ETH2 Staking Community Call #4 - Lighthouse

Transcript

thanks ru it's like 6:00 a.m. in your side I know it's 8 a.m. so it's very reasonable that's good how about for you it's 7 p.m. in bonus ideas it's been a nice day okay I think everyone's getting in here let's go ahead and get us started let me click a couple buttons this is the 4th ëthere iam Staker community call so I'm sorry I'm clicking and talking I'd like to welcome everyone for hanging out with us hopefully we'll get to involve a lot of the community and I'm glad to see so

many people are showing up I am super fizz with me is lambo she naka Guinea how you say it I've never said that yeah that's correct okay Paul how nur who is with Sigma prime White House patricio Worth officer who runs po app and anyone else that that you kind of know us let us know we'll give you give you an introduction so today we want to talk about lighthouse actually one more task pardon me for we're going to talk about lighthouse but to begin with let's

see I want to do introductions first and then talk about po app and then we'll get into talking with Paul so if you have questions feel free to write them in the chat and we'll go from there so again I'm super fizz I am a moderator at reddit.com slash are slash eats taker our goal is to promote the etherium staking community and that just means giving everyone a better awareness of the tools and the process for staking on aetherium - which is not yet live but will hopefully be live sometime by the end of this year so limbo see if you want to go ahead and book I'm working to set up this call and just

try to get people comfortable running their own nodes instead of shipping all their money over to coinbase in finance not that there's anything wrong with that but you know we want to do it ourselves decentralized and hey I'm Paul I'm a co-founder and director of Sigma Prime I've been a software developer for about 10 years and I was working on Kasper CBC for a while which I still love dearly and then started working on f2 just before it was announced publicly back in I think 2018 so I've been on earth to kind of since its inception before there was even an

official spec and now I'm one of the lead developers on light house pushing through the f2 production cool cool thank you and petruchio hello everyone I ramped up for those who don't who don't know um Bobby Xu non-factual tokens to people that attend conferences community calls or whatever that has a point in time so we want the port's to be integrated on governments in ecdm too so for example there are going to be many different variables to said when when whether how many folks do you want between two paths between something or

how many buddy there it is it's reasonable to our issuance or whatever and if we put a co-op on the hands of every person involved in the community it would be way easier and accurate together community sentiment and opinions and and that's really lacking because governing if idiom one has been a big pain and you don't really know what the community feels about power not or how do they like to do issuance or to discuss issuance of what gets on different Hart Forks and what to put in and what not and what the cost should be of every opcode so my objective here is to get power-ups in the hands of all of you even the speaker

and moderators and hopefully soon we will be launching some studying pulse I like what time do you want the call to happen and you can only vote if you have the power up so one that's one of the use cases we are also working on the East a key its takers discard for having private channels for people that holds power ups so for example if you need think it has support because you're know doesn't sing or it doesn't sing in the way you want it to sing and you are a community community member instead of having to put the post already than wait for someone to answer you could have access to a community of dedicated people so please make sure you all get the power ups I think they're going to be distributed in time soon if you don't

get for whatever reason reach out to us and we will get it the new world if you have any suggestion or you want to learn something about Pop feel free to reach out to me and please and show you the power there the call and the guest of honor we have today great thank you so much for Tracy oh I know a lot of people are excited about po app and so I received mine on the PM from Lambeau XI if you didn't again just ping him and say hey I missed one I'm working on it would you're doing great yeah look at those in the hands of everyone today so let's get on with the agenda if you if you haven't seen the agenda yet you can view it on my github

github.com slash super fees slash eath Staker it's under the agendas basically we just want to talk a little about White House we met with Paul yesterday and talked about some of the the big topics and so we're really excited to hit those today Paul do you want to tell us about White House or actually Sigma Prime and lighthouse what your your preference there yeah sure I can do a little intro to Sigma and then go into my house if you like so sigma prime was formed by myself and three others back in 2016 actually just after DEFCON 2 in shanghai so we have been working if you're in for

a while they're local meetups and we've given been getting a lot of different people that wanted us to work so we thought we'd found seeing a prime kind of on the side and see how it goes and we really found our footing in security reviews we have a pretty strong information security background so we started doing security reviews and then eventually you know we all quit our day jobs I ended up with Sigma prime so it's been I think three years now coming on three three or four years and Sigma prime was founded and it's doing really great so we had been doing security reviews and penetration testing which is like where you hired by a company to assess the security of their infrastructure by attacking it and then

reporting back on what you can find so we've been doing these type of things and eventually I got I had always been interested in proof of stake for a few reasons I quite like the security properties of it and I'm also a big fan of the reduced energy consumption so I started working on the cattle a CBC and then eventually we started then there was the shouting shouting manager contract which was kind of like wilds working on CBC then the shouting manager contract picked up and I've been working with Danny Ryan on CBC and he started doing the shouting manager contract so I we will Sigma probably put in a application to build a validated client for the shouting shouting contract which was where you have like a contract on f1

