P2P Summit - Where Are We with the Ecosystem and How to Push It Forward? by Chris Waclawek(Devcon5)

[Applause] hello it's like Rockley 9:15 everybody's here have which is amazing thank you so much we really didn't expect everybody showing them time that shown that you're either jet lagged the right way or that you really care either way it's really really good let's walk through the day very very briefly like right now 15 minutes of just welcoming wards and like shortly throws like everybody will say roughly 102 Oh sentences like we were like why you care about p2p or or maybe you don't major you're just like suppose it's about like that's also unacceptable as scientists you know you do you and and then we have project presentations and discussion vlogs in the project presentations like we tried to squeeze in as many as we could so like we're kind of gonna be my new here my new is amazing he's part of the DEF CON staff and like he's gonna be keeping the time he prepared everything he's also he also prefers the cards that showing you when you're out of time because I think like about half of the room are speakers right so like you know pay attention to what multis showing you we know that these are short time slots and it's impossible to cover everything in such a short amount of time but it's more of like hey we're like galvanizing the community now and like starting a thing that we're going to be doing for at least three months together working on mapping of the p2p ecosystem solving challenges together and you know and things like that so basically this is more of like boys who what are you working on what are the challenges and like let's work on a common vision for solving them we have - for the call today Carrie sighs here we have Demetrius enthroned stages we have September to give out reputation in p2p networks we have three trade showing wireline we have Deb Russian with Altea very all showing talking about redundancy in p2p systems I don't think she's here yet but she's not arrived sir Antoine is gonna be talking about testing networks with with Wyatt block Matthew sleepers in a give overview of SSB jaya promises we wanted to be here but I think she small accidents last minute so she won't make it this time phonemic we'll talk about al-qaeda's need for a p2p Mike and Molly from protocol lapse will will cover lifting to be entire many ipfs and I know there are like couple of people that email last minute and like what to do a shirt like to talk like we should have time for that like five or ten minutes depending on how many of you are here please raise your heads okay a piece here that's perfect and I know that there's two people that are gonna arrive soon so like like if we if we stay on schedule like we're going to be able to talk about all of this we're just going to be variable and then this fashion looks we have like you know a bunch of people we're going to be taking notes during the during hearing care presentations and like it will be great for you if you're presenting you also briefly mentioned like what do you think the biggest challenges are for us to put this forward and would definitely be great to map this out in the discussion part like what the challenges are and what the action blocks we have some ideas about that we've been working on something we all need to do on that like before we start so you know we really have high hopes for the discussion part but to keep things off we have one's own exterior too like probably remind everybody why this matters and also well I would actually actually do kind of a kind of a brief introduction of the idea of p2p and maybe maybe a better way to do this I was think about this when you mention it I would like to kind of understand what everybody thinks that term means what is it what does it mean for everybody it's here because we have a pretty diverse range of projects doing a lot of different stuff in it you know I think there is a common cause here and maybe the reason this room is kind of distinct is because in some respects some of the people here are working on things that can work in conjunction with a theory or quite separate from it or the projects like that so maybe instead of giving kind of introduction that's stating what PD is I'd like to hear from you guys in here what you think it is I'm going to I've been working with a lot of projects in this space currently we have a wire line and for me a p2p is about you know kind of a kind of a way to get sovereignty over over the sort of digital world that we're creating anywhere we don't have a lot of intermediates or any impossible you can just directly connect with the things that we care about so where would I move from here one direction how would you sir just myself and saying okay with you and peer-to-peer you say whatever you like to our device and being able to talk directly with each other and figuring out that having some utility for us what's the project through orientation and coordination through boxes coming remember Texas organizers identity and pitch me in Texas audiences for pizza so much here it's like BitTorrent and Napster but like next phase I think of hey Felix I worked on deputy for yet mate Green I work on lived to be status and beer is anything just sort of the centralized infrastructure and working on that identity that IT projects so research per Connie just exploring no algorithm founders of the context are interested in - something like hope it was hard what for really we think software Hey Gerard link that I have is a company over a block we were building decentralize get out the place Engineering [Music] I am prem KO I'm the researcher it I'm Renko I'm a researcher of zero X and looking into decentralized the chains and nearly beers importance of Roswell's permissionless and inherently Brazilian [Music] I am draw stand up so we can see you I have Chuck I stay aside when you work on a small country packages but disappear this I [Music] so they can accommodate up I'm Danny from box your voice by researcher stop I'm Matt I'm from pinata to do idea testing for structuring services and I'm here to see kind of what everybody else is here apparently episode Megan X 2005 70% of the internet was peer-to-peer and I was like 95% centralized and obviously something changed narratives are pack animals [Music] Michael also like Danny from three points for economists repeated data occasionally some developments of the different body of the system yes Chen security must be interested in attack like I [Music] you guys need I'm Bernardo I work for request Network to connect our notes you know what you both up ah hi i'm mike elder i'm from protocol labs I work on the LinkedIn in project and for me peer to peer is about eliminating centralized control points in network so that you have no single theories more powerful by the way we miss you guys what I think I want of course is just basically an introduction and also like what I'm interested in what is so oh hi Molly I work on my peoples project man each key to me is about making a more resilient network that was not connected through so that no matter where we are or how many people get together we can still make use of the tools we really care about [Music] hey Brian so tell us who you are what will teach me sorry said again and won't be the B's dude hi I'm Eric maybe have useful definition because I'm worried about data structures that playing welcome to networking y'all can figure out all right she's just hiding and it's fair enough I'm merging us with Hispanic labs and p2p well there's something that shouldn't be anyone hello I'm happy home by working on company Network Japan and my develop software for Andrew artists using distributed ledger system and interfere I like interfere girls get beer system is more natural for me thank you and when intercept I work on cell [Music] it's best that I work in [Music] taras broidy's Matthew life's working am using a new party to client and and occasion for each other with all essential ice I'm Abby I worked with radical and iOS Queen which are both building such alliance alternative to get here to here at code collaboration and not say no to the deep platform and [Music] I'm Dean I'm a researcher I think PP is about creating more accessible networks and so that's what I mainly it easy to use more accessible good morning everyone I'm our own project two technologies so BtoB us community communications and they try to concentrate and explore more about how data can be saved forever without loss [Music] the box for me TV committee to be is a waterway identity identity you know the freedom from our eyes [Music] hi everybody name is Antoine from wet look I talk to look at the bug go to this point so me because period the future specifications now rules away so willing to see more more that we specialize in testing those things and bringing them as much as possible [Music] I'm Renee and we'll solve o'clock I'm here because networking touches our program work I work with that wire line and I like a PCP because of its permissionless nature okay everyone my name is Matthew slipper I founded a company called kyojin and we help other companies build protocols and apps and I think that's this is the fear that I was thinking about it see peer-to-peer is important because it's the substrate that allows you to build a credibly decentralized applications you can't have one without the other and here we are to build a great foundation joon-ha yeah I'm Julian I think open foundation and I think that for me p2p is important because that's something that enables emergence of complex systems in a bottom-up way exponential way with the explanation of growth hi I'm rich burdened I work for a while ein I'm interested in p2p because the internet was originally designed to be tolerant to a catastrophic nuclear attack and still operate I think we still have some extent existential risks today that we need to guard against and I think Peter B is a way of getting back to the original origins of the Internet hi I'm Ashwin work for wireline and for me we do be is about sort of replicating what happens in the real world just taking it online all your kudos all your connections in the real world thinking them online and just dealing with you know people directly thank you hi Manuel and I'm basically working on a collective venture builder so we're trying to be like rocket internet but with peers that are building stuff and not only developers but also my tears operations guys finance financial help that any project needs and we just connect all these tools and bring people together what we do we do well something April's us to work with that somewhere in the middle yes who are you oh hello why do you do this hi I'm Frankie I work on madam ask what what was the question what is peer-to-peer to me yeah I'm just here to help everyone out you know [Music] so I think that's about as good of an interim generally speaking it appears one can hope to give because this covers the philosophical the social but there's one more person perhaps you'll have the size so the question is what's your name what you're working on and what they can be we're just finishing around the future what is cool when the finals work final worth now that's about it I mean like it's a quite a diverse topic and sounds like we have a number of people here who are but the right people to help us come up with a definition of what this thing is gonna go first okay so we heard that it's a community during project for secure real times distributed data sync that's like a nice way to say peer-to-peer but yeah so it's a pretty clear critical a little bit about me I worked at Google and they were close to canvas I found pandas for a little bit sold that companies in that era but I could like deal with like large companies so I just always was doing like small projects I also worked in anaconda so I came from like a science background right and I was doing research and science is messed up so at this time when I was talking about - okay everyone is talking about that half of published research could not be replicated like the data was just like locked up in silos it was like completely gone link route was an issue so scientists are still talking about this problem is still not really solved [Music] so we were talking at this time in 2013 how do we replicate research and then the credit extrapolate on that problem and talked about how to collaborate on data analysis because that's all we're really talking about right so it's a terrible they still do this like if you're having a server and you share that server with a bunch of people you know bad things happen so they started using like git or email or arcing as three Dropbox but all these have their own issues we're trying to claw grown up in analysis so I found this guy in 2014 this is max often and he was talking about let's do for data what get could do for source code like this is your little idea definitely do this because git was just not working it I don't even try to put a terabyte in to get maybe falls over so also working with me is this at this time and there are other people talking about this to look at this is wants on this computer panel this is me they're all just kind of hanging out in an archive for years ago but you may see the start an organization called that and it was focused on how we saw the scientific problem of sharing data sets that people still haven't be able to really solve and we were we weren't a company this time so he got nonprofit funding so what everyone's talking on science the same time there's like revolutions happening on the world that's part of Occupy movement we you know we're watching as the world was going into basically chaos because governments had shut down the internet and then people go to the streets and became really clear that peer-to-peer can be used for more than just scientific data sharing and we got really excited about the idea of building applications on top this so you took BitTorrent as one of our main inspirations actually use the BitTorrent DHT for a little while in the beginning so there's some problems with BitTorrent it only works for static files and create skills really poorly on really big datasets and it's really easy to eat straw and find out what people are sharing so we wanted to solve these problems so our goals are it's easy to use and it was an alternative to Dropbox and we wanted to be dynamic in real time data built for modern scientific connect with these cases and always be faster and note eavesdroppers so you take this encrypting ultimately secured that link and it never changes even as the data is updated so how do you do this you build it on top of a log right depend only loves I have to go to this but how can we share append-only logs right Merkle trees I'm not that's I want Angela Merkel just like arbitrate so this is what I'm not going to talk about garbages work because kind of 15 minutes but get okay this so wanted to really talk about real-time distributed data sync though and get isn't great at that so how do you do real-time that's what you need to build applications so every time we've been data the group has changes and this is the issue with real-time data so we brought it up to you to crypto to the rescue generate a key pair a public key and secret key so say someone trusts the public key that I'm sharing you so I assign the group of a of the rehash with my secret key and then as I add data we sign the new group right so then Bob can use this public key to verify through hash signatures over time this is how that works this is the basic building block of data file type report which took about four years of development like with your like multiple cycles of development and this is hyper core hyper log mr. hyper core and hyper ghor protocol so what people be doing to build applications on top of it you had some storage like file system or s3 or maybe browser cache generating index over the data so say you want to fetch or query that data in some way as it's coming in you you create indexes over it and that's really a database right so then people can start creating applications then you have a network so everything is built in a modular way we don't say that you have to use any sort of networking protocol you just have to somehow replicate pipe type of data in to the network and then get it out on the other side so we use WebRTC they use UTP empty NS to find Pierce is DHT it's very modular people kind of put things together I think so this is the foundational building blocks of our my paper which came out in 2017 it's part of the nonprofit called the hood for science which means fact I found it so now we're never like 7,000 downloads a week on NPM derp all the modules and about like 1,500 dependent projects that we know how to play through you know internet scraping create our own dog food also so I went often for a couple years contracted with a couple places Matthias and Max as well and started building applications on top of this and we start anticipating in the community start community grew over time so building you user interfaces on top of the stuff one of the biggest purse user interface' speaker browser which is a web browser they uses debt to send data around there's also another big one that I got personally involved in called that bale which is a collaborative map editor to take the OpenStreetMap interface put in a desktop app and put peer-to-peer back-end ready we work with monitors and Peru and Ecuador and Guiana to map their territories there's also a mobile app that uses that using node.js mobile so they connect they take pictures of forest and oil spills that are happening and they put them on a map and together they collaborate with the community so everything has to be offline we're talking like six hours in canoe they can't have access a minute there I'm accessing it and then they create maps and sent them out offline and they use them in campaigns and in legal defense they recently won an amazing campaign in Ecuador and protected a lot of millions of acres of pristine rainforest using the technology pure Maps is another one but this is not about local and collaborative editing it's about taking OpenStreetMap pickles OpenStreetMap and distributing it across the internet really literally taking the principles of BitTorrent and study at home and applying them to OpenStreetMap and and there's they're actively working on stuff it's really amazing if you want to watch them the two more care social media so there's a chat app built on this called cabal there's a collaborative workspaces called pushpin these are all desktop apps by the way and then mobile apps there's a scientific data network so dry adamant Californians internal library is collaborating with the University of California to distribute data sets across they are data centers with a cooperative posting network called Co box they're actually pretty new and they've done a lot of amazing work in the past three months it's like wild watch them live on github so you should totally check them out they're doing a lot of like multiple writers and like preserve resolving conflicts and stuff and Klaus is Simmons for the book ara it's a constant record network they've also put on much ss1 github they're more aetherium focused so and then wireline n'yar from later today that's pretty new as well as mail simple stuff yeah so people are using the stack right now to build stuff the problem that we're seeing now and you know where are we going with from now so I have this amazing set of apps and people are using it's distributed in the centralized organization and there's like no central control of what we will build the issue though that we're trying to put this on now is how how you can make this these three parts a lot easier to work with because right now when you come in and you try and build an app you would kind of have to build your own database from scratch every time which is well it's like I'm probably sure if some of you in the room to do it but like someone who wants it just build an app Platzer raise a lot of money hire some distribute engine software engineers it's like it's a huge task it's very complicated and things break so what are we going to do to try and fix up create a software development toolkit hot so people can get started quickly and prototype do very rapid prototypes and then if they want to have a lot of customization they can start going under the head after that also helps if you have like major bumps or big things that happen underneath the hood you know if they change your entire application because an API down the snap changed so I was using the non names but here's the actual names of these signs on github you wanted to check it out hyper core protocol pipe report we're coming out with these new modules this year hyper tribe pepper salon and this is going to be part of what we call that 2.0 so this will be this already these models exist but we're not putting them into a cohesive package till early next year so how much money do you think that we've spent at the death foundation no I can't yeah 1.6 million since 2003 13 so the that foundation is us-based nonprofit started 2013 but formalized in 2017 when would 6 million over the past six years mostly funded by the night in the Sun foundation although in the past year we've gotten a lot of individual and organizational contributions our current operating budget is $300,000 we have like forty thousand left in the bank right now this is like classic non profit like always writing grants always like getting people involved in getting donations so how do we do this well it's community right it's not about the money it's about people caring about this and contributing time so our values our community first transparently funded and I have open governance so we have a series of working group so you can it all comes down to making good software and giving it away for free and that Network effect of that good software and people care about the scientific values of that software the computer scientist that software is like it's people excited and together we can build this amazing applications and work together so that's kind of the vision for our future so as we go from networking to collaboration this is kind of our focus for the next year is how can we get developers from networking and I don't mean like peer to peer networking by the way and had a conference networking to cooperation for coordination coalition and collaboration so our plan over the next few months to the build a membership model and next year have to move an event and start a community grants program and you know people say we're standing on the shoulders of giants I'd say like we invite the giants and come sit at the table questions questions we have we have time for two questions he took aim there are no question I'll even ask one quick link when you receive the grants from the Foundation's what was the start they're interested because of scientific itching so most of I think like 1.

