Hashing It Out
Hashing It Out

Episode 93 · 1 year ago

Hashing It Out #93 - Sigma Prime Mehdi Zerouali

ABOUT THIS EPISODE

Corey and John speak with Mahdi Zerouali of Sigma Prime on how ETH 2 is going to look like for users of Ethereum.

Links: Sigma Prime

Sponsor Links

The Hashing It Out Social Media

Everybody got an awsome episode for hetoday, but first going to do a word from our sponsors aswell as give you a little more administrative stuff going on in theback end. Hashing it out, and things were involved in that you can um. Youcan become a part of so first off like thank you, think oursponsors avalanche avalanche lads, the highly scalable open source platformfrom launching decentralized financial applications recently raised aboutforty two million dollars through a public sale and no gering up for itsnext milestone. Next week, the launch of its main net on September, twentyfirst, that's right, they're launching their main net September. Twenty firstso get prepared also to strap the ecosystem. Mavalancheopened up a bunch of new grants for developers who want to build highperformance defi, that's de centralized, finance applications and infrastructure.They have open calls or projects like a decentralize exchange, liting deptsstat coins and with morenit every week. So they also accept applications forother disentroysed projects to join the AVALANCHICA system, so go build anavalanche, build without limits and go learn. Morat a e Labs Dotbord, THAT'SAVA! La Bs dotwork, as for W A we're doing in hashing itout Um, there's two things I want to talk aboutfirst, is hashing out is a part of the PannValaleague. If you don't know what panball is look back, if yew episodes a,we did um an episode with Neuron about what Pinball is, how it works so and soforth. It's a really awesome project that we're Havtobe, part of so Um. This round, Panvala is donating about.I think current Pan prices about a hundred and seventy thousand dollars to the ethereum community. How does itdonate those things? Where does it figure out how to donate them? Well,the Panvala League has giptcoin grants and Hasin Ois part of the PenvalaLeague, so we have a gitpoin grant that basically is a multisick and if youdonate to the gipcoin grant with pan it w'll get matched not only by the clrmatching of typicalcin grants, but also, additionally, by m a hundred andseventy thousand dollars that pennvall is giving out and then we're going touse that money that's raised through that grant. With the advice and UM decisions hold from the EtherianSecurity Community to fund security in in infrastructure projects. I webelieve that Hashing Iout, that security n new prestructure is a veryunderfunded but incredibly vitally important part vecosystem. That needsmore funs, so we're going to try and do that and you can help by donating panor whatever to ti the get Cin Grat. That's going to be the description ofthis episode, so pitor Pan, don't en it to us we'llfind a good place for it tohelp smelt the security and intrastructure of the THEOMICO systemand other big news. I I don't think I've mentioned on the podcast yet ishashing. It out is leaving the Bitcoin podcast network, because the Bicoin Podgest network is no longer a network. It's just a bickoin PODCASS, so over the next, maybe ten or so episodeswere going to be continuing on this. This fee that you're subscribe to now,but in the process there's going to be a new feed, that's only going to behashing it out. They oll need to resubscribe to because at the end,you're not going to be able to get it on the feed youreon. Now the DianPotgass, so we're going to have ther own thing: Reneo, new branding, UmTryand, add u some more resources and so on and so forth to the show so thatyou can be a little more stable. I don't know we'll see, but we're goingto have own feed check out, listen up for it check te twitter see whenever wepublish that, but at least you could just listen to us, no one else. It'sgoing to be great bick Wen pod, Guessis, going where I'm still doing that. It'sjust two different feeds now and onto the show...

...kindmet word welcome to hashing it out. APOCAST forrethoughk to the ATECH intovators, behind blocked in intrastructure anddecentralized networks. We dive into the weeds to get at Wy and how peoplefilled this technology, the problems they pase along the way I'm listeningand learn from the best in the business. You can join their recks woking back to Hasmelt on the hosttoday, Dor coripetty with Sean, and today's episode we're going to talkwith Mzrali Thik. I said that right, Morcela said that right from secretprime and the main focus tois psoeis going tobe around kind of therium two point: Oh, how it works kind, O a couple of Gachasof what to expect as a user man kind of the time ine on there so start off,John, which o say hello, O quick, hello, hello, real, quick, Helenen getHaranodov. Let's get hashed and UH sobe some cheesy Tagin to the show I'mworking on it. I've been eeinout new one forget what and h may give us anintroduction M. I don't think you've been on the show before, but we talkquite a bit so tell us about yourself: Ann Siaprip, sure, hey everyone, myname's Mettie, I'm a cofounderand director of Sima Prime. We are andinformation security cinsultancy, mostly based out of Australia. We'vebeen hiring a bunch of people all over the world. Lately, Um we've been aroundfor about four years, providing security assessment services to a lotof block. Shame projects working almost exclusively lately on M terium,providing SECR SSMEN smalt contracts on a lot of Pronocol, H, layerdevelopments, Um- and I guess we are here- Tobett talk about Easter 'cause.We are also the founders and maintainers of lighthaps, a rustimplementation of the etherium to pointo specification e thanks forhaving Yep sire real quick. What would you saythe I guess, techical advantage or differentiator? AWhite House has over other IFTHERTO imvovitations m ATS, a good question question that weget quite often we because of I guess the DN, a of the company we're anintermation security consultancy. We take security very seriously, we'd liketo think that we've incorporated security into our development lifecycle.Very early on so um we'd like to think that our clients is probably one of thehopefully on the safest, OT, secure clients out there. It's written in Rust and rust isparticularly fast, so the focus has also been on performance. Speed on ismentioned security, so that's probably what the friend shakes us from the rest of the other teams wan to give you moment to to pitchshill out. Second, prime, in the process of Tuo wifes, like Flithouse,is good nels, opendo yeah. So Li e, I guess their proachhere- is the kind of Rak down the concept of a theoryim too,and what's going on and what youurs can expect. So I'm not exaly sure how I want to dothis. The conversation's going to naturally evolve, I think, but like ifyou had to give the highest of high over level overviews ofthe transition from th one andf, two whawould, you say: okay, et's sound a crack at Dhis, so f,one is t a prof of work chain that we...