that controls the f2 shards but it became pretty obvious that we couldn't really get the bandwidth we needed for like a fully sharded proper proof of stake etherion by that so aetherium 2 was a started becaming like became in the works I guess bimetallic he had this hack MD document and that was like the the spec back then which was old was no code it was just all written it's quite difficult to understand and so I picked that up and our previous grant application for the sharding manager contract or deprecated because f2 was coming to them we kind of like just rolled with the punches and started implementing that specification for metallic and we were I'm not sure if we

were the first or after the pipe implementation or what prism were up to then but it was definitely also prison that had one of the first implementation so we we've been in there since the beginning since then lighthouse has really kind of evolved from this side project of me trying to figure out the spec by writing it down in code through to a real production client and thanks to the support of the theorem foundation of consensus and italic and all these people who have believed in my house and confounded it so my house as it stands now is very much intending to be one of the first clients of production I'm quite confident we will be we have a focus because about background security

we have a focus on building secure applications so we want to build a client that is correct it adheres to the specification that it's stable and secure that you have you know a good uptime and and people can't create crafts messages or attacks that can can try and bring us down and a part of this we're also looking at performance where written in rust which is a modern systems programming language assistance programming is like the backend things so it's a language we're a little bit difficult to understand but it gives you a lot of control over the the hardware that you're running on and it also means

that which means we can write quite fast software so a part of being secure means also being performant so that if people can craft malicious messages we can process them quickly and we don't get bogged down in the denial of service attack so we're focusing on being fast and performant it's something else were also performing focusing on now is staking so these two properties kind of work pretty well for a staking client if it's if it's fast it means it has low resource consumption which means that your overheads taking a quite low and if you're secure incorrect then it would mean that you're less likely to go down and have downtime so because of this now where we've started another project

which is to build a user interface for lighthouse so where systems programmers mainly the core team of lighthouse so we've reached out to a product design mikheil from Empire who's building out who's designing you get to see some screenshots later actually the first time we've released them he's building he's designing a user interface for staking I'm designed to be like very entry level kind of hold your hand through the whole process and then we're also engaging a front-end design team to build this so it'll be kind of like a web first application that connects the evaluator client allows you to start/stop control and see evaluators and I think I think that's that's a lot

of it anything I missed or supposed to be too much I think it was it was a lot it was good it's so I'm still learning to do this part of it and it's hard to process all of it but yes that's stored a little bit too much for my part and no no I hope you'll join me here I bounced in and out of consciousness and I'm like ah yeah so thank you so long answer

short I didn't process what you said but I'm sure it's great yeah I'm gonna I'm gonna chillax a little bit hey I would like to maybe go ahead and unmute the room as we go through this is that something that works for you at this point so you either public yeah the UI demo that's right no one if you want to share your screen and hop on the UI demo then we will open it up after that one moment so UI perhaps a little bit advance so what

we're doing so what we've done is I've asked Akhil who's doing the user interface design here to kind of share some of the wireframes he's been doing so he has some videos and stuff like this but he wasn't quite ready to share those so what he has made of some teasers and to demonstrate the processes that we've been going through in designing this user interface so it's really not a we're not we're not hacking this thing together so we started with a research phase and we went out and interviewed I think on the order of a hundred people to understand what their needs and wants and desires were for staking hopefully some of the people in the school managed to get some input into it and now we're in the process of wire framing as you can see we've got

quite a lot of wireframes down and then we're going to move into development so the UI is kind of split into two phases we had the onboarding phase and then we have the dashboarding phase so what you can see here is the the user journey flow for yeah it is a bit small but the resolution is not great I think the idea is that it was like to kind of just show the general process as opposed to the detail hopefully we can share some more detail in a little bit yeah so we're putting this is this this kind of cakes through the whole key generation phase so we want to like walk people's hand all the way through from you know like here you are you've download the application it's running now this is how your keys work this is where you're

going to put them this is what you're going to be required to do here's some links to read some more so by really really really trying to handhold through this process so there's I think there's another one yes so where where I'm gonna be tracking you know like really trying to hold people's hand and show them that you you know your nodes are syncing and avoid allowing them to start doing actions before before you're ready like if you if you deposit if you submit your you're staking deposit to f1 before you bacon notes thinking there's this kind of minor chance that your bacon know doesn't finish syncing in time and you

