hi everyone this is Charles hoskinson broadcasting from warm sunny Colorado always warm Always Sunny sometimes Colorado so today is September 27 2022. I wanted to make a quick video to talk about a idea scale proposal that I believe is passed called Daedalus Turbo let me share my screen so here's the proposal deadliest turbo 10 times quicker sync and it says the problem statement is that Daedalus the centralized Cardinal wallet is painfully slow taking a whole day to sink initially in hours to resync when used only occasionally on favorable impression of cardano for new users it's true it's a full node and every full node in the industry suffers from this problem for a variety of computer science reasons so anyway looking at the solution set because they do some calculations they say well we should improve the network utilization and say prepare a method for a more efficient delivery of historical blockchain data over Network for a new dental setup we don't have a sync problem you get with the network protocol the way it's designed information as fast as you can this is not something BitTorrent is going to solve because you can get the blocks you can have a snapshot of the blocks but you still have to do the validation and then it says things like prepare an algorithm for quicker reconstruction of wallet history for on-chain blocks chain data and we have at least a two approaches attractive indexing and better utilization of multi-core CPUs okay but again you have to verify cryptographic operations and these are transitative so you have to do a sweep of the entire history unless you Short Circuit the cryptographic operations so so you can't just put B tree indexes or hash tables or these things unless you're proposing that you want to bootstrap from a checkpoint okay create a Daedalus launch of wrapper desktop ads for Windows and Mac OS and command line utilities for Linux allowing general public to benefit from Quick syncs while the algorithms are reviewed by i o for inclusion in the mainstream code there's a lot here and it seems to be a decent team I don't know if they've ever reached out to us or talked to us maybe they have but it didn't get to my level yet so just let people know there is a plan to make data lists a lot faster and that's mithril mithril allows parallel validation of ethics and then you can utilize a lot of these computer science techniques that Alex is talking about but it does look like this idea has been funded so what we'll do is reach out to the team that's planning on doing this execution work and actually it would be really cool if this team has the capabilities to do the mithril integration into the existing data list code base that'd be great but I just wanted to bring this up real quickly it just came across my desk literally five minutes ago and I'm a bit surprised but I haven't really looked too quickly too much into all of the proposals that fund9 has approved and dental is Turbo I I understand the need but people do understand you're running a server you're running a database and you're running all the wallet infrastructure it's a full node it's a very heavy artifact and it's not a network problem or it's just slow download speeds that are causing this the issue here is that when you get the history you start from Genesis block and you have to sweep through the entire history you cannot necessarily unless you have checkpoints parallelized the data because you you one depends on the other it's a recursive structure so one block goes the next block knows the next block and you have to follow that tree as it grows we've done a lot of things to uh to try to optimize this wherever we can and you've noticed throughout the years that we've introduced new data structures and other things to make it as fast as possible in fact utxohd is about adding an actual database into Daedalus for off for or off when your applications turned off for storage because actually Daedalus uses memory database and that's as fast as it gets so this is just the reality of all full nodes in the entire industry since back to bitcoin there is no Silver Bullet unless you introduce new cryptography with different security assumptions to make it much faster or else we would have done that I am a Daedalus user I'm one of the people who's most frustrated by this particular problem and I feel the the umbrage about it which is why we started the mithril program and we started talking about a quick sync algorithm that mithril would enable now it would be cool to have a parallel development team take mithril as it gets more mature and build that fast sync into Daedalus to speed up that delivery to Market and actually with the Catalyst funds here if the team's capable it would be really cool to add them in as core developers in that respect but I I'm gonna obviously get asked questions about it and we now have to talk to Alex and the rest and I apologize Alex if you've already talked to my Engineers I've just literally seen this and and so I just wanted to immediately make a video about it we really love the fact that the community wants to touch core infrastructure and this is the whole point of the members-based organization and the concept of a technical fellows program so you can separate requirements and specifications from actual implementations and our hope is to get rid of a notion of an official wallet altogether and instead have a certified wallet versus uncertified and under the certification standards you can put functional and non-functional requirements including benchmarking and performance requirements for user experience and it would be really cool to build some protocols to make things run faster that was literally the point of mithral was to massively accelerate things like bootstrapping a full node and so forth but you have to understand a Serial process is just that it's a straight arrow and you can't really cut parts of the arrow up and run them in parallel threads or be clever about data structures unless you're willing to trust a checkpoint if you're willing to do that you can actually have a full node with no validation which is a like client at the end of the day just with more data in that respect but if you use mithral with the same trust model then what you can do is assign a mithril certificate to an entire epic then you can run all epics in parallel and each epic could be on a different processing core whether it be GP GPU or an actual CPU core and then you can validate 18 at a time and run as multi-threaded as people can allow it would be nice to have different transmission media but BitTorrent is not secure it uses a mechanism called cademlia and secure Academia even has problems and that's why we constructed the network protocol that we constructed and the peer-to-peer protocol that we constructed for cardano now there's probably a lot of ways to accelerate this but it's it's it has problems asymmetry attacks DDOS attacks Eclipse attacks these are known issues so if you create a bootstrap process that could be ddosed or hijacked you're actually creating a security vulnerability on the network side of the system uh so uh I'm we're definitely as I said going to reach out and we'd love to find a way to work with this team this shows you the decentralization of cardano we don't have any control over the distribution of catalyst funds uh so if the community wants this uh this project will get funded and uh we're always willing to work with anybody who has engineers and funding and they're capable of doing things so we'll roll out the red carpet and we'll find a way to accomplish that as for Turbo I do believe we can get uh great performance and I think it'll be end to the processing threads once we have a a proper protocol to bootstrap from so mithril in particular and that would be a great integration into the Daedalus team so uh just a quick video about this and I'll have more news for you guys my next AMA and as we have a chance to kind of circle through and talk to everybody uh and messages received about core infrastructure I and I we do understand the frustration I have it myself in that respect but you have to understand this is not unique to cardano this is literally the user experience of all cryptocurrencies that are real cryptocurrencies because a full node is everything it's a server it's a database it's a wallet it's the whole stack and it has a lot going on under the hood why running Daedalus you are running a full copy of cardano you have the DNA of cardano in that code base and any one of these nodes could ReStore cardano in that respect okay all right I'll talk to everybody soon cheers
