewhere uh in the demo you uh you saw where they're going into one single address but now they can also go to what we call uh Splitters where the rewards

will be split across the different operators as per the rules that you define into the obal splitter so it's basically a set of Open Source smart contracts that also any other protocol May integrate with obviously karon uh is there has always been and it's ready uh to be used okay so a few opportunities as I mentioned for open Beta so you can go ahead go on our test Net Spin up a cluster or even on mayet if you're feeling like you're ready so we have also a key initiative with eifi called operation solo Saker where they're running uh solo saking uh clusters as part of eifi and they even provide you with the hardware to do so so feel free to chat with them if you're interested

in taking part in thei Solo Staker finally we also just passed a proposal uh with f though for simp for DVT which is going to be the first live module uh V2 module of Lio uh including uh only DVS so uh I think we have about 30 solo stickers which are going to be part of these um clusters and that's really the beauty of DVT it's allows protocols like Lio to actually reduce uh the trust they need in the individual uh running uh nodes uh thanks to the the more added resilience of these clusters so we are on boarding the first wave of simple DVD at the moment but if you're interested also chat with the Lio team and there

will be further waves finally if you're developer uh check our product as I said the SDK and the splits that you can also integrate with uh finally I think quite critical for solo stakers uh I'm not sure the industry has been good enough at recognizing the talent of the individuals that have the skills to run uh solo sing even if it has been uh becoming easier but we're basically launching the obal technique credential which is a badge showing the industry how much you know about DVS and that you're totally capable of running uh a noble cluster so that means that protocols that might be on boarding solo staking solo stakers like ERI or Lio all the rest that we're going to work with um might be recognizing this badge for

easier onboarding within their process all right so as I said I think solo saking and DVT can really make like uh a very good partnership to really bring back solo Sak taking to the roots of ethereum and decentralize it again uh versus the the the liquid staking pool so it's time to make squat staking a reality thank you everyone thank you questions so the you're sitting there waiting for the other participants um is there a timeout or just isn't responsive like what do

you do from there I guess it's like at that point right you can still kind of just back away yeah you can hang around you can leave the long running command running until something happens um yeah could be could be a day could be however long but I think after a certain stage you do want to cut your losses uh currently that would involve creating a new cluster definition so you would not be able to so at the operator in place uh at that current stage once cluster is running and's say someone you know becomes unresponsive orable to running what happens uh so in that situation uh you

would probably have two routes that you could go down depending on whether you were in like a less trusted setup or with your counterparties that you know uh you could either uh like just exit the validator create a new cluster definition and go through that process uh again from the beginning or there's a situation where you could take a quorum amount of the key shares recombine them split them and distribute out the uh actual configurations uh without having to do that manual configuration again yeah okay thank you very much let's thank the speakers again

uh next talk will be in 18 minutes through yeah 15