start losing money so we're really trying to avoid these this type of things we're also looking at really describing what what's involved for people really want to lower the barrier to entry so people can can really understand what's required of them before they commit to dropping their money in I think the next one is a health check to so we're gonna add in a whole bunch of monitoring to try and keep an eye on the system that evaluator clients running on you can get these kind of in browser pop-ups so if you know you ran slow your disc is low you're getting errors things like this and then eventually we can link this into other services like pager Duty where we can start pushing out

notifications via SMS and email it's if you value the clients having having troubles so it's not a whole lot to see but hopefully it's a little bit of an idea the things that are coming for her bath - in lighthouse awesome looks really good to me I've said it before I love a good UI so you know yeah yeah I think makes it much more user a lot of people are came to it before we stole from the survey was a really a lot of people want you eyes even very technical like the researchers building the specification actually one of you I just because they wanted to UM they wanted to be eating that coin core had a UI and

windows implementation like pretty early on I think I store easily it was originally Windows the first Bitcoin client was written for Windows huh there's a fun fact okay so up do we want to start unmuting some people see yes earlier no no not yet can you share with us pull with them in a march spec what hardware spec would be yeah sure so minimum hardware spec is so high spec is a little bit difficult at this time so there's kind of a known hardware

specification like known requirements that we have now to run the the beacon chain in ideal conditions so there's so beacon change in ideal conditions like smooth seas so to speak uses some degree so some amount of resources right but then in rough seas we're going to be using more resources to process you know bad blocks and dealing with faults and things like this so we have we have currently have a reasonable idea about what what it looks like in smooth sailing and we need to do some more testing once we have the big multi-client test sets up we need to start doing some coordinated attacks and pushing clients up to their limits to see you know how much how much how many

resources they need because kind of want to avoid the case where you buy the like you set up your system so that you like you're just like kind of a just under like it's smooth sailing you're just under the limit and then if someone can craft some attack she pushes you over the limit and you go down so with that kind of caveat in mind we have so our latest we have our latest six like sixteen thousand validated test net so something to remember is that the resource consumption is very much dictated by the validate account so we expect to launch f2 main net with at least 16,000 validators it's not clear at the moment how many will show up but we're thinking it might be on the order

of 30,000 to 100,000 showing up so without 16,000 validator test nets we have we have some running now that are using less than 300 Meg of RAM so that's a beacon node and a validator client running 2000 validators and those two together using a bit less than 300 Meg around but we do have some cases where we're seeing higher usages up in the order of 800 so we can kind of get this this idea of the range but we really expect that number to go down as we keep focusing on producing memory consumption and in terms of see if you use I think you want something with at

least two CPUs just because this allows your system to kind of like still handle Network messages while it's doing some heavy processing so I'd be saying so all of our test nets so far have all run on an AWS t 2 medium instance so that's 2 CPUs 4 gig of ram and that's been plenty for it so far we did try to run $100,000 to test that on that but we started to run I ran because we're with yes but our latest numbers for 100,000 validators is looking at about 1 to 2 gig of ram so perhaps that gives you a bit of an idea this is a pretty pretty small weight and this is something that we're really really

focusing on is reducing the memory consumption because that's probably the biggest overhead the stake is sorry no yeah you described the rough sea scenario so what is what happens in a rough sea scenario does that mean that that the local machine stops providing attestation while it chews through the data or what is that yeah it means that so they're kind of like the first thing that you'll start to see happen is you'll start to fold behind the head so like the head is that your the latest block in the canonical chain so if you're kind of lagging behind in resources you won't just be able to keep up with the head so you'll start

attesting to blocks or a little bit earlier than the head and this means that I think you'll still be able to get at test stations out because producing in attestation is a very light on past but processing and verifying inter stations is where it's heavy so you'll still be able to get your attestations out but there'll be attesting to things that are slightly older than they should be which starts to reduce your awards not to zero but you'll start to be missing out perhaps on maybe a third of your attestation awards and you start to be missing out on is that is that a scenario that would lead to missing a window to propose a block because you haven't you haven't brought your not caught up oh and then you'd you can't go ahead I think yeah

that's right yeah so you'd start to put you be producing a block off instead of producing off the head you'll be producing it off the heads parent so one behind I mean you would have you've created walk in and network then and then it's kind of a bit of a chance that's to see who finds your block or the next block is the head so this is a case yeah where you start mixing blocks - I think missing blocks is a walk so likely to happen every few days so I am I'm mostly concerned about the attestations because they're constantly ongoing but you definitely don't we missing blocks either yeah great thank you so much all right I'm gonna probably summarize that make a reddit post so it shows up on Google okay great

let's unmute some people actually earlier Joseph C has a question I'm gonna go and here I believe already has in fact I might let you manage let's let's for the moment give everyone access to ask questions as they see fit will open up the conversation and if we need to go from there we will my god she secret shared validator client do you