...all love and use Um and, as you allprobably know, the plan was always to transition to a proof of stake andsense mechanism, so, namely M Casperif G Um. So the Research Shene has beenworking on thisnfoundation research has o been working on days for the pastfour five. Perhaps six years and the dementors have been m developing thethe spestication turning it into actual usual software. For the past two twoand a half years. The main difference as mentioned is theconsensus mechanism. So moving away from prove of work, which is you know,quite a westeful way of coming to consensus. If you ask me, I know this can be quite polarizing inour community, but this is basically how Paul Mico founder got started. Umon on Lithouse, I think etherium currently uses as much electricity asCosta Rica. If I'm not mistaken- and we think at Singnar prime that there's youknow much better ways of coming to consensus,Um and CASCRAF G is one of them. So stuper excited about the environmentalimpact of the transition from provob work to PROFOR, stake, Um and h. Thistransition will happen gradually. Um will be shipping, though in variousphases, so fay zero is the first phase that will be introduced. Hopefully, ina few weeks I a couple of months, the latest seventy Tis year. That's whatwe're aiming for Um and this introduces the consenstis change right. Thisinstroduces the proof of stake consensus mechanism CASPERATEG, and weessentially call this chain the beacon chain, and you can see it as thisorchestrator Um that gives sort of the pulse of the network to Valeydaitors,giving them tasks rewarding them when they do their job correctly andpenalizing them when they don't et Cetera, et Cetera Um. SO THAT'S PAYZERO! That's what I guess all client teams are working on at the moment andUm. These one will hopefully come shortly. OUTTOP and phase oneintroduces Um Charge the concept of shod, so etherium Jobono will be ashodted blockchain, meaning that will have sixty four sub blockchains. You can seethem, as you know, oindependent block chains that are all linked and tied Umwith each other through by that consensis mechanism in the beacon chainm. So once we have, that will be effectively using m those shards justas Thogta availability layers, so we won't efvectively have um state transitions or user sort of smotcontracts or transactions, as as we are all familiar with in eath one that willcome in later. So there's talks about a phase, one point: Five, which hopefullywill um get ef one into east two and have thatas a separate, dedicated shild and then face to Wi'll, introduce, I guess,propler state execution and small contracts, and this is when we'llactually have a blockchain, a usable blosh in for developers as WEALL WE'Rll familiar with Wi, H, um. We need to hear him one Ho that makes sense yeah. I guess, if we're done here, then rock up fruast Um. Maybe can you can tell us a bit moreabout Hoizero like wh? What do you mean when you say Beagan Jane? What does it do? It's likethe bays functionality that we get then...

...something I forgot to mention is thatto transition from Um East Teorum, one toetim two we're goingto have a deposit contract on Ethearem one. I think that's quite important toh mention for our rlisteners Um. The wayyou're going to become a validate is by depositing thirty to ETA, into thedeposit contract on the theorium one proof of work, Cha Um. As a result, youwill become a valleydete on the beacon chain and the beacon chain iseffectively a set of Um comprisive set of bollilators there to propose blocksand vote on what the canonical us of chain would be Um. The way it works isby leveraging Caspar, F, F G, which is the consenseus mechanism that we use oneater Um and allowing us to yet jstsorry, allowing us to essentiallycoordinate the work of of these valley daters, so Um vollytherers areincensivized to be online to perform their duties. So there's ffectively twotypes of duties, perhaps three, but let's just talk about these twoproposing blocks, so the beacon chain has a schedule that we call slots. So aslot is happens every twelve seconds. It's just a time window right andwithin that time window you have a vally adal. That is expected to proposea block, as you can imagine. If that valley, there is off line well thatslot won't have a block. So I think it's quite important to try todistinguish and understand the difference between slots and blocks.Slots are just a time window blocks fit in one slot, potentially if thePROPOSAR is doing their job, correct Um. So that's one of the duties. What'sin a block in neath he hob, I theres Wat vesierecgreat question ees a whole heap of stuff in a block, so we don't havetransactions per se as mentioned, but we do have information that availablefrom the Erium onchain. Remember I mentione as that deposit contract. Weneed a way to effectively translate some of that information from at one toEster and there is an ethone daughterfield. Each block and vallytesare responsible for effectively processing deposit logs from at onechain and translating that information into the beacon, trank, and thathappens using that particular field. A another interesting Um, I guess messageor field. That's in a beacon block, is what we call atte stations the stations you can basically see themas votes. So remember how I mentioned the first type of utes. The production s proposal of blocks is another type ofduties, which is the at thestations and having vallilators attests to what thetheir view of the Chines and those votes or ather stations are effectivelyUm, aggregated and put in beacon books. Ifthat makes sense, what else is in there Um we have therand out so a great way. So one usual sort of question is well. How do I know?Which Block am I supposed to vote for or which Um Slot am I supposed topropose on? And this Goms with the this is usually um explained by referring tothe Shuftling algoritm, so shuffling expectively, you take a set ofValeydaters, you shuffled it, and that...