3 million of that money came from scientific ditch there's a lot of money in the reproducible research case that are still going after grants in that space in brain they found us they're using them open source modules and they're contributing to the code it's all open source community so we have given them any money that's just they found us conferences yeah that's it yeah thank you guys so I'm between an origin of its status this song is going to be mostly permanent people need to hear from the front of resource constrained devices so sentence is a company that primarily works on a distributed chat like application which targets phones it's currently heavily multi-building flies forward with its major focus being that we help define the 2.93 inspect and there's also that pursuit of something called what about he could be which is the tool in to build to get to build the p2p solutions like application level so how does statuses it's a distributed chat application and it was robust peer-to-peer stacking order to will feel it's from the weepy to be is no no sorry the problem is that the current atrium solutions peer-to-peer solutions are not necessarily didn't necessarily feed the status require separately so we are going we need to be which is a wide alert actually in swapped components in and out it's a transfer support you can use this B is the new base oh there's so many men in the wheel oftentimes when you're building cured of your application you tend to sort of build up your career start from scratch it's probably not a good idea so and there's also lots of buying from from other projects again itself hold it up and the five point which is from Langston regional creator so our implementation of the big build using something called name rather than that which just want to give us a little bit of props instead of making language haven't sinned try it out it's super fast it's got mega browser support and it is they call it super productive but and it also just got version one took a while but it's here so I want to talk about the challenges so both feet I'm sorry peer to peer in general and those arts scalability so maintain stable connections on in terms of your networks it's very challenging that's one of the biggest major issues scale in hardware being very diverse you can find them a few servers or you could be running it on a browser or on your phone so I'm not convenient you have to go for that this coming of this also a problem there's different solutions for dad there's heaps T residues etc but out of those is the only sort of discovery mechanism and it is relatively easy to Samaritans although there are ways of going about it g'day Emily a house and fiscally only entity providing some workarounds or solutions when you get to mitigate the seasons that's our problem a problem as well the what else yep bandit defended it's a big one usually in intronetworks there's a lot of traffic and discovery and and then privacy and anonymity you can build that you can encrypt data and prevent from it being inspected and things like that is still weak a lot of metadata and usually or sometimes that's enough to basically to invalidate every other security measures happily and then in order to support the work to run a large computer during a break you need some sort of incentive Asia I think that's the thing we haven't really saw yet [Music] so this is some solutions to some of this problems that I just want to bring it up and bring them on there obviously the tree gets to build clients that always contribute something to the networks there's this idea of my clients and Realtors and usually that implies that the client is basically the most cases leaches data from other bigger notes in an effort or more powerful notes and they're usually that doesn't speak scale very well does he have to end up with a lot more like lines in the never again than full clients and they're only attendant it's very hard to scale during effort so one of those things that I believe that should that peer-to-peer network should have is the ability of beauty line that every single resources available to hear a light client and you can steal serve some chunks of data to other peers in a network if we're talking about data specificity and the probably should do that it's gonna help scale them and we also wish it also sold the patientís there are several projects but and obviously this two things are not too exclusive there every weekend yeah that looks other time which is great we're back on track now and my question is from the room there already I like that something so satis is one of these one of these projects which basically kind of good men estimated 1/2 as a result I guess you could say the original problem [Music] are you looking at statuses and its current development is a way to sort of match that whole thing is it is it it's not explicitly yeah so it's basically trying to conceal the like in scalability or okay so yeah status is basically trying to help those scalability issues that get run into trying to sort of build its its WeChat like application and yeah now it's pushing forward was that the question I was asking [Music] my name is September and I work on the cellar protocol today I'm going to be talking about I can trust which is a protocol that I wrote back in 2002 with a couple colonies of my mario schlosser and Hector Garcia both and I could trust is all about trust and reputation in peer-to-peer networks and I want to talk about it today because at the time we worked we were thinking that there was going to be whoa I'm gonna walk through basically the context where I didn't trust when I wrote it then I'm gonna walk through the protocol itself then I'm gonna walk through why is relevant today and then I'm going to talk about some interesting research that we can do going forward that's a lot just a little bit of time so I'll just grab it jump right in so we started to the context in which we thought about I can trust was back in 2002 peer-to-peer file-sharing networks we're getting popular and at the time the record industry saw this and started flooding the peer-to-peer file-sharing networks with bogus copies of music so someone with search for say an anime 8 Mile Road and then you get 500 razones latter ok you get 500 results 400 of which were inauthentic copies of the song and so the way that they would do that as they would come into the network with with an adversarial node and start spreading that information on the network and we saw that at the time and and while it was the application itself was somewhat interesting what was more interesting to us was that we saw at the time them we felt like peer-to-peer networks would start to get very popular and this problem was not a problem that was just about file-sharing this would come up in lots of different scenarios and we would want to have a situation in peer-to-peer and networks in which nodes adversarial nodes could be identified and people could be for good good acting nodes can be prevented from interacting and so there's a couple of ways of thinking about doing that I mean the first is you can imagine just having a proof a cryptographic proof that this file is the right file like this song is actually in a lot of cases that's difficult to do I mean in the case of in that case it would be difficult to do without say a central authority actually having to copy a song and that's not possible and so the first thing we choose to chose to take was to assign a trust score to every peer in the network so for example if if this peer was choosing to download files from either that geared up here it looks at the pier where that a trust 4.9 and chooses to download the file from that gear that I mean it's straightforward but the big question is how do we come up with that score and I'm going to talk about three approaches each of which building on top of one another culminating in the eigen trust approach so the first is just using past history and so the idea is that you have interactions with a peer and if the interactions are positive the scores go up and if the interactions are negative the scores go down and that's a pretty straightforward way that's how we we interact in real life and so so we can mathematize that by creating a vector of peers we'll call this peer peer I and then therefore this this trust vector would be CI the vector of opinions that this PRI has of the other peers on the network now the problem with this is that you only have with a small number of peers on the network but there's a lot of peers on the network and so you don't have enough coverage with just past history so the natural thing to do is to ask friends ask a friend of a friend this is also what we do in real life when we don't have enough coverage and we can mathematize it pretty straightforwardly so we ask a friend what they think about their friends and we wait their opinion of their friends by how much we trust them so this is CI the opinion vector of PRI and this is CJ the opinion vector of this peer and if we want to put it in if we want to put in a matrix form we say okay well let's ask our friends J what they think of peer k and wait what they think of your k by how much you trust in matrix forms and look something like this so this is I called the ask a friend operator that you have a matrix C transpose C transpose which is the list of opinions of everybody it's the network graphic of all of the opinions of everybody on the network and then you just multiply it by your own opinion and you basically take one hop out the problem with this is twofold you either have a lot of friends who have a lot of friends and then you get good coverage but you have to do a fair amount of computation and storage or you don't have a lot of friends we don't have a lot of friends and then you don't have to do a lot of trust you don't have to do a lot of computation and storage but you do end up having the same problem that you had in the first place like you don't have enough memory so what we really want is we want lots of coverage and low complication so one way to get lots of coverage is just to keep asking friends so you start to ask your friends and then you ask their friends and so you just apply the ask to the friend operator again and then you keep asking until the cows come on them so you have the trust vector is equal to C T to the N kind of CI and there's a great benefit to this and it's not just the cows is that this trust vector when you compute it in this way actually converges to the same thing for every peer on the network so what we've done by asking friends ad infinitum is that we've actually computed a global trust vector for the entire network and so therefore you each peer can collaborate in the storage and computation of a single trust value for each person so a single stress vector for them [Music] and the way that the way that works like the way that we initialize a trust vector to any any any vector let's just start with the uniform mentor and then we just keep repeating applying the ask a friend operator to this iteration and you keep asking and eventually I'll converge and look what it will converge to is the principal eigenvector of that trust matrix the in principal eigenvector of the matrix that represent the opinions of everybody everybody in the network so this is nice and if we have a centralized system or even if we had a block chain this would be sufficient we could just take the we could just take the opinions and have a centralized computer or lots of computers operating in consensus compute this computation but we can take it one step further because in a true peer-to-peer network you won't have any single machine and you single node storing the whole matrix or even is storing the whole matrix or being able to compute the whole and supplying in fact what you have is you have basically each node storing is own row its own basically his own opinions of other people and so how do we compute a principal eigenvector when you don't have any machine that knows the whole dangerous well one way we can do this is we can recognize that we can take this matrix computation and actually break it down into a single element conversation by saying that this element the ti element is actually equal to this row times this column and to be a little bit less abstract about it what it means is that the trust my own trust score at any given time is the the opinions that people have of me weighted by their trust scores the trust scores of people who have and so in that case everybody can compute their own truss core at any time was they just know those two things so you end up having a pretty straightforward protocol were you for each year you compute your friend trust value by taking the present trust about the present people who have opinions of you their opinions of you times their current trust values then once you compute your current trust scores you send your trust fund and the opinions that you have of other people to them and then they aggregate those that compute their trust score and then you wait until it goes around and they send it back to you and that iteration keeps happening over the course of the network and that way lots of peers all over the network end up collaboratively computing the principal eigenvector of a matrix that none of them knows the whole matrix so when we would be when we wrote this in 2002 is really you know like there's going to be lots and lots of peer-to-peer application just right around the corner but now is the time I mean I'm seeing that there's lots and lots of peer-to-peer applications right around the corner and I'm seeing that those peer-to-peer applications will need notions like reputation identity the ability to identify in a decentralized way adversarial nodes and in a decentralized way be able to self organize to avoid those out adversaries and and the exciting thing this is not really about music sharing I mean you take a look at lots of different graphs and you can imagine lots of different computations that can be did trust computations that can be done among those graphs so for example I mean one thing that I'm very interested in is active stations and by and large I mean people think about identity computation in the front Matis access station so people will attest like I will plant that I have a certain property and other people will attest that I have that problem but the problem with that is well I mean who knows these other people who are or attesting that I have this property and in fact I can make five hundred civils but all the tests that my name is John wait a second and so though but if we wait attestations by a trust score then all of a sudden we have a test issues that are much more justifiable similarly if you imagine kind of a decentralized finance application you can take the graph of loan repayments and then compute a credit score in a similar way in a decentralized way another another application that I really like is actually I have two minutes left and I wanna I want to move on to questions so let me just so so to me this is a very exciting time before trust computation their point of view networks and I think there's a lot of really interesting research opportunities and this is one of the things that I'm excited about area of the community I think is really interesting in the current context to compute these trust computations without anybody knowing your opinion of them and so that's one one interesting area of research a second is we can imagine tweaking the trust computations so that different people's preferences actually make for different levels of trust so we actually go back to rather than a single trust vector multiple transfers and then finally we can look at dyadic graph structures for example if you have buyers and sellers or two different kind of people who interact with one another but not with with one another and those will lead to trust computations that are more around the see evaluative singular value decomposition rather than the eigenvector calculation so that's all thank you all very much questions people question yes yeah I mean so so absolutely I mean the way that I see PageRank is as an example of trust computations and so yes absolutely I think it'll be really valuable in the context of history I'm sorry yeah yeah so basically kind of if the message complexity is the is linear to the norm nodes of the network and the reason why is because each peer only has a certain number appears that they interact like say ten or something like that rather than every peer every peer interacting with every other peer in that case it would be nice right that's right exactly yeah and that testing yes it is a paltry make of you generally I'm somebody's a good Pilate algorri is it a flighting algorithm yeah where Palomas is our printable no no are you you can actually send the note as you can send the message only to the person who needs to get the message and you know who that is because the I mean you can run I can trust in the same way you front page rank with just a few notes I think the important thing is you want to make sure that there is the number of adversarial node so it's not the majority are you trust how do you trust the plus support line you're appearing near neighbors good news yeah that's great that's a new camera I didn't have time to talk about it today you you don't in this so what what we did we can is distributed hash tables to have a score manager and that store manager ISM and actually it's three score managers in a boat and that that make sure that those are trustworthy yeah yeah so basically and this is something that I also did gets you into this mock what we have to do is we have to add a regularization parameter so right now what could happen is there could be a click and that clicked you just basically end up never converging because you never get out of that click and so basically what we have is in page rank what they do is they say you know so the way that all offering that is there's a probabilistic interpretation of this the problem is the interpretation is that you're jumping around from node to node following people's local trust of one another and then after a long period of time your likelihood of being at a node that is the IB trust for that now in order to make that work you can't get stuck in it away so you have to have some probability of jumping to a random note and once you have that regularization parameter as a