ETH2 Staking Community Call #8 - Joseph Chow, Lakshman Sankar, and Aditya Asgaonkar

Transcript

welcome to uh east acre community call number eight um we're glad to have you all here today it is thursday september 10th 2020. um today on the agenda we have uh we're going to talk about the studymaster program we have joseph with us jc with us to talk about the studymaster program we have blacksmith here to talk about the madosha data challenge and i also see stephen gilbert here and we have well if he hasn't shown up yet we should have um aditya to talk about a secret shared validators program that uh sounds very interesting we'd like to learn more about uh our call will take about an hour it's

here great uh we want to be as inclusive of the community as we can let me tell you a little the background before we get into the call the east staker community is housed both our at reddit.com and at our discord um east acre we run a a twitter at eth2scce2stakercommunitycall.com that you can use to keep up with us we are generally a bunch of people who want to support the onboarding of new users into the ethereum ii staking ecosystem we recognize that this is a major shift

in thinking and ethereum and we want as many users to be on board as possible you may see that we engage with some staking as a service providers but as a community the e-stakers do not endorse any staking as a service provider we do encourage solo staking while recognizing that there is a place forsaking as a service we cannot vouch for those services because we never know if one of them is going to uh disappear and we don't want to be responsible for that so we're always going to leave that up to you so we have this interesting relationship

with staking as a service providers we do not endorse them but we do welcome them uh yeah so we're working to hold these calls every two weeks uh let me introduce my team really briefly and then i'll ask them to say hello i am super fizz i run uh the i'm a moderator at the rent.com eatstaker and i work on the uh discord my pet project right now is the studymaster program uh lemboshi is my wingman uh patricio is here with poet we also have our dear friend buddha who is our devops guy who is on vacation this week and we're welcoming a

a new guy uh michael uh who goes by and vatica uh do you would you guys like to take turns saying hello yeah i'm michael uh this is my wife emma and we are uh together on vedica we're just uh you can classify us as enthusiasts for the time being uh hi i'm nolan um moderator at east acre and on discord enthusiastic participant hello everyone i'm patricio i'm running

pow-up i'm powerbeats using the e-stickers community skiniapix for doing your features and products so you will be hearing about our product maybe more than what you want to yeah so we're we're a little obsessed with coap so i know that i talk about it too much but i'm obsessed with it so it is what it is um let's talk about my obsessions uh you can find the agenda for today's call at uh github.com superfiz slash east staker uh i run an agenda there and they're sorted by dates today we're going to doing this introductions and warm-up

we're going to get into a study master discussion we're going to talk about the client co-apps for a minute if if you know that the eth2 client validators had a um a po app opportunity for each of the different staking clients and we want to chat about that a minute then we're going to get with laxman about uh the uh madasha data challenge and finally we're going to hear from aditya about kafta and secret share validators a lot of what we're hearing today is going to be brand new which is kind of exciting for all of us so uh without further ado are we ready to rock

and roll let's let's say hi to our guest real quick um blacksmith can you tell us a little bit about uh yourself and your background yeah sure um hey everyone i'm i'm lakshman i'm with the ethereum foundation um i've been kind of involved with ethereum cursorly for the last like two or three years before that i was like a software product person at like normal web 2 companies for for a long time um and uh yeah i guess i've been with the ef in some capacity for the last year and just generally really excited to see this ecosystem and community around youth 2

form it's been a lot of fun trolling the discord and i'm excited to actually have something to share today great thank you joseph can you tell us about yourself sure hi um yeah joseph and yeah been written different things about ethereum um made some talks different times wrote one of the early smart contracts called btc relay bridge with bitcoin and right now we're just trying to stop all this um zoom bombing happening here thanks great it's all yeah so whoever is helping with

uh with that i really appreciate it i wasn't even aware of it so thanks um and we have aditya hi dijia how are you okay so i i'm not sure if you're muted if you're not catching me uh but aditya is here uh to discuss saying on the chat that he can't unmute himself yes i'm sorry i was uh so lem uh laxman is host right now if you can give aditya i just asked him that mute should be good

okay well we're going to take a minute because i do want to make sure he has an opportunity to get there okay i also clicked on you hey did that work hi great here we go welcome how are you hi i'm good i'm aditya it's my first time here on our slash eatstaker uh been around in the ethereum community for i think two and a half years now and i used to work on casper cbc with vlad zamfir before and now i'm doing eats to research and development with the ethereum foundation and uh yeah excited for this call great great glad to have you uh very exciting

