Eth 2.0 Minimal Execution AMA by Alex Beregszaszi, Casey Detrio (Devcon5)

[Applause] like the concerns of this proposal is immense to adjust mean we have to do with kind of other things that people have been complaining of Annabelle at four months the biggest one being like few markets and real labor markets so I will kind of explain the problem first so the problem is like status quo is basically a bunch of shards and each of these shards have claws but the shards only fear above each other had a kind of theory delayed schedule and like this is because like it's somewhere in the middle here you have a cross linked into the beacon chain in this case if that was about one block and so basically like at base in the status quo proposal had basically here for one shard to worm-like state birds of other shards it takes like about six minutes and vo this is obviously bad for if you want to be able to have any kind of cross straps and the purple small w pants just thing to get around this basically has to do with these connect lofty mystic states so basically the top of us have optimistic mechanism that says oh well if you're over here then you can make a kind of conditional block that said I am this conditional block just that access all I think this team I think the state route of this chart actually is this value given is missed that is going to do this and you just have these blocks identity exists as something on way or two and that there would be these different at a protocols for achieving this and at some point later there would be some leader process over here and then I kind of storms through all these small X figures out which ones are kind of correct which ones are not correct and kind of threads everything together red so and if step one block silver sphere or kind of packages over here out of publish ahead of time opinions about what happens in other shorts and these opinions are not necessarily over two percent correct where things have it based off of those opinions until things have dependencies and clients can like basically figure out ahead of time like all these walks will get one yes I kind of glued together but then there is a process that kind of whose these walks to cut together one chain and up in the future and execution environments are supposed to rather than be able to implement signals the reason why this is because this is not very nice it's basically because if he wants to look at this from a point of view of like I am a transaction center you know hi I'm Alice and I want to send some points to Bob and I Alice I have points on this chart but it turns but what I do what someone says it's over to Bob and you have Bob over Sierra and Alice well successful defensive Wednesday law and then Bob Ameet wants to be able to do things with those points immediately like sending them to Charlie or putting them into Eunice lava which one would automatically send some die to generally but if Alice test planes and Alice sense wins the bottom then the coins like the underlying based protocol of ether we're not going over here immediately and so Bob does not yet have protocol level either Bob has kind of gone from too optimistic ether which will become protocol no mediator but I mean what because like it's voodoo ether is only your two things like we can't trust what proposers to understand what who Demeter is because what reporters are not expected to speak all of these execution environment languages and so block proposal I basically bought has nothing with which to take love proposers until media stuff with later in the future is either becomes real which one happen over here and so Bob like theoretically has coins but do not pay transaction fees so this has historically even kind of salt with the scent of magic catch but it's really our markets but we've been kind of thinking about the labor markets recently like realizing more and more that first of all this kind of design like it seems to minimize away or one complexity but in passing a lot employer to complexity around all these optimistic games and real labor markets get monopolized easily there's and a lot of other nasty stuff so that's the problem so here is the solution the solution basically says we have that we have a bunch of different charts and what we're going to do is basically have a cross-linked kinetics lab so these charts are going to get across with a good block and then the short blocks of the next slide are going to be aware of the output of the beacon water they're gonna do things and then they're going to be included over here and then you have another region plug then an overseer human have more sharp blocks and so you have kind of a much tighter coupling between the shards where each shard like any block on any lunch lauren is aware of everything that happens on the previous charts so the reason why you this is nice now is because like on top of this we can also add functionality for Andrew may be there between charts within a single slot and so if Alice has peso coins over here then Alice over here I can just make a receipt that says AI wants to give points of points the fog and the you can even have functionality where a head of the underlying execution environments can even much from one chart to the other and then Bob who was so there's gonna be over here the blog will I basically get me able to include the received and its blog actually asking you know because well actually has Ethan over here ball can immediately use these to do things like BAM transaction fees so I guess like when we look at like what really or what really are sort of what execution environments are forth like execution environments kind of serve different roles and eath to so one of them is they dis provide like Java semantics so basically like is this like balances Beauty xhose receipts and they also described by and they also kind of prehistoric system and they also have to handle like basically cross sure we also have the support things like optimistic state might be without the support they would have to come up with some like special purpose like vehement strategy is so be keying a kind of full execution environment like have all these components spread and this proposal basically says you don't need to do this anymore in you know you do this anymore because