...gives you their roles andresponsibilities for each one of them perslot. So in order for us to do so,we need some sort of entrepe right and the entropy is basically a rendal typeof Um entrpig mechanism. So Randa is, wherebyis where you have a group, a selected group of potcipants that each propose,but each give Um ou, know come up with a random number and then the actualrandom number used by the system is a result of you w sometimes just exhorean excusive or of all the inputs from each one of the participants and that'show we get our randomss and we then use that randomness to essentially yetassign vollulators to various dudies, including blockeposition and Um. At thestation Ogivemet imeant of what shards a particular validator willbe validating, Alt assigned at that point Eh. So that's coming into play. Abit later, as mentioned, that's that's phase, one pase one point five phasedupe, but yes, the idea is to use that M entropy, that sorce of entrepede toassign people to what we call committees and committees areresponsible for validating M transactions happening on specificshots. Absetly C can any any one valiator be assigned to multiple thike,multiple slots or shards at the Santime, so each follid addor is guaranteed totest at least once oexactly once per epoch right. So a epark is thirty twoslots, one Slo is twelve seconds, so anypark is roughly six point three sixpoint four minutes and within that time frame you are guaranteed to attest towhat you see is the cononical chain at a certain slot. So an in terms of block proposals.There can only be one block proposal, er slots, and this this probably Um youknow, might starp talking about slashings and how we h police thenetwork and make sure that e vallulators are doing their jobcorrectly. If you see someone proposing e twice right, U found the bloodpropose up on that particular slop and I'm caught proposing two separateblocks on that same Slok. That is a slashable offence right. We gont havethat we cot have people proposing blocks, are right, left and set ar. Soif someone catches you doing this, they can simply submit a proof of you doingso, which is kind of trivial right. You've been signing to conflictingmessages at the same slot. You don't deserve to be part of this anymore. Wewill slash you and kick you up of this great network, so we can perhaps talkabout slashings a bit bit light off but Um yes, so we can't have two separatetwo different Valeye thetters on the same slop, but we do have multiplevalutators UM, t sort of securing and vogdatingtransactions on each shard good Tono. That's KINDOF, curious aboutthat in terms of the h kind of the committee selection around thedifferent tasks at had so this will be hardened with Um whatwe call vdfs Um. Oh dear VD, stagfor Um. Oh my God, I'm sorry verifiable! Thankyou! GERIFIABL blay function so because...

...the Rando Um sort of entropy mechanism is kind ogamable right like if you're the last person, you can essentially say well,because I'm the last person I can see what everyone else produced and thendecide to reveal my vor according to my own interest right and we pabt havethat, alas, obviously extremely difficult toactually pull out like this attack is not so trivial but satisfifoble offenceto not reveal not really yeah. Well, it's it's just like it will.It will kind of come into Um a an OWF linse ofe penalty, so yeah 'cause, likeby the way you're like. Oh, no, I'm not interested in producing this particularblock and revealing myrendow. If I'm to block Te Poa, we cout realy tellwhether that that was the deliciously or, if you were just offline Um, so itisn't. Isn't that L, like I thought, that's PARTB, of the job ofvaladatingis being off line right, e's being online right is so you are you incentivize people forbeing online and you penalize them for being offline and the penalties ofbeing offline are basically kind of the same as the reward of doingyour job correctly. So you obviously lose the you know: theres the optiunitycost th the weenings that you don't make pull us an actual penalty die thatmake sense. So I think we should at some point tryto clarify slashings verses penalties for beingoffline. A slashable offence is essentially triggered when you canprove that there has been equivocation right. There has been maliciousactivity. Well to be to be perfectly fair, it does not have to be malicious.It can result in a software bug right if you're running some software that',you know having difficulties, keeping track of the head of the chain and hasbeen sort of testing letting you a test to. You know separate sor chains at thesame time this and you you're not even knowing that's, not really melicious isit? That's probably just you know, result of a software bug, but this is provably Um, an equivocable sort ofmessage: Asper Casper, FMG, slashing conditions, whereas if you're all flyinwell you're off line you're, not Y. AH IT'S TOT software bug, you're, notbeing malicious. It's just kind! O! TOUGH LUCK! I don't know your Internetconnection went down or you rint out of this space or for some reason, Youre. Idon't Kno Electricity and your house went down, so we did a way to incensivize people to keep their nosesonline, but at the same time we don't want to penalize them too much ifthey're missing at the stations or missing blocks proposing blocks fromvarious reasons. So that's kind of the trade off, and this is where I think weshould be careful when we communicate with the broader community, aroundpenalties in the eat to space, there's two types of penalties: isiy slashingslishable offences and then the offline n, the provable cheating and likeunprovable Mon participatation. That's right! That's right, and you also ave carrots too. It's notjust all sticks so like yeah. What what do you get on the beacon chain in terms of carrotslike? Why would whywould someone besendifized to do this and that willthat's going to shring board Meino conversation later that Um? I wouldlike to talk about around kind of user expectations, but h, like whate care. Why? Why Wo LD want todo this? 'cause, you would get some sweetsweet rewards Soto, botstrap, the beacon chain. Weneed some ether right to be locked in this depotic contract and so turnedinto eitherinto peaking chain and the way we incencivize people to do so,isby giving them rewards and those...