convergence is quite quite fine and it depends on what that probability is you can actually tweak the different you know next half dead rat from Altea and there's IP address on the board so I guess level another yeah so here is the there's a fee amount of the five gigahertz in the 2.4 and you're welcome to log in to enforce them so internet connected to it but you can go to the dashboard and see the the interface that are in users see with a [Music] okay hi I'm Debra I'm one of the cofounders of TN and I'm really excited to share a little bit about how it works today and then I'll be sharing more about our experiences working on the ground in my community setting up one of these decentralized internet service providers on Thursday Wow so Athena is so Athena is a software that runs on readily available consumer routers like the one we have here that allows people to set up centralized Internet service providers in their community because the routers hate each other for bandwidth so this lets homeowners and business owners host some of this long-range Wi-Fi hardware and earn money automatically for doing so so that instead of one is he running all of the infrastructure and owning all of the different pieces of the network and extracting all about profit many different people can get all these and and earn money and all these does this by using price aware version of the Bible in this protocol to select routes so what this means is that instead of only selecting routes by the best around trip time as we would see in the local match protocol this also takes the price into consideration as well so that it's been automatically on a second-by-second basis select the level in the best route but the least expensive sort of adding in that audition and it's also like a prepaid system right so essentially you load your router roll it up with aetherium or X I or recently we've added in a bridge so that it's easy to get a theorem that you can get from point A B's you can send out to your honor wallet and that it will bridging the music unit spot to a design you'd be a single currency to know how those price fluctuations that aetherium has for and users just trying to pay for Internet [Music] okay so what this what this does and we have lots of different people learning infrastructure is that a case on automatically competitive bandwidth marketplace so this is sort of the picture what we have with the the internet currently right you you get a an Internet service monthly subscription and they bring in a specific hardware and it's very siloed but with Althea what if you can automatically compete if your service provider on like a second my second basis if your router was too may not work for you so as we build these networks and different people on hardware the the links will will select the path Ralph so this is a this is a map of one of our this thing that works in Oregon this is actually working promise we're all community here and you can see this sort of distributed nature that we have here each of these different things are what we call them relay and so that's at home or a business person or you know a farm in this case that is hosting some hardware and getting painted on there but further further routers do that and so you can see we even have some really dynamic leaks that this is that way in the valley there and then routes to a person up on the hill and then from that person on the hill all the way out to the river and the reason why we did this instead of going straight is because there's actually geographical challenges to that this sort of interoperable kind of agile configuration lets us do as well and then we'll be adding in the worth more redundancy and links as the network sort of gross another thing about Alethea is that all of the all of these interconnections between and homeowners are subject using wire art and that is then decrypted at the exit what we call an exit node which is just like a VPN server Center and this also allows for us to verify that the that the cost that each link is reporting and the round-trip time is verified instead of them just sort of potentially misrepresenting a fact [Music] so here's one of our relays so this is actually that farm on the in the valley there and then that correct meaning green connection so what we can see is be havin a ton of coming into the that it's receiving signal from a and actually this is from like a wholesale fiber not call and then she's hitting an incoming connection and then the two smaller antennas are connecting up her nearby neighbors and those three third net cables are then routed to her consumer router in her home which does all of that filling automatic agency this is her either scan of money put in and out and suddenly they're in the Miriam and it also handles all their routing and then it provides for her home her home Wi-Fi so for her she she just gets it or not and actually gets extra money that she Nina that about a hundred dollars every couple of months from about six connects attic science that she should use as point basis and extinguish not to extract that to be on instead of the other way around just pretty ball and then she also we have this sort of bio oil drill so she'll go and talk to our neighbors and try to get more people connected we should have one relay that will take a bottle align those neighbors and see if we can get the kids and connected to the internet so you can see it as I sort of like incentive for for viral growth as well and there's always that incentive because it is sort of meter that we want to add in more more bandwidth better service for folks so we do realize though that even as permissionless and us automatic as we would like these two things to be there still is some friction with the hardware the radio hardware and outreach part of it so we have another incentive for what we call organizers which is typically like a business or in the case of that rural network of it's a cooperative or sometimes it's a nonprofit as well so it's kind of different legal structures but this sort of entity provides for the customer service that things go wrong helps solve people in many cases and kind of helps curate the network as it grows and they and they earn a monthly for way this description but for the in users and subscribers we simply load their router up with aetherium or X die and then it's rather automatically to the people who are forwarding them bound with or in a pro-rated way for the monthly subscription fee the organizers address [Music] and so this is sort of our next iteration of the project we have the pilot program and and now we're helping other folks get started with their own networks so these these are what we call Thea feds and essentially it's just sort of a set of marketing and pre-registration page so that organizers have time and kind of front-load the user acquisition so when they're ready to build out that sort of initial fiber backhaul and get that network anchor then they have all of their you know user acquisition and marketing or whatever already in place and so we have a few folks we have a couple of members starting in the United States and one in Nigeria here in the next few months and then about another ten networks that are working on starting up 20 I'm just to kind of recap and then I think there's some time for questions yeah well so kind of the benefits of course is that it's owned and run by the community that it's serving and I think we think the another benefit of a kind of any peer-to-peer system is that the money stays within the economic system that it that is serving so instead of the many telecoms where it's distracting walls was to have these developing countries right where we have foreign companies coming in and distracting all this that community's able to reap the economic benefits and again that's that also that built in competition where your router is is basically choosing the most competitive link it's a sustainable way because of the incentivize links we see the network factors well neighbors hanging neighbors for bandwidth which is like very poor p2p great and yes so so do you have any questions what's the cost I guess put that infrastructure in place is obviously an all Hardware and I guess to create a relay to install that with what's the cost associated with that he has a really good question and if we that that relay in the picture there has about four hundred dollars for each of the outgoing antennas and then about three hundred for the incumbents on it so that's actually a little bit more robust than some of the relays I would say you're probably gonna be up to 500 and then and a thousand dollars depending on how how many people it's gonna serve and then what the routers is it like a specific router that you have to do their communication routers like that you can get specific so we do have supported models in its open work based our open wrt and these are consumer routers that you can get from Amazon and we have about three or four supported models [Music] yes it's not a single radio of match that you saw everything appointed coming in the radio and without boys we have combined reciting drives percent packet loss her performance offered with links so the one that was like two or three team you know we're still seeing 50 megabits per second cuz it does the like infrastructure for breakdown or like people throw a parrot like how does that work yeah so this is why we have you know the concept of the organizer right now they're kind of prosecutor a but unfortunately the radio hardware is it completely in a permission the state where people can just leave and pop them off and they just work and everything doesn't even configure we're excited about what some some of the things that are coming with the informing radio so that there's less of the aiming and kind of maintenance but yeah all right now the organizers do that well when they get painting with um comfort than the issues will the service providers I expect that that might be perfect mission to resell your fans which to either exam well we do use a whole cell connection so we are just operating as an Internet service provider at some point in the network order gateway however what's been really interesting is that some people some of the Internet service providers have actually at first been really business in it and then thought this might be actually a pretty cool idea so I think we'll see how these networks growing to stop the space they're gonna come up for this out the other gateways like what are the reasons for that you think this is like why do you think they would be killed bad um well I think there's always legacy resistance to change but I think I think as time goes on that they'll see the market potential that's less lost give a plaintiff and I ask advise you to agree we would hope that they would just run the okay how do you sustain your work because you said an adult extract value from this community that also start the caduceus than your yeah so we're working we're building a cosmos for change so that people will pay with the period and then it will fetch to our token so no seven areas minor five percent of the transaction easier to get token holders actually well so for end-users they'll still have the same process of buying ethan routers but then our companies actually yeah this dog is savvy team provides because i was invited yesterday fat i i'm given a longer stop at working swans recognition a soprano so little bit about me Tirion Athene I was doing research on the racial codes for cloud storage and we have to be a systems I did my PhD in Switzerland I also I did disorder center at UC Santa Cruz since during my PhD I produced the alfonseca macros and the since last year I started collaborating with the people from the fever swamp to see if we can implement this coach in Swan to make the Swan more robust and since this year I joined the BB change Norwegian project at University of Festa Wonga which aims to develop disease assist them to verify documents and the first use case is like diploma so so so we can have the digital diploma verified that this is a real drama and all the process the process of the student I'm for supervising the work of these two students were seeing who is working on the actuary and the implementation of autonomy calcium swab so he is here a second comes he will be talking with me is optimal another your who is focused on the part of their contract and verification so this is very very short idea of what I am working on that you can get more info so basically the main method of them simple way of getting redundancies just copied your MacDonald and this is this is a very popular and easy way of just keep replicating data so this is basically the man from behind this is the more the better which do you have more more replicas you can tolerate more crazy machines if the network is not good the nodes disappear so you can still save your data recover your file but this has disadvantage which is the space overhead then we have another approach to achieve redundancy which is with your racial coding spacing the MDS codes or read sort of accounts which have the mantra behind that is the less the better because you want to save a space and buy it here a circle in your data but basically you you do is like served in your file and then creating some extra blocks but now you use less space of a fat but the problem with is with this method is that man Dana's is very hard and many peer-to-peer systems failed because of this in the sense of faith pay to protect the data because you can you can create this redundancy at the very beginning but then it's very common thing to keep well when they noticed are disappearing when when there is a failure Devon Tanous takes a lot of anguish and you and also it requires that many other nodes are available when you need to do the repair so I'm working that's what I developed in my PhD on Allen recommend course or in tournament cause the idea of entanglement course is that each time you add a file to the system you again you charge the data so you have like blocks of of these rights and you exert these blocks with other blocks that are in the system creating a bottom creating like multiple chains of real-time data and this chase are intertwined together creating a larger structure of a conical lattice where this lot is what creates many different ways many different paths where you can recover this data and by doing this in a very efficient way that and so every time you have a failure a single failure you only need two blocks to recover the data and then you can do the repairs in parallel because this lattice you can you can do some repairs and repair some blocks that are missing here or some some blocks that are missing here always by reading two blocks and only doing X or that's also mentioned these codes are normally the aromatic suffix girls are are based on karma fields which is a little complicated so so to make it more efficient you have to have a team say show some some of the incisions are so that their patterns become the center so this this was a very simple to implement and facilitate their repairs and the space requirements and also can generate a lot of other failures and we show that instant assistance is superior than average Salomon so you will get worried for this afternoon if you want you have to swamp recap such as will mr. about former cadet autonomous and though you can also read the last paper to publish the PSN and then I want to mention something I'm working on the and a proposal a workshop proposal to submit to a very top confer upon systems this conference Swiss about systems but in Sheridan I'd noticed that there's a lot of focus on centralized systems but it's a system conference so I would like to discuss more all these top personalized in also in the academic environment why because you know in the to sauce it could be between $2,000 and this is all sonic man there was a lot of research and there was very very serious research of European systems but on how this disappear immediately change the name that was not cool to talk about peer to peer because it was a city with with illegal networks or elegant allowed to the lasso and and then I mean there's a steel research of course in Europe er in the academia but and it also takes different names but not as much as I would like to at listen and also we have the problem probably researchers that we are interested in to in search of these different systems it's very hard to I had to attend this kind of conference because they are developing conference and when you have funding it's only for academic conference because we need to publish so it's my aim is like to to mix but both communities to buy in this workshop where the purpose of this workshop is to have some keynote speakers from fraud academic units a speaker from the decentralized community also have son for I know that this and that people from the community they don't polish the dog have the time to publish an academic paper but but it will be very nice to have demos to have posters from the community so to show the students what interesting things or what what what problems can can be good for research and also to contact each other to continue to to have the contact of the students with with the savannahs community and and also to have a plan so I just started my associate professor position so I still don't have them much fun but so I need much more help than that what I can provide but if somebody is interested or or or no have contacts with other people that could be interested in participated in this we need to submit a proposal at the beginning of you November I think and the workshop will be in April increased in the doctor conferences so that's all which course is like patents and and also in some cases emits I don't have comparison in the papers with food workers we have comparisons with others like with low locality codes with Prince aluminium also now working on the on a system that will be useful for anyone to compare my goals with others so you can implemented workouts there but I don't have measurements of two rotors thank you so much everyone my name is Fatma slipper I'm the CTO co-founder of Killian those of you who don't know what we do or we are we're a blockchain needed a software consultancy in foolish these a bunch of notable folks in the space business pop see why we did mathematics do you I also have a grant and we rented an audited dr.