the protocol is kind of tightly connected enough to just be able to do this both natively so to kind of poke a bit more deeply one it's like basically the way that fee payment could happen yes you have Alice over here and Alice is just going to send the transaction and the transaction would have wasa execution and the ones of execution would be see what have one of the outputs that it could do is it could give an opcode the B C's we dyslexia payments but in few payments is basically an opcode that says take ether out of the current execution environment and paying ether to the current proposal so look we have a special upload this is basically the PAS proposal right you have like this kind of top-level wants and Wako that just said our welcome and the five that just has to keep her out of the excuse environment and kind of give it as a fee and Baja whoever is the one proposer does not need to understand execution environments because bog can just run code and sneeze politics made the fear and Bob Noah said both that Bob will get paid so is there enough off the proposer no us have to propose that the proposal now there's kind of like there's stuff that you didn't do that you can do on top so like for example if you wants to allow transactions to be merged to share morkul groups and all that nice stuff you can do this but like that's more detail so question like the big question here is like what sacrifices are we but do we have to be together that's great so if we do this without changing parameters then basically the overhead is of the b-team is going to increase by a factor of 64 and increasing the media welfare specifically the at the station processing overhead is going to increase by a factor of 60 now that's kind of and the other problem that we would have is that the number of active validators that we wouldn't eat for salat in order to actually be able to process everything in every short light goes up right so basically over European and shards then like you would want mm 28 or 128 back then and then if you want like to basically if we're still like if we still have a good luck length of 64 then we would actually need a hundred and twenty eight times ten times 64 and as 1024 and so the grew would need there to be like basically more you throw them actually this validated to be able to process very anything everything in a single salade or else we have to cut down the ePub length that if we cut down the ePub length that we have kind of compress things have been bored and really blow up the chain so the second part of this proposal is that we kind of do something to some parameters weekend in perimeter it speaking that I'm suggesting is short counts goes from 1024 down to 64 shard block size goes down from goes up from 16 to target 64 max to 1:28 sorry Gibble's biomass and so this altogether this is obviously divided by 16 times 8 and so we obviously get times 1/2 times 1/2 charge throughput or total system throughput but I'm arguing that this is fine and the reason actually we also we're also doubling the what well instead of having the smallest one times who would have larger more clients so there would be another platter by Tears so we would be kind of one quart basically cutting down the throughput to about a quarter of what it was and my argument is that this is flying because select realistic weather is not going to be a thousand X increase in the number of users anyway and we can just kind of start these numbers below and kind of conservatively ratchet them up over time so what this you know there's kind of other optimistic of optimizations that we so I guess so first of all just to clarify like in this architecture even though you have a cross-linked every SWAT from the point of view of FFG you would still have ebooks and fge blocks will still have a length of 64 swats I mean if we want we can make of 32 but let's tear it up when you reduce that okay so the treetop who for used Scylla so kind of overhead analysis so from a overhead point of view right there is basically kind of two major there's a couple of major components to overhead one is PC pairings the other is you see additions and kind of T reads and a third is like and of EPOC kind of balance updating so I am basically ec pairings are there actually EC pairings are proportional to the number of cross links that you have but no lot to the number of total validators and so basically EC pairings are the thing that this proposal alright might end up going up by a factor of floor but I have an idea for instead of maybe going up by factor for like maybe only be to go up by a factor of two the idea here is basically that like delayed at the station is just going to get merged all get emergency up there to publish something else for the wait at the station and then you see ads business proportional to the number of Ella this is like basically validators divided by equal a blank and this is also no health leaders divided by you mark life so if the employment goes down then a work walks will be up these end every block walks to take a rolling large amount of time the process you know we be what we basically kind of have a have someone have a kind of trade out to meet here then you would you agree in that like at the status right now and then people walkable walks are and of closer to being like taste considerably longer to process that like hundred zatia blocks I think it's actually the other way well I think that even have even then muscle comes from the other stations at least in the lighthouse benchmarks that I've seen the you proportion of that even at like 4 million validators is a fraction of the actual bottom so that's actually not as big a concern okay I mean if this is true we don't look at what went down from 64 to 32 at that time actually to it for other reasons like it cuts it cuts in half the ability that we wouldn't even for of the study but to actually be able to process I call charts of a single spot another parameter is right and I think in general I don't like I would not support reducing you for a validator like if I was coming up with a nice firm elevator number from scratch I would say 64 instead of 32 but maybe and so I feel like the mean value of 32 is worth more than 24 right like if we have to yeah in the value and the mean value of luckiest