plan for that or secret sharing values that's not something that you know so you mean like a multi-party computation kind of validator yeah yeah yeah that's right and I wrote it's basically you know in case some finance offers free speaking you know if there's like you know if you also want them to just not have unilateral control of the validator but we also want to say Sigma prime and prismatic collapsed also at least that I

had a second signature to make sure that it's you know what they're proposing is you know not attacking network yeah yeah that's good yeah I think it's a good idea I know the research team has been designing the system in a way that that's possible it's not something that we have focused on the moment just because to kind of have our hands full with the minimum requirements and that's a little bit advanced but definitely something became to get into it and that's the sort of complicated interesting stuff that Sigma Prime likes to get involved in so I think it's probably a future consideration at this point but I don't have much I could I could tell you in the way I've been I'm sorry no problem thanks Pablo earlier in

the meeting asked a good question and he was wondering what's the best thing people can do to assist each to developers well it's a good question it's it's difficult I think I'm one of the things that'll be handy so when we get the the new UI out is to have people beta test and give us feedback on it it's great when people try and run tests that's share with us the things that are difficult with them and share with share the the the bugs that they have in good descriptions these are all very useful and I think something that people can do too is just to learn things and then

share share the information online because you know we spend a bit of time answering people's questions and sometimes their arm they repeat questions so people that are out there you know like understanding things and writing blog posts they helping people get get across it and kind of reducing a bit of load on on developers and researchers and stuff for answering questions so some of the way that I can think of offhand if you if you're if you're a rust developer then you can definitely contribute but there's a whole lot of orgo developers or NIMH developers anybody you can jump on and you know start closing issues too that's another way is the discord your primary social contact yeah that's right yeah if you want to get hold of

goodbye that is good as far as questions question helping and providing support with is that also through discord yeah it's really definitely through discord a lot of our team is based in Sydney Australia so our times are a little bit odd compared to everybody else but we also don't we're not nine to fivers and so our times are also odd compared to Sydney so just be aware if you don't if you don't get an answer straight away it might just be a time zone thing this is one of the same questions we talked about yesterday failover node failover when so you're running your validator

keys on on one station with a beacon chain is there is there a planned and I know the answer to this but I want to give you a chance to say it again a plan to allow failover to a second beacon beacon chain node yeah sure so there's two there's two things in this question so there's two ways you can failover one way you can failover is you have one validated client and then you have multiple beacon nodes so you know just one value decline is connected to say the first beacon node and they're talking together and then for some reason it detects at this speaker node it's not giving it a good quality of service so then it's swaps over to the next week you know this is something that's quite straightforward and we haven't implemented it yet just because we kind of want to like nail

down the base cases before we start swapping around things like this but that's that's quite straightforward and by design our value declined I think everyone else was validated a client is designed to be able to handle that but what a lot of people I've noticed thinking about doing is the different type of failover where instead of one validated client multiple beacon knows you have kind of multiple validated clients each running the same keys and then you kind of like have one that's you know in the background off ones running then if this goes down you have like an automated failover script to swap between the two of them so this is a different kind of failover system and I think this is the one that's kind of worth considering a bit because the risk

that you run when you have these two when you have to the same key in two different pieces of software to different machines is this is the perfect scenario for slashing so if if you just have one valuator client we have some some typical we have we have some we have software like protecting against slashing and it's as long as there's no serious bugs in that code it's impossible to be slashed but when you have these two disparate systems and you can never rely on these kind of third-party ways to like try and swap your keys over like with your own scripts this is where you'd like it definitely running the risk of slashing this is like textbook case of how people will get slashed I think so now we

understand that if you have these two validated clients now we have this swapping between the two and it's taught me between two of them is risky and the risk is on the order of losing you know half your stake in being ejected as a validator so that's that's quite a large risk if that scenario goes wrong so then we need to look at by having these two validators what scenario are we protecting against what's our risk versus reward here so our reward is that if we do manage to successfully swap over between values we've reduced some downtime so then we say okay what's downtime worth in f2 so I have a little stack overflow maybe I can share it with the latest I don't you know get my mechanical keyboard making much noise but um when you're offline as an f2 validator if you're offline now this

comes with a big caveat because there's rewards and penalties in f2 and not straightforward it's not like proof of work like you produced a block here's you know here's your F goodbye in f2 the rewards and penalties like there's divisors and things are going on and it's like the amount you get is relative to how much there is but if we take say the sandy case we have 16,000 validators at Genesis and the network's kind of chugging along as normal if you go offline for a week you stand to lose 0.3% of your stake so then so then we go back to looking at this scenario where we've got these two validated clients and we're now we're hot swapping between the two of them and what we're trying to do so if if we didn't have hot swapping