so we're gonna begin i guess talking about uh studymaster and uh though i'm not aware of it i guess we're having some some trolling issues going on right now i'm very interested to hear uh maybe through the chat box um questions from people who are participating in uh studymaster what you think of it i'm going to give you a broad overview and then we'll get into it a little more um so studymaster is uh an idea that i had to incentivize learning more about the ethereum two ecosystem basically we look for articles that are introductory in nature to ethereum

too so that people who are interested in staking can read an article learn a little bit about it and then take a quiz the big picture is uh we're going to offer 10 quizzes over 10 weeks if you pass 10 of those quizzes with 70 accuracy you'll earn a co-app that is basically a badge that says i'm familiar with ethereum too and i don't know where it's going to go in the long run but i sort of have the idea that looking at someone someone's po list of badges and seeing that they've been through this course would be a strong suggestion that they know what they're talking about when it comes to staking

the quizzes are uh intended to be at a 10th grade reading level so that basically anyone who's interested and is willing to put the time into it can learn more about ethereum too and we have joseph here who wrote the the first uh article that um our first quiz was based on uh can you tell us your experience and and and how it's working out for you joseph sure sure yeah so um yeah first uh yeah i guess the first experience it was a surprise pleasant surprise um being uh

like selected as the author of the first first article so so that that quiz um still it was pretty interesting because in my opinion fizz does a really really good job of um writing these quizzes and it's like uh you know some of them do definitely get you to think because you know sometimes it could look like there's more one more than one answer so then yeah you always have to then follow back of which which answer do you think is is is best uh so even though i i wrote the first one and i knew quite a bit about things already so it also helped with the second one um we learned

learned a lot in the third week with jim's jim's article it was a great way as well to just um you know i yeah pretty sure i probably saw jim jim's article in twitter some some sometime but never got around to to reading it yet so it was a great way to also um yeah actually read it and then uh you know read through the whole thing and then actually yeah see the quiz and do it thanks i i find it the same thing myself i many of these articles i've read but it's the close reading that actually brings you to thinking about how things actually work

together because there's so much to read about ethereum too until you really have a reason to do a close reading at least you know i'm an end user i i i'm not required to use that knowledge for development it's easy to let it wash past but when i have a reason to focus on it i find that i i learned a lot more i'd like to uh to bring up michael michael is uh a developer he's he his business is unvetica and he has he's joined us to help us improve the study master program can you tell us what you have in mind michael yeah just on a real surface level so i'm a devops engineer by day and some

sort of applying those skills here in making it contrary to what blockchain is sort of centralizing some of these features uh for the studymaster program where we have um both the study materials whether they be articles or a single article or website and then package that together with the quiz material in a learning management system so you can learn and study the material and then that will be linked if you will to our discord channel so that any comments or questions that you have as you're studying you can make

within the study material itself and that will copy over to the discord so there's you're sort of always sort of in the discord ecosystem and then you should be able to have at least on our end a more manageable system with the quizzes in the sense that we can track per user um exactly by your discord username the progress that you're making you know we can send little nudges if you've only gotten a certain percentage done if it's the last day you'll kind of get notification automatically you also see sort of scoreboard so it's just a platform overall to better interact and sort of consolidate um

more user-friendly features to sort of further incentivize um the the over or what is it what i'm trying to articulate make it more welcoming i think um for the overall east state study master program so awesome i i'm really happy to have your support i i had the idea for studymaster for several weeks but i never it took me a while to get it off the ground and finally i kicked it out the door and said joseph's article is great let's use it um i didn't i didn't expect a hundred participants the first week and that participation rate is staying strong so

uh i'm excited to think at the end of 10 weeks we're going to have 100 people who have uh a very strong foundational knowledge of ethereum too this right now i'm planning to be my multi my year or two multi-year project um looking forward to sharing it if you want to contribute please let me know uh i i i don't have a strong roadmap for it what i do see is is people's strong interest in learning and i want to support that um really what they're interested in is the poep but i'll tell myself it's the learning um speaking of poep uh patricio is is our