the composability is also pretty high so sacrificing eight degrees the other seems good yes yes to what extent can I rely on a cross linked because shots can fall cried so yes presence of a cross link doesn't mean together anything is final at least okay so one thing that this proposal is not like in this one like like one thing you might notice here is that shard chains don't exist right now in the proposal I have char chains do kind of exist in this sort of limited ephemeral form you know the in shown kind of winders value but okay so basically in this model the like if I publish a transaction kind of Sears the kind of increasing levels of confirmation I've experienced right so step one I publish a transaction so Alice publishes a transaction is floating in the air right now Bob knows that unless Alice double Spencer mob is going to get paid great like this is like it so this is the same as like Bitcoin ease that you know like zeroconf or whatever step to this transaction gets included into a block into a shard boy proposal now there is a bit more confirmation steps are being then you have cross link votes and crosses and clothes start accumulating on the show proposal and then you get a bit more about prevention step for the short walk gets proposed and this short vlog actually like two-thirds of the votes in the char block for this party or two-thirds of the voters over here I'm actually are for this block in which case now the basically the region chain matter stands behind this block so that's option one option two is maybe less than two-thirds of the committee votes like support this book and then there's two options option 2a is that two-thirds of the votes support something else and if two-thirds of the votes support something else then this guy then this guy's gone in this transaction should just wait until it gets included sometimes option to be is that maybe there's two blocks and that's maybe none of them got two-thirds right so maybe like when this one go 40 percent this one about forty percent and the other 20% were just offline if this happens and these could be like to compete conflicting vlog so the proposer gets slashed or this one of them per deal at the zero block right so if the product of choice are published a little bit late then like you might have a 50-50 if this happens then what we have or the third possibility is out there our 70 percents but the well beacon blog proposers like being a little evil and deciding not to include things so if this happens then our channel just draw it out let's say you have another beacon block and include it weird it would include but this is not getting funded great then what happens is that a proposal over here but basically it would be pointing to this and it would also have another pointer to this right so this is like shards 0 slot 1 2 3 this would be a short 0s to left for a somewhat poor proposal but basically a salon or proposal if it's pointing to a B pitch in which we are in the previous sensible the previous included as long as less than 3 M it has the proposal has to contain an opinion about what happened in small cream and then sure look what's suppose what loss of generality this nice legitimate then he is going to include 1/2 of these things and he's gonna include crossly and that cross link is going to get you in the indian about this and its making about this and so this gets immunity so there ESP I definitely didn't worship functionality now so the only thing that you lose here is basically in that over here to cross charter can be like cross right communication gets delayed by one slot because this isn't like this is not aware of this but otherwise like you have to wait a you have to wait encourage in its life and then the final level of confirmation is one like this tie gets finalized and that lol is says such transactions about is finalized so the only poor thing that's happening with respect to a shark tank actually contain words because these sharp block is containing a previous reference to do so only only if you reorder that you can change what you actually could you okay and then in this case really kind of an American chart change you do have less confirmation but it hasn't yet communicated with other charts so you don't have the issue like shards me right we assumed that one of the shots was completely oppressor of books and protections that actually better than just a transition but since everybody Sebastian any will it be accepted by them becomes a no good actually do some sort of validation on this stage well it depends what we mean by like cruise day right because so a list who the actors are one accurate sharp reporters in other actor s committees so if the proposer is malicious and prim proposes a bad walk then generally the committee will not accept it another possibility is the proposers malicious the media is also malicious the community points delicious thing and and then this he didn't want and if theoretically points to invalid starchy did it even yeah it's really good right look you know so if this happens right then first of all like this is already violating like the security assumptions so there's like to recover from this kind of situation like you would be simply need to so I wrote a separate proposal about this actually like what but so that's not basically there's two kind of possibilities one of them is that it got this work then eventually it'll be discovered through fraud proofs and when it gets discovered through fraud books then like everything gets her bit and you can't really revert lost everything because the reality in practice the dependency graphs especially between four sharks are going to be super tight the second proposal that is basically that you have this kind of roll off lights iam where you don't like if it's unavailable than you referred but if it's invalid that if you don't revert like you revert than each chart state but you did not revert the chain and you kind of revert the shard state with image within the chain the idea for this behind this is that this allows us kind of brought your assumptions to be a kind of a subjective at more of the clients that choice instead of like forcing a secret assumption on everyone but these things will only happen in the extreme case where a community gets broken