penny plasma MVP bunch of other stuff you can talk to me about that later because today's talk isn't really about what keoki knows about something that's important to me and that's peer-to-peer decentralized social networking so I'm gonna kick things off with a little bit of a story so there were two most memorable moments growing up that I had when I interacted with the technology and the first was when my level nine dwarf hunter first walked into Ironforge and saw the dude with the glowy sword the second was when I signed up for Facebook the reason these two things were so memorable and important to me was that it was the first time in my life that had actually seen technology be used to connect people in the case of World of Warcraft it was here's these people that are immersed within and changing this incredible fantasy world and the second one on Facebook was here's all these people you know maybe folks from the weird private school I went to when I was younger that I hadn't talked to in years maybe folks from the you know there was a whole film group just called the slipper clan so it was other people with my same weird last name and we could all get together and like talk about what it's like to have a weird last name one of them was a member of the Australian Parliament so I'm gonna bring this up well SSB is one just a protocol it's a set of applications that let you build decentralized social networking so I've got a bunch of slides in here on the talk about like you know how it works and things like that and we'll get there but like the first thing I want to do is just like show you who has this people so let me X out of this and come in here and drag it over here and that's go like this so this is patchwork patchwork is a client application that talks to the wider SSB number and the resolution was kind of screwed up so it can't actually see the channels on this when we do that councillor shine on the right all right well I'll wing it guys homing so anyway what you're seeing is Patrick mr. Klein that connects to the larger assets we network so I'm most important to know is that you know if I Mouse appear you'll see that my Wi-Fi is off so everything that I'm currently interacting with on the screen is local on my computer that I've synced from a network of decentralized peers all around the world if I were to go here and let's take a look at what people are talking about the closure closure channel well I'm just gonna scroll here you can see that we have a sort of very typical social feed here but the difference is no one is actually posting the speed it's all local and I'm replicating it from other people out in the ether as it were and no one can take these away from me there's no product guy whose responsibility it is to drive engagement eyeballs to this that's going to say I'm not allowed to talk about X Y & Z the only person that has any control over what content goes in there and the people who are posting it and me the guy who's viewing it and so it changes the web from this paradigm where we go and post stuff on to some platform somewhere and some other product guy decides where that's going to be rendered and somehow the rhythmically design queue of posts instead I'm creating a viewer onto the web now lets me see the web the way I care about and the things that I care about forever and no one can take so you know if you guys want to download it yourself and take a look at it it's relatively straightforward but you know you've probably seen this type of UI before but that's why I think this is important and why I care about it and what the end result is of the technical protocol that I'm going to be giving you an overview of next so anything you guys want me to cover against quite launch into how it all works all right cool let's go back to mr. PowerPoint here like this who said that I all their alternate user interfaces oh yeah there's lots okay so there is patch Bay there's a whole bunch of CLI based ones there's always two us are you that's patchwork I like patchwork because it's a little bit more polish yeah yeah I write in the nice patchwork but it's dis one the one you are personnel using the day-to-day okay yeah I don't wanna tell you my account Issa is one of the fun parts is that it's like actually pseudo anonymous and so uh yeah sorry about that but okay so this is how secure scuttlebutt defines itself as a protocol so it's a family of protocols from the security naration and replication of append-only data means you're gonna see a lot of local office and parallels between this and you know what was discussed during death because it's a very similar idea you take an append-only feet of the content you replicated to a bunch of other people and then you allow client applications databases etc to be built on top of it and then allow people to index and display the data in a meaningful way and when I say it's a family protocols also very similar to doubt they're designed to be motto am I already known at three minutes Jesus so yeah they say it's a family of protocols that they're designed to be modular it makes it a match so they already know here we go the SSB family consists basically of these layers at the bottom layer you have a set of keys so a key pair that's used for user identity there's no canonical naming on SSB so you name yourself and other people decide what name they are going to use to describe you that have your feeds drug the append only feeds as we mentioned earlier a secret handshake which is how peers talk to one another we identify one another and it grabs a traffic between one another probably I'll get into in a second and applications which is something that you are in the sauce so Koch work is an application this example what a key looks like so in this case soap dog the guy who was talking about the go back up tool this is his key anytime you see a feed that has messages signed by this key you know it belongs to sub dog but sub dog doesn't actually own that I have decided to accept the fact that he's called himself back on SSP this is kind of what feed looks like just like a blockchain this content hash linked so this previous here is the hash of this definer message before this is how you're providing authenticity to the messages as they get replicated people would need my key and they would need all of my messages going all the way back to sort of Genesis message if they wanted to Forge anything that I'm trying to say unless is B so there's also probably no networking that's included in the secret handshake now I'm not going to go into the details of that because it's not enough time but if you imagine a TLS handshake between two peers where the identity is the key that I both recently discussed that's the way the secret handshake works now we get to some of the interesting stuff so peer discovery you can pull stuff in over a local network so that's Bluetooth Wi-Fi whatever else the guy who's grounded SSP actually lives on a boat and when he gets back onto Shore he connects to a local Wi-Fi instincts all of this stuff you can also use an invite coded message and pubs are important because following has meaning this is there's one thing that I want you to take away from this sort of talk about SSB is that SSP has baked the concept to play the social interactions between people into the protocol and the way it does that is with the pub and you when you sign up for bug you now have a pub in the hub follows you the pub will be replicating your feed messages to existing users on the network when you saw the demo apps coming from me connecting to a pub and the pub giving me everybody else's messages and me cashing them locally so you end up the system that looks a little bit like this I took this from the SSP website basically you have a user and you are seeing messages that are n hops away from you this is kind of interesting because it means that how many of you have gone on Twitter and like to spin bombarded with trolls or people you don't like and sort of things like that sort of see some nods and things along those lines on SSP when you block someone as a protocol level meaning your client will refuse to replicate any messages that come from the blocked identity oh I'm out of time can I go over for 30 seconds yeah you should you should and thank you so much for doing that person from SSB did was not able to come to Def Con last seconds and in math math it's like nice enough to fill in and the brave enough to fill in like on a very short notice and some are SSB so yeah yeah yeah Simon by the way seven seconds I'm a fan of SSB I'm not a core developer but I just think it's super dope and so I was really happy to actually be able to share with everybody anyway so flocking has meaning following this meeting you see the things that your social graph wants you to see because they're likely to be like you and follow people that are more likely to be interesting to you and blocking the same way so I think that's the speed is the potential to change the way we think about social networking as well as just social content on the web in general I think it's definitely worthy of your time to check it out because if there's one thing that's going to help and the stigma of decentralized social networking doesn't work it's actually having content on there and your friends on there because that's the number one objection that people have to any piece exactly and you should talk to Frankie about SSB too she actually works on that's all the time I got identify yourself thanks everyone oh this is gonna be really exciting because I'm we're all used to demos failing well haven't seen the kind of catastrophic failure scaling all morning so it'd be good to get this out of the way now I'm gonna do a mainstage talk on and I'll go quickly through a couple of things so I'm - rich Gordon I work on a project called wireline we're building a new kind of peer-to-peer network and they're gonna demonstrate today quickly how we can build applications that might enable us to stop using things like Google Docs and I was involved in group building technologies that Google Docs Google I'm going to go through all these slides just quickly building that's it's hard building apps of internet scale is really hard and building decentralized apps it's much much more difficult so I thought just as a start we do a thought experiment about the kinds of technologies we're going to need in order to rebuild with internet scale applications so we need new kinds of networks data is probably the most important and difficult thing to solve where do we run long-running favor intense computations and how do we deal with identity so quickly going through these if we were running Google Docs here today all of our keystrokes would be going through a server in Mountain View California or perhaps Tokyo they're not and if they're going to be going through infrastructure across the internet and this makes absolutely no sense so we're starting to have different kinds of technologies for peer-to-peer we've got the BJP with WebRTC there's still a big chat challenge getting this towards seamlessly in the browser so we need at least additional infrastructure like stun and turns did not traversal and to do with firewalls but we also need the HDS and signal service and the different peers can find each other we need to do this in a way that doesn't reduce any centralized dependencies if we were building on the cloud today we've got dozens of different databases managers MongoDB DynamoDB Google cloud of big table s3 but we have to give up all of that if we're going to build this on a peer-to-peer network and we've got now you know starting ten years ago maybe before that victim demonstrated stable file exchange will that require decentralized index and now we're moving rapidly towards okay the peer protocols like that and secure scuttlebutt to give us access to real-time date so they can be securely exchanged but we still have a fundamental problem which is if we've got concurrency we've got multiple people editing the same document at the same time how do we merge the data so that we've got a consistent view of the way Google does this pioneer technology called operational transformations thank you I'll try okay apologies sir one at the front good Google pioneer the technology little operational transformations ot and this is actually from a project some might remember called Google Wave ot is an eventual consistency protocol that requires a centralized server where essentially instead of obviously we don't send the whole document to the server becoming edison we send a mutation and what OT does is a lecture that the server is able to rewrite that mutation so it can be applied commutatively in any order and that means we can have lots of different peers somewhere across the network some online some offline keeping a local log of all of the mutations and then at some point they can push those to the server and the server will reconcile them kind of like a dynamic get but since I said so how do we do that in peer-to-peer network well there are new technologies called conflict free replicated data types and trouble saying that sometimes CID T's which enable these kinds of algorithms to work in peers but they have some challenges there's scalability problems and security issues and so there's still a lot to solve there you said to us compute everything's happening in the cloud today as a obviously for problems we have where we need global consensus we're gonna use blob trains but that's not a proper appropriate for all computers cases I mean even given layer 2 solutions that we're developing right so as a general principle of decentralized compute we want to move as much as possible to the edges that means of doing more work in the browser that they're going to be many use cases where we want long running processes doing compute intense things for example if you wanted something to index all in your document so you search them you need some other kind of system and we're taking a cue from Google app from Amazon lounder and Google Cloud functions they're very scalable technologies to probably make those run more peer-to-peer network and in a secure way and then finally identity there are many cases we want to be an anonymous but we're social beings so how do we represent identity for us and for the systems that we interact with but we have single sign-on today but that's a terrible thing and that's the central authority track everything that we do we've got new w3c standards who are self sovereign identity where we can create up multiple identities that we control the private key pairs and we have verifiable credentials where we can exchange signed proofs to give each other access to different resources so um what we are focusing on is building this missing peer-to-peer network that joins together the existing cloud technologies with the emergent and what train these centralized technologies because we think there's gonna be on that the web will evolve into a hybrid of different decentralized systems so can I do a time check yeah six minutes okay great so in eight minutes I'm going to rebuild googled well we'll see how far we get so I'm showing these slides on an app that we've built on our platform called Launchpad and here it is here Lord Launchpad is a is both what would be called today an app store although it's a actually permissionless blockchain for any developer can freely publish their applications onto this block chain and so Launchpad shows those applications and then I can sort of create one and add it to what we call a party so parties are like lightweight replicated database that is constructed from that fields and the user CID tees to create a consistent global view about all of the participants of a particular set of data so we've got very simple spreadsheet starting to work with third-party developers to build out some of these applications we've got a very simple calendar and that's a document now one of the things that fell are still about resilience earlier so we've got some Wi-Fi resilient issues let's see how far I can get here so this is me here typing in this text editor so far and then over here in another browser incidence which will pretentious somebody else not me so I call that non trench up here I'm gonna send an invite which you would send out of band or once you get to know somebody you can now use an address book and hopefully these two clients will discover each other I'll move this one over here and close that off yeah and so these are two peers on the same network did it work yeah okay and we're going to the same document well okay you can see the if I go over here quickly you should be able to see me typing over that so what's happening here two clients swarming with the same discovery key on the same local network this will happen across the internet if we had our Wi-Fi working and then the the mutations are getting exchanged in real time so we'll go from this into I'm gonna try to build an app now and remain five minutes see a father yet so it's check we've got an SDK called the ys tli that the first thing I'm going to do is build a new pad we call these US components pads so I'll use an incessant and a fork an existing github repo that has my starting point me so why I've had to create and then I call the name of the template and now if I go into demo pad we're gonna code here I'll see hopefully that it's just waking up okay here we go [Music] all right just try thank you this they say backs are more reliable I don't know we need we need a keyboard I think you have that much stuff let's know that we commute he doesn't have to stop there so I don't know okay I'm worried tight without let's do it life I'm gonna do I'm referring to that book yes Jay so this is why we're gonna stop building apps and electron and start using the press okay okay can I have another yes I was gonna say when I was a kid Microsoft was treated as the enemy in the real sort of worst contender oh that kind of repeat rehabilitate themselves but going back on that that guess code this is pretty good okay so look bad and what we have here is this template pad and what I'm going to do is go into the terminal here and say first thing root is register the time of the block trace and what part of the register register okay and then yarn pills with the Kratz permission okay so what should happen now is that we're running the launchpad in dev mode and so this this template is a simple tasks app if I create a new instance of that inside this party which is this spiritual database I can now I usually type things a little bit more thoughtfulness but given the time I can now see that we've got this very simple task list if I go over to the other browser instance over here suppose that should open up again here there's my task lesson I check some things up so automatically out of the box it's collaborative so let's let's change that alert I'm going to now close that happened since we're in Japan I'm gonna build shogi yeah so show his Japanese chess I did this really quick so edit this task list and I've got a couple of these things couple of the import statements already ready we're gonna change the entire render function if those things only act we're going to render this Shoei board using your current game state and let's just try it and then inside deaths I'm gonna change the icon somebody give me a car black could be a plan amendment shogi say it and let's see so it should get the changes should get picked up automatically and we're going to show you what in green I know if we create a new instance of that we've got somewhat of a show you what this is here right I can I can move pieces here but obviously probably to refresh the screen we're gonna lose all of the stain so the next thing I'm going to do is go in here and add a add a little parameter here for the drop and then add a front end to that effect handle drop so I'm going