Hashing It Out
Hashing It Out

Episode 22 · 3 years ago

Hashing It Out #22: Casper+Sharding - Danny Ryan

ABOUT THIS EPISODE

Danny Ryan from Ethereum comes on to speak with us about the Casper Sharding v2.1 specification. Casper Sharding is a combined effort between the migration to Proof of Stake and Sharding for Ethereum. We go over the motivations behind the new specification, the research efforts underway, the architecture and design, and even how Ethereum can maintain liveness during a major world catastrophic event. There's so much information to cover, and we only scratched the surface. The future of Ethereum looks exciting!

Injywelcome to hashing it out poast,for we talk to the tech intevators behind block in introstructure anddecentralize tatonds. We dive into the weed toget at Wyand. Help e build thistechnology. The problems they face along the way come, listen and learnfrom the best in the business. You can join our REC episode of twenty two passing it outsay: Helo Callan, you hear with me todays of always hello, Colin, I'm herewith you today, as always as always Um. So we tried a couple of episodes backto go through the castforplush sharning speck that has been recnt a release fora theorium, try and figure out exactly howthe infrastructure, how the baseblock chain changes with Sharns, and we weren't terribly happy with ourunderstanding at the end of that. So in order to fix that and maybe help youout, because you possibly weren't terribly happy with our with ourconclusion, we brought Danny Ryonon to try and help rock through all of thestuff complications of it. worts going how far it's going and GE Om allclarity to the situation. SAMES UP TAI Ey how's it going I'm happy to be hereUm. I spend a lot of time ironing out the details of Thi Speckworking, onimplementations, O epect, helping other teams understand thistect and kind of coordinating a lot of the development around it m. Sohopefully I'll be able to answer yau question yeah. So actually, I think agood place to start is since some of Hem are liscers haven't heard. Theprevious episode tell us what this is and what are itsgoals, and where did this come from? It felt like it, kindof dropped out ofnowhere and then ite was like. Oh, we could do this new thing and it actuallymakes more sense if we just did all this together. So explaining me thebought process that led us to creating the SPEC in the first place and what itis ye h absolutely so we had two really in a main scaling efforts going on Umfrom a research perspective and and beginning to the inplimationimplementation perspective. This was the Casper FFG and the sharting Um Cashrap, F G was going to give us Croupof stake on the existing AF therean botching and sharting was going to giveus Um cale ability gains by breaking kind ofnew layers, wof the block chain up into a different short Um. Both of these efforts involvedvalidators. Both of these efforts involved M reward teams, penalty schemeand both these efforts at the time we're going to be utilizing systemlevel contracts on the existing evum Um. So I was spending a lot of my time,developing the FFG contract and working with the team that was pomably vericingthe contract and building out the EIP to Spetifi the whole thing and workingwith people t as they were building it and we were, we were making making alot of Parger UH. On the other hand, other other membersof the team and the community were working on the sharding managercontract, the SNC, which was going to be doing a lot of the sape things, butfor managing the new sharp h. There are a lot of reasons that weditch these two efforts: Um for the new design, Um think three three or earlythe TALP relevant one um one is that processing crypographic fignatures inevium suck the EVM is very efficient at certain things. Wella lot of things H and proing finitures was going to be amajor bottleneck, so signatureis in the sense of I'm a validator, and I want toyou know, sign a message saying. I believe this is the canonical chain, or this is a recent block cash from theShard Um. If I'm going to put that message intoan EBM contract, I have to process the transaction like normal andthen also Valida a pretty complex signature in the BM. This is going tobe a major major bottle, nextfor both of these system. So much, though, thatyou were seeing that really high um ether stak requirement, so we're we'reon the order of a thousand and fifteen hundred eath to be a castor validator.This almost entirely was because of the limitations of being able to processsignatures in the evium. So, by putting that high cap, we ensured that if allthe eath in the world stake, we'd still...

...be able to protess the signatures intime, um so signatures were signatures are a bigdeal. This new peakin chain of Lomitation we get to alol boo, moreuses a different inshrime signiture scheme that allow aggregation off chainand because we're pulling it out of the BNLLOWS for efficient processing onthese aggerate ignitures Um. Another reason is when we had, we had casper of the system lovelprotocol and we had charting a esistalevel Potocole. We hadtwo really kind of competing games in a sense of like a economic game thatthese players can come in and play, and with that we had, we had bastalydifferent requirements for each game and, at the same time, different different reward schemes.Different penalties means, depending on what the requirements o the name were,and so if we had two competing games, the worry wasthat we might ene get large acymmetries between the two. Where a you know, ifit's easier from a systemt requirement or the reward schedules a little bitbetter of one of these games, we might get way too many people validating thecort protocol and not eenough Thso, where we get high shart security.But the corrotocol is not that secure because of the a cymmetry and then itit it kind of unnecessary complications and there's a lot of work and a lot oforganizational stuff going on on both Sade Um, the third and and one that's a littlebit less talked about and something that les talked about, because it'ssomething that came up in my work Um and was relevant to me being gung hoabout the position, but wasn't something that had really been out outin the community very much, and that was that the caterog required that castervotetransactions could be processed in parallel with existing normal bok transaction. Thispremise at the CASPOR contract was written, was lad and we were going tohave to require a major reway, rewrite kind of not the bulk of the castor contract n. TheAm but is Gan rpire a lot of surgery. Um moving things around hiding certainthings changing when certain things happen, she'll allow for this votetransaction trelization, and with that there was going to be a lot of the form of arrification workthat had that had gone over the past four months was going to have to bepretty much completed, reworks nd, so that was that was going to set back theFG effort Um on the order of at least four months, which at that point you push that back and taking all these other things intoaccount for the reasons that we might want to switch. All of a sudden. It'sreally starting to look like if we have a better design lets to go with thebetter design. Let get this thing out: Let's do it right Um, you know. Do it right? First, do some short termshort term pans for a ultimately mega long term game. So, if you had to Geta,you had to guess like I'd, say the Vindiagram of overlapped work thatYouyou get to use from both projects. Cowhat percentage is that across boththose projects, how much? How much of the stuff do you actually get to reuse major portion, so ther was overlap in the project inthe sense of we're managing we're both both of the commoist of the project ARmanaging valadaors and having to do things with haladeirs and Bondervalidators Um. Both of these projects were being worked on as corprotocalprojects that were being built as UM protocol level contract in the Abium.Neither of neither of now the current effort is not using a protocol levelcontracts for the managing of things, because it's quite frankly, tooinefficient and w want to break clean from the EBM. So in terms of research tons, everythingthat we've done in the past is informano design in terms ofdevelopment. This is totally new develoment Soth. This is this makes a lot of sense, and I'm gladyou guys want with this approach. So in order to sort of like get the communityinvolved in this, the person you did was drop out at two point: one: a Tobl,Zeo Speck for the caster plus sharting Um Initiative. I don't know what to callit at this point and uh, and you know a lot of us kind of wentthrough and started to read it, and there were some Um questions that wehad and things that were raised of and Tel tells what what kind of thingshappens between two point. Oh and two...

...point, one that we might want to take.A particular note on. I think one of the things that we particularly pointedout when we co- and I were discussing it- was concept of a proposer in theamount of power that the proposer has and I'm kind o curious. You know howthe evolution has come from two point, oh to two point, one rigiin today andit's funny as you asked that question I'm like what was two point. Oh, whendid we change the name to two pint? One there's been some I'd say. I think thatone of the big things between two O and two one was the refinement of thedifferent roles in the system. We have pin the Deacon chain, which we haven'treally talkd much about what this thing is I, as lose to take a step at Likea, abirds, Ey view of the architecture of what this is right. Okay, so we have, we have a block chain. Wecurrently have an EBM block chain. It's a Prouf Ho work, blatching it it'spretty good it. You know it does what it it does, what it does. It has someissues with the evm in terms of efficiency. We want to move to es andit has an existing proup o work architecture h, but what it doesn't do is it I it's kind ofit. It's a plane and flight. There's a lot going on. There's pe tons of peopleare building on this thing. Tons people are relying on this thing for real realword stuff. So what the Deacon chain does is. It allows us to cleanly break freeof the constraints of the existing evium to buildin parallel. It's kind ofa conservative approach. Het allows us to kind o. We have the architecture ofthe existing system and allows us to kind of build this new component of thearchitecture in parallel Git the new component kind of stable and startadding to it, adding the shards, adding more functionality and then eventuallytake the existing edium and kind of roll it into this new sharter than toUm. So the the beacon chain really is the the core system level chain, likesystem protocol level chain of the new sharting system. It's where thevalidators exist. It's where the validators do they finalize things it'swhere they get organized to do all their duties on both the Deacon chainand on the new shard chain Um, the general architecture there is, it be, can chanase of blatching thereare proposers that propose blocks and which is a subsed which is on theseduring these cycles to the subsede of the total valot ierstep that proposedblock during a cycle. That's kind of the unit of of time, like a block oftime in this in the system, during a cycle, all salitators get to act as a testersto these blocks. To then finalize them in the sense of proof, osake finalityUm, it's Pard to kind of HARDF. The core caster thing is roting on things,comin consensus and finalizing points in the chain, so the beacon chain doesThi um four validators. At the same time,these validators play a game. The game is still a little bit. The design isstill there, so ti being worked one, but they gaers play a game to alsobuild Um, to create randomness, to create a a source of a source ofrandomness. For this whole of this whole system, um without proup of work,is interesting and that it kind of th the process really does kind of createits own randomness. In terms of like who, the preticipants are how they getto play Um with group of stake, you don't really have that extra protocolsource of Entrope, so we have to design an R ang, a random number generatorinto the protocol that allows us to Um parket orchestrate where the valtersare and what their doties Ar aret any giving time. So so, there's AK caster mechanism forfinality. There's UM creating a random number for the organizational componentUm and then using this random number. We organized the VATUES into theproposers and tattesters, but we also organize the validators across short.So if I'm Vad ator zero, it might be my duty at this current time to bebuilding charde h. One hundred and the rand g is going to slowly shuffle meand my duties around and everyone around such tat. The system was kind oflike low, balanced and UH controlled controlled in that way, so the beacon chain back to really whatthe PEACON chain is. It is it's the core: it's the core, intrestructureforeast Twoo, and it allows us to uh by kind of breaking three from theoriginal architecture. It allows us to...