and one went down for a week and came back up we've lost 0.3% of our state which is really quite small so then we start to think that you know like if you just profit focused you start to think well is it even worth it for me to have this this slideshow everything did this swap over thing because like I can be away on holidays for a week and and then come back and then I've only lost a fairly small amount of money so there's definitely a case there's definitely an argument saying if you're if you're a small stake a hobby its taker then the failover system is probably a little bit more risk than its then it's worth but then you have on the other hand you have

just kind of we have another problem on your other hands where you say so everyone's a hobbyist a guy everyone decides it's not worth and not worth it for them to um to set up this like redundant system so then everyone goes away on you know like whatever type or like holidays end-of-the-year holidays yeah people go away on holidays and then someone attacks the network now everyone's on holidays networks down no one's got failover the network's just kaput so we have this problem where you know like we need some people to make sure that they have these sophisticated systems so my feeling would be that if you are a sophisticated stager perhaps you're probably more likely to have more

stake on the line and perhaps you have some pretty serious engineering resources behind you then I think it would be smart for you from a profit perspective because you have more stake to lose to run these these failover systems and I think that you would probably be more likely to be successful at the failover systems then maybe someone who's just a hobbyist and is likely to maybe make a mistake it's difficult because I'm wondering to myself am I a hobbyist or do I have a real stake in it like I don't it's difficult to guess you know is that is that someone with you know three validators someone with 330 validators

someone with 3000 validators and I know there's the answer is different for everyone it depends I guess on your your actual worth but it's you know I'm like well I don't know where I fall yeah I guess maybe you look at you say like okay I have said three stake is 0.3 percent is you know how much money and then do I know how to build this do I have a confidence in myself to build this failover system and if you can afford to lose the money and you don't really have the confidence in yourself and you definitely be in let's not fall the case I would guess that's that's a great calculation to think about what what dollar value point three percent of your hold would be because that's that's the balance of whether or not it's worth

getting that either hiring it out or doing it one thing that that keeps coming up and I love it back to hardware and I know you know we don't we don't have the full client yet so it's still difficult for hardware but we see the range from things like raspberry pies if you would address raspberry pie specifically and then something like AWS in red and let's ask that separately can we just begin can you can you show your perspective on staking from raspberry pie I know that a lot of people want to that was one of the things from our survey was that people really really really want to stake with raspberry PI's and I think because that sheet but I think there's also I don't know there's

something the raw surprise are cute they're fun and the idea of being able to do this decentralized blockchain on this like little open source computer you found I think is appealing to people and so people definitely want to do it and we want to make people happy we test I have applied for around here somewhere we've been testing on it we really we really want to be able to support it it's looking good at this stage but we do need to go into scenarios where we start you know doing malicious testing of these relative prices like before I would like to say yes hundred percent you can stake on around three five so that is like is that the final answer yes it's a it's a firm maybe and pushing

so I if I had to bait if you made me bet on it now I would say would say that we're gonna be able to stake on the other three pies but I wouldn't I'm not ready to make the call yet until we've seen multiply and test that with sophisticated malicious actors on it so it's something that we haven't seen on existing public text that's now is sophisticated malicious actors that's what we're gonna be scattered and malicious so there's a follow up to that and you mentioned the minimum AWS requirements do you one of the things that I talk a lot about and I'm not even really sure if I'm on the right path is I discourage people from planning to take on AWS because of the outage

problem if if AWS went down which we have to kind of consider to be a real possibility then that means if everyone's on AWS it's a large swath of validators down and massive leakage do you have at least that's the way that I've viewed it do you have a perspective on whether AWS or any kind of VPS is is a good idea or whether people should plan to stake from their home or or local area yeah I think it depends I mean so not everyone is going to be able to stake from home I know like myself

I've lived in rental properties for a while and you don't necessarily get the internet you want you don't have the control the ability to control the internet you get or the reliability if you're in share houses and things like this just on University campuses it's quite possible you're just not going to be able to stake at home at all and I think that's taking on a cloud platform is acceptable I think we really we definitely want to have a mix so we don't want everyone like everyone helped me on AWS like you said it's a bad thing but having some people in AWS having having other people's on you know like Google cloud platform other things like this I think is it is a reasonable idea like I know myself um I would I'm able

to stay from home now because I'm kind of permanently in this house but in the past I definitely would have staked on AWS um just something interesting is that so we do a lot of testing on AWS and one of the reasons we chose AIDS because prison does all their stuff on Google cloud so we're trying to already trying to get the teams to just spread out across the cloud platforms right that was a perfect answer thank you all right let me see we got and I'm sorry this is something I don't actually understand i plan to