to get the events that were published by that we are control validate that's a reasonable move and then I'm going to append that to a law we're using these higher-order components which are right it's part of a react idiom it's going to inject the database from this party into the control give me this view object that I can then append changes to and then in order to process inbound changes from other players I'm going to get the log view go through each of the messages which are just mutations process them against the local state machine which is this show you third fun show the object and then return that as a state object so let's see if that picks up here now if I make it should stick so why don't you just finish off this part is we can go into the chat application which is another app here great a chat room and I can say hey not rich surely if I go over here and then press ENTER if they go over here into the chat app which is just shown up magically I can see that I've got that link and if I can click on here they go straight into the show me game and so now not Ritchie is playing the first move and then like here rich here is playing the second move and then you can see that everything's working right so automatic no 10-millimeter woman okay woman right let's build an AI to play all right boss creates see the demo going to be I equal Fraser it's because you trying to open yeah I'm gonna kill it okay alright so let's see yeah this will be can I get from the crowd two minutes we take your folks now these the next speaker alright so next speaker do they feel anything at this point we're super curious okay here we go all right so let's see fine yeah that's great so we've got the body on I'm gonna chain citizen that's a configuration file I'm gonna change that to show you one I'm gonna change the name of the view which is the virtual I'm gonna send you to the name of the box to show deport in deaths I'm going to change the view to the same database name that we chose yeah and then inside applause I'm going to same standard library beast earlier and walks processed the log so that's the inbound law we get this is a lambda function gets called whenever this irritation on the feed and I'll skip ahead now and we'll say boss a pen move so well we'll get the set of valid moves shown the standard library and then we're gonna do this really smart thing where we're just gonna pick and move a random and append it to the feet so in the terminal we're gonna say quiet walk register and the wire that's right so I'm running the bot now in its death mode and it's debug mode on this local machine so bots are humans too and so I should be able to if I click on here I'm going to I think this is where we actually do give up what would happen is so what what what what would happen here is that publishing the metadata of the plot to the local blockchain I can then discover it by clearing the chain from the browser send the signal of discovery key to the bots in swarms that joins the party and then it gets all of the mutations and we would be able to play in this case unless okay since we since you are so generous and we got this small let's just try the one at fault register long dong Qiang okay so we've seen that's more promising while okay the block times about five seconds so now if we clear these she's setting listen tell me the settings ladies me yeah wait wait you said that beautiful one this is exciting oh right okay surprise them okay so let's create a new game and we're going to members well we can now see shogi bar on the registry show you Bob I was joining the party if I click on the debug panel here and see the different members of the swarm eventually now if I make a move for us or what happened oh it's unhappy okay so that's my name I'm French so the exit is warranted do anything about it 24 years sorry if you don't want if I say I'll be good understand myself but you can ask me questions let's go go so here's what I'm going to talk about the other side of the equation so everybody hear stuff give up new innovative ways of doing but chain also few puzzles come together I need to tell you about the risky side of what you guys do which is natural desti you're lucky right so the issue here is that we are not creating too much pains for people who want to make sure and things that you work as they should right so that's the old wall you know reading reading reading complaints much that simple communities cluster you know you know convenience like lemons instance spray flexible be asynchronous mister time lots of web series this week software is not pretty well test it's pretty hard to test at scale any which test things me into just this like just go for test net try for 50 weeks see if things stick spine up the handle and it's a new stack a new stack we will need to start from scratch again creating new testing tools a specific level of change if you recommends networks this one here we have almost critical asynchronous not always final even fidelity itself is fun to test there's reorder chains that's something you need to be prepared for any quiz so you don't have like the instance information something you know hours to actually test available network we can now see that between proof-of-work for state trade-offs performance three incision since approaches so we see all sorts of different ways this playing out of investing and it's just not the blood seems like everything built on top of it starts to it's great to civilization you're at Norco we have a breach all of a sudden have things you could take for granted start to misbehave Unwin and that's you know thick victories so we'll talk about that if you want to learn more real we also have we also have new metrics to capture before you just be like what's a delay how much do you replace to buy now we have all sorts of fees we want to capture which are more critical to the purpose of the network in top of that approaching specific so if we failure out there some of what we have to capture for every love for every child actually the size to get the center of zooming we have to get all the time sensors to compute the median time and average time we have all the information on the blog's also like traditional one also emphasized collection exercise most efficient time it's gonna be tricky especially when we start doing double engine testing and the head of an addition doesn't kind of fighting with your attention terrific troops that you're getting the right headers but not least this is something that because recently the m48 to ensure that we get like a good operation instead of returning a favor so for that we just choose an open source project which is also complementing back tools all sorts of data extraction and analysis so Genesis allows you to run anybody you know agnostic environment where you can actually control every aspect of your blood chain from the tiny wants to have between the four nodes that meet you can have some class you can do all sorts of tests which are famous torturing any blocks out there this is a great doctor chamber it's out there in the web you feel like and on top of that so we plug them into our platform which allows authority comments other source of data tends towards his feet to develop chamber you can read all the logs of data make sense of it if they're that between different runs we'll have that between blockchains pickles solid pencil so tuck those two sweets we start very small we just want to make sure you know works just on obstruct justice with work on your machine so nice and even then start doing it for testing so we know the parts will be juice event we we make you know those messages CT to be personality with little bit of that testing how much tend to stick to stick it out we do some opportunity testing so it's getting to work which that version party I was indicating exactly the same way we should start testing all the UPS codes did we see that the extra behavior 77 we don't have a refill that's negative to the positive stuff things you're some secret testing and things someone mentioned earlier keeps attacked that's bread and butter for us we like to do that double spins between two hijacks and those I mean a man is very bitter things little testing with just one milk skin to get completely lost like thousand three collections a second for five minutes big thing even night recovered from that when we do this bridge testing everybody just dispatched for five minutes and look at the result of the the damage upon such testing is like we do at all we're doing this receive to talk about that we'll do testing where we'll take the note state will get your point to transition to the next block and we'll make sure that's able to get to the exact same result is expected as long as we can determine all the stuff that's outside of like the term test the market leaders all the space we can actually do this so this kid listen gates right so you're a talker you want to use some level of testing to make sure you're both entering work we can give you a month and distant you're going to take your fancy words now you'd like to make sure that actually works is another client that you don't control that you have Miller hear about that's a big step right so it started working this point intensities you don't really know how they work Laxmi sure can't write lots the right time now you can also try this by checking the new version of the time so you can hear is it outside you can start doing all sorts of things it becomes interviewer privates Roxton and applications on that you reserve to your stuff so you think you see the speaker that works for me well iam before on as well holder like you know permission networks much more fun yeah we also have these stress testing like once a week in on Saturday to just deploy your feet and then do your smashing it work although we didn't even signal to sustained effort and you can compare results with time to make sure that you're getting better performance as you please so just couple samples for what we do so we do let them see testing very easily that's monitoring these tests we take your network you start pushing the notes apart and seeing this to communicate we can even like do some geographical replication so if you had a note in New York in a building talk to you know here's an identity we get into those two locations immediately make sure it is syllable to communicate properly under something but fine we have also started my favorite this one we take a nice definite we're ready to to wait five minutes bring them together so what are we going to see some networks some protocol will be able to know how to consensus for example for the five minutes because they come get to consistently one but if you have to work for example you start having to change the content of each other now if you bring them back together the beginning you have to give them some time and eventually into some consensus reach back some margins don't actually support that paper and this is what you've done recently to interrupt so big seven different clients then a farmer of your soup of a destined end and make sure is able to come to consensus right so he's innocent of figuration files knowing about each other this is had fun a bit of a challenge of the column where they need to credit date sometime get the right blogs or post boxes right it's a dining with those things and this is something we need for for find where they wanted to distant name key and then we test it in the safe manner so that you could get four digit results that there if he wasn't week to break the maintenance so we to domain that we change the plan to be so slightly to allow larger collections and that gave us the actual results in terms of analytics and we could see the uncle rate was breathing good of that material to system he pushes what sort of scale you guys that tests soul artists are reproducible that's enough yeah we run this interview and then close the ramen so we can make sure that you get the same results at a time we can do up to 100 notes right now we're trying to go much bigger but frankly what I've seen in that 30 notes I already know that you can really mess with my stuff so I was wondering can you do some tests on we cannot look like yes so we can really it's really very much for picketing and running that completes doesn't it so you thinking very much employs contracts on it we have some test that involves my contracts we collectively try to reduce them even use the infrastructure elements that are injured article which we will be able to run to test destitute certain breach and then we can actually do also saw flight but it's just easier to see full coverage intercept we're Network dinners so we always end up like me back down the powerful Network first and we're doing yes do you have like I don't know we wait for about that success I just think like I go to your blog and then I like I have very some of my things with like started coming around that's different becoming so yeah I would like to be able to have public datasets so not just results right the lecture giving you the whole data set of everything going on about investment and make that public visibility to our users to the public platforms people down low like 1.5 gigabytes of blogs and number crunching even like you don't work press this to do blood testing in tomorrow so I do know what the doctor ordered back to dysfunction days where we spent quite a bit of time actually we do a lot of discoveries of clients because it's a very fancy so important and bunches of these different portals different approaches permission regulation all those fees so we end up like Lowell is having a collaborative discussion about the right metrics know and being able to test for so you know initially when we study about people just yes just get like a bras for boot of the platform I think he's changing a lot more now I hope you can stay for the discussion part because actually plot yeah we're going to you guys I'm sorry sir so we really decided to see those summaries on I hate my name is Mike gelser I'm from the Olympia peat project I work at protocol labs were one of the sponsors of looking to be here's my partner in crime rollerblading say on the floor there so then so what we have planned for you here at Def Con this week is a kind of escalating series of talks this 10-minute intro is going to talk about what Olympia Diaz what's coming next in the project then in what is the two hours from now I think and Rudy roll is going to give his talk on 101 of to talk some blue P defeat that will take it up a level so some of you I know are experts on Limpy to be some of you have never heard of it before that's fine we're gonna take it up a level in two hours and Rudy Royals going to give his introductory talk on Limpy defeat and then tomorrow we're going to go roll is going to give a second talk which could dive deep into one aspect of loopy to be that is of interest I think to the etherium to community especially which is gossip sub and so digging digging deep into that so that's kind of our plan for Def Con if you have no idea what lifting if he is let me start out looking feet we're out of the ipfs project ipfs is a project that actually I think there are out of quarter you're going to hear about this in a few minutes so I won't give it away but it's a project to build a globally or are even cosmically distributed file system based on content addressable content content accessibility and so let p2p was the networking layer and ipfs and what the builders who I think has realized as they were building idea test was that hey we could factor out this networking layer it could be useful for other D apps or anybody who want to build a peer-to-peer system and so today Lib p2p is its own standalone library its own standalone project it has the implementations in multiple languages as we'll see in a minute is the foundational layer for many different projects before ipfs obviously those are came from for aetherium 2.0 that's what you're going to hear a lot about this week here at Def Con and then the sort of main source of interest at this at this event and then for many more projects I don't have time to go through them all but this is a rough list of many of the projects that are using lit p2p today okay let me just explain really briefly what lit Fifi is like I said it's a foundational layer to build peer-to-peer networks not build peer-to-peer applications where you need an underlying pure peer networking layer and so we have the different abstractions that provide you with different functionality that you will need to build one of these types of applications so an example would be transport a transport could be something like a TC TC P it could be something like UDP it could be something like radio ham radio it could be a lot of different things it moves packets from one place to another we then have these notions of crypto channels and stream multiplexers I won't go into the details of these right now but the basic idea is that inlet p2p you establish some type of a connection over a transport and then once you have that connection established of course you want encryption you want authentication and you want you've got the ability to send multiple streams data streams over that connection at the same time and that's what extreme multiplexing is about of course like any peer-to-peer system like some of the ones you've heard about earlier today or some of the older ones like BitTorrent and so forth you need a way to discover other beers so we provide pure discovery we provide several mechanisms to do that you can discover peers on your local network through mdns you can also discover peers on a global network through a distributed hash table which is a I think we've actually heard about these a little bit this morning so but it is a the most common data structure for finding content on a globally distributed peer-to-peer network that doesn't have any centralized lookup point the way that the central centralized applications do nationalism is another big issue many users are behind Nats they either don't know how to or don't want to go to the trouble of opening ports on their routers so we have logic in Limpy P to help with traversing Nats and getting the actual machine they have to appear to appear as if it has it's on the public internet publish/subscribe is a whole topic in and of itself but basically once you have a set of peers connected to each other you want to be able to you want to enable them to communicate with each other in one of the ways one of the most common communications paradigms in distributed systems is publish/subscribe so we have several mechanisms for a published publish/subscribe and one of those gossips of is it's the most oh the most efficient one we have right now that's the one that Raul is going to give a deep dive on in tomorrow so you should come to that and lastly were in the final stages of security auditing Lupita Pete so it should be I believe it'll be one of the few peer-to-peer foundational layers the lilly open source that has a public security audit and you can see where our vulnerabilities are what we've done about them as some i think is f1 and maybe some others have mentioned there are some generic attacks against peer-to-peer networks that are very very hard to mitigate against in the general case like eclipse attacks civil attacks things like this so there's a kind of a mix of attacks that are very hard in the general case to mitigate against and then all the way to specific codes bugs in our code that we have patched or will patch the tongue we publish the report that's coming later this year oh yeah so so in a sense we like to describe the PDB as a modular extensible peer-to-peer networking stack I mentioned earlier that we have implementations of limb p2p in multiple different programming languages so this came about largely because of the etherium 2.

