hi this is Charles Hoskinson I'd like to make a video tonight to discuss the launch of Shelley this is a very complicated thing yeah it's not like launching a normal cryptocurrency and it's not like launching a normal hard fork the reason being is that we're going from a federated model to a completely decentralized model and that has to be done in a very systematic careful way and so what I wanted to do is make a video and try out my surface here and do a screencast presentation and kind of walk you guys through with a white board a digital whiteboard of exactly how we're gonna do this so let me just clear the canvas and let's talk about it all right so in September of 2017 let's use black I get to be Bob Ross today okay in September of 2017 we released Byron and Byron runs in what's called a federated mode so federated means that there are collection of core consensus nodes and these nodes are basically responsible for running the network meaning they make blocks so since the beginning those nodes have been operated by three entities the Cardinal foundation of ergo and IOH kay and then when people install Daedalus Daedalus would be a full node so it would have a full history of the blockchain and also all the ability to validate the history so nodes would communicate with each other they'd send transactions to each other those transactions would be made into blocks by the federated network and then the full nodes could validate that so there was this beautiful interdependency between the users of the system and the core consensus nodes of the system and technically we could run this system indefinitely if the three entities wanted to keep doing that or we found other people to do it but the whole point of Cardinal is the centralization and one of the things that we've done is we've invested about five years of research into building a class of protocols living in a superset of protocols called proof of stake and basically the idea is that we want to go from a federated world this world here to a dynamic world so this is static and federated and we want to go to a dynamic and decentralized world okay so we had to figure out how do we go from the biron design this is Byron to the Shelly design where we have a dynamic and decentralized consensus algorithm okay so the first step in that was just seeing if we could actually get a group of people who would be willing to run card on outside of that federated set and that was the point of the incentivized test net and that was launched in December of 2019 and basically the itn really is its own cryptocurrency it's peer-to-peer it has staked pools and as delegation it's a full version of Shelly in a sense the difference is though it doesn't have the Biron history and so the idea of the ITN was to get data on some parameters in particular we were looking at things like how much ADA is staked and I think we're approximately around 12 billion here which i think is around 40% of the total supply and then we were also looking at things like how many state pools we were looking at things like Network quality so basically is a network stable is it up when people here are trying to send transactions are they posting and validating quickly are we making blocks and with the expected block schedule these types of things and here's the reality the ITN was an overwhelming success our decentralisation target was instead of having three entities to have 1,000 or as close to a thousand stake pools as we could get it's a pretty big difference three to a thousand and we've seen over a thousand register and I think around 300 or regularly participating making blocks which is a tremendous start tor it basically a test that that has some incentives behind okay so first step was to understand this model and make sure that lots of full nodes exist and we are there thousands and thousands of them exist the next step was to verify that there's lots of people who would be capable of running the network and we verified that with the incentivize test net and then the final stage is to begin a phased transition from the Biron reboot to full Shelli so let's talk about how we're going to do that phase transition because it's currently underway so step one was to get everybody on to byron reboot and built into byron reboot was basically all the bridge code to start the process of upgrading to Shelli so we launched this March 31st of 2020 and built within it is or Boris BFT Oh BFT you're going to see this term a few times okay so what's gonna happen is we're going to go from or Boris BFT Biron but still running in that federated mode this this federated mode up here these guys here okay we're gonna go from o PFT Byron you know boo there we go and then we're gonna go to basically Shelley upgrade period or phase then we're gonna go from the upgrade phase to Shelley hybrid phase okay and let's walk through what's happening in each of these all right so with Byron otft that's what you guys are installing when you installed Byron reboot and that's gonna be available next week for the main Daedalus client flood clients 3 was just released today basically Byron otft is gonna run for a little while and in parallel we're gonna start releasing a series of shelley haskell test nets and there's really three of them so there's the friends and family shelley Haskell test net so this is internal test net that we're starting to turn on and some of the more prominent state pool operators people who really did a great job here are going to be invited for an alpha test and the goal here is just to verify that everything we think about the node and the command line and these things are working correctly and that's not going to run for a long time that's basically just a burnin before we do the actual node plus command line shelley test net and this is really the the first opportunity for the stateful operators who participated in the itn to start rebuilding and redeploying their infrastructure one of the reasons why we're doing an internal alpha