deploy lighthouse via infrastructure as code planted oh yeah so can you explain that I'm sorry I'm just not aware right it's just a question about DevOps um do we well actually I'm going to show you there so it's a question about whether we saw about DevOps yeah okay it's like so we how we run our test nets at lighthouse is by ansible so ansible is made by Red Hat they make a distribution of X and ansible is a way to create AWS servers deploy software yeah DevOps Automation deploy software to them and then manage them so we have an open source repository so all of the all of

the DevOps stuff we use to manage our test Nets is open source at Sig p /s - danceable there's a pull request that has the main has the main like master at the moment isn't what we're using it is a pull request that has what we're using so we we can so our lighthouse test nets are to commit so we go from from nothing like no AWS systems nothing exists and then we run one command to deploy the deposit contract and then we run another command to deploy all of the infrastructure across like n machine we have sixteen machines four reasons at

the moment and then we have another script which installs white house on them installs everything on them puts sensor configuration to them connects the boot nodes and then starts and so we have like a fair bit of DevOps automation going on the lighthouse now and our plan is to eventually turn these into a set of kind of reusable Hansel modules so people that are familiar with automating their DevOps stuff can just pick up the code that we've written and go forward and deploy their own service with ease excuse me okay Petruchio did you have a question on the

chat so one that I've been asked in the chat but I still wanna notes um early on when there were going to be 8,000 refurbs charts I heard that some taker may have a slight advantage if he could manage to get a huge yes sir buddy they don't have this chart I'm not sure if they I'm not sure on the attack of getting evaluated on each shard I know that if you can get a validator on if you can get like a large amount of validators on one shard then you can start to grease the shot you can you know get it to you know you can stop it

from from following ahead you can get it to choose Forks and in the worst case you can get it to decide on some chunk block so I think that that was the attack that I know of was that because you know you kind of take the whole value set you split into one hundred thousand and twenty four and if you can manage to be a majority of one of those small committees and you can cause a lot of grief there but it's not current anymore what's a current number now this ya 64 yeah and then it kind of it it grows and shrinks with the evaluate expand but I think 64 is what we can

expect to see when we go to Maine then she shrinks we'd buddy it'll come to ISA on this peg the doctor gross like programmatically yeah it is in the spec but it's not written in a way that says oh you know it's not written clearly it's very deep inside some Python some code so yeah it's it's it's the spec itself is it's quite difficult to read and determine these type of things from it hopefully as people go on they can start to write these more human-friendly versions of it basically the specs thing changing or what - so it's been hard stable do that but now it's become fun because people

writing human readable things that's one of the cases that I think having something they are both difficult because I don't think Cortland developers really have a perfect grasp on the actual people using the network maybe you know because your company does not support on is one but sometimes I feel like people that have dubs on main net the really happy voice on the choices that are made for it - yeah I think that's fair I mean the it's a very technical discussion that happens around

the protocol and the barrier to entry is quite significant for it I know that when I first started I was very scared of asking questions because I felt like I might feel I might be asking basic questions or making myself of like an idiot I'm spent in perhaps this is like something people are concerned about as well yeah I agree I mean we've seen some great stuff I read it last Def Con in Japan we had some panels we got to see some really interesting stuff from from deaf developers I guess I don't want to I don't I definitely I want to put the blame on them but I would say to them that if they do feel like getting involved and I should just actually just start asking questions because there's like it's a cliche but there's there's no dumb questions and people in have to

understand it's highly technical and we're working to hear from people so yes that's easy to say but if somebody said to execute because there's no single place for someone to ask questions like if I made a developer which I am and I want to ask something about these two my oceans are either go to one of the client developers which might not be the accurate place or go to there is to research is called what I may feel that I'm kind of an idiot the Stack Overflow that you were it's not Paul what you're talking about the eath to Stack Overflow seems like the ups seems like the right place for that yeah so I've been writing like I checked the Stack Overflow every couple of days I answer any questions that come up so I'm trying to help there

yeah I agree that I think that F two Rd is probably the pace the place for it but I know it's definitely easily said but you really don't need to be afraid of looking like an idiot in that because you know everyone understands that it's not it's not simple to get your head across the spec we have another question by Joseph C Joseph is asking everyone you know this is for everyone mainly not for Paul okay thanks yeah I would say I wouldn't be willing to do that it just I want to do it myself it's decentralize Joseph do you wanna go ahead and explain your question and then

we can we can banter about it for a moment oh sure sure yeah yeah I know that everyone here is mostly do-it-yourself so definitely if everyone says no that's the sentiment of this group but basically if if there are companies that that offered on staking without fees but it's also secure for the network because you know you use things like secret shared validator clients you know would you be comfortable you know giving those you know like one company a share of your validator private key and then you know another company and another share of it or even multiple even more it just seems to me like it's one of the easier