our co-op guru and i want to tell you that today's call does come with a um an east stake or poep you should receive a private message regarding that poep um patricia do you want to tell us about the program yes thank you very much so i think by now all of you got a message on your zoom inbox um the zoom in is is very limited in functionality so you need to make sure that you open the link before the call finishes because after it finishes the chat goes proof and you cannot get it afterwards um that link that you got if you never use pow-up um allows you to get a non-financial token which is a

kind of token we use on ethereum and the tokens prove that prove that you were part of this meeting and and that may sound like something that doesn't have much value or it's something that you shouldn't care about because it it's really not a big deal but once you keep collecting these tokens um they become a collection and that collection can represent things you are invested into like ephesians 2.0 and it might happen and i say it may happen but it's already happening because we are working on this that people holding certain perhaps getting incentives or opportunities that others couldn't so make sure you get the pop of this call if you have any interest on

learning what are the things we are planning or or what are the things you will be able to do with your pow-ups follow us on social media mainly on twitter um we are about to release some medium articles with new things that are doing with poa there are very exciting there's a very exciting roadmap with fact of functionality and features um the roadmap is slightly delayed because we didn't expect to be forced to move to a layer too this early so so now most of most of our development resources are focused on moving to a different chain likely the xi chain which is an evm compatible

they call it a sister chain of ethereum so there are many news coming um feel free to stop by any of our channels um one teaser um like what would be likely the most exciting thing for us is um among the people that has the medasha po up medashapov is a problem you get once you propose the vlog in the medacha testnet um by holding that pop you may be automatically eligible for a raffle where the main price it's a highly powered computer for running an is 2.0

node so so there are monetary incentives for paying attention to this um although it's not the end game there's many more to come but i think for the context of this meeting the call to action is check your private message inbox claim your po up you don't even need to pay for the gas of the minting you can just withhold it until we have the layer 2 solution ready you will get a message about that and and stay tuned for what's to come thank you all right so do we want to move on to the data challenge yeah that's great oh and so i i wanted my uh i had a a hang up a glitch i'm sorry a network

glitch for a minute um but i i'm really enjoying all of the engagement in the uh the chat i it's a yeah sorry for that but that's just like a lot of fun you know when you know i that's exciting stuff this is the internet you know so it's okay well so it's yeah so um laxman is here i don't know uh lamboshi are you yeah so i didn't get my my discord message um so laxman is here to talk about um the uh day of the day challenge uh lakshmi do you want to give us a broad overview and uh let us know what's going on oh he's muted it looks like he's been

bumped uh okay sorry yeah um and i am the hey i know i my connection is a little spotty so i i okay joined um glad to have you yeah yeah thanks for uh well first of all thanks a lot for the time guys um it's uh it's it's really great being able to kind of like reach the staking community pretty directly um so i can talk a little bit about the broad goals of the challenge and so maybe some of the specific details but i also wanted to spend a little bit of time talking about some of the paths that i think you could pursue um there's a lot of information on the flyer and there's also like a whistle stock i'll share all of that on discord

so i don't want to go through everything but um hopefully i'll be enough to kind of like spark some interest or questions um i don't know if this is feasible but if maybe we can take i can take questions and chat i'm i'm worried about uh maybe maybe that the troll no ishness is a little too much of it no it's fine because uh the video record only has our our video uh it doesn't have the chat so if the trolls are going to troll then that's what they'll do if we can still pick up good content there then it's right we can filter through it okay yeah yeah so if you have questions about and chat feel free to ask as i'm talking um can i share my screen i believe you're a co-host if you're

okay let me cool cool cool do you guys see uh see my browser um so here's the here's the flyer which i'll share again um just just really briefly uh our hope is that basically we want more people looking at midasha data um and there's kind of a few different pieces of that one is kind of interesting visualizations one is new analyses and one is kind of new ways to get data from from the test net i realize it's not it's not trivial necessarily to get like network data but if you have ideas around building tools that make it easy for other people to get stuff um that's really valuable

as well um like blog posts or analyses or open source tools that are you know kind of meet the uh criterion kind of like the our quality bar um which which we talked a little bit about in this flyer um are eligible for up to like 15k usd um and so yeah there's some there's some financial incentive as well but i think uh it's it's also just a great way to bootstrap one really important part of the ecosystem um that is still kind of nascent so it's a really good way to be a you know be a contributing member of the community um so i want to talk a little bit about i i've gotten some inbound questions because uh on twitter and on