before doing this is that a lot of the people who participate in that curl alpha will help teach the remaining state pool operators and give them advice on how to configure we notice that there's a lot of cross-pollination in the ecosystem a lot of people participate on telegram and so we felt it was really important not to slow down our ability to deliver the product to make sure that we train a group of people who could then retrain other people so we're gonna do a forensic family then we're gonna do a Shelly test net in Haskell and that's a node with CLI and then finally we have something called a balance check and basically what the balance check is is it's everything it's gonna be the node the wallet Dedalus and it's going to contain the byron and the shelly consolidated history so everything together so effectively that looks a lot like what Cardno is supposed to be when Shelly is turned on and the primary validation of this is going to be that the ITN rewards plus the maintenances are consolidated because what happened is that some people participated in the ITM approximately 12 billion ada give or take a state in tens of billions of dollars worth of rewards been paid out for that test so those have to be consolidated with the existing and remaining maintenances and so that's what the balance check is all about when the ballot check is done this also is going to be end of life for ITN so the ICANN will be shut off at least are part of it it's already decentralized enough if it wanted to to run for longer because it's practically a cryptocurrency in itself but anyway no more rewards will be honored at that point that consolidation is going to happen after we've reached the end of life so balanced check will be the last haskell test net and basically the point there is to make sure on the consumer side that all the rewards are consolidated and also this is going to be where the exchanges will be given a full listing experience so basically that means that all the libraries and all the things that we've been planning for address dia will be done ready to go and if they integrate against this then they get some certainty that when the time comes for Shelley main net that all their integrations will not be broken though they'll be forward compatible from that day okay so after the balance check runs for just a little bit then what's going to occur is we're going to enter the Shelley upgrade phase so what that means is that if you go to the card on a wallet websites and card on Oh dot org and all these other places and those are being rewritten in anticipation of the launch of Shelley or Daedalus wallet and then other providers for example providers like Lloyd providers like your ROI basically what you'll be able to do is download a new wallet and that wallet will actually be Shelley you'll have all the logic for staking for delegation in the Daedalus user interface you'll actually see the state pull registration Center all of these things will actually be there but will still be running in Biron meaning that we'll be in a federated consensus and it's going to be running Oh PFT okay so there's going to be Shelley clients we're going to replace all of the core consensus nodes with these Shelley clients and then you the consumer can download the new Shelley wallet and that's going to run for a little bit to give a nice upgrade window because the exchanges have to upgrade and we also need make sure that people upgrade in the network and then we're going to trigger a heart Fork and once that heart Fork is triggered all the Bayern relays are shut off and that means that no Byron code is going to run anymore Byron's done so if any exchange is still left on Byron the wallets would be locked until they upgrade and at that point no more Byron blocks are going to be made and that means we've then entered the Shelley hybrid phase so let's talk about that this is the last phase so we have this thing called a deeper ammeter in our system and you know d4 decentralization and basically when we've entered the hybrid phase Shelley is going to have blocks made in two ways that Federation that we mentioned up here is still gonna make some blocks but not get any rewards so they're making blocks for free almost there come on okay great so this Federation is still going to make some blocks and it's going to make it with Oh PFT however because everybody has Shelley clients now and everybody can understand Shelley clients stake pools can start registering and operating and what happens is that when we break an epoch down so that's a period of time where blocks are made some of these slots are going to be made by the state pools and some of these slots are going to be made by those core consensus notes and that's controlled by that D parameter and it's ranging from 0 to 1 so in the beginning it starts at 1 which means 100 percent of them are made by the core knows then epoch by epoch what will happen is the will gradually decrement so it gets smaller and smaller and that means more and more of the slots will be made by the state pools that are registered so there are no limits to the state fools that can register that K parameter of course we'd like it to be a thousand but technically you could see five thousand register there's going to be competition there's going to be delegation and regardless of what the D value is set one hundred percent of state pool rewards will go to the state pools operating on that so we're going to do some video series and lectures explaining exactly how these work and these triggers work but basically D is in either-or situation so one of two things is going to happen either we hit some metrics and those metrics are gonna be things like if we go up here the amount of ada state metrics like network quality and other things that we were measuring with the incentivized test net and if we hit a collection of metrics that we feel really good about we say okay it's so decentralized we can just shut this off completely and let 100 percent of