...be really aggressive and take all these new ideas and implementthem and a componer of the protocol that live kind of in the same world,but slightly parallel to the existing protocal and allows us to go in totallyarchitec totally create new consenses and then from there totally recreatewhat what Um the execution layer chards are and andhow to communicate and at the same time the BM can kind of exist in it's in itsown world and only when we've reached kind of a stable place figure out howwe want to Lope muvium back on. So we interviewed Um. We interviewed decenity well back andI'm assuming that, based on the current Spector be read, the randomness is donethrough Bel s, signatures, which is what I he does as well, and it was aninteresting conclusion that we came from that interview. was that thepurpose they're doing it is your separating stake from randomness, whichallows you to be very um open and how you innovate. The actual state 'causewith proof of work. The randomness that you that you create is dependent uponthe stake updates of each block right,especially the final Hash and all the transactsion includ in single block,whereas the elastico Cos are always going to be randomlyt deterministic, but don'tdepend on the actual state being updated. The state being updated, getstvalidated in the pens on the the metial Randomnane of the beacon chain,correct, so yeah there's an issue right if you Ersaywere to use a prof work,block Hash, a a source of strong, randomness, um or or state yeah th the state s as asorts rendod. This gives a large grinding opportunity, sosomebody who might be able to profit off of this randomnes going in certaindirection. If I'm a block proposer- and I just want to change- I I can you know- grind on my pr. If I have a largeenough mining pool, I could grind, I could try to. You know, make a block.That's not quite the has show out. Let me make another block an me makeanother buck, and I you now. I might miss some blockor word there, but Imight have the opportunity to really h manifellike things, the other anemissso important. Our system is not using so where you can be able to aggurgatesignatures or dige signature aggregation, and we havemassive gains there in terms of the minimizing the amount of e required tostake and th maximizing the ofvevalites Li participate, but weare not usingtelles Threshel Chreshold teneratures for RMG um, esrase still randal, so that this is for the implementers. Currently there'sa number of teams in blunning Orangey we're just black woxing right now, Um,so they can, they can just assume they have a. They have an rg and build thesystem. Um Justin Drake is spending he's aresearcher on the FTM is spending a ton of time and resources and effortworking on the RN design Um. The current direction of that is to useUm Randal has a weak source of intrepe and make it a stronger, lessminiputable source mentry by larring of VDF. On top of it, which is averifiable delay function. I'm sure you've seen a ton of stuff about thatand we can get into that. I'm sure you've seen tons of murmurs ofvdfs and what are these thingo, these new random tools that might requireweird hardware and stuff Um. There there's a reason that we're shying awayfrom cocinity esign and one of one of the design requirements of Ither in two point O is that it can survive world wortre, actal, ture efirement. They can surviveworld wor three and they can survive or it's able to change when it it's able to bequantum secure in the five or ten year time, risin Um, teas, stressold signatures H are not they're, not quate und secure, and they require a inprotocol thresholdamount of valaders to be online, and so I think t fifty percent is their number.So if for some you know if, if the case of the networkis majorly partitioned world ware three, they can't then create their R andgtheir system halt, and they you know they probably hard work and coordinatearound that, but they, you know they do lose liveness Umin that. In that scenario, that is a very awesome set of Cols, so that Isavitious, but it's reasonablly tambitiously. We are designing the next stage of...

...trust, and if you can't trust thenetwork will survive major catastrophe, then it's not resilant enough toactually mean employed. So that's kind of the interesting things t Ipersonally thought. Tothin Yo would be a good layer, two solution, um ratherthan a you know, havindepinity their Dunkaron superofsom project anddifferent different design. Gols and you know different trade offs in termsof their RNG, we're just going down a different H, different rabbit hole anddesigning our remgy. But we come to the same conclusion like you have to have astrong in protocol rng to Denineng systems appropriately. So before wemove on and get a little deeper into the beacen chain, I'd like to then movea little bit more into the architecture of the shards, and I work currentlywork for consensus. A I work for status, sorry and doing security as status and as as wellknown. We recently made a a sharning. I amentation called membus.Yes, O that t been working quite a bit on on trying to implement this thispeck,and I asked the guse if hey had any specific questions for you, the mainone was like with Wi the beacon chain, taking up so much of a mindshare ofdevelopment ats. The plan for the rest of the shining up pro structure than hese put in place right so um you can change, can be rolled own n Tas,which is Um exciting from a kind of iterativedevelopment perspective. But those first couple of phases: Don't havedon't even have charging so they can chan can exist. The valitars can beorganized, do their duties, kind of stimulate, finalizing Shor, even thoughthe chards don't exist and finalizing the decon chain and creating Orangey,but that can all kind of like a Phas zero. Where we've gotten the core proupof steak architecture and Orangy in place and from there we can add the thecharge Chan Um. The next phase is probably going to be adding some amount of Shard chain Um to thethe deacon chain, the vadeitor duties these sharchains.Actually, at that point, won't even have state execution, so they'll bewhat we've been calling Datablov we've in this design, ar moving towards thedecoupling of the data layer and the state execution layer. So first, let'scome to consensus on data. Then we can come to consensus on theexecution and state of that data that that th that data brings us to so in terms of the architecture, then wecan have um charge chains that are just change ofdata. Uh valutures begin building these thingsthey stuff the data they can do cwo one or two things: Allball Shar, bloccb N,the same size, so they can stuff these blocksfull of H, Crap Zeros, whateverthey want, or there might be a secondary market in terms of utilizingthe data layer of thes Chare, an for maybe say a decontralized twitter orsomething so somebody could h begin to pay for the utilization of thisConsemsa, tate layer, then- and I I mix up the saes. I don'tthink that ASE members are truly that meaningful, but I think that N andabout a phase too, we bring in a state execution. So we bring in a state astate machine like like VM, but instead we're moving towards Um, ewis andconstruction. So a lot of the same goals there, but UTILIZ is a lot moreefficient underlying architecture Um. So then we bring in tat execution atthis point. This is where things get interesting Um. This is where you startactually having h what looks like in our mind, a functional block Han insense that we can resolve the the resulting state of bundles oftransactions, and at this point, when you start consultating that CrosssharCommunication Um, although the focus right now um ofthese implementing teams, is on the Deacon chainge, because we're going tobe rolling thing, O Ta Thereis, definitely there's a ton of researchgoing on on the SAS one face to and actually the thegeneral design of of Dase, one in terms of data blobs and committing sensis ondata and dat availability, proofs and things like that. The this design hasn't actually changedsince early this early this year, um the that base one faith, two of thedesign existed with the sharting managercontract, so the Sharg manager contract in the previous time was really kind oflike what the decon chain does now m or serve some of the same role and that itorganizs kind of the Coro frestructure, but the e actual shards that we'regoing to be existing on the trate manager contractore going to operate ina very similar way to what's going o n and what's going to happen in pase onepace to so we do we're feeling...