...rewards are effectively if aricalcorrectly, I haven't seen the calculations in a long time, but withthe threshold, the very minimum that we require to bstrap the Beacon Shain. Ithink we're looking at about twenty to twenty two percent return than yieldper year, which is pretty decent. Obviously, there's a lot of some risksassociated to we this and w. We can talk about it, but this is h. This isthe carrot that you're referring to carry the UM quite juicy returns forearly adopters early participants that are willing to sort of race cor committo securing the beacon shop. was that a function of what dis that returnpercentage or function of it's a function of the total member ofValleydaters Aka, the total number of ethol locked into this new system, sothe more ether we have, the less yold will get if that makes ese. So youreally want to incentivize people to do so too to actually look in their eat.So the very first Um people to to to to t have done that will effectively get avery, very high yield, but that yield will you know slowly h decrease as weget more and more valley daters on to the Beacon Chap so there's a there's, a fixed budgetlike sa like block reward Um, maybe not fixed, but like it. Itdoesn't go up if there's more validators right so in in the firstweek. The plan is to issue some amount of eth to and reward the validators andso the more validairs there are the less ess exactly exactly, and we can'treally predict how many valed leators will get right. Um we can only. We can only be sure thatwe'll only bootstrap and kick off the beacon chain if we have more than acertain number. I can't remember the number off the top of my head, but itis obviously public information and it's on the Tro Specrebo. I mit 'll definitely putthe spectorybal on the sescribbent. The show, where does that yeath come fromthat's brand new eath, and if so, what does that mean about issuance on themainshain? The interesting question? Daddy's Bran, new esh Um, and thatwould, I believe, be part of the total inflation of the etherium currency orcripal currency O, whatever you want to call it these days. So for a whilewe'll be running two parallel chains right, we'll be running here in on Boyo,the prof of Werk chained that we all know ano and use today, andtherm twopoio one pointo has a block reward right. This is how we incensilizedminors to do with that. Job Spend a lot of electricity, valutating transaction on top of obviously transaction fees,which I think these days account for the Vos majority of the actual. I thinkhat'se. First Time we've actually had a blockchain that has equivalenttransaction fees to district t o issuance wow on a public network.That's that's HCR. Col Is Hok for users that aren't doing defi stuff, but yeah.It's pretty neat. I guess I suppose yeah. I guess it's Kindof working as eintended Um, but yeah so that that's the inflation of the proof of Webchanges, I believe,to ether per block and found not mistaken, keeps changing every now andthen, but Um should be cweeter, Porblok piracle correctly M, and on top of that,we'll be adding the inflation o the beacon chain, so that issuince that wewere referring to will effectively come out of the infllation or the issuancethe total overall issuins budget of Etaria. Maybe when the next phase happens,we'll do this again an talk about how transaction fees and cross sharredtransactions playnto Um...

...rewards because that's o going to besomething, but maybe not all the details are hammered down yet so Iagree thits, probably for another conversation, Ert Cool, let's see! Where do we wantto go from here m? I can change to Steftike entirely unthiss John. Youhave something you' like to continue here. I don't have anything OTOF my tongue, I kindo want t transition now that wehave, I think, that's O, maybe the best v verbal overview ofkind of what you could expect the, how the Beacon Chang to work and what itsfunctions are in terms of phase zero and within the context of all the thelarger phases. What I don't think is very good. Public knowledge isexpectations from a user standpoint. What's the process of me getting avalitator H, submitting my money into the F one deposit contract and then beingprepared correctly for when the chest thats start for whene, the mainchainstarts W we've seen from previous testnetes. Is that people love lookingup test not either they have enowaaors, but they hate having notes, they areready when wnes life tetenly. How do we? How do we? How can we get users to understand whatthey need to be doing and the pressociated resources that arerequired from them and likewarlee optimizing for in orderto pet there's a few things here? I think in your question cory. The firstone is probably the challenge of running and incentivized desnets. Sopeople are supid can help, and you know thy'ethey've got girly et and they'reKein to looking their girlly a into the potit contract, but when it comes tactually running their nose well, you know someone else is going Ta. Do that?It's okay! I don't really need to 'cause there's. If I end up losing mygirl Weit's not worth anything. So doesn't we don't EALLY CET? So I thinkthat's what we've been facing, as in we the to DevelopmentCommunity Um, is that problem of like insensivizing people to actually youknow, help and run their noses when, when we mon these destinetes one of theone of the great, I guess incensive mechanisms that areout there is thepopes pull ups. I never know how to pronounce that, but it's great the Provof attendance protocal. So if you do your job correctly on a Testnet, youget a nice nft. ICH IS ORT of a badge that shows that hey. I was one of thepeople that sort of held bostrap the east to network by testing by runningBali beiters on certain testnets. This is not necessarily enough, as we'veseen with MEDAA DASHA and Spedina Sospadina is the latest was thelaghtest testinet, Um, cold or um, described as a dressrehearsal for valley. Dators to you know, essentially play with theTDEPOSIT contract and generate their keys and run their clinets accordingly.We know a bunch of people that took these very seriously. A lot of peoplethat are planning on running a bunch of vollydators were super keen on tryingthese things out, but for the change to come to finality and thit's a conceptthat we haven't actually spoken about yet, but we need at least two thirds ofthe network to be doing their job correctly. And if you have, you knowmore than one third of the network not doing the job correct or being off line,then Youre, you have this Um Shan, that sort of lingering and not finalizingand fenality, is this beautiful concept that we have any to that allows us thatthat that is responsible for so many optimizations in ECAL client. So if youhave to keep track of all the potential...