every epoch be made by the sync rules or some amount of time regardless of network quality so there's no way we sign up for an indefinite commitment basically this is just a phased way like training wheels effectively to gradually open up block by block to centralization to the system and it's a control valve because while we can move deep down we technically can also move D up in the event of some catastrophic event a bug or that the network parameters don't look we're like what they're going to look like so giving some numbers epochs they're gonna range between four to five days we're still working out the final parameterization of how long we want to do an epic because if preyas works a little differently i and if for example were you decrement d by point ten per epoch then this would be about two months an example or we could do it faster like 0.2 every epoch and then it would be half the time it would be about 20 days so this is a nice control mechanism that will allow us during the hybrid phase to make both opf T blocks as well as blocks made both or or Sprouse and it's important understand that these are not test nets these are real estate pools these are actual people making blocks and these are rewards going through so the whole point of this system is to make sure that we can responsibly retire this Federation here and it doesn't result in a catastrophic issue to the network so people have confidence that availability will always be there and we have a high uptime the whole point of running the incentivize test net and we had to pay rewards to make it a real life test was to get a good sense of how decentralized the network was going to be in the beginning so given that about 40 percent of those who could participate participated and if you look at for example Tron and Tasos that puts us about in the middle taste those is a little higher I think they're over 60 percent and in Tron I think they're sitting over somewhere around 20-30 percent participation so given this is a test net we said between two main nets that's a pretty good number so there's a great indication that from day one there's going to be a very large set of people willing capable and ready and given our phased approach where we're going to have these three events that give people plenty of time to be able to upgrade and by the way these won't run for months there the situation where they're gonna run for weeks each so it's just a it's just a courtesy for people to move infrastructure over we have every indication to believe that this deep period this hybrid phase will be on not very long lived anyway as the network gets more decentralized Wendy equals zero then we have achieved full decentralization okay and full the centralization effectively means that all the blocks being produced are being produced by Sting poles and there are new blocks being produced by any core notes and those have all been retired and shut down and what's really cool is that we've developed a new construct that allows us to orchestrate this beautifully it's called the heart for Combinator and we're going to be releasing some information on this this is another innovation that we created for Cardno that allows us to gracefully heart for from one consensus system to another and actually be able to live with a dual consensus system with still strong security guarantees and no loss of availability or performance so it's quite a complicated endeavor mostly because there's tens of thousands of nodes that have to upgrade we have to go from a federated consensus system exchanges have to upgrade wallets like you really have to upgrade and we also have to make provisions for these operators to basically migrate over get prepared and ready to go and then we also have to build a lot of infrastructure and turn a lot of infrastructure on like for example the Explorer infrastructure and it's non-trivial but overall I think we have a really good plan and we have a lot of evidence that that plan is going to work correctly so one of the advantages of this design is that we've been able to actually test and verify this design through simulations and formal methods so we've already run a lot of simulations about these switchovers when we go from a no PFT mode to prowess showing that those boundaries are ok and we've been able to really think carefully about this hard for accommodative design and we've been able to really think carefully about how we're going to handle the decrementing of that decentralisation metric so we have a very graceful entry into shelly there's nothing worse when you build infrastructure and then you have a jarring sunning requirement for upgrading and it kind of breaks everything and we didn't want to have that happen we wanted to have a situation where we can grace we upgrade everybody and this is going to achieve that I think with a high degree of probability we're going to release a lot of content about it Cardon org is going to be completely redesigned it's already underway there's a lot of designers from McCann and I which Kay Amer go in the foundation working together on giving advice about how that content needs to look we're also doing a reskin redesign of a lot of our GUI assets so in the coming months you're going to see all these things trickle out and card on org is going to be released the new website commensurate with the release of Shelley and when I say Shelley the Shelley upgrade phase so you're going to ask what do you have to do as a consumer if you're just a user of Daedalus and you want to stake then the only thing you're going to have to do is when the show upgrade face begins just simply go to Daedalus wallet download the new version of Daedalus like you did the buyer and reboot install that and you'll just wait and then basically when the hybrid face begins all the grayed out sections for the stake pull interface will be turned on and you can rotate your wallet from a byron wallet to a shelley wallet there's going to be a new dress format during that time period