discord about kind of like ways to start and so i wanted to just kind of share kind of like what's in my head at least around um what you can do um as i mentioned earlier there's kind of three pretty broad areas that are interesting um one is kind of like basically supplements to the visualizations that we get for block explorers like there's many more things we can be looking at in real time one is kind of doing some kind of uh analysis of existing data so like if you go and look at like the kind of the wealth the wealth distribution of these two over some period of time and like produce like the genie coefficient or

something like that those kinds of that kind of information is very useful um and then the last of course as i mentioned just kind of like new tooling um so this is this is the part i think that a lot of people have questions about is like how do you get data like how do i get started um and they're sort of like uh i would say this is kind of like an increasing difficulty like if you if you don't want to get too technical i'd i'd recommend starting maybe just with like block explorers like the block explorers have apis or you can kind of just scrape various pages and some of the things you might get there are you know like uh wealth distribution kind of like how deposits come in over time i think this one's really interesting

like uh what uh what's the conversion rate funnel look like from proposal to inclusion um which you can get a little bit more difficult if you're comfortable running a node you can look at like the you can hit the hd apis of given nodes and i think this is really interesting for looking at like comparing so those apis are mostly just like status data or like syncing data but you can do kind of comparisons between nodes for example pretty easily with that and so that's that's also useful information um for us to look at and then if you get really fancy there are some tools uh which are used i think pretty extensively within like the research community um

for kind of like monitoring network data so like gossip data like directly um in in uh on and giving these to the node or on okay it's like uh like monitoring for potential dos attacks and like gathering some data on that or something like that um and then finally i wanted to just kind of like mention some of some of that some of the things that you could do on the tooling side largely around helping number three happen are you know like scripts and setups and basically data ingestion stuff i have lots of ideas around there so if that's interesting to

you definitely uh definitely dm me um cool i'll share this this short document in discord um but that's all i really had um let's see are there any quest questions or things let's see um [Music] one of the one of the things that i i caught um i have two friends who are basically uh data visualizers but they don't have a background in ethereum at all so they have like a you know maybe they hold a few coins but they have a passing interest um how would you suggest that they that they get into the into this

challenge with with a very heavy focus on visualization and almost no background in ethereum apis yeah that's a good question um i i think probably the easiest way right now is to pull like i think i think it'd be uh actually this is something that i'm i'm going to spin on a little bit i've realized it's like a pretty big request like basically are there canonical data sets that we can create for people to to kind of play with even if they don't want to get the data that are relatively simple to create but in absence of that i think the simplest thing is to somehow

consume data from block explorer ideally they can work with like a like a web api or something like that and just like pull data from that api um that's probably the easiest way to get data right now um but i also i see this as a great onboarding opportunity for uh people who are yeah curious about the challenge but it actually gives them a reason to investigate the apis more so yeah exactly hearing from those people really excited me um i think i think if you're uh in that category would also love to hear what data you want to look at because this is definitely something we

can either work towards internally like kind of making some data available or you know maybe someone else in the community can make some data available and then someone else can work on it and both people can kind of partake in like a bounty around that um so if people have ideas definitely definitely share publicly and uh it feels like it feels like there's like two dots we can connect here if we know where these two dots are but i'm pretty confident we'll figure it out i had suggested earlier um the idea of creating you know either one or multiple like packages of information so things like you know all the different validators all the different attestations they made and then trying to wrap that up into

like csv files or something and actually just uploading it as a torrent so we could all just seat it and distribute that around because i think it's going to be i don't know all all attestations would probably be like into 100 gigabytes right yeah it might be interesting to go ahead oh no i had nothing else though it would be interesting to now that we have a big snapshot of the meltdown the madosha meltdown from a few weeks ago it would be interesting to uh play that back in real time and see uh just a better visual understanding of how the network responded and

uh that might give a lot of greater under a great a greater idea of how things uh worked together even though i know much of the data much of the opportunity lies in imaging present data i think there's also a lot to be learned from uh that that meltdown that that maybe has oh i think that i think that would actually be huge that's a that's a that's definitely something that would be really cool to look into i think we we we all have like a like a cursory understanding of all the things that went wrong but there's probably many other things to study there um historical data see a little 30-second animation of it yeah yeah exactly right just so make it i think things that make more intuitive