by an attacker which in general requires more than conscience yes I'll change what will happen it's not possible for to deflect a charging watch to be included on this a baby chick oh wait my different-looking so what we're seeing is that there's to be in Chains and then there's like the shark chain which gets included over here and then over here let's say this one gets included here and then this is the zero block and he gets it to live here right is that what you're saying I mean this is too busy to shop job changes yeah so basically okay so in this diagram right like the the zeroes was a hole between the barbecue chain watch began the the the Pullman's are searching twice right like is this wigger is basically saying or are you saying some other fashion pointers Oh like these are the hash pointers I think yeah so that pork at the end you the the dot at the top is one shark lock that Obama's different rope yeah they've entered into or yeah this is fighting in inflicting but what's the point to get my picture James it's a matter of so eventually the task that most the majority to go goals become whoever the Viva which ever beat me Shane wins this fight so when you're when you're in tested with a sharp lecture also testing to the head of the pain so you're giving the weight to the result by other for choice by just five bucks so much because we have some sort of consensus which music company offers many new chambers sensors cavities a coordinated system so I'm also it doesn't mean to of such an approach or become Jacob question so so I think in this context that's that's be able to fly the peaceable kingdom chain run some pls give any weight of the initiated members or if you really wanted to someone asked a question about the previous proposal looking at the yes okay you shared before they interrupt longtime leader on focus and it was probably an alternative case - yes with removing persistent state groups from shards and executing code on while the court in beacon box I think we almost liked that proposal because we've got to offer start transparency great innocence soon and then put then in this radical new new proposal is it we're returning back to versus insurance or is there a way to combine the two so I actually think that you are in a kind of similar Cola softball direction and if the reasons I would give for that are like one is that because you have like short chain like the shorts are kind of tightly connected to each other like theoretically you could have designs where I have things happen on one char but let me just get moved over to another like the kind of secret easier makes that easier and the second thing is that this is proposal a kind of execution aptly the proposal of having execution happen on the beacon chain like if in this particular case the move would be delivered yeah but in any other case I don't leave any other names so like basically the execution environment is going to be or every one of these lies can be aware of the post seat all of po state routes over here sorry you could just get a second instance that you can you can have a PE that swaps yeah to someone else that's way so this like each general the thing that EE execution is aware of is that OPA aware of the like it's always aware of previously route of its own chart and it's also aware of recent statements of other charts but like that if you like the way that you've used that abstraction to be a lots of things like you could yeah and I'd say like be softly the facts of this proposal is kind of biasing toward more layer one communication like basically it does kind of oh more toward this kind of idea like shards is being things that you just do computation you want to keep the state from the shirt just we have to fit missing it's an option oh you mean over here yeah yeah yeah so so he's right if you want that feature offender need a caress well you know like well every sharp block is read like it's given points for this pretty ansible life so I think what Keyes was asking better be could they really just move this state from the Stratus Oh removal yes okay so this proposal right so in this proposal I guess one question is like why not simplify even more radically and just say like if this thing does not get included then this thing is just great and if you do that then like you really are removing state from chunks completely because like all the state like when we say that exists a state that gets faster to beat the chin do you give a lot of power to make yes basically because I can I can just say I don't like that block and now that block in the camera we're going some sort of topic but if we want to kind of make things slightly better we could say something like oh we have a period of two selasa move in which to move these things and that would turn the power to a wonderful to go like I'm scared an adjusted on his head for it you know I did not think in the board so it's all I can answer actually yes so we need sorry its I don't know if this is kind of a second but in this new model if you dose the block those are the breakfast is one of the difficult is all the shards falls short see why you of a cross short communication across our communications won't repeat drums like yeah what I guess where our thoughts are out back this is one of this is one of the reasons just as excited about see group leader election I mean so thank you the optimistic stuff as we might be avoided in the short term I think will still be valuable secret leader election protocols I are not too scary I guess but like especially 41 most based on written signatures do you know anything is the toy integration so basically it's like a two-step process the first step because this pre-registration when you send a message wishes which is secretly actually identity but people don't realize that I forgot to take that and this has to be double to also when you don't do linking Zebulon he produced a block element so I guess like it's not useful to do secret leader election until we have look at the tour backing right right okay and that's that's good rationale for delaying and until phase two or three what do you mean by crushed recommendation attempt always basically so let's assume that all block recorders become bet for a while then you have like short walks and then