...a lot more and more confident about thestability of the DECINTA design and the fact that RPAs one data data kind ofdata sharting is has been so stable for a long time. Um We're gaing go acomplidyt about that in terms of the face too and bringing the executionlayer ut the there's a ton of really active research in terms of that execution, Tela Pate execution,ewam, crosshar communication. This is superactive discussion on each research,but there's actually now a team. I think a a subsection of thetheorium J team and maybe I think, th Y- They KINDOF- have some crossover withmaybe teterm JS teams, or I can't remember, but they have a lot of crossover with the UASN tem and they are working on Blackboxing, the Ba zero and even thephase, one of charting and they're they're working on a prototype of paseto that would be okay. We can assume likethey'rethey're they're. Turning from the other side, they're saying, okay,assume, we have consensus on data assume we have sharchain with data,let's start building the execution Leyr, and so in that sense I imagine members of the Nembers teamand members of some of these other Beecan Chan Te to teams they probablyth y. They probably feel like they have a little bit of blinders on andplusthey're, like really digging into the research about the future stateexecution stuff, but it is definitely actively being worked on Um andsomething that I'd like to see over the coming handful of months really have alot of progress and start to get a lot more clarity on that AFSO IG. I got onemore piece to this: This bird's eview puzzle of what the architecture lookslike, and that is the transition mechanism from the current evm to thebeacon chain n how that works for both underlying ether assets, as well asmaybe smart contract in that state, correctly, Pi ads in tha, PostoRalequippenos, something I notice that you're still requiring to stake in theproop of Workchain, and how does that really tell this and howdoes that kindof work out right right? So I'LL ANSWER COT question? First andthen we can talk about the general kind of difference, EU systems, how they relacto each other now in the mid term and and maybe in the future, so uh currently no eather exists in the easttwo outside of the protocol. All the ETER exists in Avium. We know that hjust a fact, so we need a way to open up this new version of the Protocole,this new component of the protocal and in doing that, similarly, to how we'regoing to have this like system level, contract of the FF G or the M C we'rejust going to have a system level registration contract where itrobalygoing to have one function register it's going to take my initializationvalues of being a validator Um. You know my Bila public key, the thirty two e Um, a withdrawal address etcetera, and it'sjust going to be. It's GONTA. Take in that information, Takein the tiry TeSalidate, some of the MPO and then broadcast Eris, Um er, a log and and doing that the deacon chain is going to be a alight. Client validator. So the beakon chain needs to at least know about h.The Sharg Chaine from a light client perspective and need be able to readthis read and know about those receipts and those receipts. Are they going tobe used to induct vality into the the DECON Chan? And so there's there'sdirectionality there there's and to note the withdrawaa address of avalidator is to be to a Shard Um, and so this mechanism there's adirectionality. We have these two systems living in parallel, but theethe can go from system. One Te system too, and notnecessarily back at least on the short comedium terms, Um, and so that's whyyou do have this. This contract still coming into play. Um and C. Your question was more on the UM anew Tom any you're. Basically answer like what wasah you kind of answered in a way, and one is like how do we get Eser from themaeither? It's also s up in workless current state of the evium into the newsharts. Yes, so actually answer our question on each research about the theEVMSETO thing, so the validator h, that's the directionality, an kind ofthe validation and that, if I'm a validator, I can move e out of EAP on Oand you sewo. But there's a question you researched this morning. That wasokay. We can move thirtymove thry to...

ECS chunks this way. But what about? IfI have less than that, or I don't want to be a valilator, how do I move myethroever M and I imagine probably after the the Deacon after the phasederolaunch Um uh, the dcan Chan, we should someone shouldride an Ip that proposes a a different system. Ovil contract. That's just ayou know, one way, a one way: Teposit one way: Transfer from Ech one O of esfrom each one O to eat too. So. Similarly, then it would broadcast thareceipt and instead of and it would have a destination, Shar anddestination address, and instead of being picked up by the Corse System,Becan chame protocol, it would be picked up as a transaction on thesesharts, and so you would need that save to so you need estate execution on theShard to really have that eter over there, so that that's going to be alittle bit further, that one and hit's the same thing for contracts an thereand ther and their stay s well, uh contracts a little bit complicated in the what I haven't spent too much time inthinking about actually porting state. I can imagine I I can imagine taking asnapshot of a given State and then I can associate an contract on the new,whatever shart it exists, and then transferring that state over and thenullifying the original state on the DM right. It's hat's like n right yeah,because who, who who have the right to do that O wery owner 'cause? That'swhatever contract is basically but well. We can't apper AO. I don't Kno.You know just like mechanism that people just KINDOF use will get us overUm. In these short term, the state of etOneo exists, independent of the state of Eto Nasterm, probably as well in thelong term. I would like to see one or two think two things happen, um thesecond of which I'm wor excited about. The first. Is You once you have astable, shart edlandscape with state execution, you roll the EBM and haveyou know the the lashard Um as a exceptional chart? It you know if I'mgoing to be, if I get organized to be a validator on what becomes Tho evm share,then I have to Um. Do things in the EBM way I have to run VM execution andprofesband transaction, but I could live in the same landscape itss alittle bit dirty in the sense that we still have to th n. We we still haveAnvm, but now we have uwasn't to U and it's Gonto, be you know in terms oflong term, managing the software- It's not probably not ideal. I would like tobreak free, ultimately Um. The second I haven't sent too much time thinkingabout it but B t. This is a little bit more more of an exciting idea for me:'CAUSE, it Cllase us to cleanly brick grieve. The architecture F, the BM andthe idea is to take the entire state of the existing evium and drop it into a contract on a ShartUH. I can't say: WAAM is a beautiful,really cool thing and you can do a lot of Tran filing between state anddifferent things, so this is theoretically possible. I can't saythat I've tully vetted the complexity of doing this, but it it definitely ahandful of researchers of the time are like more excited about the notion of working, the entire evm current evmstate into a contract on a single shard than to rolling rolling the rolling itin as it son charging Um to somament on a single Sharg. So BloodShar would be Legacishar, basically well right. So if you, if you, if theBM became its own chart, it would be a legacy shart, but I might just take theevm drop it in to Sharg, zero or Shar, onethousand and twenty three at a certain h address, and so then you have to justmake an interface to hit this contract so uh the eating aeties o the talkpoint,but that address would have beenthen be at attackpoint because he could maybemanibulate the contract just tomlar. I mean it feels like if someode gaincontrol of that particular um ability to WHO's, deploying that Adrabasially Deplaine that contract like O. No, that would be an that, would be ahard for. If you would fork in, you would take a scamp shotiy like thecommunity, would decide to take a snapshot of the existing vm chain andto place it into a contract on a shart chain. Mov A coon IILOPA in a regularstate chaine to change to do that. Similarly, to deploying like a the AFFC contract or inthese system, oel contracts, it would happen out of pork.Toud be an a regular state change. It would put it out a particular address,um B, there's a lot of weird design, things o think about like what's theinterface to this contract, and how do you golidate things? I mean it's not...

...yeah like I feel like if somebodystumbles upon the key for that that te that would own that contract, like yeahSoow, a they o Chuck that Alli or Research Asi. What we're saying howityeah, if no, I temble upon that key. If it's that address one, no one has thekey to address one. No one ever will have the keyd address one cryptography.Our entire system is based upon the fact that you can't arbitrarily getAyto to the Randal. That's right: your new system's can be qualte of secure,correct, so will the Poun Bee Quato Secure Um? We want to move in the reeses toallow Quanta Wol for things to Deqan, secure, Um the signature scheme when we have the everything is kind ofdesigned around the idea that a lot of these signitures can be wrapped up intoH, stark friendly Hashes, which are Ponsiner on the order of the five orten years Um gothe current, am it's being dedign in such a way that itcan be replaced very, very seemlessly with on hem to crover them Um. The starks are not uh on the short term.Th they're actually becoming a lot more efficient, um and in terms of size, but in terms ofjust the ability to build them and to use them we're still h. You know threeto five years from really getting a dall attraction on that, but but in notfive year time rising, the goal is to definitely switch out a lot of thesecomponents, wher a more startonpoe. I was, I think, one of those things thatI was let fuzzy on. When I first looked over this eees and other people had thesame thing with like was a number of sharts dynamic and if not, how is thatdetermined M in the first place right Um, so the current SBACK has onethousand and twenty four. This is a function of the amount oftime. Tho Shards have to be cross linked pack into the Beacon Chane, sorecent references from Sharg chains are brought into the decon chain to befinalized Wen a recent reference from a Shar chain. It's pinalized in theDeacon chain that serves is the basis of its pork choice, rule and so thefour choice rule of all the Shart changs is has been premise and linkedinto the decon Ching Um, the sorry missd, I totally dropd. Thequestion is Namic Iswa reqired, a Packa te BEAter, not dynamic the function e, the function of the soap. We have to sign messages about theseShar chains and it takes a minimum number of valitators, depending on thetotal validators, to bring these messages back in, and so if we have atheoretical maximum valiulators, which is a function of that thirty. Two eastdeposit and the total et out in play. We have now the maxium numbersignatures. We hat the process, which is also going to be a function of atotal number of charge. So we've limited the number of shards that onethousand and twenty four so that, if all east participated in thismechanism, we would still have enough time to process fignatures Um. It isnot dynamic. It is something that, in the future you could add to it would behard pork, if that's something thit the community wanted to do and somethingthat we hadvebted as possible in terms of pastting enough figutures. What kind of throuput does that? AllowI mean withthat, withthat Um haven't done any sort of like what isthe maximal scaling? How many transactions do we think we'll be ableto get through with this, giving a certain blocksize just ice in? You knowyou're looking at that on the order of thousand XS, but you're, also replacingthe the virtual machine so assuming you're doing similar work and you as amachine d you're also looking at some multiple game, so you're looking atthat thousand extimes, whatever efficiente games you're getting fromthe the AZA machine, you also get a little bit of Gams in that Uh. The timing of events are Um more in lockd up because of the PEOstate meature nature. We don't have that puaon process anymore and sothere'sthere's, some marginal gans in moving to the proof stake as wellthey're on the order of one to some multiple of one thousand X, N Wel. Sowe're talking about like millions of transactions in a SOTIM, five hundred thousand right now Sabbe,but um five hundred millions e Comin SariPerda. What you're saying per day Herdan a! I remember last year at Devcon Tree Um, when thetolic was giving his shartingtalker like of the future of the road...