it's called bec 32 it's a great format it looks very different about half the size of the current addresses that we have and it's got a lot of cool bells and whistles built into it so you'll rotate your wallet when the hard work begins I and at that point you can stake to anybody you want if your staple operator you'll get involved if you're not in that inner group right here to help us kind of alpha test a few things you'll get involved here in the first shelly Haskell test net and that's coming very soon and basically at that point you'll have a node and a CLI and the idea there is for you to just get used to creating a state pool running a state pool communicating with other state pools etc etc so basically taking all of that infrastructure for a ride and if your everyday consumer you can get involved on the balance check side especially if you hittin the itn to verify that the reward consolidation is correct if your staple operator in exchange building infrastructure against the node when the balance check begins that's when we're gonna have everything ready to go for you on the address the aside so you can begin your burn in phase and begin your integration phase so that you know come Shelley Hybrid all your stuff is going to work okay once the Shelley hybrid phase has begun no Byron infrastructure will work at that point so if you have an old Byron wallet any of that stuff it's not gonna work you have to migrate so you have to download the new novel node and restore your old wallet to this to this new thing and we'll create a lot of guides on how to do that migration people have done it before if you participate instead of ice test at your start your wallet there that's an example of a migration so there's gonna be a lot of content coming and because we have an upgrade phase that'll give you plenty of time to be able to understand what you need to do more often not it's just simply a restoration and rotation of credentials not a lot of work but it's important understand that you must upgrade to be able to continue using card on Oh once the Byron phase has come to an end because no buyer in infrastructure will be running anymore and will be in the hybrid phase will be running that hybrid infrastructure so it's important understand that this decentralisation parameter in the Oh PFT this is mostly an operational component and it's a safety component but that safety component it actually comes for free for all users of Cardinal the operators of the michelle yeoh 50 nodes are uncompensated we get no rewards we make no money from that as I mentioned previously a hundred percent of the rewards are going to the set of blocks made by the or Sprouse stick pull operators so as a consequence is in everybody's best interest for this to go away as quickly as possible so our hope is to decrement the rapidly and of course that will be determined either by a time or will be determined by metrics one or one or the other are going to occur first and personally I believe that we're gonna hit a pretty high degree of decentralization within a month of launching the network if the trends as we've seen with the ITN hold so hopefully they do and there's no indication they won't we've had overwhelming participation and we're pretty excited about it what's also nice is this hard for Combinator model that we've come up with gives us a very clear path for doing more upgrades to the system and because we will never again go from a federated to a decentralized system all future upgrades will actually be less complicated than this so there will certainly be hard Forks in the future but those hard Forks aren't going to dramatically change the system as much as this we're changing the way the network stack works we're changing all the ledger rules we're changing the consensus rules it's the most significant upgrade from that perspective because we broke it into phases where we have the byron reboot as a major component of that and we have these nice phases for going from you know the early alpha to a Shelley test net to a balance check and then we have this transition phase we think we've built it enough built-in suspenders and have checks and balances that we can get through this upgrade without a major disruption another thing is that these phases because they're formally described we've been able to test them aggressively with simulations and formal methods and we've taken a look at a lot of different cases and scenarios and we don't think that there's going to be any major issues there but this is something that we spent quite a bit of time well over a year actually thinking about and carefully working our way through so to recap Shelley is going to be launched as a series of phases here we go from the byron reboot era to the shelley upgrade phase to the hybrid phase the reboot era is basically getting everything aligned with our boards BFT so all the infrastructure can talk to the Shelley infrastructure as people are upgrading the Shelley know so these two things can basically work together during the Shelley upgrade phase none of the Shelley functionality will be turned on but you can a full shellye node so and you're supposed to do that because if you don't when we get to the Shelly hybrid phase shortly after the upgrade phase has started then the buyer infrastructure will be turned off so you'll have a window of time exchanges will have a window of time manufacturers will have a window of time to upgrade and then when we get to the hybrid phase then the hard fork has occurred and 100% of the blocks are made with the Shelly logic delegations been turned on stick full registrations been turned on and actually an increasing amount of the actual network is being manufactured made advanced by the state poles themselves not by the OVF daemons and then we have this nice little control mechanism this D parameter the D phase that basically decrements epoch by epoch starts at 100% and it goes down to zero it'll either hit zero naturally over an amount of time or it'll