each other in this direction like if no if we think what you might be saying is that like you could also have communication graphs that are kind of partially intertwined so like for example like gonna require charts you could require this block to have an opinion on like these two and you could have a look at way like in a dependent anyway I guess the problem of the air is that like when you have things actually have to come together like you would have to kind of first include these and then based on that activity to be from that it would be so the opera's of pre-caching out would be more complex maybe so my understanding is that so in short what is it really good boy it's awesome yes yeah absolutely in that case is that me talk about some of the research somebody can always use how do you exercise at all just about sir yeah how do you resolve the crusher transactions in that next block are they personally done or would you have the options so here's the proposal that I have the total the total internetting is forcibly done but the receipts you have to go to do yourself and there an execution whole thing you must be is curable successfully so basically so the approach in my proposal is that you basically require each chart blocks here so like if mrs. say shard index one then over here at each of these short blocks is going to publish a more poetry that just says like this is how much people are sending to each other shark and so this chart you're going to have to quit the morkul bridge here and will have to include the moreover actually Sierra will have to include that same level directions here good versus one proposal I mean you could also have it make it more forgiving the following question is like how to percent motive that salsa post in the latest case one of the gloves down not I've become black and later shamrocks yes and so that means they're mostly converted that's for an opportunity through all these transactions or shanks are crossed out messages yeah it's possible so that's not it's not gonna be forced but even on these ones there's still be a little bit used right yeah so if there are some limit to how much you can in this wall okay so in this one I'm generally assuming that I've here a large portion of transactions will be cross shard so okay the be a chain is not required to like literally include like a lot of one data for every single cross your transaction great like most of it happens to receive passing like basically in everybody thinks you need is you need a shower chance to be aware of from all these other short chains you can do things with receipts and if he wants to move across char then you need to have up some kind of basically betting capability we can have different execution targets on their charts so we're here today to talk about EPC execution in general and this is highly relevant just because this reconfiguring of the architecture is a changes how we might need to keep an eye on so I guess if I'm accuser then one thing you could do is you could say Oh boom I am a good I have like full state on this Shore and I can just like self package another thing you could do is you could just like talk to a folk like you basically use like my client protocols like I see two spots but I do it yeah yeah you use a length length protocol a blank blank protocols right so basically likes you talk to if you don't have full state and some charge for some the execution environment then you would just talk to a server but does the server gives you branches so that you have a package of broadcast right so that simple great supposed to be market but there's still a notion of stay providers really layers I don't know that term that really applies to all sorts of things but there's still an idea of like other like weird general general purpose hash preimage for lovers that market for just market participants one concern with this we had was in the previous one it was being a bigger concern because you had a passenger port charts and the same provider would have to have an eye on each of those and I think the reasons we like the the pre interrupt red was that you could maybe have some kind of scheduling figuring out which charts given insufficient marketing so state water could just respective you yeah I mean here you can also have your exit environments funny that like I think that maybe like the issue of just like in general aggregated transactions for I retains that state isn't something that the specification actually addresses right so it's more about like I feel like you initially at least like i mean e-to is mostly about achieving consensus then just you know like making state and execution is an abstract thing that is so funny because I can eat one all we've ever been doing is like people always just been worrying about like state in execution I can't think what census portion of each want is that like never Bryce so like I mean all that never breaks in each one is you know like say that execution so I feel like this something where this is all there's definitely kind of bad in good aspects to this like one good aspect is that it basically mats that like if we want we can just like bring state an execution and it rules almost as they are inside of this the in absent Prosser and stuff but and another and about thing obviously acceptance means that kind of work is being done to and have allow lots of different kinds of things to work but the kind of less work as we have done to grieve one specific very good thing that works but like that's that's parallelizable it just means that we need more two more teams that get it's a greeting of your specific they know thing that works in one benefit is that it means that I think if we if we do have a kind of good solid and census layer that provides kind of enough base functionality that people want then you can see more you know we keep working on those other things so I do feel like we actually already have like a couple of things working with one good thing and that's this one yeah yeah I think like making if one like be a popular like making eath one plus like simple cross harbor scene passing like be a popular execution target at the beginning it seems to be kind of fairly Network effects preserving a conversation with some mystery teams and suggesting depth and the