...map to the SYREA. They talked aboutShaes and Eintedil shores, as kind of being the testinet for future upgradesthat could be rolled back in the so each endiple shark kind of hasit like a on rule set associated with it thattenfeeds back into a big chance, that's something that's being scrapped or oror re Y. U Planning on all shors having the same core architecture, excludingmaybe the legacy Shart. If you do end up doing that, yeah Um samearchitecture, the complexities that arised from heteoogeniuts sharting isjust massive Um and dealing with different rule sets and and thecontensus on all these different shards and moving validators acrossyhard Um.That is not h. The test bed for layer, one and the andthe long run is probably going to be later too. You know if there's someeradical often thing that's happening in terms of things that we want tointegrate. Ind The layer, one, it's probably gonna after sharting is rolledout. It's probably going to happen in later too, and maybe we'll stay inlator two. But if it's something that it's so good that we want to integrate,then it would probably move in o latter one rather than using. You know, Shard,one thousand and twenty four as a test et or one thousand to the test. Ten,it's kind of 's kind of a fun idea like we could imagine the functional showerthat, like I completely functional and like people use for certain ways or youknow, but it it, the complexities are just it. It becomes quickly a prettyintractable problem, dealing with the consensus on tes thing, Calln the question I'm trying to like figureout what what has been said and what hasn't. So I don't Wantta I've got likehighlights all through the the farting, but I the beauty of just plained. Byear you don't have to worry about hat on, though alred I go through, try tofigure out so got over the validators in the Cocte, it Sasbin the main poschain C pw, cn Um. You know h. How do you so? I think one of theproblems that we originally saw was that proposals had a great deal ofpower in the previous system, meaning that they could actually seem likethere's ways that the proposers culd influence things, and I know that wasaddressed I'l. sorry. If maybe you can talk to me about how you wadress thatwat? What kind of proposer issues you noticed in M, you know rig. You knowwhat what the changes were there, because at was one of my meain concernswhen I read the Topono Peck and I monitored Oug Research Enough Othi sawthat and addressed it right, so proposers previously Um the problem. The problem, the mainproblem becomes a proposer of the peacan chain is the player in therandomnes game. So, during a cycle which, right now e cychois sixty fourblocks, we have show previously chosen. Sixty four proposers they're going toplay this randal game M, where they're going to be essentially revealingrandom seed that they previously committed to and the m the randomness for a future game isthen going to be dictated by the axor of these revealed. JEEDS UM. There is ends up being a pretty massiveproblem here and that if I'm the last proposer, I can decide h what I can look at. Whatthe previous proposers reveals were, and I can then decide to reveal or notso I don't. I can't just make up a number B'cause. I probly committed tomy number, but I coul decide to show up or not and that's going to influence,Ranima no Om. The last two proposers I now have two options to do that Um lastfour I have more so and then, if I can, if I can manipulatethe random, is now might be able to continue to Mi I fee, like the're aneosin the future, and if I can manipulate the randomness of this system enough, Ican now m maybe allocate a lot of my resources to a single Shar and a singleshard has only you know, one thou one, thousands of the validators, and soit's a lot easier to attack and mot. I usually make control of Um and so hardening up the heardening upthe randomness from the Basic Rand Al Team. I s very important, but also- andso we've been moving in the VD F direction- to harden up randal h butanother another thing that the tow clinning a lot of time on is okay.Let's say that somebody can get hold of this. No canby that latacoseor can likemanipulate the randomness to a certain extent Um. How can we prevent h? How can we reduce their ability to doso and reduce the effects they have um and so you've seen? I think it wascalled. It was called R PJ, which is Fourth Choi proposed, fork, choice role.It's recently been called change. I M The ARCA J, which Recurseo Proximatyjusffication Um img is immediate...

...message driven. Those are the samethings if you see these emrality, it's actually just a renaming Um. So thissport choice rule is to try to prevent the power that any proposer can have, andalso the power of proposers being able to kind of h run shortly, in short,ranged, forks and change the beacon chain protocol. I changed the peakinchange kind of Porchoice, and so the design goal is the current fork. Choice should be a good predictor of thefuture forcoic an then I a block right now is included in thecononical chain, as as people see it, it should in all in in most scenario,most likelihood be included in a future version of the portrois t's, a that'sstability, and so this h, the sport choice rule, is designed forced abilityand ends up being a lot harder for um proposers to kind of bypass each other.And if I have a string of proposer, try to manipulate the por choice and try tomanipulate the the bandofet Um. So that's one of the components is thispork choice rule another component is the UM vdif hardening of Randaw, which isstill definitely up for debate Um, and I know there were some other designdecisions around h. The separation of concerns between what all the power theVALITER has versus the committee and a lot of that have been Um. The combination of t, a sport choicerule combined with an assumed honest majority, ancommittees which are the subsection of validators thathve, been attest tobluck. I've taken a lot of the power out of the proposer's hands, Um andactually the you know, the fifty percent. Honest Committee is not acrazy assumption in that H, if we assume, through third, honest majority and ortwo two third, even rational majority, in our- U valid or set, then these thesecommittees are small samplings of at least I think, a hundred, an twentyeight Galidators of the the total vaoti Er set h you easily get to that. Halfthat fifty percent in the committee, like you, know a billions of a chance or something orwanted a billion chance, or something for to not have that six percent in thecommittee so again removing power from kind of like shuffling in these tweiksremoving power from the indirisual proposer and kind of shifting it towardUm are more Li and likely more likely, honest, larger committee. So let's talkabout war war, three, then so in the event of a major. More! Uwhich I mean hate to be an ative Nancy here. It would be really unlikely thatin all human history o like his testnixists, I think it's interestingthat you guys are definitely designing towards that. The Internet will be cut off betweencertain nation states. Now, if one of those nation states has a significantamount of state death in the validation scheme that could allow for other nationsates to sort ofmanipulate or the rest of the Worldsmn negilate what's going on in the actual theirversion of the chain itself, how would we bring things back together in theevents that their proposers have more power than the the nation seats? Proposers have morepower than another nation states proposers and can nipulate things soitsr t so in yeah partitions. When we talk about well war? Three: U We're talkingabout Um being able to survive from Majorin longterm network partition and also to be able to survive h having to be alive inthe sense that, even if there's just one validator left on the chain cankeep can keep nomping like one one Dolta left in their own partition. Youcan keep building the chain Um. This is solved primarily through. Oneis to not have any component of the system that requires some h, some percentage of valitors to beonline for Livenan, so the the chain can continue forward. So that's whywe've gone a differn direction for the the orangey another. Another thing that that saves us is we.We need two thirds of validators participating in the consensus tofinalize the chain H, but on the order of I know it was recently changed in thistack. I wonder what this tind. I E E look s real good...

Oka, so it's currently Stt tooapproximately twelve days and that if a chain split happens for Um, twelve or greater days, the we have twosplits say. Fifty fifty on both sides, their ghost they can both still bealive. They can both still build a chain. They can bosth still do whatthey need to do, except they cannot finalize, but we have this accinentialbleedout or drop off of offline validators. So from the rerspective ofsplit a you know, we have the online fifty percent validators. On the Orderof twelve days, they will become the two thirds majority and beginfinalizing again, O on and on the same side, O on the on Split B. We have thesame thing happening where uh, the other fifty percent they'regoing to continue to guild the chain and to continue to pross estransactions and Continu Tof Validat and the offline H. fifty percent fromtheir perspective, will bleed out on the order of twelve days and they'llbegin to. Finally so World War, three happens: We have a massive networkpartition and we have validators on both sides.If the pretition resolves in less thantwelve days Um, these vallators can then build. Theycan rectify and build on the same chain. You might have this kind of th someinteresting things that happened that might happen there. They might actuallycurpse to pork out h what happened? R, there's all sorts of crazy things thatcan happen, but if they do nothing, the chain will resolve and be one chame. Iftwelve days pass. The chains will now have finalize H, separate histories and will now do two differentshames Um. In that case, a year passes the network partition ends. We have, wenow have two of their engine Um and from the Etheorim Devin decisions thatis th. That was the expected scenario is to have two of therimchangbecause wee prioritizing liveness Um. Otherwise you have the chain go down after twelve days. Imean you, have the chain go down immediately because you don't haveliveness and you probably fork anyway and probably have two Chan anywhere, and so this is kind of an in inprotocal meghnism that allows for liveness in the case of that Werpectition Lt'slet's bring that scope back a little bit and look at what itmeans to be a single validator and right now I guess Heaven thenest put, Ithink in is in this week in aserium that I think it takes around sixthousand dollars currently to become a valinator on the beacon chain. That'sgone down a little bit or it'll, be you know, depending on the price? I it'sit's reasonable right, but one of the resources and like like liveness? URequirements for a single validator and ote or the consequences to say, like Ihave a power outage, am I then my valet or chet off for a certain amount oftime right right? So a really cool thing about t design isthat we've moved to fixed size depositsrather than a minimum deposit. So it doesn't you can't you don't play withthirty two eath or more you play with thirty two egremen. If I'm a Validator,I am thirty two EE. If I am two vallators, if I want to play with morethan thery Y, I can do two vallaear since I was sixty four and so on and soforth, Um. But we've done this for a number of reasons.Um Internally, it's actually a lot easier in terms of accounting in termsof organizational things. If I have, if I just have a list of Ilicialilators in the protocol, I can just move them across the protocol andassume that wherever I put them, theyre providing the same amount of securioweight, a s, a waiting, associate it a lot o thing. You have to do correct,correct to make EITER omx perspective, but it also has a really nice propertyin that the h resources required to validate scalelinearly with the amount of Ech I participate with. So if I say avalidators on Um Avalidator, it's required to bebuilding at any time on the order of one to to charge chane. If I am now, two valadators require tobe building on the order. Toporshart chain Um there, there's kind of theycan have a little bit f overlap in that shuffling. You can just think of it asone, so a validator is always rsponsible for one charge chane. If Iwant to play with sixty fourty, I'm not responsible at any given time for tooShart Tam, but my resources that are required or scaling on nearly, and so I'm also by adding more e that Iwant to play with. I have to add more resources to help out than ot work tobe. You know- maybe that's maybe now I have two separate notes, or I have onenoe that has more resources or Um,...