things that are happening on the network um are are really really valuable um so so that i can see that being really useful i think historical data may get more challenging the more uh the more especially the more granular you get if you're trying to look at like what what the attestation history was like or the attestation propagation was like at that time in the network but definitely historical stuff is super interesting especially incidents great um are we set with daily challenge are there questions or anything else that anyone wants to bring up are we ready to move on oh someone asked a question

can we connect the data so shredder asked can we connect the data to outside of the test net um [Music] can can you elaborate what you mean by outside of the test net like connect it to data outside of the test net i'm going to give shredder unmute if you're ready okay sure hello hey hey yeah um so the question was kind of um like i was thinking the test net correlation with something like uh defy and theory and price kind of like seeing if there was any correlation

between that so yeah oh that's interesting i just hadn't hadn't thought of that um i know oh ben's here too um i would be interested to see you know when we first launched modasha teku was sub-optimal but then as they've been releasing updates it's they've just been getting better and better for a testational conclusion i think it'd be really cool to see like average attestation inclusion and then annotated with the updates and new releases from tekku team i think that would be fun to see and just see how you know

fast and consistently they've improved yeah that would be cool yeah definitely um i think i think uh seeing things like progress school i think i think uh like correlating with eth price or data outside of the test net could be interesting but i think it's a little less um yeah maybe we can we can talk offline if you have ideas it's not something i would like really want to push on as one of the main areas because i think for the most part we're trying to just understand what's going on with the test net um but let's uh let's catch up about that offline shredder uh it looks like ben also had a question

ben i'm gonna give you unmute oh hey thanks great um yeah i'm just casual casual interest really as a future staker i work with data i do a lot of statistical analysis myself on different kinds of data on day job but um yeah i just interested to see this um call about uh data analytics so but i have very little grasp of what i have no sense of what these two data data structures you know the um kind of data looks like so yeah i mentioned a minute ago about having some kind of package um that would be super helpful because then you just start firing the analysis rather than having to deal with

sort of you know scraping the data and selling anyway it must be anything what sort of uh like functionality would you want on that package set of curiosity oh i just mean um just the data right um just something to work with i would work with statistical analysis packages like r um so you know you can suck in the data but something that would avoid having to get into you know writing scripts to interrogate apis and things which is something that i have very little familiar with familiarity with uh does that make sense yeah so like

extraction and munching basically like extraction into like data types that you can kind of like understand yeah kind of yeah like if someone else does does this deals with that sort of the side of things and just represents the data just so it just makes it a lot more accessible for someone like me to not have to say something something like you mentioned that i thought was maybe a neat idea if if we could get a large dump uh whether they are uh individual uh multi-gig packets but even if it's a hundred gig dump we could all uh create a torrent swarm and share you know a large torrent of data um that that might

accommodate a lot of these solutions something that that would be analyzable in r or importable into a database that that reaches people who are not quite ready to query apis but have a lot of experience managing big data sets yeah actually collecting the data i think is a big roadblock for a lot of people myself included i mean i can you know wrangle excel a little bit but as far as scraping data from apis and all that i i have no idea yeah this is really good feedback um i think this is kind of like what i've been hearing from a lot of people last couple days as well so um yeah i think yeah i'll i'll uh

get back to you on that i want to spin on a little bit i think that there's probably some stuff we can do to make that easier awesome um any outstanding questions regarding uh the madosa data challenge oh is there any money involved oh yeah yeah there's uh yeah yeah so uh you can uh if basically if like you're you're what you present is good enough you you can win up to 15k is that a submissions deadline oh uh october 20th i believe

so six weeks starting oh that's great this past monday is that is that 15k uh the total prize available or is it are they individually judged and each entrant is eligible for 15k individually judged uh i think there's like an amount we sort of budgeted but at the same time like if there are a hundred things that are really worth 15k i think you know well maybe not 100 but if there are you know some number of things that are worth uh 15k uh then then that's something we would consider on a case-by-case basis and so if an entrance was like if their goal was to

speak to your ultimate goal what is your ultimate goal like if if you're saying uh what are you trying to accomplish uh by setting up this this challenge yeah that's a really great question it's it's really to make i think make the make data analysis more accessible and also get more people looking at data so part of it is part of it is this thing that i think we've all uh we've all realized which is just like pretty challenging to get data if you're non-technical so part of it's that start building some ecosystem around that and then part of