...forks and reorand heads, a F, F,multiple chains that are out there, it's a heavy load on eastro clients onthe actual blokchain notes, but when we fitalize a anepocorous lot, webasically come to the conclusion where we say: Hey Network. Everyone hereagrees that up to this particular slot number, those transactions cannot be reverted.This is now the cononic call, Chay and well forever will be so then clys cango like okay sweet once we all agree to this. I don't need to keep track of allthis information. That's you know, exhausting my resources. I can then use-or at least lighthouse uses, two types of databases, a hot BATERBAC and a coldbatebase. The Hot Databas is where we put all the non finalie stype stuff,and once we het finolity, we can be like okay, sweet, I'm going to transferI'm going to move a bunch of that information to the cold thoght of it Um. So I think Um. What we've seen on those tetinets islong periods without finolity, because we don't have a we to incentivizepeople to actually you know, do their jobs correctly. On these marke networks,there's been chats about having incentivized, destinetes and whatnot. Ithink afreven said that, while fat zero is the incensivize destint for eastu Ecuse t's, it's an interesting way to look at it, but tro sorry to go back and beetberailing a bit, but I thought it was quite important to to mention that glyo that that's in a very important aspect that I think people inunderstand is that if you don't have the right amount of participationacross all the valevators that are registered, then the chain doesn't cometo vinality and T it causes a tremendous amount of havoc. This was probably the best thing thatcould have happened to those deat nitts Tomatasha 'cause. As a result, everysingle cliet team has been pushing hard to optimize their software to handlethis particular case, and I think that by having that very low aicipationlevel, um closing havoc on the test net client teams have been working reallyreally hard to make sure that if this ever happens on Mynet, which is, if youask me quite unlikely, going back to the carrots people like carrots, peoplelike working thoars rewards and people will most likely keep the invody tatersonline. But if, for some reason these happens on May net, we now have a setof pretty hard an clients to face this Um situation. So Yep, sorry, you go wan.I want Jus, like I t Wano, be sure tis clear, like people are, so people aresigning up, they're depositing their eath. They are registered as validators,but then they're, just like they're turning ou their laptops and walkingaway like that's. Basically, what's going on and they're gang slashd andthey just don't care so w, would it would he e? Would it be better if therewere fewer people signing up like regulain, alidators and y version?They were Betsin more dedicotally, totally t tol, that's a great questionso to to do that. We've been rining private sotest nets, just east oselvs,like th, each ceam would have a bunch of private networks or we spint up likeone thousand ten thousand folty Dars, just to see how m just to basicallytest SOM optimizations and you features and whatnot when having a large numberof valley. Theaters is very interesting and important to test is basically Thoy.Try to replicate the actual load that those clients will be facing in mainnetUm, you know chances are. This will be quite a popular in new network. A lotof people will be willing to actually look ther their eat and what happens isthe more so the actual resource consumptions for each client Um, effectively scaleslateally with the amount of voleuytairs,...