...and so that ends up being kind of this NiceLennier scale property that we without taking that into account you, you kindof lose like with with provof work, the more you're participating adding to thenetwork, the more resources you kindo have to. You have to add morecompetitional resources and probably notworking resources, O t the nor M andtil. We get that property by scaling, intereally Um. Now what happens to those are ourrequirements? I need for every Valadagor. I need on the order of sresources where s is like the standard resources in a computer um in aconsumer computer, and so, if I am validating with H, you know a thousand times. Thirty, twothirty, two thousand an eight. I need t on the order of a thousand cee so onthe order of a thousand sandard computer resources. Um. Now what happens? If I go upine Um. IfI go offline, when I participate in the protocol, IGane I slowly gain reward Um when I go offline or if I'm censored. U Can'ttell the difference from in protocol standpoint, I slowly lose reward. If I'm off, if I'm offline- and I finalityis not occurring, then I I slowy Ros those rewards and it ramps up overtime.So I, if I go offline for a long time, I'm I'm losing a little bit AFM livinga little bit and I kind of hit that curve that xential CURV and I bleet outpretty much entirely. So if you have a machine at home and it goes offline foran hour- you're- probably fine, youhave, a machine at home- A it goes off linefor six days or a month, you're in trouble, you're going to start losing alot of money, Um and so in terms of being a Valey losing a lot of money interms of what my deposit is or losing alot of money in terms of potentialgains. No in in terms of losing your depositors, slashing at this point, you're not being so flashing, is you'vedone something very nefarious and we take your money and adject you from theValot ator, so Um being offline is a lot slower of ableed, but but it is taking away from the depositand not PTOTA gains like. If I, if I like turn my computer off and walk away,I lose whatever 's a part of that state, instead of instead of like having allthe money that associated with with Thi with that Valitator, when I turn itoright and if a high portion of the network is still validating andfinalizing you're losing a lot less money than if, like fifty percent goesoffline at one Um, but you are losing money overtime bynot being online. So there's a lot of consideration. You know what, if a good,vality or set up wit like does it look like one computer at my house? Maybe ifyou have one valiator and you you know you E G,Yo 're you're around a lot and you can you monitor it, but maybe maybe you'reset up looks more like I have green machines in different locations andthese machines have you know c talk to each other and come to a consensus forethey find anything. So you know. Maybe I need two of three ofmy machines to Asigne messages and then they brought like to agree and thenthey broadcast and if one of them goes offline and fine Um that I I'm kind, I'm really curious to seewhat the what the different that ups are. What the different solutions are Nand Kindof how people choose to address the problem of being live because it isdifferent in the the P work? PROWORK, you don't have any you have you haveopporcuuniy cost. If you goffline you, don't you don't have the h the losingof your indement? You know your machine doesn't slowly break down well likeright, but rt long enough time, but your machine doesn't slowly break downjust 'cause, you stop Baladin it and, whereas, if you're tirty to eat stuff,stok salidating, it's kind of like your, your machine starts traking down, Yeryeahit's, a diffeent t kind, O Ik, I'm still clear on. maybe you CN I', Invest, I stake andthe proof of Worchin. I stick thirty two meeth in a contract Um. This is this is myvelitie. I M thereforgood or is t our ability to stop valadating Orgo in we can set up that. If I gooffline after this impound, I can automatically pull myself out right, but we because we're doing thiskind of assumed, turt es accounting, where I can assume everyone's kind ofthe same H. we've been discussing a mechanism that if a Valodat or dropsllo a certain threshold, they are kicked out Um and in that sense it's it is alittle bit more protective mechanism for the Validator, and it also is more,is protective for us, because we can. We can still operate n under thatassumption that all my Valueis, the same the same so in that case, say yousateat, Tha, twentight TWAND and...

...seventy billed to be determined. Thisis, but this is kind of because we want the property of beingable to assume everyone's the same Wa. We also need this kind of Ojectionmechanism and I did mmispeak a little bit. Um valuewards are not you' everDrivento Zerre you're, driven to that minimum stake, um required to still bea perdicimant and then you're objective. You Tat, Ms Take Thirty two egth. Whyvery ice take over thirty to me and then my minimum is is like thatthattome makes more sense. We Bet Bot yes out, aty Tobes, because that's yourstake in the game. That's your network! That's a LIK, saying hey! If I dropelow this I'm eradicated like. I need to make sure that my stuff is an order.Otherwise I'm a bad validator right. So we can. We, you can imagine what were profesing pretty much thesame thing. Ind that say, the absolute minimum you can stilloperate with is twenty eight, but you have to come in with thirty two UM. IKnow Justin is s at the proponent of being able to top off. So like say, youwere offline and you've gotten down to thirty E, but you you want to play andyou don't want to be at risk of being ojected and then going through a apormonth Witha drawal period, which would just be like a standard Lago Um.You might then have the opportunity tof top off and like add two or three andand go about thirty, two orn thirty, three Um, so you have the Yeu have the rightingtuition that there is there's this theoretical minimum, but we wanteveryone to operate when Yo want to start people above that theoroticalMinimam yeah, because you know all your mouthsare based around the thirty two eath representing a certain percentage ofthe current circulation of of Tobal supply. I guess I should sayof I circuialy Suppli total supply of either available on the network andthat's all based off Tho. Thirty, two point right. So it actually wouldn'tthit need to be a little above thirty two to be a minimum staking and thenthirty two 's, the threshhold you shall not go below so that we can ensure thatthe math is so correct. Where am I misunderstanding that so th the math issuch that Um yeah, so the the math is, if all hesvalidates at thirty two, then you can still process all the signatures. Inthe correct time, um nouadar still have to uh to participate. They still have to.They have to start at that thirty two, but if they are, then, if they bleedout to twenty eight and then leave the validator pool, that foor is gone, it'sburned. I suppose I want asknt o no one who actually IV you actually reducedand Eit t's like why wonl they district, I, that is network orwards, then like.Why is it burned whys? I not liks like re circulated, Um. Well, there's aproblem. If you, if you start giving that all to say the validators, theother Valatrs, then you have incentive to censor Um and you can now griefvalidators like if I'm a I'm a majority coalition. I might then be like wellfuck sorry I'Kar Te Sa when every on loo thought they p all the time Bo. True, this t e other one third likelet's take all the RORD LE', just censor them S, R, having prettyperverse andtenis there reases that that increases the possibility of, likeyou know, gain seory around it right so like ifyou had a large spellidator set. I somebody who's just like I want Ta Shitload of Alaniaters and they say I would like to then get more money. They can.They have the resources to them to censor other people, aeiwe think anmaybe governmentes are doing this. They can cut off the Internet right and thenthat that particular subset of people canonlgervounet than all of their resources get redistrubuted across thenetwork. If there are a big portion of that network, then they get more money.So it', you don't want to you, don't want to give people the option to havethose games right and essentially by Y reducing supply. Um a burn is Equa. S is at leastproportionaly in distributing value to moly holders. Yo know, whereas a givingto the validators is is only giving to the valaors and giving them anintentive to do bad thing Um. You could definitely imagine the community comingaround some sort of proposal to make some sort of Dev Fund, or you know likeI. I don't there's, there's tons of options there um it's a matter of ifpeople want to do things with that that I've, I've seen EIPs to be like H, leftuse all the eatthroughout. The burn address to like make a community fund-and you know it's interesting- it's not something that I'm it's kind of a cor.Protocal developer! I'm like! Let's just get everything right. I don'treally want to think about. Like the government around Uh, you know redistributing burned e, butyou know it's a it's a open platform that the community could decide theforkin degran and things like that up. There Tho incling has this becomeparticularly dangerous, T um. I love the wholewort three scenariobecause its s like in the event of a or...