it is literally just uh we want smart people and data kind of like thinking about this and sharing their ideas with other people great um all right um that's exciting i i looking forward to the the outcome of this um especially knowing that that each entry will be judged individually because when you have this uh 115k prize you know the the big dog is going to take that and everyone else is like oh well i even bothered right uh so that's fantastic yeah thanks a lot for the time yeah i'm so glad to have you uh and hopefully uh we'll do a follow-up in a couple weeks if this is a six-week program

uh we'll have a lot of time to to see how things are progressing um so aditya are you with us well before we transition um i guess the question here was or a comment is um the data about validated privacy as well our visualizations around that would would be um some something that you'd be also interested in yeah definitely i didn't really mention that but yeah if like things like figuring out like if it's easy to figure out to identify a valid like associated validator id with like an ip or something like that uh if you can figure out ways to do that

uh and you can kind of like demonstrate that it's easy to do that that that's definitely something that it's worth looking at yeah and also if you if we find out that um you know guesses or heuristics on what what infrastructure that they're using things like that and i think this ties in with what is gonna talk next thanks yeah uh thanks again lakshman uh thanks joseph for helping out there um i did you reached out to me uh to ask if there was a spot for secret shared validators and this is a really hot topic right now so um i was going to attempt to explain it but

you know what i'm going to do i'm going to let aditya tell her what it is tell you what it is okay thanks uh let me just begin with explaining briefly the why and the what of secret shared validators so uh there are obvious there are some obvious risks involved with becoming a validator so the biggest two risks are one your validator goes offline maybe your machine goes down you have network issues you have power outages if you're running thousands of validators on the same machine maybe like sabotage whatever happens you don't want that to happen because you're losing out on rewards and you might even get a penalty for being inactive and the second part that you're worried about when you're a staker

is your key being stolen and if your key is stolen then someone else has um you know someone else can make faulty attestations on your behalf and you can potentially get slash so these are kind of the two things that you definitely want to avoid and with the current setup of how validators work yeah your validator is just like one machine sitting somewhere and it's not redundant and it's not secure in terms of keeping your key safe so um in order to address the two problems uh there are some pretty well known solutions for distributed systems the first solution is redundancy so redundancy allows you to tolerate some amount of

outages faults machines going down crashes all those things um and basically what this means is instead of running just one validated client you run multiple instances of it and all of them are doing the same job uh so even if some of them go down you're still fine because the others can survive and you know for example what you can do is uh deploy some validators on like google cloud someone like staking service a someone's taking service b run one on your own and stuff like that so that allows you to get redundancy and tolerate uh outages of your validators the second part which is the most interesting part is uh secret sharing so what this means is

that you take your private key uh of for validating which is like one secret and you break it up into small pieces and uh you assign each piece to one of the validated instances from your redundant setup so let's say you had like five return validators you would break up your validating key into five pieces and assign each piece to one of those validated clients now how secret secret shared keys work is uh the private key is broken up into small pieces and these become like mini pieces of private keys now when you all those validators sign something with those pieces and once you have enough of these pieces you can

reconstruct a signature for the same data so you need to have uh signatures from at least three or three out of five uh pieces of the of the key to combine and make it into a complete signature so yep that's that's the basic idea behind secret shared validators uh allows you to tolerate faults uh prevents outages always keeps your validators online um yep and the basic problem uh in implementing these is you can't just go ahead and you know deploy validators do share your key in secret pieces and so on because what might happen is let's say you have five validated uh five validator instances

and three of them uh are following a different fork and two of them are following a different fork now the problem you're going to have is you can't reconstruct an entire key because well you don't have four out of five pieces to piece them together and create a new create a signature so you want all your validated instances to be following the same fork now this sounds awfully centralized doesn't it because you want them to do to follow exactly the same data see the exact same messages and so on um yep so the way you solve it is uh i'm sure a lot of people might have guessed by now is this is a consensus problem so you run a consensus uh in a consensus algorithm between your validated instances so before your

validator goes ahead and signs something it confirms with the other validator instances that all of us are going to sign the same thing right and once that is confirmed all of them go ahead and sign the same attestation or the same block proposal um you gonna share my screen real quick okay yep uh can you guys see this yes okay so this is the program that i implemented uh this last week uh and what this is going to do is it's going to fetch data from the from a lighthouse client that i'm