0 project and one slip p2p got locked into the specification for how aetherium 2.0 nodes are going to communicate then we had a situation where a theory of 2.0 wanted nodes in many different programming languages and that meant that we needed to have limb p2p implementations in many different programming languages and so some of the people who some of the authors of those are in this room some of those are some of the more advanced folks I think probably all of them are at this conference somewhere so you can certainly feel have opportunities to run into them but in the beginning was jazz flip p2p and go lucky to be those are the first two implementations rust that came along that was developed by the folks of Paradis and then since then we've had four new language implementations of Lippe 2p and we're really excited to have so many different interoperable programming languages all implementing the same live p2p specifications I have a little bit here it's give them through this but just the kind of the most mature implementations right now rjs go and rust they have the most features we often treat goes sort of the reference implementation we're trying to move away from that and toward a model where limb p2p is spec first so lip AP itself would really be a specimen not a not a library but a specification for creating a library in a language or on any device or platform that you want to implement peer-to-peer functionality and then these are some of the new relying as I showed this on the previous slide but they're they're coming along each one is releasing on its own early cycle and and they're all making good progress now I just want to talk a little bit about what's coming next in Limpy tepee we have a lot of things that we're working on in the project so I mentioned the new languages are at varying levels of maturity we had to mean a minimum or sorry the implementers of each language had to meet a minimum set of requirements okay one minute minimum set of requirements for each 2.0 now we're trying to build up on top of that and enhance that testing and a visualization of distributed systems is a really really hard problem it doesn't matter whether they're create a peer or or or or centralized it's still a hard problem but it's particularly hard at your computer so both it's 2 & 4 there what we're focusing on in terms of trying to make the system less okay spec first I mentioned we also want to improve documentation and we have a variety of features that we want to add I will not go into but I think you'll hear more about that from role so here are the two talks I mentioned this is the my plug for roles to talks one of them is coming up in two hours you should absolutely go to it it's indeed it's and then the second one the more advanced one that focuses on gossip sub will be tomorrow at 1455 in v10 check the schedule this is from the official DEFCON scheduled but if design changes will be on there Wow all right as their question time worship is just delicious okay ghetto so do you think the other pyramid not even sees that you want to be oh yes we're this is a subject we're very interested in an anonymous DST privacy issues yes sir I wouldn't say that we've solved it yet it's uh it's a subject of interest so there's a proposal and it's basically using my morality to climb on the TCP layer yes I have seen this of a new space well yeah yeah there are a lot of different approaches to to achieve this end yeah we would like this are either broken or easily as I go by even moderate in size nation states so not actually there so Billy's building must work those systems are deep those right Mountain foreign fighters helping but for now like because the thing is you gotta you gotta figure out which Frenchman can use let's manage making a privacy or a that but you in a fuse or be aware look like probably thank you yeah I think that's a really good at demo okay thank you for your time everybody I'm Molly I'm the best project week and I've been talking to a little bit about ecosystem growth and an option my slides are definitely longer than Mike's so I'm gonna go real fast through some of these examples to try and get this up but if I don't come back to me afterward for where we're heading next that's the part that moment that missing if I haven't gonna cut off so feel free to chat to me quickly what is ipfs it's aiming to make the web work here to here so instead of HTTP you are all talking to a central server or connecting between all of the different notes in the network directly and and it's going from something like location with addressing where you're going to the example.com to contents addressing where you're finding content like what it is not whose hooks together any of us aims to solve a whole ton of different problems around public with the inefficiencies of all trying to download the same video at the same room at once to break links to about security models to censorship resistance and we're we're in different stages of completion on various ones of these I think censorship is actually an area where I think this has been used pretty effectively in things like the the work in Catalunya where they were using IDs to put up the various polling location where people go in and vote about whether or not they wanted to be a dependent but hitting a number of these areas we still have a lot of work to get to our end goal as mine was mentioning IP MSE is just kind of one component we've done a lot of work to factor out the different pieces of IQ recipe you could use them independently they've grown and taken on their own open-source projects so the PDP is one of those but there's also a multi-format and i Kildee if you want to talk about IW go talk to Derek is or your master about everything linked to data related about schemas but we kind of sit on top of this amazing staff and get to utilize all of the wonderful components that folks like let me be and I feel like everything on Ibanez is a super huge community there's four thousand contributors around the world we actually just had our first meetup in Japan yesterday which was super awesome great to eat this slice of the community have been a lot of other people from Hong Kong and trying to also be a participate in that okay I don't learn for today gonna talk a little bit about the growth we've been seeing and some cool new projects that are being built behind EFS and then hopefully phone gets time to it I can talk a little bit about the future and the cool stuff that we we've done this past year and what we see going forward from a growth perspective we saw huge growth this year we saw a 30x increase in nodes connected to the network in q2 so now we're up to hundreds of thousands of nodes accessing itams every single day super cool super exciting also a lot of performance and scaling related challenges that come out of that as you probably expect some definitely new challenges for us in terms of being able to spin up test nets that can validate the new releases of IQs are gonna be able to handle that sort of load on a day to day basis the IMS HTTP gateway that we run there's a lot of different gateways but one in particular that we're running is seeing 300 million requests per day from very accessing about one terabyte of data so that's just one slice that people who are currently using HTTP trying to get content on my ipfs network and they're getting a lot of content there's a lot of people trying to utilize this infrastructure that's another thing that we support and have to make sure it's performant and we're trying to reach all of these new people gets you the number of people who are building on top of IP FS and building new types of data related out with you to social media marketplaces content things things in the finance space governance really there's a ton of new things that are happening and something that's really exciting is that now some of them are reaching millions of monthly active users so we have multiple apps they're all in the millions of monthly active users space which means there's some really exciting in the use cases you need to make sure we're supporting people who are utilizing ipfs for kind of their normal browsing needs for PSL content or videos or other things like that that's an increased scale that we need to make sure we're supporting as it's sort of follow that people build on top of all right talking about new new projects so what are some of these three three categories I'll touch on I mean the last time we're going to say grab that in February but mobile was a new thing this year where new mobile apps are being built on top of NTFS we didn't expect this to happen so quickly I actually think like it's hard to get one of these nodes running on a mobile device where you have like pretty significant environments in terms of like not draining battery how how much you can actually run in the around all of those things but the folks at textile have done a ton of work to make this actually really easy and accessible there's also a lot of stuff happening in identity space people build in a cup of IP Avesta to support variation custom identity work and apparently other really cool projects that Braddock will and a couple of others I'll mention that are building on top of my advancement cool new ways as I mentioned textile they're building a number of SDKs they have an iOS and Android SDK right now for building on top of IP FS to make mobile apps that you can also do web apps and build things on top of their network of cafes their documentation is beautiful I highly recommend checking it out it's a great way to get started on top of IP FS they make it really really easy another example is even this is the mobile app that the open Bazaar which is a marketplace for buying and selling goods on top of that if it's network it's fully here and they have a really slick new mobile app built on top of I think both for Android and for iOS building I might invest that you can use highly recommend it it's like beautiful thanks for them thank you they move a lot I would say and then there's also birdie this is still in the works still a work in progress but they made a new awesome Bluetooth transport for the p2p which is super cool so you can actually do fully end-to-end encrypted offline mobile chat so just here with the people next to you and then you're able to sign messages back and forth no internet connection and no anything I was also I think I would if I confess camp and we're continuing to push on them they're planning to make it a fully open source but I don't think it's quite yeah it's something to watch out for on the identity sighs you ports been around for a while they build on top of a PMS for looking up identity documents and there's also now free box which is super cool and they're doing kind of user app profile data stores on top of IP address use a Normandy B so that's kind of another database style tool on top of IP MS that lets you store data and this is super cool I think some of those guys are an audience up go chat with Mark and then finally there's the Identity Manager this is kind of a spec and they also built a proof of concept of it working in practice this is trying to do this is trying to do and so that you can log in with many different identity types say you have three box identity and you poured a movie and the IPS identity of the some other somewhere identity you can use all of these things in one identity wallet so you can pick which I guess it you want to use in which different application it still this is just a proof of concept implementing the spec that they've designed but super exciting stuff all those on top of the IDs and super interested in collaborating with more people to utilize in real world use cases my locks image it's just there's so many more things here but I want to call it three just cuz I think that super cool there's matters news which is doing an a Chinese kind of new sharing application so that you as a person who's writing the content contribute to this website you really own your data it's kind of like medium but you can actually pick up all your data and go and take this full site with you it's radical I think someone from radicals in there right yeah go chat with her about radical probably super cool she's gonna know about a lightning strike on this awesome even better you can hear more about this in a second and finally I'm also an identity space Microsoft just watch this yet even thing called an ion okay quickly in my last three minutes accessing IVFs has also gotten a lot easier desktop and what UI are too kind of and user tools that we created in the project to make it easy to access and run an IP let's note highly recommend getting started with desktop if you don't currently around an IT investment it's really easy you can add IPS to your tool team you have a persistent like widget in your address book earlier whatever top bar that you're going to play with and you also get all these beautiful visualizations at the files you're pinning and what nodes you're connected to and all this other exciting stuff you could run it you're connecting with IBM s companion which is a browser extension we've created for Chrome and for Firefox which now has 20 20 K people using it super exciting speaking of companion it's also now integrated with braid directly so that you can just install brave go to your like little settings area and I think this companion is default can't really see it this time in Java but you're just toggle that button you get a companion at it and it also gets access to things like the chrome sockets API so if you do some really cool stuff so expect more more exciting news to come there but it's it's letting setting you up where you can do kind of mdns connections to many people in a room over brave and directly share with each other they're also working on adding some nice IP best colon slash slash so that you know when you're accessing it might be FS URL directly within the browser itself and leading the charge opera is also working on IQs and the NS support they announced this back in April mini team something like that the aim hopefully is that have support in the end road I ended here and then Firefox we've been working on that for a long time they added support for decentralized protocol handlers in Firefox 59 and Lib D web which is that kind of generalized supports many many different web handlers is currently working on being integrated into the Gecko browser engine there's also really exciting stuff happen in the gateways and pinning space there's a ton of computing services not over somewhere in here from pinata and that's where anyways and as I expected there's a lot of stuff next I think a big focus I'll skip to that slide it's effective a big focus is testing and so we should chat later but one of the areas were working on right now two simplest application it's very hard to test we want to be able to spend up 10,000 nodes in a network be able to emulate having some of them hidden behind apps run a series of test cases back and forth and try and validate that the hundreds of thousands of musics connected to that because network are going to be very happy the next time that we want to go ipfs code op 5.