...so the more value thatters you have onthe network, the more cryptographic operations you have to perform tovaludate atestations Um et Cetea, et Cetera, so essentially BLS signaturevarrification. That number of operation will, you know, go we'll follow nearly the number of Umnodes or sorry boy daters. On the network, so it is, it is quiteimportant for us to have a trial run with a very large number of valleydaters, who we can effectively see whether we can handle that load m. Whatwe've been doing, as I mentioned, is you know, run a few nodes with a bunchof Vallin batters on top of them. It's also great to have a greaterdistribution whereby we only have perhaps you know a couple of vollydators for each node on average. Jest do make sense, then continuing on to the resources that a user can expect tohave. What's the thing that first like one of theminimum requirements that someone should have for running about eer itHes porphase zero. For one thing like well talk about this first, what arethe requirements that someone should like a reasonable requirement so muchto have ben trying to run a valid of proface Europe? That's a great question and it's atreaky questions. Well, Um! I don't want people to think that therrequirements for Pha zero will be their requirements onwods. Oh I I we're goingto talk about that. Yehit's probably his misconception that I'm just goingto buy Raspberry Pie an Oll rasberry Pi with you know two or four gigs of Ramand that'll, be it I' just Gonn run my body better is gonna. You know, do itsjob on the the beacon, Shan and also you know, valleydate chart transactionset Cetea Etcetra. Please don't! If you do end up running your voly thatI on a rapberry Pi. It is certainly doable right. We've been runninghathouse on rasery pies, it's been running smoothly for and a bunch ofpeople have been doing so, including Paul Mico founder, but do expect to have to upgrade yourhardware at some stage. So back to your question clory, I can only talk aboutlighthouse, I'm not entirely sure about the other clients. Perhaps you couldchime in with status in o status. Nymbus is working hard for making Um alight or not like client, but a client that can upright under Um sort of typeconditions in terms of theyare epecially, more focused on resourceconstructeve devices. That's right! So in AKA UM K W A standard. As I said, araspberry py, eigh gigs of Ram is kind enough. Um, your storage is quiteimportant as well like distorae, I mentioned the cold and hot databases.So having you know, N SSD, that's connected to it. That's up to a hundredgigs of storage is, is plenty M and I think, as a general rule of thumb, anyrecent hardware will be Wi'll, be more than an offer. Pay Zero, at least for alithouse. The memory footprint of Lihthouse is very low, at leastcompared to a lot of other clients. We we reallytried to make it Um optimized memory fprint into CPUconsumption as much as possible, because at the end of the day, that'slike say: You're running your valley, daters on a clouding for Shopo. If youcan hole your resource consumption in terms of ram and CPU USAGE, well,you're co, almost doubling your R. I right. That means that you don't need asuch a large instance on, say, aws or...

GCPR or whatever cloud plaform. You usethese duys Um and it's obviously very appealing for people to not spend asmuch on their boly thet esel. So it's been a focus of hours since day, oneready trying to minimize that food brint and, as a result, yeah we can runcomfortably on et tee, to instance on ws teacher medium instante on as for abouta bandwith and power like uptime, isincredibly important here and so is consists, the Internet connection, whattype of bend with consumption do we have and what are there like wher thewhat', the uptime expectation of this of this device EC you're, not runningon your Laptop Youre Usit for work I ate recommentit I mean I would recommend that 'cause.If you end up, I don't know if you end up switching off your laptop orrestarting 'cause, you need a to apply security patch or whatever you you're, basically going tobe off flying for a bit and probably be you know, penalized for a Bidor atleast losing the opportunity of making Etho during that time. So I wouldpersonally not recommend using the same hardware as you use for your day to daylife. I would certainly try to use a dedicated piece of hardware Um.Obviously, we kind o want people not to use cloud services as much the wholepurpose of this, the engish to be they're really sentralized. So if weall end up running those on n, a Lus, sorry well, yeah Mayb, it's sad so certainly uptime is quite important,but gotta be wee. Gotto, Te careful that a lot of people have been thinkingabout backup plans and redundances and, having you know, some sort of two,let's say two or three nodes getting fired up in case. The primary node isoff flying, keep in mind that if you're a flyingyou're not actually slashed right back to that Um explanation that we gaveEarlyuh, there's two types of penalties, slashings and all flyng penalties. Soif you're off flying it's actually not that bad right, like if you, if you'reflying for a whole year, think pold the math last time and, like I think,you're losing like ten percent of your stake. That's if you're offline thewhole year! So it's it's really not that bad. Butas a result, if you stop having all these complex backup plans andredumbancy set ups, what can happen? And what actuallyhappened on, I believe Cosmos was that you have this complex set of redundancyset up, whereby the two nodes actually went online at the same time andstarted producing the same message and start of producing the same blocks.Remember the proposer slashing that I was explaining Girlie whell. We can'thave someone the same volley. There submit two different blocks at the sameslog chances are, if you use a Repundancye set up that you know is, isnot perfectly finely tuned well, you will end up at some point: either atesting wrongly or submitting blocks wrongly. So really. I I think we shouldreally be careful here about about those trade offs. Obviously, uptimehe's critical, but what I think even more important isnot producing slashable messages. 'CAUSE, if you get slash not only youlose either, but you also have now this the remaining of your balance thatstuck in a state that you can't actually sort of, withdraw or getaccess to. That's now, stuck until we antil everyone in this new chain hasaccess to their eath, which should happen in face one ASE, one point five.So...