running and kind of i'm running seven nodes uh seven uh secret shared validators right now and it's going to form consensus between them about what is uh the hash of the block at uh slot 100. so let's see what happens when i run this uh awful lot of things in the log so you can ignore most of them i'll point out the interesting parts so this if you can shift your attention here um yep so these are uh logs for node six uh out of seven nodes and the first message here says oh node six received a message this is these are not eth2 messages these are messages of

uh the kofta consensus protocol um yep so it received a message and blah blah blah some metadata and the interesting part here is uh the hash of the block at slot 100 is uh this value right here so uh this message was special because after note 6 received this message according to the consensus algorithm uh it it processed the message and it had certainty that oh after seeing this message i am definitely sure that all nodes uh that are running uh this protocol are going to decide on the same value which means all of them agree that uh the the block at slot 100 has this hash

and it sets its state to decide it which means uh yep it's it's done with the consensus protocol and it's ready to sign uh the particular block so that's what um this consensus algorithm does um if you want to play around with it i'll drop a link in zoom chat yep and this is basically a very very uh fundamental setup for secret shared validators and i hope uh this helps normal stickers at home build a more resilient setup um

do you plan to keep maintaining this protocol so it's uh it gets robust enough for for us to really rely on this for our home validators because it is nice it is a nice experiment but i would like to see more barrel testing and more like responsive support than just dropping in on a github oh yeah definitely so this is this is only a proof of concept this is not meant to be used in in an actual validated setup uh of course we like we would need more resources more like minds thinking about this to uh bring it to production uh maybe i'm thinking about maybe reaching out to some client teams and

seeing if they have uh some people to allocate to this uh it's not a especially hard thing to do so i i don't think it won't happen it's just that it'll need a push from the community to make it happen so so what are the next steps you are you're dropping is for us to test it and what would you like to see so right now it's like secret sharing etc is hard to set up in the validator client themselves uh i'm not aware of clients which currently support it i'm dropping it here so that anyone interested can join me in an effort to build secret shared validators okay i'm getting it um this is a good place to do so because

um i feel like many of us here in the stages community would like to see a product based on this concept so we can push the agenda forward with client teams and with the community at large so so we kind of expressed that this is worth allocating resources to i'm curious if if it introduces any latency is uh so especially if it's widely geographically spaced or on slower networks uh does it increase the likelihood of missing an attestation no i mean yes obviously it it introduces some overhead but uh the this protocol is

actually quite fast um in the sense what happens is there is a leader validator which says hey guys our slot is coming up uh what do you want to sign and you know they run through with the protocol and for whatever reason if the leader fails there is a leader failover that happens like immediately and you elect a new leader and like yeah you you're right back at it so in that sense this is probably one of the best uh overheads that you can get for any secret shared setup um yep can you i had a quick question oh god um can you compare this to a testance dirk software because they sound

very similar is that i'm not sure what dirk is uh the last time i glanced at it yesterday i think it's a key management distributed key management thing i'm not sure if it's uh secret shared validators though okay all right it would be interesting about breaking wallets up but love to look into it more yep thanks for the heads up i'll check it out uh who did i interrupt oh those that's me um okay sorry that's not good i was just i was just

curious like what in the ideal productization of this like uh what does it actually look like would you would you expect people to kind of be running distributed nodes that share a single secret for a single validator um and just kind of like because they're listening to different parts of the network there's like i don't know like can you talk a little bit about what it would look like in production right so there's there's so many different layouts for um implementing a let's let's call it a resilient validator it doesn't have to be secret shared if you only want redundancy but you don't want

uh you know to split your key you can do that you can you can kind of run lots of beacon nodes and only one validator client and you can run this consensus algorithm between the beacon nodes um just so that uh you know you're you're you can tolerate beacon node faults uh beacon not going down begin not being disconnected from the network uh beacon or existing in a partition and so on got it just out of curiosity um how i know it's a proof of concept how how ready is it to implement for uh

a mere mortal right um so i guess um so this repository that i uploaded it has actually nothing to do with ethereum it's just a plain old consensus protocol uh so for people who are um kind of uh aware of distributed systems and kind of are okay with concurrency and programming in concurrency this should be fairly easy it shouldn't take more than like two weeks or something uh but taking this to production for secret shared validators is a challenge because we need to come up with a solid specification so that validator clients support it uh beacon nodes uh work how we expect it