0 and it doesn't have any issues about to that especially since that contains a number of changes to the DHT to make it way more performant resilient those are hard to test and we need to validate that they rather worker really thank you so much the key to kiss that so before I have a very short overview ultimately but basically consists of these multiple chains that because apparently it's short for parallel chains these are connected together through a million chain which basically these pallets are connected together to a relay chain which is securing we have a set of validators that are securing and a chain and we divided them into disjoint subsets that are assigned to each one of its energy so here you have the tire chain e you have error tree when there's another side to this you have the small collectors you that are producing the parity laws so how it works is that users submit transactions to these encouraging these collectors produce the block here for the Karachi the meditators here validate these dogs send them to the reader chain and then you have shared security for all these tire chains chains now the networking that we use here is the usual custom thing that is used it was of the network you skip to to be here these collectors basically I wanted this merchant block so that our children it enters and then they accept the summaries to them the set of all validators produce the relation blocks again everything is constant however there are couple of problems that have occurred that we have noticed with using just gossiping and so one of the problems is that because we have this sort of charting or this basically number of step of of chains we don't have this pool of validators that are securing the whole chain but we have to also assign them to pilot chains these are these substances for helium 2.0 that would be the same case now here your attack surface is going to be wastewater for adversary' that then you have a huge central validators so this is assumed attacker wants to attack make an ellipse attack on the helium or argon polka dot and we did if we didn't have any packages he would need to attack one of these alligators but he would not be able to do much harm at the attack surfaces I take however many from sharding awarded the case of Aden these marriages the attack surface gets very small so if the adversary's flips in history Paddy's acres is going to basically compromising the validators are going to say we have a fallback which are which we called fishermen they should notice any wrongdoing of these three people or disregarded answers however it would be good to have some sort of network eclipsing this is also very difficult so basically security and the networking against it its attack is one of the big and our priorities for us another problem that we have is clear discovery all of these chains are going to be in a big condemn via Network but the problem that occurs from us is that these meditators are not set aside to entire chain and stay right there forever and will be connected to this collectors but we got it and the reason for that is that if a random assignment at some point three validators found malicious they don't keep to compromise this inspire chains forever but this patch is going to have some challenge fighter the shuffling of the panels however when you change the agitator for example if this validators right now it decides to create a the later on is going to be assigned to eat the question is how this disguise difficult to find the new collectors in the new pattern the prepare discovery becomes very complicated and we still are working on having basically this charted system and also having this rotation of the validators the second problem that we have is that a lot of times we have to communicate things that don't have to go to everyone so it's not like in nearly oh man you have a block and you have to send it to all of the notes but you have sometimes you have to we have to set things that are going to a specific validator in that case gossiping becomes very inefficient you don't need to cause if something that I need to send to one specific person to the whole network and assuming and since we have a large amount of data this becomes inefficient very very fancy so using Pacific for a lot of parts of the protocol becomes not very suitable yet so what we are looking into right now is some sort for alternative routing that would be more directed so basically if you want to send to some validator and energy coded piece that we are using you don't have to gossip it to the whole network but you want to find that very different address and you don't directly send it there however there are problems there too we wanted to be reliable because a lot of the security of the system relies on T's and data to be successfully received so these are the main networking challenges the short you if you have any questions - what's your current thinking around how to solve these are how to do propulsive mortgage one piece I guess number two but we are trying to do right now sorted by vector routing so they are looking into policies that this is this nose could have together and look into a car type of gravity strategies which will apply to these operations we are not going to be reduced in therefore part of the protocol but for the part that we are not going to send a specific ADA to everyone but to specific validators therefore to do something there are a couple of topologies I mean the easiest would be just products to everyone - however if you have a larger center console they all have two polishes that are a bit more [Music] in fact or theoretical discovery discovery defy that you can look at it oh do I need to define them I don't know okay whatever look into it and I'll ask you later oh thanks money getting calls for this mission your twice enough to get any attention that's the person simply psychometric foundation before to learn whisper [Music] I think it's like a year ago it was announced in detail and so we have a project that we are working on that is implemented is privacy-preserving messaging and they are looking to respond to it as a collaboration with staples and but that's in Paris that okay okay yes that was bad okay thank you Thank You radical of a network for code collaboration we've launched radical last year it's built as a Pico collaboration staff that's built on ipfs and we started the project with at least that code collaboration and centralized to one platform is to go collaborations to critical to be centralized to one platform that is expressed in this tweet that I included in my presentation specifically having all free and open source software which is like critical to all of the functioning technical infrastructure in globally right now so we were inspired by existing decentralized protocols networks like I speak that everybody who's here speaking today and we try to learn from other previous attempts like in SSP and get torrent so the current state of radical is basically a secure and offline first in terminal native alternative to centralize the code of collaboration it's super cool it increased with distributed version control systems like get includes a high level language for reprogram level of semantics you'd be network for sharing application states to this super-awesome replicated state machine language inflexible c'mon who's command-line tools so basically makes code collaboration of primitive and is building an entirely peer-to-peer developer experience that a Kurtz experimented experimentation about how we work and develop together but what I'm here to talk about in the last three minutes of my talk is how the code collaboration is more than just and what works exploring beyond peer-to-peer post collaboration at radical so I think it kind of is embodied in this tweet like you know code collaboration is more than code network effects are a thing there's a reason why most code collab is done on github right now and I had a really funny slide here that says this peer here and watching kind of had this like awkward relationship but we think and at radical we are exploring how blockchain can be code collaboration trans dental experience because we think that Foss is a really great design space for adapted formation I mean it's really common based peer production at its core and so we want to experiment with distributing rotation collective ownership and community stewardship of value I'm so secure an offline first is peer to peer that's what we built the radical step but now we're exploring what it is to have a community owned encrypted native peer to peer collaboration effort with watching so watching as a bunch of things that I cannot talk about in this one talk so please come talk to me after about all of this but for example one hit ko collaboration beat is we have verified wallet set I love direct peer-to-peer code blogger ocean ecosystem stewardship through block Awards inflation funding what if we had a chair a teacher dependency graphing between projects maintainer xand communities so we can understand how we distribute value from a system so we don't have to rely on building a business model for Foss and make an instead created a sustainable economy under free and open-source software and also what if you had a global project registry of foster projects allowing global canonicity and discoverability so those Network effects that make it come so great with us also add a really awesome security layer that audible and temporary project histories kind of up the security that we can have on any centralized platform so all this is really really cool and we're exploring it at radical and really we really think that code collaboration will bring about our new vision of what is free and open source sustainability mine Abel trustless workflows or LaValley exchange and global canonicity and those network effects that people think is only you know prone to github so cool come find me my name's Abby I'm on Twitter I have a tick home we're not doing a talk but I'll be around so please come talk to me we're going to be building this really soon so I'd love to yeah so I'm hosting a workshop personally on that's called beyond aus that's actually in this is like what it's about that's really great that this is learning here in our shop tomorrow at 3:30 where we're going to be diving into like doubts as a design space for experimentation so would love to have you guys all here it's on my Twitter I've been posting about it everyone Michael for three bucks four-minute lightning talk is too fast to cover everything but we'll touch on the important things cool so as the ipfs team mentioned three boxes building identity and data tooling for me to be apps and really some of the needs we address in the market were like how do we bring p2p systems to blockchain users and developers and let them together to create a better application I'll skip this and so yeah we've built a framework and really at the core of three boxes and identity protocol and developer tool so these two components together allow us to connect to my PFS storage that lives on top of or below distributed databases and connect those to the clock Jimmy Mullins and so really we looked up the needs of DAP devs to start building and people really that want to implement a service p2p architecture really also just want to build useful engaging apps because they still need users and so on the other side of that looking at the needs of users they really don't care about the tech just sort of what they can do with it and so we built this framework but more importantly let's talk about this cool so the decentralized identity stack really what we wanted to do is utilize the existing wallets that people already have it's like if you have meta master wallets how can we let those connect it to a control user data that's stored on the distributed weapon so our first implementation is or meant to be instances that are manifest but really we can abstract these and sort of let users control and connect to many different databases that are sort of on different storage systems and so here we design this protocol 3 ID which really an enable system so it's like an access control encryption data authentication protocol that allows many different accounts and keys to connect to the same ID which is actually what authorizes data on the systems and so all the other stuff we've built we have SDKs so if you're an app developer you can use the Box SDK to do all these things so it makes identity easy authentication user profiles for social interactions generalized storage for our content specific data and messaging and to make it even easier if you're lazy we have a hub which is already an out-of-the-box profile application profile numbers which is a drop in component that is like Twitter when you hover over name you get the social profile and vote comments rejecting demo was just the commenting plugin kind of like discussed but built on the sentiment systems and chat box there's a little tool box for the corner and identity wallet is actually what wallet developers implement to add three ID to their system and kind of like ties it all together we just talked about this agenda new all SDKs like why would you do this this is actually what implements three ID kind of like web through junior ethers dance does for the etherium protocol identity wallet does for three ID so it sort of works the same way it's with existing keys it works in different contexts react native nodejs can support different transports between wallets and apps like wallet connect or directly in the browser yeah and it's more secure for signing keys cool so here's how you can build app so three locks there's really like three different models so if your app has a built-in wallet so it's either self-contained mobile app or if you're obviously posting it for your users you can do something like this where you implement identity wallet and three March to s if it runs in a browser and the browser is facilitating the communication such as they can meta mask it's also running that meta mask would implement any wallet SDK you would pass the provider to the app that's integrated with free box tasks and also external wallet so this is like a mobile wallet that uses wallet and I'm sure some other transport to connect to an application yeah here's some plugins will make it really easy to build and that's probably all we have you can make a profile thanks okay so I'm actually proposing something but I'm basically a collection of do to prisoner since it ever going to share what propose oh that is pure long finger brother Research Board which is someone close to have the humming browning on your theory it's not a so so Who am I the gym I'm jutsu from Taiwan Taipei I'm a Washington TLT researcher and so I basically do a bunch of different different researchers including comes in support of ghosts and unfortunately one of my article is twitted by Potomac like one week ago and just to talk about Casper MFG explainer so speed also so and today so yesterday I just published the article that talked about privacy function and I'm kind of routing such Adam Koch to avail about it so I big cared about privacy analogy so people sometimes confuse about these two ideas its privacy is something that people know where you are but then don't know what to do what they're doing but but not know this isn't the other way around this it's like people don't know anymore but they know what you're doing so I mean I don't even see his wad I'm gonna do this one so it's actually has something to do with it being like lime because it's pretty critical to charting right the Charlie you don't want any node to store order data so we have to explain to them they'll into two ropes one is the Pony Express process and store data and the other one is like climb but the problem is how can you prove that data is available on a cool note right so the so what if solution is to put the troops and and and and the didn't know who's supposed to do that on a smoker track and so that the third party can just access the smoke might read ham senators action and then for attention to data ability right so but but that's a problem so the solutions that the problem is that the full nail probably can just a sensor the electron so this little solution to that is is to make the client anonymous thank you and so this game is to use the iron rally so this is how the right user can construct so basically you have to choose the fool the field hops right so these are basically the EEPROM notes so you have to like contract this circuit one by one and this what what's specialty here is that it's not just these two notes are not just in two keys but they're actually using some three or five of rhythm functions to generate it written over here and then so they're not really operate so that the bread number is unpredictable to the next relay so they each gain over 800 number at a time and finally Bureau circuit so that is it no it's actually it's anonymous to basic to anyone right so this is for nope so the opponent doesn't know who is their true unlike like yours so that basically protect this is our ship attack right that's it so I hope this really pretty quick introduction and happy to dig more into this concept thank you hi I'm Alex Brown and I work in the CRX I'm here to talk about how we used the tippy and webassembly to enable peer-to-peer training in the browser 0x is a protocol for a decentralized exchange that anybody can use to trade digital assets including a or C 20 or C 721 your transitional assets with anybody in the world to trust this way 0x protocol features an off chain relay and on chain settlement so what that means is that you share orders offers to buy or sell samatha with people and have an off chain and then when it comes time to settle the orders you go on chain and use smart contracts to do entrusted sub [Music] Xerox mesh is a way to exchange orders and it totally decentralized appear to appear fashion it allows users to share what air is directly with each other using computer network as opposed to the alternatives day which is using a centralized service to collect orders and I create a field Xerox masters grid didn't go and we decided to use webassembly which allows us to have an isomorphic average intern in one language we can compile 0x mesh to run in the browser or sera as a standalone node or annika back insert other couple reasons for this we think they go the PP is the original implementation would be to be as a lot more active also some smaller factors are more familiar with language get better performance characteristics or a million better tools the cool thing is that most Lipke to be components were crowned on the box everything that wasn't here from piles to webassembly works in the browser right away without any additional changes within within th to you for discovery things like Auto dadada relay and gossips a number of other things there really only two components that we needed to make modifications for door Houser the first of the transport layer so let me - Pia Transport is networking TCP or WebSockets that allows you to send messages with their peers so what we've done is we wrote webassembly bindings for the WebSocket transport and the p2p that's already done there's a PR the dopin we're going to be marriage off street but we're already using it in our own applications we're also working on adding support for web RTC transport so whoever TC can enable - browser-based appears to redial each other