...three scenario you have all thesevalidators in a nation state unless ous saying that it has nothing to do withinnoer partition and the people are cutting UPF, but literally, what'sgoing on is that network outages are going all across t yo the glowvalinators are getting burned and money is being hit pmedirectly byactuallyit's, directly being hit through violence, but or maybe evenmesoul, say: okay, let's just remove the whole rwar think eything and talk aboutClimate Changeas, something I'm particularly worried about oeo majortentry happens, yelostone national part blows up whatever Um you know, Elsonwhatever and like we have major, you know: Superbowl Cano events orsomething network, couges or random, beave, solar flars or whatever. Theseare things that can happen and can impact a particular nation or even ofthe glow, and it's not even related to on or three I that happens that thatwill eventually happen will be a catastrophe. MATATROPIC fenaraaall.This value that is being stored in validators are N. by the way, I foreseea lot a lot of people investing in validators Um. It will be burned and that will hit theeconomy in someway, and you know the total. The total value of Pereath willgo. I guessu Um and so individual holders of money whoare using this as some sort of way of paying for their Averat teothertheirtheir milk, because the value of the O has gone has kind of been hit. Therewill be some immediate economic consequences surrounding the the Um.The fact that this is burned- I I Doni, don't think you coan do the system ofall of rules that covers catastrophic events completely. That just completelyignores the decision. maging process of the aftermath of those KINDOFcatastrophes, like is scenario say it is burned and the pricemeather goes up.I would hope that there's DA social poplice social UM rally around gettinghelp to those people basi on the new money they just got from all theBurndes, I would hope, but like and and sure, if we burn ten percent of Eashwith nothing else happening. I would I would think that the value ofindividual each would go up, but you're talking about east burning N in acatastrophe, yea terhaving to the relatives like there's so many variablecoming in a play. N support so be Maybea Po. You know that might push itup, but who knows hat the health ongeon a now see I see there possibly. I justdo see that it being reasonable to consider the idea of instead of burningthat we put some sort of noat mechanism in there for recovering that eath on aregular cycle, yeah Ma right, Riridin, the IP AES coplated, Sir Ita Oa Jupiter in these things anyway.So that's that's actually another question cuning do cleron and navy canhelp me out if we're staking in the proof ofworkchain at least initially, and I think you're you're, depositing fromthe PROOWOR Tan Y, not that you're really you're in the beacing chain.Like you, havent exxited, the provertiain and you you entered theDeacon Chan was a mechanism or general. I thought it was a burn on th on theEVM and then a basically a hay. Congratulations! You Bur! We have proofthat you burn something you're now Homa Beapan, shn yeah a and burn in thesense that it no longer just anavium and N existinit other separate part ofthe protocol. I like to use the term directional deposit or one way deposit,but again so th. The question of having is arewe setting up two ethereums. Howis this different that some sort of fork, meaning that the Pfo Work ChainStill Runs The prof? You Know Pre Work Base in K contracts? Are we going toforce everybody to start operating contracts on the like 'cause state is not going to beautomatically transferred, is th what you said. So is this not like twoseparate theoriums at this point? It's transitional it's! I C it's a way. Ihad gradually upgrade it's transitional at first it looked pretty separate andmore for purposeful perpocal Reasonis, because it allows us to really innovateon the EP Turo M, but in the Indien the shorter medium term, we can actually utilize the deaconchain for positive things. On the PWORTHING. I didn't mention that oof,the dican Chane Protocal, is to when a when a black is proposed. We bring in a proofof work, Hash, Referende, the recent pro for cashefen and in doing so, whenwe finalize the PRDICTION, we can choose to defer our group ofwork for choice to the profediction Um very pretty much in the exact same waythat the passwor contract was going to becomethe root of the Prov of Work Fort Choice. The beacon chain can thenbecome the rout of the privil work portors and though, in that way, whenwe'V, when we're happy that the Dacan...

Chan seems to be a stable protocol, wecan then gain. We can take some security games from the the new BEECONChun Protocol and kind of begin to more tightly couple. These two things and inthe long run, yea t I's going to require some sort of pork to altenately,decide what to do with the BM chain and where to place it in the the ETO chain.Um and and like I said, you know, the e BM CHAM is, I t t the plane. An flighttons of people are building on it and we wantto like do we want to keep that?Keep that momentum going and at the same time build out this new infrastructure andthen handle our community. You know expect the community to keep thecommunity going strong on EP N O and then Um on the medium term starttransitioning and moving the community over and cor Yo. You brought up acouple. You brought o a question earlier that I didn't quite get theanswer and we answered the simple answer was uh he's one o Sta doesn't exist, anhe tooitp, maybe in the future h when we rolled into a contractrold in it a onshort, but on the on the medium term, when the to a TAT exribution exists, I might sayI'm Auger H. I have a token. I have this prediction market contract and Ihave a user intervace that builds these contracts and allowd people toparticipate in the prediction market. Um I can, I can do. I can easily redemploy that contract and eat too,and I can make that that user interfase now show the EF n o prediction marketsthat are going to be moonlight eventually when they all resolve, but Ican also show thestwo o one, and I can also when, when people make a newprediction market, I can make the default point to this new contractnetoo, and so I'm like gradually transitioning my community over. I alsohave a token there's some complexities there like. How do I auger? Actually, Ithink I think they hard forked their token before right. I think they had anissue I hin they aaga was written in serfent and they they orchestrated acommunity fork to essentially take a snapshot of the Balance and create anew contract. Um Augur might choose to do something similar, they might say,okay, we're moving like we're doing it as the community we're going to go toEatto, O we're going to make a new contract and we're going to take asnapchob of the balance on this tate and so there's there's a lot of likesocial complexity there from a technical standpoint, um it's not it'snot Mathiv or ar to say people knows great. You know we're happy here andwe're going to wait until we get rolled into sharting on the on the longer term.Inthe Ara each project can address that type of thing and, respectively, to thecomplexity of their of their platial uty yeah. It's a beauty, but it's alto.It it'. It's a pretty serious inplexity in the sense that, as long as you getgo laus, if you give them options or Hass upgreat, then they can do thatappropriate okay. So what if you have to pause your at for a week in somecircumstances, if you're doing a massive upgrade or you you like choicesto? Do it gradually ou in a way that you know as seemless of the enuser andthere's a lot of ways in which you can do it? If the paths are there and itseems as though you're agging them correct, and I I hope that tools, dustpractices and a very you know, rich community discussion on how to do theis going to Gointo emerage when when we get there Um, you know it' as still kind oflike perplexing me andmaybe Yo did adjust IC. I just don't understand it. Yet is Um. It stillfeels like you're creding, two wittheriums Um. So the value on thebeacon chain is separate that the economic value on the beacon chain isseparate from the economic value of the proper work chain, at least initiallysorighto, enbindpeoplaces and another. I be basically ETO token Ett have somepeople tha RTHAT are concerned about the the difference between that Raght,because we have at least on the medium torm H, a directional movement ofeither from one side to the other side of the protocol. Um Now so, first, it's just throughvalidation, and so you have. But at that point which is kind of another interestingthing e can talk. aboutabout O can can actually withdraw. They can't goanywhere. They just need to keep valadating not for Wa taction. Theartidition doesn't exist in starting, but when that ecusion existencecharting, I would like to see US Yobil contontract that allows people to alsodeposit over into the shiting Um, so at least from the standpoint of there's at least a directionalopportunity for arbertrage Um, in the sense that, if, if h to o, if, if, if, if the EEN t ETO LMTP is uh overvalued nd compared to the Iteno Um Youwere Gon you're, going to e veryquickly see people transferring over that bridge into the too Um...