you know even easiest ways to to run things aside from a yes or no and I'm it's not my place to answer this question directly but I worry that what appears to us to be two companies upstream is really one company with two faces or whether you know company a B and C are all subsidiaries of a different company and you're giving each of them a part of your key in reality giving them full control over the key okay yeah well the the ones that I was thinking were things

like yeah prismatic absent maybe sigma primers but if they if they do something like that oh you know at the beginning they'll they'll at least be independent I think from by Nance let's finance should be that I shouldn't joke because someone will take it seriously some will drink the bleach but say finance should be the the safest place to stake your etherium to you let's send it all there yeah Thank You Joseph for that that question sure sure just to UM your if I said his name wrong sorry - and he's comment here about the withdrawal address is yours would it matter um yeah

so my question is more about like the the Do It Yourself because even if it was I like easy then the do-it-yourself people will still want to do it for themselves but yeah you can keep your withdrawal key to yourself you know and just give the validator share sort of about a private key to others and to me it seems like pretty good you know the benefit and trade-off Thanks patricio has another another questions I is it is a more of a nice dev question but it I'm interested to hear a lot of different

perspectives about how the migration of each one onto the f2 chain shorter might occur do you have thoughts on that oh yeah so I this is something that I don't forget particularly involved in at the moment is the migration thing I'm very much head down delivering phase zero but I do keep an eye across it affects the project in the medium and long term I like the current philosophy of trying to get f1 into f2 as soon as possible and taking what we have is f1 trying to get a purpose take as soon as possible I

quite like that there's some talk about perhaps the existing f1 clients being reused in connecting to the the current beacon nodes to get the economical sense of what the beacon chain is but then keeping all of the execution and transaction processing inside f1 because we don't want to do rework and they're pretty good at it so this is appealing for me I don't know if it's just appealing to me because it means someone else has to do it but even without that I think this is I think it's appealing because it's fast and it means that we don't reinvent the wheel and we keep using battle-hardened software instead

of writing new ones so I I quite like that idea but this that stage of the specification was like definitely subject to change yeah I in fact it's funny I don't even I don't now similar you are I'm so focused on beacon chain right now that I haven't really looked at the spec for phase 1 and phase 2 and probably over a month like and of course you know I'm an end user but I guess you're probably similar you're so focused on phase zero implementation that those that it doesn't really matter right now for you what's going on in phase one and phase two is that is it accurate yeah that's right we kind of like try and keep a side eye on it so it

doesn't blindside us but I think it's like really I think the best thing for us now is to focus on phase zero and get back down and the research team the every such team you know dining in finally can adjust them they do a good job of highlighting the things that are relevant to us so we don't have to we can focus on it so yeah done much in the same boat as you yeah so in a different direction we've been looking upstream looking downstream we we had brought up rocket pool yesterday and and maybe even other services a test ensign Oh is is another service that we know through Jim McGee

who who provides a lot of support what are your thoughts about using let's let's do rocket Bull specifically I we know they've integrated lighthouse do you have any uh any thoughts or perceptions about stagers using that either as as a node user or a node operator uh yeah I mean I I know I've been around rocker pulpit was there also from Australia and there's not a whole lot of us in Australia so he gets to know each other I quite I like to detainment rocker pool I think of guys and they've been there quite a while I think say can Bulls definitely have this place obviously you don't want everyone to jump in once taking pool because then

we have a big centralization risk but I think that I don't think that shaking pool is inherently bad and I think they definitely have value to add to the to the system I think you'd probably be a shame if people that you know rocket pool is not new everyone doesn't just run on the rocket pool server it's something it's not fully it's fully centralized you know they do spread their machines about I'm so given that I think it would be perhaps a shame if we were to miss out on stake is where the only reason that they would stake is if they could do it in that way so I think they definitely have a place in the community and I think they look forward to working with them and you know

supporting their their application great yeah thanks [Music] anyone else want to speak up with a question can you give an update on where we're at with slushy and and the multi client test nets and maybe we're you know when are we gonna launch the one so I see is every from one slash parity um paraty he has been jumping in and getting involved and running clients he's been really invaluable um I really really enjoyed his input him testing thing is really detailed bugs so he is

he's definitely pushing /e slash lessee there at the current stage he is kind of just going around and figuring out which clients can run just trying to connect the things together you know just kind of sitting on the floor with legs crossed plugging things together and seeing what happens in reporting bugs but I think that it shows it's a very promising initiative and I think he plays a very important part in bridging the gap between the teams those are the teams are you know like hostile to each other anything like this we are very much heads down and focusing on our own

thing and so having this person who's kind of independent and it's like you know bringing things together connecting things together is really great I think I freeze last comment was that he needs maybe six weeks or something they expect expecting clients to be ready in about six weeks for a big multi client test that this sounds reasonable to me I think we'll probably see some more longer-lived multi-client sets that's coming in the next few weeks and then once we can prove that those little little test sets work together then we can go and shoot for the big multiplayer test then so what is it no it's Mme yeah May forum I was fourth yeah