context people wanted me he is like still like crucial portion or some diversion or iteration cousin is a crucial portion of the stoppers that it is that view into the state that you and I see immediately well there's a lot of the system basically a lot of things about each one points that are like outside of the he specification like one thing that having there's obviously like those transactions and laying rights they have increasing there's actually no need to specify these things because there's always better ways to like them in any kind of software and this is just like a software problem it is moving in a circle it's really beautiful that you took the step and actually like so to say that you know this is not our problem with dougies it was just it just means that I could never get someone else to someone's problem yeah yeah we do actually have like a pretty good idea of like what that problem is because what we've never been doing it like we've been living inside this one specific to eat at least before right now so it's kind of like we already know that what's going to happen yes if that model is based on the problem here I you'd like is being used for FSU yeah so you still divide slots and 80 bucks all validators are called upon to do to attest once per that unit if you pump to go cross-linked but also to vote on that management so the cycle of effigy is still like large units of rather than the other thing it does is essentially it takes that shuffling of committees and rather than having a set of shards a set of shards instead of charts the different set of charts per slot is just the same chart each slide address the mechanics repair so actually an exercise right now essentially just like removing the notion of a crossing of a zero so that space zero in terms of for choice in terms of the validate with validators doing in terms of like voting on it's all barristers we have final questions one limit communication my being answer to this would be it that like then you would have to deal with wipes easy level reenters and stuff which just seems like weird and complicated but like if people want me to do it is the like the post state execution would then be like two favorites well I guess the idea would be like right now like if these cans call each other it's more like you have like a and then B and then C it but then here you would have like a and then you might have it and then he just decides that oh we're gonna call a inside a CRNA decide so we're gonna call the inside of here and then he ends and the dance and the pants again then you have like C and then B over here like there is benefits but legs you can see how this is a potential a security nightmare even if you don't have that secrets could be in the same shark lung and if they don't you know so what are the same part everyone first yep so one thing we can't do is win I think we can kind of allow a temp data free like basically free receipt passing between from what executioner's of second ones I think that's uh don't do that you can parallelize so there is actually one very good reason to allow receive passing between kind of separate institutions of any EU which is basically that if one updates the market area that the other might have witnesses that are all they might have to update them generally kind of it's absolutely good for elimination explain oh you had well this is like if you have new ability to kind of communicate between different is that you would just know excited it was single yeah then you could just say what one thread brought us all the a what thread runs all to be one thread runs on the CI that you had at the end yeah like that might even be here good because that way we would be able to get watched produced and verifying the more you can store your essence in a different thank you kindly that's true right this is true so me me a lot more communication is due we'll use pension should ease be open for any we talked about should be easy open for anyone to deploy or originally put joggers are no face to adjust phase one so we thought about this I guess I am the mean read night one of the big reasons face to is nice is that we have a goaltending market I mean we could do like this 1 plus mu in a multi you market but if we do that then like a minimum fee of like that requires you to be more diary than the old by smart contracts and so yeah well I mean there's two does this mean that you know the the current keeps changing running in parallel river you know what just means each chain becomes [Music] yeah it's no need to execute it I don't like even if the under roll-up chain right from whose context into that because proposers need to know and what the hell am I supposed to include what's actually going to get me pious but like if you if we kind of go roll up style that cross-linked committees wouldn't would theoretically not need to vote on the execution results which could be interesting like you could have a setup we're floating on execution results that's kind of delayed a bit you know this occurred immediate well no that's not good because that currents like fresh start why do we even need shards she would say my date availability throughput in an Amish are consistent with the ratio go to gate availability groups I guess because that kind of centralized proposing [Music] well I think that you to question that and if you say you have no execute among chives and their shirts and you don't then you don't you chose this is well I guess like when you put it like that they're kind of the data availability chain proposals like there has been a couple of them recently read like they're it's an interesting idea like you just go like political let someone the charts and then you just like have you just say the whole thing is supposed to be a rollup then you basically like the proposals that I've seen like they do kind of allow different people to submit like different bloggers routes in parallel and I think that is showing but in any system that I like I think any system that doesn't kind of bottleneck proposing so that one person needs to be aware of everything is - are by definition there's definitely a trade-off between like how much you have strategy versus be more complexity and that's one of the big things anymore there were a lot of six