...and in my prediction, in my opinion, it's not GOINGTA that gten mystery'sprobably not going to be another direction Um, but you could imagine it. You couldimagine there in asinatry between the two, because Ofr kind of makes mecurious. Why you're going to release some phases rather than just release ona test at work, Um and M A and basically just do everything on I mean.Obviously you could be doing a test nutwork, but like y Don have like toeat two of a test network and then, when everything's spitalized, we knoweverything how everything works, how it all interacts itwit one hundred percent,where we need to be, then you start, you know 'cause, like I'm, not sure Maka Misunersto theprocess of the phases, but to me it seems like if you were to roll this outin minor Ao. Our forpases like that, wouldbe kind of problematic in Um n, in the sense thaty you are fracturing the value of etherium on the poperwor chain tobasically create a new currency Um. So again I I'm not I'm not quite aconcerned h. You are about that, but because you don't learn a lot of thingsuntil you have a a real network M, a lot of these thingsthat we're going to be doing we're rolling out a new PDPLAYER UM,essentially like a sharted pet aplaer, where people can um tag their messages.So, like am I talking about this short or that short an? I am I aingthis. So Ican then, if I'm assigned to be a valilator on Shart line, I can thenlike ubscribe to the Sharg on topic and so there's kind of partitions in thefutplayer. This is, you know, we're using some really interesting software,probably from LE PP, and it's like H. we've been doing a lot of simulationsand a lot of experiments to bet, but a lot of this like it it's new stuff, andif we wait until everything is ready to go on on a on atest net Um, I don't think that we're truly going to be able to have vetted alot of the stuff at scal, whereas if we do it incrementally and we do it in away that it's pretty isolated from the original protocol, but still areputting in value and putting in the opportunity to make money for thesevalidators we've venalizing Galitieers essentially join the real network.We're going to see things that we're going to see these newpans operatedscale, a and figure them out. You know figure out the the bumps and issuesalong theway, whereas if we wait until tha ecusions, you know sh beacon,change, ready, Sharchain datavill willy is ready. State etitution is ready,crossark communication is ready and then we roll it out. Only then are wee probably going to see.Some of these problems really come come out an scale and you know we're goingto be dealing with problems from the fly and all the problems on the flyrather than diterative problems. I um another interesting thing is Kinda cool that the the valiletorreward Ale uh, with the square root of he totallysalidated. So these are numbers are very much up in the air. This is aneasy number to think about, but say ten million e when ten millionaise isvalidating. validators are making approximately five percent returns pera year. If only two and a half millionaie isvalidating valters Nike, ten percent Verer. If forty million ESIS Aida in gother, Ike, two half percent, so the the scaling. So we have this kind of earlyadoption of the Deacon Chain where Vale there's a lot of risk. assotheater withvalutors can't even withdraw yet. So you have to really be like an earlyadopter, Eger oiatine system. So you might not have that non. Have nearly asmuch ease come up to participate at this point in Crotocol, but the valitisthat dupertiiate Mikea crazy game compared to a year down the line or two years own,the line when the Protocalis more stable, more futures you can withdraw,then more is going to show up and you might have much lower return becausethe you know the risk the risk. An time horizon is much different Um, and so I you might have, on the shortand mediumterm, strange aemit, dymestries between the values of these two sections of thesystem Y. I keep saying you know sections and subsections and differentcomponents of the protocol, because I I truly do I see this is one protocol Um.I see this is kind of a a road to get from point a to point V, but all underthe same umbrella, Um, and you might see some there's going tobe some growing pain and there there might be some ACI mintury between valueof the so the compunersection but rolling out iterively is, is the moreprudent and conservative approach t I think, going to get us. Basically,where we're trying to go. I was Calin that I'd call it like buildinginfrastructure for an entire ecosystem,...

...like I the're trying to build differentpaths and different roads and upreading them in various ways, but ultimy we'reserving the same people and that's. I think what like we need to keep in mindis the Hend of the day. This is like a group of people and we're just buildingtools to help them do shit. Yeah Yeah. I really like the ideaofvaliteaers Ingusting in the future of etheorium. I think that's one of thefeelings I got early on 'cause. I bought a three in between two and threedollars I felt like. I was investing in a network that was, you know, going to go places. Ibelieved in it, and you know right now. It just feelslike you're playing with Curmenc, you knowplain Tok, an that's actuallyfeeling agailn you're, the pesting in the future of a theory, you're takingyour money and youre putting you o know putting it wor your mouth is and you'rehoping to gain some rewards as a results of that. I feel like that's areally wise way to look at it, and so they had a GE there. Early earlyadoptor of the VACON chain, validation protocol. Have you know the opportunitytoll probably see higher games than Um than the later doctors. So Yeah deiheda couple more questions and there are more implementation questions, UM, thythey're they're, not really dad complicato. I don't think but o I aflot as selected as a period of eight seconds and I'm just kind of wonderingwhere that number came from h. We like Powerstoo, if you've Kno, pretty much everything'so power too. It makes h Ma Fun and Nife, but h we are based off or NE NetworkSimulations ANDB base of the requirements in Noork. We believe thate is probably going to be Um at this point. We believe it iis going to beappropriate member. When we actually get a test nit up. We might find that h.You know under maybe optimal conditions, it's okay, but it maybe a degradesquickly. I, if network's, not operating under copial conditions, and if we findthat we'll ajust that number accordingly, it's our best cast rightnow, okay and another thing that I think we need to kind of go into forour audience. ECAUSEWE, probably hould have gone into it. A little earlier isthe DIFFERENC btween active state in crystallized state N. What that means. So I think you kindof have an understanding of what crystallized is right, so m, crystialized tate is Um. We have this notion of a cycle thatkinwe he we have flots, which is eight second we have in which a block can beproposed per slot. Sometimes you cand have missing blots. You know where youdidn't have a block of post, but then we have a cycle which is um fifty fourslots for accounting purposes of updating thestate. We we only really update, like the big bulky state, which we call thecrystallized state every cycle, and so every tycle were like okay. Whathappened? What do we ha? What happened in all these boks? What what actessstations do we have the which which operate kind of a Caspogot to like whatw? What on these cycl boundaries? Can we finalize justify? What crosslink can we update like thebig, the big work we do on these cycle boundaries and we update it in thecrystallize state, the Chrislitic, you know much larger state. The activestate is really just the accumulation of all the just the little things thatare happening in the block, though, like a block comes in. primarily the bulk ofthe work of profiting a block is, are all the actestations in the SBLOCKVallad, so we propiss the signatures. They are. We add these actestationsinto our active state and we do the bulk update on it. When we get to thePCYCLE boundary and so by separating the state, it actually helps us Um. So by putting these things into thestate, it helps us serve light clients and helpthe people kind of like followt the protocol without having to validate everything M, and so we kindof chunk everything into the active state so that we can realeays serve itto people and prove things about it. And then we do the big update h in theChristallife state, and then we can also surf. You know surfing he of Othat state route to other people, the perten about it Y. U Stacheck, twenty thros got Chrislisely kind of grows with thesize of the number of valid enders in notice, and it seems, like my basicmats, says it's going to be probably around the size of six hundredmegabites or just tha. Every crystalization is that accurate um that sounds roughly AC. It dependson the the number of validators Um and we were talking about this theother day in terms of what say, I'm a Validor, or what do I have to keeparound Um the naive that ieve answer is. I only have to keep around everythingsine it was last final Iz Um, but that actually isn't going to help me h catchbad valadators so like if people do something tevarious, I want to be ableto notice it and H. prove it to the chain, be a slashing condition, t slashing message, and so in thatsense I probably want to keep more on...

...the order of the things that havehappened in the past four months, and so I'm ou o a lot of a lot of the, and primarily I don't have tonecessarily keep the crystallized state. As every CHUNP like I don't have tokeep exactly Thi Crislifeid, exactly Thi cistlide state, I'm actuallyprobably keeping references to afte stations from the past four months,crystalize Tay Rte and I can rebuild each Chrislie tate as needed. If, if Ith somimplementation baciils a Ou, not but the state doesn't h because we're not doing you know Causeothe TR, the lack of transactions and kind of Arbitrarye tthe state doesn't really blow up in the same sense that Um, you know the evm or or even a sharkchain could overtime. Does this this whenyou say at TeStation you only care about the beacon chainside. Do you care about what'shappening on the shards itself? Do you need to keep all one thousand ND?Twenty four was OES atorshards states of update, inor Caus Exists on theBeacon chain. It's going to be tosomally, sharts and Yeh also bringsup the question that we probably should have adjrested e so cross sharkcommunication. U, but like how? How are these valitators validating you know?Is it across all charge or just te Shok they're interesten right, so I'm whenI'm, the rang is going to shuffle me Um s, going to tell me what Sharrd OrshardI'm responsible for it on the order of one to Thu shards in any difveent time.Um and my responsibilities on that Short areto build and attace to that Shart kind of do like a kind kindof Meri what's going on in theBacon Chang, and so, but my my I'm going to be shuffled much slow,much more slowly onto these shard than I'm shuffled around on the decont chainin terms of my crosslinking, so I I might be on a charg anywhere. You knowfrom you know, two to two weeks to a month and a half I ma actually be onAshard and so N, and I'm also shuffled nowators ar shuffledcontinuously and slowly. But I don't just get to the end of a moth and belike okay, everyone, switchd Shart. You know it's like okay, Valadator, zeroyou're now over here and then sometime PA, IE BOT it or one you're now overhere, and so doing that you get a lot more stability and who has the state ofeach Shart Um and my my role is to build these shards, but also as acommittee when I'm attesting to these shards, I have to Um. I have to sthink the Shard from the last cross. Lank I have tostink the Shard and I have to say whether the I have to the things thatI'm crosslinking in Um. I have to actually stake in a test to the factthat the data is available so there's there's a whole 'nother kind ofcompoinent of the consensus, there's a a game around attesting to theavailability of Shard data, and so you have now h the validators not only saying this isthe cross link and not only saying we should finalize decon train here, butthey're saying Oh and I stake my money on the fact that the data is available,and so there are my requirements because I'm I'm always vadating Um, I'm always validating on the order. Onetoo Shard and I'm always attesting to h Ishard at a time M. I'm kind ofconstantly have I need to have the resources of one to two to three. You know on theorder of a few sharts of of data around M, andso my requirements are going to always be. I have the full beacon chain. Ihave a at least like what would be considereda full think tout. It could be like a prune, ful, think of a Shard, Um and H,and I have these nipits of tate from the various shards that I've had toattest to and also attach to the availability of the dataand IM. You know I I still have on the order of see resources requirements,but not um. You know which, if the Chardes aregetting big, I have onlh, I have to you, know, be able to handle the the Shargstate, but in a in a a pruned way you don't have to Neeab a kind ar I no ony short and to justtoclarify your youryoure you're, believing that each charg would haveits own state. So youwould play a contract to a aged ARD in only one Shard Taret yeah uness. I wanted to to play acontract. A multiplshor you cant do that, but you'd have to then rely oncross linking crossling and mechanism statually share value but tween thosetwo yeah crash ACR communication, what'sat's, looking like supera research, exciting area ofresearch Um...