May 5th May the phone that's right yeah it is coming on there with you yeah I think next month we can we can hopefully see a bigger multi-client testing I think ice I'm sorry I saw some of those timelines I you know we were all we've looked at it's like July was was the big you know we get single client test nuts and April and then I think people speculated that May through July would be

multi-client has has the window been pushed back because it appears that things are running very well it looks like for for one of the first times and I didn't mean that derogatory but things are just humming along and improving very quickly but it also appears at the same time that the window is being pushed back is that is it accurate or is any misunderstanding yeah I think so I mean I think s2 is being pushed back quite a few times you know we could there was this kind of idea that perhaps would have something like a mane that in January December or something and that definitely didn't eventuate so being pushed back is something that's happened I think all declined test ants are happening a little bit later than we

expected I know for us the delays for us are coming through on the networking side the networking stuff is really complicated and getting it to be super solid is difficult so yeah I would agree that it's being pushed back a bit it's live p2p still on track working and every I haven't heard that in a while so is it yeah it's doing really well for us okay so yeah there's another we all need to update the gossip sub is kind of a protocol in Lib p2p that we're using your next to we're in version one of it now we need to do an update to version

1.1 so that's something that's kind of kind of come down the track before we can get like ful fully magnet ready specification so I think this is like something that keeps coming up it's like fully main net ready but then like there's always a caveat to it so we we're getting close that's just you know there's just still is still a few things to deal with great so Mariano Conte asks how the audit is going oh yeah good question so we are going to start so we're doing kind of a two-phase audit so I mentioned before gossips up we had need to bring

in the new version of gossip so Joseph servers like very high-level overview it's how we broadcast network messages across the network and broadcasting is a scary thing because it's kind of just like yelling at the network and everyone starts yelling at once we don't know who to listen to so you need to be very careful about broadcasting messages so gossips like one point one adds in some more security mechanisms to try and stop people from all yelling at once so we need that before we can hit maenette and be ready and so lighthouse we are doing kind of a two phase audit so we we have a security audit walked in with trail of bits we start the first phase of it in

two weeks - one day and that means that we attack everything from like so if you imagine like you know how to network at the bottom and then you have the you know like the database and block process like verification singing and verification above it we kind of cut it off of the network we explained to trailer bits and assumptions about how the network works then they're going to review upwards and then we're gonna go into multi client test sets we're going to implement concepts up 1.1 and after we've learned some things from multiplying and we have got a sub 1.1 we're gonna get trailer bits to come back and they're gonna do the bottom half entirely and they're gonna get what's going to go across again the top half of it so yeah we have we have our

first review starting in less than two weeks which is exciting and terrifying yeah it all feels like it's so close even though you know we're still looking at months before rollout but it all feels so close right now yeah it's getting close there's a lot of momentum things are happening we have you know like this that's running and and things happening it's just yeah like one of those one of those things where you like I only ever see the things that I need to do in front of me sometimes like taking a step back and looking how far we've come this is pretty impressive well that's interesting because one of the struggles at of as you know I refer to myself as an end user is every day it's like how much progress has been

made how close are we and it's like riding in the backseat of the car and you're like are we there yet are we there yet are we there yet really no shut up yeah totally I feel like I think a lot about that so I think it's time for us to begin wrapping up Lambo Chief Paul Patricio do have anything you'd like to add if you didn't get your token if you didn't get your token message me okay so Lambo she still has the poep tokens you can probably catch him later in discord if if you didn't get one here if you can prove that you were here there are a couple business notes I want

to bring up we're interested in getting the timing of this meeting - the best time for everyone so I plan to send out a survey the to the people that are in here now you are the people I'm most curious about you stayed to the end of the call so you have some interest in this so if you will look on reddit.com slash are slash eats taker for this survey and let us know what time works best for you I will probably take the mode of those times or I'm not sure we'll work it out but we just want to know when works best for you I'm also very interested in a new platform I don't know if it's new a platform I just learned about called Remo Co REM o.co

and we're going to potentially plan to move our meetings there it it allows breakout rooms after the main talk so that groups of people can communicate a network and talk about whatever so I'm looking forward to that Paul it was incredible to speak with you I love not only that you know so much about it but that you can come down to the level of the mere mortals and and let us know what you're thinking it's it's been really enlightening for all of us I'm sure it's really been enlightening for me so thank you so much for your time tonight no worries thank you thanks

everyone for listening and for sticking with us while we push this across the line great all right good night everyone good one or good you good morning Paul have a great day that's it