...again be careful out there. I understand thatpeople wan maximize their profits, honest an that people really want to beonline as much as possible. It is cright, but when it comes to actuallyhaving redumbancy plans think about at choice. What, if uh hat? If I screw up on accidentsomething happens, I have a bug, my software and IT subfits the wrong attastation or something two questions here. Can I migrate? Is it easy tomigrate to a different client? If I don't like the one that I'm using and can I top up my my th acount so that Iculd get back to the right threshold, so I conontinue participating like agood valiator best question. We have been working.Michael FROMOUR team actually has been putting together an interoprableFullmat for slashing thout of his protection, so that, if you move sayfrom Nimburs to lighthouse or from Lhouse to prison, then we are all usingthe same Datoba schemo and we can import were slashing protectiondinabase, so pretty much. Every client nowadays has a slashing protectiondownof vice, which is effectively checked before your Valy dettor isallowed to sign messages so, each time there's a message should be signed. Theprocess goes: okay, I'm going to check in this Thatabas whether this messageis going to be an a slushable offence based on what I've submitted previously.So the problem here is that if you start changing clients, then youpotentially lose that information right and if we starp running, as I said, a valleydator on lighthouse that usedto be on Nimbus. Well, we CA. We now don't have access to that informationanymore, so we can't protect our users against it Um. So now, thanks to thatEIP, I think it's in the Ip now Michael put together, every single client teamis now building the same, slashing protection thought of a Shchema so thatwe can easily import them and export them. So, yes, we will be able tomigrate from a client to another in a relatively safe way. So if, if you doyour job correctly, the slashing database protection is is up andrunning correctly, you shouldn't be producing any slashable offences. Thisis something that we should be. Perhaps broadcasting a bit more o know slashingis quite scary, but if you do Um, if you follow the recommendations from thevarious client teams, if you do run the software that Clinteems promvide withthe slashing protection, then you should be fine. We will check thosemessages for you before submitting them and broadcasting them on the networkand effectively before even signing them in the first place. So that's the first question and Ithink, what's left for the community to do a you know, there's a lot of workgoing on so but Thi. This should, in my Pedan Ginda, be a priority of the nextfew weeks is right, Easi to follow guides to migrate from a clint toanother Um. I think there's a couple of resources out there. People have beenasking US questions on this cord, but we haven't arms, have put together acomprehensive guide on how to import and export your Um, your your Vodydaters. If that makes sense, can you remind me where your second questionwas I top up my ethicount? Why would you need to do so? EVENA EERclashed O. I had as AST. My friend is KINDOF game over so like. If I get slashed hat WHUNI'sgone, I can't top up my other account and continue to participate back theniwork that SA pet for plamage. How much does the slush AING PIL? Is it thirtyteys? No, no! No! No! No! No! Thank God! If more than thirty two eth Orisinlating one validator their bestron one boy, TOR, thirty, two weet. That isold. If you want to stake a lot more...

...than that say you want to Stak threehundred three hundred and twenty Ethoug make things simple. Then you have torun ten body daters, okay and bacater you're question gory. If you getslashed, then you enter you're into this state lingering state, where youreither is locked in the beacon chain and not accessible until again phaseone pase, one point: Five and you're: You CA participate in the consensusanymore, Os', Aralatet or Complat y thanks for playing. Yes, like they pausated again, eot o eat theget on the on the Fone Cha, and so like was n ther asumption that there's likea small slashes, so that when you git under the threshold, though thirty two,you can no longer participate and you can then top it up. THAT'S N! That'snot the case! That is not the case. Okay, good tonight, there's onlythere's o there' ths, multiple types of of slashings right, but there's onlyone slashed state, at least that make sense O it's a flag. It's Notyes! HOW MUCHMONE YOU HAVEOR! I meanof you have it exactly: It's a flag, Wen his guy'sbeing slashed. Sorry C N. have you around anymore doing doing stuff but ev,even if it's a penalty, even if what we actually slashed Um, what the protocoltook away from you is minor Um. So it depends on the offence to back to yourquestion. Some would be like a third of your balance and some would be muchmorer than that so Itof, I I um say I start with threy to eat andI'm I'm on like 'm I'm around like. If I just if I serv thirty two et thou ou,my computer, I'm going to start losing money right away and I'm done Buti Ori an I can balid out for a weep.I have SOMB OZ and balance. Go off and I've earn some vacation timeand I can turn off my node Yeh, remember deference between slashableoffences and offline penalties. If you're a flying, you're, not slashed.You said that flag righ. If I've been a fling for a bit, let's say I've missedfor some reason. I was too busy an drinking bees with my friends. I wastoo. I missed the actual genesis time right forgot to turn on my computer.Well, you're, not slashed, it's fine! You can. My ballance can go below thesuas do the slash exactly and actually it went below thirty two for pretty mucheveryone on mnasha mm, because we couldn't reach finality and whathappens when we can't reach fonolities that everyone- everyone, includingpeople, that actually are doing their job correctly, have their balancesdecreased. This is to basically force people to come to come online and alsothere's this concept of De iactivity leak. That kind of forces, people thatare offline th, it's kicking them out of the actual vally dator set or theActiv validater sets that makes sense good to know ot accus about, thoughthat was that's good qaification. I think for a lot of users h running low on time, so I don't want to divetoo far NTO this. This is something that I've been relatively knitpig aboutover the past couple of weeks management. What could users expectabout Um, the safety of their keys and what keys they have and H wherethey're keeping them interesting. Interesting question Mperhaps worth mentioning that there are two types of keys in Eastto, your hockeys or signing keys as thekeys that are no effectively used by your valuedators to sign consensusmessages, blocks at the stations, etcetera and your withdrawal keys, which are the keys that you will sendyour eath or two once you exit...