hit zero once we get some predetermined metrics that will broadcast after we've got our final post-mortem on the instead of ice test that so after that occurs we've entered the fully decentralized Shelly phase and that means that card on O is being completely operated by a as we stated in the very beginning of all of this a dynamic and decentralized consensus algorithm and that's pretty much it ten kilowatts of power thousand operators a hundred times more decentralized in Bitcoin tons of great security guarantees and benefits super-fast performance in fact our belief is that over time or bourse could transcend well over several hundred transactions per second I've seen some beliefs from our team that if we could hit over a thousand transactions per second just with this consensus protocol and that's not optimizing for Hydra or anything else that's just the base protocol and because of the hard for Combinator idea and because of all the prep work that we've done it's going to be very easy for us to continue upgrading the system and adding in new for example the upgrade to gobin is just simply an upgrade to the legend rules we go from you take so to extended you to excel and we're also going to have a special upgrade for multi-asset support and that's going to have a dedicated language and other things to make it really easy to issue assets and because of this process it's actually quite straightforward for us to roll these things out very quickly it's also important understand that Shelley Goggan basho and will terr we're actually working on all of these in their own ways in parallel and so as we do this upgrade to Shelley the Gogan team is still writing Gogan specifications extending you TXO specification the multi-asset specification building all the Plutus tooling there is a full-time team working on Voltaire right now and they're gonna be making some announcements rightfully when we talk about Shelley when will you start launching Shelley and rock Cohn is the 1d risking basho getting a deeper understanding of what Hydra is gonna look like when we need Hydra and create some stubs for that so that Hydra can be implemented by future teams it's not really necessary now because we're not even going to saturate the base protocol for several years more likely than not but it's nice to know that when it's needed it'll be there and it'll be very easy to turn on so all of these are operating in parallel and basically this upgrade path is de-risking the most significant and Lara piece of work which opens up an easy transition to go in and voltaire so it's it's been quite a privilege to get here this is more than 24 months in the making it took over a year to design all the components to a point where we felt comfortable getting them where they were going to go quite a bit more time than we thought because there were a lot of education and issues that had to be very carefully taken care of and also because byron is a cryptocurrency it's not a test that it's cryptocurrency there were certain things that we felt were sacred such as the transaction history had we not care about the old transaction history in the old ledger rules and we didn't care about people validating that we have been launched to market much faster but that would have meant then that byron was not a real cryptocurrency so we treated it like one and we invested an enormous amount of time and carefully thinking through backward compatibility so that for the rest of time all nodes can look at history from September of 2017 moving forward and validate every single transaction regardless if it's under the old SL model the byron reboot model the new shelly model or the coming models under Gogan Pasha and Voltaire so so this is basically how it's all going to come together Aparna is going to do a product summit product workshop at the end of this month at the end of April and she'll talk a little bit more about this and then hopefully we'll be able to attach a lot of dates to these things you know there's a lot of people seem to believe that everything's six months out it's not things are coming very quickly so this friends and family thing we're already starting to set up Duncan has a Shelly node running himself and now we're just wiring some things to it so this phase is coming very soon and Aparna will make some announcements about it and then once we've reached the balance check getting to the upgrade phase is very straightforward and I'm not one for leaving large upgrade windows I don't want to have a 2/3 month upgrade window I think that would be insane I'd rather have an upgrade phase of about four weeks and so we'll take a look at where we're at where people are at if there's a particularly huge actor like a massive exchange that would like a little bit more time than necessarily open for discussion but outside of those possibilities I think it's entirely reasonable to have a short upgrade phase because it's not really that hard for the user to download a node and because the balance check predates the upgrade phase for enterprise integrators it's going to give them a few extra weeks of lead time to be able to build against that so this is how Shelly's coming out you know this is hopefully answering a lot of questions for people of wondering about specifics and everything I've discussed here is going to be broken down as we get closer to in exhausted details so there's going to be videos made blog post made tutorials and guides made some biossom by the community somebody foundation going over each and every one of these things and giving a lot of parameters and so forth also we're trying to get a single source of truth with the Cardinal work refresh so our hope is once that's ready to go you just have to go to one website and no matter what you're looking for whether you're an operator you're in the governance side that developer or you just want to install a wallet you have a place to go and it'll be a one-stop shop and it'll be very well organized and we're also rewriting Cardinal