to uh making the spec is definitely uh the biggest challenge there are no research problems it's all implementation now great um i'm actually looking forward to checking this out this week so um we'll see where that goes uh callum had a question um he wonders if the cost for multiple modes of redundancy would cost more than the amount of loss on a single node with expected downtime with a single validator um and i think that's that's more of a general question yep so this definitely depends on what your threat model is so for example let's say you're you are hosting your own sticking as a service uh business and you're like allowing people to come in

and host their validators on your setup and you have like uh you know hundreds of or thousands of validators running then you definitely want to be resilient you want to be you want to have redundancy and you want to have security against keys being stolen so for in in that case yes um it is cost efficient to do a resilient validated setup but if it's just you with one validator and you don't you know you have a really high grade internet connection and you live in a place where like power outages aren't common then it may not be cost efficient yeah and i would say also also uh that's exactly my thought it's also the number

of validators in general that you're running because the overhead for one the risk of running additional software on top of that one validator is probably greater than the reward from what you're saving by those by keeping that validator going right and and and with as with all consensus protocols uh it really depends on what your threat model is uh on what your expected downtime or expected crashes or security risks are and so like coming up with a highly configurable specification for this protocol

is definitely the goal great yeah thank you so much for your time aditya um and so that's great i'm i'm so happy we've we've i've learned about two new products today while so i knew about the madasha data challenge but i i learned a whole lot more about how it works um and we all learned more about kafta today so that's an exciting product to try out um we talked a little bit about yes thank you so much we talked about uh the eth2 study master and our future plans for that looking forward to people earning uh a poep

for investing in uh their education um we welcomed our our new team member uh michael who is uh working with unvetica to support um support us hopefully everyone got a poe app if you if you lost your poep link today let me give you patricio's private number and you can just call him uh he's always available by phone and he loves to to catch those calls um that was a joke is there anything anyone else would like to put out today i guess i'll just clarify since uh

ben gave us a good answer dirk distributes the validator key whereas this would have but it only connects to one validator so you don't get the redundancy there but this kafta gives you multiple validators for more redundancy so that's the difference between those two different pieces of software what one um thing to mention about secret shared validators and it's more for um relevant if a lot of them users are using um you know staking services is that uh yeah a dream is to be able to be able to use more than one sticking

service that a single staking service can't um just sign your validators um just by themselves maybe someone else can rephrase that thanks i thought that was really interesting yeah i'd love to hear that expanded on but did you can you explain on that yep so um so let's say you like want to run five bar you want to run a validator and you don't have a machine to run it at home and you want to go to a sticking service which is i don't know custodial may be and run your validator there the obvious risk there is if the staking service doesn't implement their setup properly

or steals your key then you're kind of screwed um so the ideal setup here is kind of you utilize multiple staking providers like provider a b c and so on um to have redundancy and also you don't want them to like any one of them to be able to run off with your key so you secret share your key and then you make a redundant set which is exactly what secret shared validators are all about so if all of them support the same specification for a secret shared protocol then you're all good you just distribute your secret shared key among like five uh staking providers and between themselves they'll run this protocol for uh agreeing on what to sign and uh yeah your validator is basically

uh voldemort and uh you've distributed horcrux everywhere nolan and i are going to set this up together now [Laughter] we've got it all worked out that's awesome yeah that that is a i i i my mind didn't get there yet so that's that's a great a great way to look at it um i'm six times more excited about it now than i was just a moment ago um great any other closing thoughts okay um i'd love to thank everyone here this this is actually i have to say been maybe one of my favorite calls um because we we focused on some newer news

things which was just really really fun so i hope you'll mark your calendar we're going to try to get back here in about two weeks if you're listening to this call and you've got something you'd like to share please reach out to us we we love new developments and new ideas we we put a little bit of distance with staking as a service providers but that that's not to exclude them just to be cautious uh so you're still welcome to to share your products with us well thank you very much everyone see you next time and

stop by to our discord yes please uh don't forget to check out the eats to eat staker discord as well as reddit.com r eatstaker we work really hard to welcome new users and help you find out where you fit best if you want to make a contribution that is time labor to our uh our setup then we would be glad to find a place to plug you in so thanks again for your time everyone all right thank you very cool thanks team bye y'all