...in terms of a synceness car shagcommunication solved not a hard problem Um. But you have you end up with kind of alarge latency and that if I'm on Chard A- and I want to communicate to Shar, B or transfervalue to short charged, be then I essentially say so on Sharday uh Ay receit is created that can then beconsumed on Sharb, but it can't be conpue, Shar id be until what I, my transaction, O Chart, a it'spinalized Teo, the peacing Chane. So until he cross link is brought into thedican Chan from macharde. Once that happened, my receipt can then be consumed onCharb and then that Crossshr, that communication can happen. But then, ifit Hav to communicate back, then you have to do that. You know do that againand and these things are happening on the order of a cycle or multiple cycles, which isyou know, on the Order of eight ten minutes, and so for a lot of things. Maybe it's OK.Maybe it depends. It really depends on the yuthcase. It depends on how use'refacing some of these things iin the requirements o system, but o athinkerscausio communication can happen m not too difficult. More exciting. AROF research is singerscrossercommunication and it's not something. I spend a ton of mytime. Thinking about. There are a lot of people thinking about it, but thethe idea o can we do better than that Um. Can we there's some reallyinteresting work in the H, like espetially probable of sixstate execution like maybe the receipt will be consumed inthe future, but can I probabblistically at this point assume what the state is going toresult in and how much can I assume that you know my nine nine percent sureNi nine point: Nine Ninine! I if I am then like okay cool this is som thathad happened, K or it's kind of like Cotio iy updates. So I you want to lookat like the Front End World F what s how that works rightrbut then there's there's also other potential things her.You add stuff to the protocol in the sense of like, I might add, there's areally cool ease, research post recently from Batallic, that's likewhen I make a transaction, that's Gong to be a crossyhard transaction. I alsospecify everything that' going to touch, and so by doing so, um I can kind oflike isolate this transaction from other potential Crofsshad transactionUm and the kind of the validity of the transaction coal be premised upon thefact whether it was included in a block that didn't have any otherthings that wase touching what it was mo to touch um again, I'm not that's about as much as I'mgoing to say on the thinkerness Coss Arc, musication, not something Ispending a time at my time. Think about so that that brings up kind of another thing that that's kind of. I am going to address a contract inthis system. Um The contract exists on Asshard Okso. I need to saythiscontract at this Shard or is the contract just known across the networkin in the beacon chain. It has some sort of reference to where it be wherei.The earte confact exists, Um the address space and how these thingsare addressed need to be locked down Um. It would probably be a combination of Chard ID and address and that's wherethings would live. Um there's a lot of things to consider in terms of USEIRinterpase. A lot of this like. I don't want my users to think about this. Idon't want them to am s at all Um and the next phase of like you know, webfreed j oweb through dot Pi. I need to really uh consider what we want to expose to ourdevelopers and the develovers really need to consider what I wantto exposeto my actual users Um and there's a little bit out of my area inin terms of user experience, but there's a lot that the community needsto m start kind, O digesting about what this might look like and start thinkingabout what might be Abe best practice. In terms of like exposing this to h,you know how how are People Goin to interact withte, sharte system, um yeah, it seems to me, like all you need-is a Shart Lik at an address and the beacon chain can keep track. Ith Theaddress of the contract and tell you where the contract presides. It's alookup right, it's a distribut, it's a DHT, basically and- and you can nd noteven that, like youcd literally just say, hey beacon chain has naming systemsort of like kind like yeones, where you just say hey. This is where all thecontracts are right. Now. One of the...

...problems I icotic see happening is: Iexist on Shard, five hundred and twelve and then for some reason, acryptokities POPs up on five twelve and they're bogging down my transactions,because I'm all five twelve and my application is negitively impared bythis other application. WHY TRETI TO RT my application to a different Shard andto me that, should just be a civil kind, like state, swap request put into thebeacon chain that just kind of soks everything over right? So so some ofThi, the Bicanchain at this point, it's likein terms of the state of the Shard, is very decoupled Um and for uh for simplicity in design once you by decoupling, the state ofthe shards and just thinking them through these Um crosslinks and through the dateavailability proofs to prove that at least everything's available on theShard. We can have a really clean, beconcane design, but in doing that, welose the ability for the beacon chain to be this, like load balance mhm inthe sense that, like deecon chain is not responsible for Um, like you couldimagine a system, and I don't know the relative Complexogiv, the system ut.You could imagine imwhere this core beakon chain thing and not onlyfinalizing all the shards and final iing itselping making the rangy. Butit's also like monitoring the load on his shard and like shifting thingsaround Um. That's a something that has been discoton inhasbeen thought about,but it it's very, the the complexity seems to blow out pretty quickly Umright now. The idea is more of this um economic, low, balancing in the sensethat you're right, a criptocity, shows up on five twelve, and now I'm likewhat the hell transaction fiews are so high. I might want to get the hell outof five twelve Um and is there a mechanism to do that H,potentially one one version of CROSSAR communication? Is this idea of likeYanking or locking contracts where Soo slids, yeah wherei might say, I'm GonTa, I'm trying to book a a a train ticket and a hotel room, and theseexist on the hotel. The contractot represents the hotel room and thecontract Orupson the train ticket are on a onseparate Shard. I might yank the so there's like a hotel contract and ithas individual contract ore trim. I might yank the room contract over tothe Shawer that has the train contract, and now I um atomically book both. Butif we're N, this is this is a mechanism, it's kind of kind of like a lock butkind of kind of a Yank, and then I can then I can spend the contract back over.But if we have a mechanism for yenking and can give and can say, a contract isallowed to be Yang, then we could. We could imagine you know if there's anowner of a contract they could, they could be allowed to move it around, andmaybe that's the only thing that they're allowed to do m. It seems to bea little bit of a dangerous design decision to start allowing people tomove things around, because it's like a lot of people are expecting. Somebodymight expect your contractto exist on this Shorg, and so you might end upsomething. A Sharg lookup would be easier that he could say this coushurexist on the shor ut. I I I understand Um tit's, interesting, something I needto think about a little more but you're right like if you have a like a manormemory management table or something Li eersinal memory table O al like yeahshouldn't. Be that big I mean like it wouldn't increase by a tremendousamount. I mean O wed creased by the number of contract er deployed, but notnot like a ludicrous iolt and hey could all exist in the peacon chain, and hecould just say this. Shart exists on this place. At this presentcryscnalized state. You can only swap over on chryslize sates, yeah interesting Um, probably not going to go in the initialdesign, but something I'm goingta Yo all right. I think I think we shouldstart to wrap here, um first off thanks far coming on andespecially going longer than normal and answering a lot of H, mussions. I thinkit's going to bserve a tremendous value, toh the community as a whole guess I know a lot of people have verysimilar questions, because they're asking me them M is thereything that wedidn't get to that. You think we should have or anything you like to say thecommunity overall Um. I think we coverd most of itwithout like getting into the crazy, crazy nidygritty. I would say that we do um. We do east two implementors call everytwo weeks e d. We have one tomorrow on...

Thursday, I don't know when this call'sactually on come out, but Thurdsday September twelfth Um, there's like fiveor sixteen now implementing the new protocol. I know Parrty's actuallygoing to be on our next call. They're getting excited about the stack I spolitified and they're, going to be they're interested in terting emblement,O the new protocol Um with a lot going on. There's a lot still to do if you'rea developer, get involved. A lot of these guys have are using the tag whenthey're get hubed a good first issue. That's a great that's how I gotinvolved with he e in general. I just started working on pipers ithonricos'cause. He had good pirst issues and it was like Oh cool. This is the Greg Wadoet started Um, so get involved. Help out this Shi doesn't build itself Um ifyou're more of a community member watch the calls. You know if you'reinterested in that kind of stuff, pretty cool stuff and we're going tohave a lot of Um, exciting things talk about during death. Concactoug therecheck it out and if not ther're going to be super cool live streams. Um. Youknow the this stuff takes time, but we're doing it right and uh, and it'scoming awesome denny makes a lot. Ma'am Te Looki, Rokry, Ni, ei a I really coe, I'll, also see adepgone, so hopefully he can buy you beer.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (108)