Docs at the moment so all the documentation will be redone as we enter in the Shelly upgrade phase so that people can use that documentation where they see fit to build their infrastructure so this is how Shelly's gonna work it's an enormous amount of work on our side almost all of its already done to get to this phase we had to write about a million lines of code across all the teams and as I mentioned about two years of development effort went into getting to this point this is also a great way for any cryptocurrency to go from a federated model to a decentralized model for example you could replace up here Cardno with ripple so if x RP wanted to go from there federated consensus to a different consensus model this would be a functional way to do that theorem is going to be a little different because they're going from a work based system to a state based system but had they been federated so you can actually move the system from the existing system to this so it's really exciting that we've been able to innovate a lot during this and what's really cool is that we've been moving in an evidence-based way throughout this entire process so the itn for example gave us an enormous amount of data on basically what those metrics should look like for how we can lower D from one to zero and also gave us a good sense of the overall level of participate network we were expecting maybe 10 to 20% given the nature of it being a test that the fact that we got four times more than the low-end is pretty amazing and especially given that the advice test that had a lot of issues in the very beginning so that's a super impressive thing and it's a super exciting thing to see that go through you know another thing is that the high byron reboot the the ability we've had to navigate that codebase and quickly change things add things in in just such a graceful way is a good indicator of the overall code quality that we're entering Shelly into so Shelly really doesn't feel like an alpha product that's going to require months and months of debugging and tuning to get to a point where it's usable it really does feel like a production system as we pull these components together I and it's just a testimony to the processes that we followed so they were front loaded with a lot of work but now we're starting to actually see the dividends on the back end and there's so much complexity when you start thinking about the hard Forks and things like that that have to be done and then there's always these questions of well what could go wrong and the fact that we can run simulations and use formal methods to verify that things can't go wrong or there's a low probability of something going wrong and also talk around these events is something that is unique to Cardinal and it just shows you the the magic of what we've done so it's taken a long time to get here it's been it's been a hell of a ride it's been exhausting but I'm glad we're here now and I'm glad that the communities with us and you guys have some work to do so look for the product workshop at the end of the month Aparna we'll discuss this in more detail some dates will start flowing in particular we'll give some dates for when the nodes CLI test that's going to hit and we should know that completely and then when that hits we should know when the balance check is going to hit reminding everybody we're either on a weekly or bi-weekly release cycle so these things come very quickly now and highly highly highly highly recommend that you guys check your your wallets out and make sure that your ITN rewards and meaning that rewards are consolidated together and that the balance you expect is the balance that's in that state because that's the last opportunity to raise a red flag and say there's some issue there during the upgrade phase it's not going to be that hard for most users it's just simply installing a new wallet and then just waiting for the Shelly hybrid phase to begin and when that begins then we're in Shelly all the grayed out areas and Daedalus will turn on and from that point on you'll rotate your wallets and be able to use Cardinal how and when you see fit this is a skeleton of a lot of different phases and as I've said repeatedly each of those phases are going to be discussed in more detail and we're going to provide a lot of content videos and other things if there's something you don't understand just make a note of it and put it either in the comment section of this video or put it into reddit and we'll make sure that when we create that content we release that content in a way that try to answer your question it's also important to understand that this process is not a slow one it's going to be a very rapid one so normally when you do test nets and phases you run them for months we're just simply not going to do that we're gonna run things as they're ready and get things out as they're ready so there's going to be a lot of pushing and if you miss a date for example this note CLI forget about it go on vacation if you can and coronavirus times you might wake up tomorrow ins and see that balance check has already started so be very very attentive watch our Twitter feed watch our YouTube channels I'd highly recommend subscribing to the emerg O'Connell foundation and I which K official Twitter accounts and I'd highly recommend subscribing to the IO HK YouTube page and making sure notifications are turned on because stuff is going to be pushed out on a weekly basis in some cases daily basis as we quickly March our way through this through this upgrade cycle so you can stay as informed as possible if you have any questions to put them in the comments you can always tweet at me too many of you of course will and other than that thank you so much for listening this is my first whiteboard video I've done in a few years but I felt it'd be appropriate since we're just about to get enter the Shelli era to to do that and I hope everybody has a smooth and graceful upgrade to Shelli I know I will thank you so much