months the execution of those like actually happens unlike the shark client himself early we need small external clients the so like running the code would be done by shark clients as you have to like check the state routes doing kind of user specific activity like maintaining States and like generating witnesses like all of that is definitely right upgraded terrain that make things like definitely we need to who's this for challenges three labor markets challenges with the markets [Music] and you know what you're up to points about challenges over labor markets the I I think the challenge with the relayer markets as basically a bit like and this scene this is like a common opinion I've quite so and I begin I had like two or three months ago as the dungeons good that we have like optimistic roll up in the C key role of happening on this one because like that means that there was more people like T thinking and experimenting about real labor markets and that community seems to keep coming to the conclusion that roll up chains are likely to have monopoly promoters in the argument basically has been like as soon as there's one party that's better at sequencing that others that they need to be only slightly better that everyone else and this potentially has censorship risks and so like that means to funnel your tuna later to context you can obviously switch the differences that was what annoy or what contacts what you get so yeah a layer one context like you want more security of Mexico gliding so the next one we kind of hug below this morning and is just among some of the participants should E's be open for anyone to deploy or restrictive I'll give you a quick on the range there so on the one end you could deploy the system such that these can be employed by anyone with a certain economic model to deploy these probably very expensive to do so but people play with them let standards emerge and just kind of see what happens on the other extreme there are some participants that argue that maybe we should deploy one very good EEG that feels like ethereum it kind of gives a very clear user story and allows people to like build up and make it operate in this rashard context and then eventually open up be used to the the deploy and let people experiment the second proposal was called dictatorial at this point but at the same time provides like a little bit more guarantee that we get something out looks area so it's a it's a to be debated still the second proposal that does about what you eat still as you can Billie's and top of this nice yes so I guess the question is what it means to have a flip top level T versus deploying will all recall beat each other contracts that's the person in question it's probably that would be what one of two bees that some big things are going to develop at the so what would be the benefits or not nothing anymore experiment with some other one nakiya nerve distinction has to do with like state so one of the remaining rules of an EVD is kind of like describing a diffuser almost state structure if we have one and try to even and let abilities Italy in that idea then like no matter how successful this integrity is are there because they depend on the original year whatever the state model of the original you know steps like a statement like that status state that basically everyone will need to have whereas if these are more parallel with each other that falls into this yearís that had a fewer people can store it state again you were then and so fewer people think like it might become like slightly more difficult to do things what that but like other people don't need to be bothered with that data and so I've had to a parallel approach like at least in the short term it gives us freedom not to commit to a particular state structure and to a particular trade off between involving things away events versus storage status versus like pasta filling that confers resistance date there's also risks and how a CER and also with respect to potential bugs someone brought up this morning like the omen from mallero perspective the only thing that going to keep her like validators enemies and then you actually as a user own beings within the context of the account structure of the night but if for example there was a very popular ee that exchange is recognized and someone exploited a bug such that the account structure within me thought that it had a bunch more even the layer one still loved that it only had attending than that he someone's account now that the exchange recognizes it's invalid account might have worried so this is kind of like I can do a bug where rap teeth might print rap teeth but there's actually not that much ease back again so there's there's like all sorts of complexities on like how this concept of domain issues rolled out to users rolled up to exchanges rolled out select the various participants and the various all sorts of other company software that maybe go with interfaces so there's just a lot of complexities and a lot of conversation still have around what is so what committee about the ease since issues to the universe of the network from walking because right because if they are for what ve this 20 premonitions they commence would be ready Nations the ears will be written about assembly right so there's a like there is no notion like it since ceremony it's just the code that runs okay so the energies have yeah it gets to a point to the beacon pain as a code it's code it's a comment so I understand the risk that if anybody can be pleased some pain pulling to misuse and then you have to stay on the beacon gene there was that maybe you look up each or to deploy it yes and you may need to wear time you may need to look of course the kind of like event but if you actually burned that then it's gonna stay there forever if there should be something or exploring because I'm really sure that the risk one of the risks you guys really think of is that it's blow to the engine like allowing the e's which may spell intimacy and to be welcome out of the room as long as other participants and pitch em to lock up and that there's like on your time because you wouldn't want to eat it just is it definitely zijn you throw that base that there might be a funny your friend actively operating system yeahbsolutely