...the Beacon Cha or was to exit the Valydereset, so the first one, the hot siny Kys,they're, hockeys l, you need your computer, your Vollue, the actuallyeffectively excessed in every at least every Um six minutes to atest. So wecan't have these or iould' recommend thinking of having the in a cold WalletRight Lut. These are hot for a reason. You readly need to access them quicklyand frequently, whereas the widroll keys, on the other hand, can be seen asthis cold walnet right that you will effectively never touch if as long asyou're partiipating nvaluvating on the bikon chain, you will never needthem right. These are this is where tha e thought that you're using will besent to when you exit the Um. I guess the Consensus Al The world thatyou're doing on the consensn mind iside Um, so I'm very excited about Ledger'supcoming support for bls. You might have heard about this, the I'm! Hopefully,I'm not leaking any SENSIIV information here. I don't think I am, but there they're they're working hard and an andthey they have there like. It is no working and it's in the final or stageof testing. So that means that we will be able to use hardware wallets forthose Whih, rokis whichs, quite exciting. It's not completely fully integratedonto the clients yet, but that's certainly something that will belooking at doing before Ma net Um. Otherwise, the way it's currentlyworking for everyone you've probably seen thelaunch fads, which is a website put together by tther infoundation. Thatallows you to generate keys um and it's a very simple FA, friendly website thatsort of Um hooks into a python software and t thesteps to follow are very, very simple, and you basically use that Um softwareto generate your keys and then you use the website to submit the keys or sorryyou submit the deposit data onto the one smart contract and the launch potletax all that if that makes sense. So that's what users have been playingwith over the last few weeks months, with both MDAIAMIDASHA and Um Spit Yep,and it's also expected affor them to to use that once we have mane net, buthardly while it support, he is coming awesome. I'm pretty happy with that in terms ofa overview and expectations of m coming lauch with the beacon chainmain knit. Is there something that you think I should have asked you that youwould? You would have liked to have put out there that I didn't hm? We did cover a lot Um, there's, obviously a lot more that wecan talk about. So what are you worried about? What am I worried about? SEG? That's, a good one, so we're security expers at SIGNA,prime. So I think one of the things that we're extremely worried about issecurity incidents and obviously you know vulnenrably is affecting clientsand people Um. I think one of the worst things that could probably happen is AH is a fork like clients, not agreeing onthe cononical heads and perhaps having a discrepand seen their statetransitions. That's pretty bad 'cause. Resolving this M would most likelyrequire fork, and you know how we know...

...we all low the appetite of ourcommunity for folks. So if we aoid that that'd be great, so we we are taking steps to making sure thatthis doesn't happen. HT At least reducing the likelihood of of of thishappening. One of them is perhaps something we should mention. Is thatwe're we're running this project called Beacon, flubs that allows us to willtry to find bugs inclint implementations before we hit mainnetand beacon. Fas is a set of fuzzers Um, perhaps with I'm not sure how much timewe have yeah running out of time. But buzzing is simply a a software testingtechnique that a Laancov security bugs re, not only security bugs, but bugs ingeneral and software and the idis to flod certain functions in your software.We a lot of different types of being, puts and monitor themfor crashes rightand if it crashes. Well, you probably whant. I expect that particularynputcarefully and make sure that your software is patched accordingly. Whatwe're doing with beacon, fazs we're taking this one step by having what wecall befrential fuzzing. So we feed that those inputs we can think of themas like azilions inputs per Peraa, and we feed them intoall the different clet eplimitations and we compare the outputs and we evenwant to make sure that the output is effectively the same across all theCOIC implementations Um, so that we can't actually trigger Dese, crepanciesand forks consensus, splits on the beacon chain, opefull that makes sense,e'R, probably getting a be too technical here, not in this adience atthin. You Find Coole hat what we shrive we strike for O understands wha. Youjust said this episode may be broadcast a little bit more based on it's more high level overview of concepts.We tend to go very technical, but I wanted something that was h. just gavea complete package of what people can expect and how things work about, and then from here I thinkanyone who has questions is able to bat down in the appropriate place formordesails John Tyo have anymore given morequestions. We wrap up- maybe I'll yeah, just toss it to the rap up like. Wherecan people go if they want to learn? More, where can they follow? You etCetera, yeah Um, so Um, East Zad on Tweedolf people are welcome to ask me questions there, I'm happy toanswer them. We have a discord channel. That's super busy! These days. A lot ofpeople are trying a lihouse and asking questions. So please join R. This courtchannel, if you have any questions Um, we are sp, slash lighthouse on Gub andYouill, have a link to our discord. Um and U Yeah, there's a lot of readingmaterial out there. If you're curious about the specification, the the eachrespect, Repo that I believe cory will be linking as part of this episode is agreat resource is optimized for readably. So it's it's written PotonSui he cod and it's it's fairly easy to follow. There's a bunch of otherresources on that repository. That will help you understand, bette. Perhaps therationile of some designd choices that were made for the Beacenchan all right. Well, eyercime betty. Iappreciate that in H, let's cross their fingers and hope fora successful watch. Thanks for having me and Yeh, I thinkit's CROSSD...

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (108)