and there are risks like an exchange having rap teeth as you're looking as good as he in there exchange if there's there take Congress I mean if exchanges are gonna create a separate token for beacon [Music] especially given a dis proposal maybe the different user be synchronous we prefer multiple children yes but then it doesn't bargain [Music] but nonetheless let's get straight away or maybe it just becomes okay either pound said like outside we go hey yeah so are there bounds on he's like so soon we can deploy our own is it deterministic you sometime like is there a bound or there was through what we're assuming that got system-level Castle okay that's so you could deploy any that like say just burns your computation they don't know you don't be able to make a block on it if it always capital the gas so this issue with like for techies potentially accumulating on the Deccan change sounds like you might just want to go to the next level and have like stateless beacon chain where like the ultimate gonna exist somehow access events feel a computer well I know it's more about like I think we already kind of like like we've seen that problem happened before you know what the stay for a blanket but I feel like I need is you my opportunity not like to avoid yeah you can make the tea I like the ideas you generally want these to be very expensive and you wanted a I'm show you so blocking up bounds state size if there's like a second on a block for size 30 or and burning also found this so like we don't explore those rather even if it was inexpensive but just a fee model like to see your rent or something yeah essentially has to have is the bacon chain it has to be synced by any production on this fully sinking into our chains they have to sync this component server yeah like if we set it so you have to like the Walker for any ether per kilobyte and people decide to form 30 30 million to create to blow the state by 30 gigabytes that I think you maximalist a lot he is going to be which how often do we need to over 70 design and hope and business meal so here like these are just crossing committees so they adjust everyone Willie set their own gas that's actually more of this layer one concept is like the gas against the gas limits by time when launched solid we have all sort of stuff moving it's moving as fast we have 64 choice the ultimately this and one that will allow some good things out Bastion seriously similarly with the ease it's a simple blanket rustles so you know it service like in the other reason why I like this as simplification games like basically if a lot of kind of sharp chain logic that we have been kind of wrangling whatever comes unless needed also like a lot of groups like this crew of custody complexity can be better produced down further because well basically you would be able to reduce everything to a single kind of front groups where the single front group is it just is importing an entire Charlotte and chin and that's something that was not possible with cross-links because across was just too big a importance in the beta channel [Music] the last one answered have a Mississippi lesson what will the transaction throughput being pressured yeah so I mean by these numbers I estimated at one point for me it's a two point seven megabytes per second of data so if we allow like computation and if we assume what transaction is as vehicle injured bytes and that's well worth a billion divided by 30 that was good but depends on the construction yes all the constructions it depends on whether or not you're using a rollup it depends on like what the gas do many people write [Music] in no version of the poses do we not open up I also I just I just want to speak to the two ranges of it I don't firmly [Music] and they're going out to meet up to have you for a multi-state it's gonna hold one it's a yeah so when you make a block you'd probably specify what he eats or or what side of BT's and that would interpret the block data as such against likely be asynchronous on the order block sloppiness proposal but there's there's definitely room to discuss intra be within lunch our last question tiny about other pups epistles of Paul talks about this today validators don't select a variable oh yeah so I don't seem like you're running to your ease right how does ostracize me yeah so I mean if you're it's it's a free market so in general like if say one block can have one evening there are very high about a beauty store lots and lots of hiding transactions was being picked up more than say an underutilized might only have demand to be processed every handful block so but that the actual mechanics are there but wouldn't that just needed an illogical team not necessarily and the that was an extreme case where we just have one eating her but if you allow multiple easier than yet likely like you might fill in the gaps at the end with like underutilized but actual last question but is it I mean is there rinsing would like to be saying that to the crystal isn't sponsor clock so it's proportional to the amount of contact that's not well everyone so if a proposer similar to oppose or anyone gets to kind of bundle things in as they see fit it's not a matter of like all the other validators coming in on coating you're not so but again if you're in a very underlip utilize e you might pay so very high i receive it will check if that's somebody include it's almost thing that those bees have some kind of catch to save on local boost I was just wondering how this [Music] and this proposal indefinitely I do it reduces the amounts of kind of crossable locks a thing that you can do but like you get some of the benefits back you penciling the one good your wife's get bigger so but they in terms of like having it the cash generally we usually think of these as having like a stapler 32 bytes per chart but we've discussed option opportunity of like when deploying the e depending on the amount of capital of earned like you can specify you know state refer short or maybe like some larger chunk of data which means that being like it like my cash so that's still kind of up to debate the actual mechanics and economics around how these things relate but it's something we're definitely I think we're gonna call it cuz I think we're over time and great [Applause] you