directly without relying on everything out the other major component that we had to replace is the storage layer so what we're doing is we're using level Eenie we added webassembly bindings for level gb just replacing these underlying storage component laughter we did that everything worked pretty well the pdb already has support for using level 2b as in the datastore so we just had to put the pieces together we were on the browser we use something called browser FS on top of local storage to shim have a file system so all the Spoils down to single time script or JavaScript library on NPM that users can download to build applications on top of mesh it's super easy it feels like you're using any other JavaScript or typescript lender ok hopefully if we have time I'm going to do it quick demo my minutes left cool so I have to give a disclaimer which is that United here is really wonky I'm currently going through my phone a hotspot connection but hopefully it'll work right now my browser is running - it's running webassembly we're hopefully connecting to the DHT and finding peers to connect to and that process go smoothly we'll start to see Xerox orders coming in there's a lot about yeah I'm using my house but hopefully we're really really out with our Butler graduates teaser when everyone else in have a backup video method so I just can't this would be running in the browser so hi I am here to talk briefly about IP LD yeah this is one of the interplanetary and family projects the IP is for interplanetary the LD is for linked data so the IP part sort of just like shared identity with the ipfs project which is where this comes out of but also of course part of the nation-state we want to build data structures and linking conventions which are good for when we're building applications that work at extremely high latency offline and interplanetary galactic whatever so to build data structures for distributed applications on this scale we need to get a couple of things that are like kind of a great opponent foundational concepts picked out so one of those is of course Merkel ization and linking that's such a gimme in this audience I'm not gonna talk about it anymore of course we need actual things let's just do it I filled the air some specs for how to do anything since we want to talk about data structures the other thing we have to figure out is what indeed is a data structure and so an IPL D we're going to sort of aim for eventually getting towards some data definition languages like the imagine schemas for databases but now for decentralized serialization concepts but that's going to be the aim and we're actually going to start somewhere a little simpler and I feel deep and we want some sort of common data model that works regardless of how you're going to serialize things and regardless of the application logic you're about to bring on top of it so maybe serialization oriented thinking about the data model but not like implement - fine so we want to have a model which feels like JSON so we can reason about malison lists and strings and so on but doesn't need to be defined by the white 4 byte serialization problems up to JSON so we should be able to bring something like C or and have an amount the same item or JSON what you do want to keep up or cut above when you all make this other strategy entirely this really fast we should be able to use all of these things interchangeably this is what I peel D is about standardizing this data model and making it connects to everything codec should be able to come and go from here as they please now once we have a state model we can also take those concepts announcement lists and we can build functional operations on top of these like any kind of traversal and fullness or selections that you've ever heard of from any functional programming language we're trying to write a standardized way to do this on top of this table so if you have any structure that you can map into the data models with your serialize formats then we can give you a bunch of libraries where these charges or thoughts and songs on top of all this we are now trying to also build schemas which allow you to have a plan the way you interact with your data name sometimes even if you don't see realized schema free data analyst Chema's at the same time our research I'm a researcher from total debt top Institute of Technology so I'm reading in Japan that they welcome you back and that conduct is not a network research and I give you an information of our product names in rock in a public forest blockchain network simulator so you see now that when you adopt 2015 and we are taking Internet emitter and wrote a parameter for as of today so anyway you see make dots of course state represents between old and drive truncations of rocks transmitting between nodes and why why did we develop this kind of software and it's very kind of easy to find this demonstration test by searching read the QR single as I am Yosuke and visualizer you can access and see on your browser and now let me both to you might go talk about boys of course performance and we're doing research on securities and we give a talk about our security search about eternal III m in Thursday writing talk session in the evening so your Sunanda decided about the sniper and this is a public vote senator straight so first target is a world Bitcoin right when those coin and naturally of 43m and as I said we didn't talk about consent and we provide riser and we provided demo and so that's right as you do routine network research by Josefa I seem to find on it on the type interchanging that you do your research say like we'll do you like around 25 minutes or 20 minutes and small groups what if okay that's it smugglers and then we'll come back here report back [Music] okay so for the small groups but it's like basically popcorn office' like raise your hand and say something about like a topic they'd like to discuss so one topic I would just like to provide for myself is a network light what is there like networking like DHT funding the funding boom okay any other topics this browser seemingly that's building browsers or yeah data availability right don't worry about much we get more granular there's actually a lot more people here than I expected so if you probably just do it yeah I think so yeah that's really like a like a big group yeah yeah yeah come over here boom s cars maybe I put it yeah sorry we can have someone take back the life instantly do a raise our hands on this hot picture there's that red petals red first thing you little next you're single if you own the interjects purple said again 100 next side let the person finish in a little Kaltura okay you'll need to checked on something with two thumbs up everybody the person will see you they said everyone cool yeah that's nice yes so I really liked that a lot of the talks had social and political implications and likely to talked about why their topic why their projects were important to system outside of just like here at a whole stuff which is important also but there's no limitations I personally and I would say that in country said the authoritarian countries like my country Venezuela and see that there are some lot of people that are actually using a lot of peer-to-peer networks there's one that is going to our chat network and you know these guys have nothing they have no money so they're just doing all this together they you can find it on Twitter it's called logic Network and this is what we do is use a mesh network and they're just communicating between them and also they're listening for Bitcoin transactions and I find this super super amazing that these guys are doing this news because you know they're just trying to find ways of communicating and avoiding the government because personally I was targeted by the government in Venezuela when I was doing a revolution kind of and like they look on my social media they become all my stuff so it's like very important that we use these kind of networks in order to communicate even more it's important in these countries where we don't have freedom it's more like a question is it important for these countries that these systems work on mobile 100% because they don't have computers how much does it cost the computer 500 $700 $800 and the mobile phone in Casa anyone $80 $90 so yes it is super important and also in Africa in Africa the African Regency is super important that we have these kind of connections Java 6 sorry Java 6 yes hundred I'm also we have this kind of phones that are very light phones they call it they're reaching like $40 50 RS phones Chinese phones yeah that's what people are using currently doing a lot of mobile stuff Network glue worth in Ecuador Peru big mobile stuff there and actually the issue is auto updating whatsoever yes there's some sometimes like different there's like different firewall like this actually happened in real life like we were about to do a deployment and it or enjoy training and then like you know some house some of the phones had updated select Android ten or whatever might go to us they're like didn't they have added some new firewall things so then you might then like to stay to connect to the emptiness that was an issue for really could be very big so this is holding cosine reduction so that's that that's the point right for me for out of this workshop I would like that's like the smallest of the things that we have to do projects themselves right we're just like may finally be censored Americans but probably is it probably there are more things falling in fact like I'm thinking for example okay so I am a developer and I want to like create for example you know about a lot lately where do are you you know do I just run at least jumping like whichever project has better the foundation of the time or is there like sources to move somewhere is that gonna be a white look report of like which stack actually works and works if we scale and like you know where do I go orientate myself around how quite great news things in any way that actually works right so for me I know how we would even call that right and maybe you guys disagree maybe there is a place in the web like that I just don't know about right that death for me would be interesting but Bunny General might this is one of my points like this source it through but but bigger one is like who has ideas for what is actually only not only the beauty probably expect right thus we be headed for a long time and by the way like I I really like the Harrises talk that that there was that there was there was listening that there was one point six million since 2013 right but that's the money and I'm also wondering like if you tried counting contributors time light starts very qualified people so their time is where it belongs so like how do we you know met with them who are swiping so I don't think what book or Forsman out that much but I think this is something that lifts me to meet some point like kind of like an ordinary level things because it's not just your app it's also gonna be sent to G to do things for example that's a very good choice if you need to have Street central pub just watching you just want to like be put in place out there investments to understand that you were not tied to instant change can change or nice Main St crisper integration song things so I think it's something though and we've seen a mobile change because absent that means with more sensitive difference rehearsals yeah you want to interpret but just wanted to say that I'm not even talking about dungeons as much because like that there are verses for multi-protocol right right so that I can't imagine where I'm going to go on the web like but I'm talking about like more like like new projects like that where I go telegram like get started and it's a discount of some place where people go through for new developments I [Music] just so did I get your question right so your question was I um I think like my application that this tactic works is that what you were asking or is that I'm thinking about so what is holding me to the projects back so like if there are more developers building on a smart developers are going to be contributing and like thinking about barriers for developers to use a blank developer recording and knowing what's out it's like one of them looks like you know that you want to be afternoo like the stopping in 60 to 90 minutes and have something working but that's not the experience that you're going to have to both of those projects at least not in an easy way right so I think where do I go to orient it myself enough research whether it's myself because usually you are trying to lock the information and this is [Music] my problem quite developers which have been pruned away monetization a little confused by the question or extra the whole thing or read history of discussing the yeah because I hear this because it says sorry for we're attitude like I was just curious about what we should like what the poll was - yeah I feel like right now really Chavez video to just travel to just like debrief I'm gonna be this mess but if people like no no I'm sorry yeah I'm just trying to like Congrats I do you think that people just need better documentation to answer your question about 7:45 fact-finding Commission and I actually don't know this is why I'm asking like you know season to me to people that's like this really is is it worrisome that like the barrier to entry and I like work on these things is very I was like a sort of like a privileged feeling why why provisioning it's the closing minutes and they're everything is like has really big words I feel like there are a lot a lot of it we've had this discussion I had this is settled look yeah go ahead in the vacuum cocktail hi so I always thought it gets to have a lot of it works cause really complicated that were starting out with a factor so if you can dump into how to use significant lessons but it's not going to so that then it's like the minutes of documentation it's an occupational training so it's a lot of it like there's a lot of stuff you need to know who is he a secret document did but you have to train your transaction first thing I think one thing in New Year's and highlight areas of knowledge in Interlaken one problem that I think is pretty core to the whole issue is that p2p just doesn't seem to be able to just pick a single horse and like go with it so there's like people are specific networks and wind are incompatible and communism right like I mean everybody was into communism Vegas reading to communism but you can still totally disagree about you know like trance Q over you know like it's they just don't get along right and turn on you I feel that it's the same with like a lot of p2p discussion for their site so many opinions and like technologies using but they're all like slightly different and it's just really weird I feel working much easier like project design decisions and trade-offs that they are making so I'm familiar with some of them but some others I am not familiar enough for example I'm going to think about every two P I am not too familiar with all the designs and how they compare to other systems and what they do and I think there's my knowledge is not like a peer comparison and the talk again just like a table remark some features but maybe some more the quality description that explains the Ginetta pretty sacred like what are the trade-offs and different features for white wines people decide to waste time for perspective but now driving into the people you are one of the we keep talking about adoption I think that p2p networks have a different value proposition but that doesn't mean that they aren't able to capture value right so like when we keep talking about like security convenience / security and how people are risking an inconvenience they don't think that that's actually true I think that we can start finding design spaces and people experiencing problems that resonate with the PMP value proposition and building for those problems and I was going to say something that was relevant to this I love this kind of yeah okay yeah they're all yeah like the education I want to make this you know it's me actually on that point but their particular use cases that uses as Springs for this like it's a movie like what would it work electability be a clear economic message consistent per se like a conflict something that's very precise like it hasn't it needs certain properties together to find this case just getting a sense in those contexts like what the difference is in trade-offs are usually different pension systems I think like priceless or some examples were useful just in from our survey appearances thanks yeah even wasn't a dad ecosystem like they're not necessarily compatible to each other so unless they're like directly like using metric or inline and so but I think that in some ways that's a benefit because the project can then succeed with Pentacles so like not use it it's like kind of like this design you start with the design of like what are the requirements for the app and then you pick yes I shall lot of people who come to dad try it out at the best first because that's like there are many and then they figured out they like being the content addressability or something and then that doesn't have that big into them they come on my but like I think that when you're doing us unfortunately it's like currently you have to have an understanding of distributed systems that have some computer science background so like really name these like properties of the database what you want and then find that which tools you need yeah but on that point which I think when we're designing to be a radical new building within a certain design space right which is creating a good career Network for resource projects to collaborate that it's like you can easily create a list of design requirements but it's so interesting because we're looking to scuttle blood right because we're trying to understand how people have designed before to inform the design of our actual network and so we're learning off of each other but also trying to create new requirements and so we're running into questions that people have solved before that have no actual guide load if for example like how many hops do we need to go grab projects scuttlebutt is following so you can see the post we just need to access the repo so we can pull the code so it's like really interesting to start seeing those overlapping it not overlapping and I think that there could be a lot more information share how these networks are designed before the size just like scraping through documentation yeah so they can tell something talk about war zone kid networks there's a talk by Enoch and then it's called go strike do you check it out to make sure things go it looks where I like the history books check it out and you guys are game by both cannot cease neutral zone right now in place for some time too long this time if you want us to help everybody in the same room and socialize with salsa feedback comments discussions one get to connect works we have a discourse in sense that means they're simply can also have to do a published report means everybody collaborating on that that's talking chest that's also thing [Music] let's later so one problem that I see said it's not easy to like like if you're building a network for phones and usually the place you go to is like pharmacies and you can just like there are so many other seeds you can read them all you can read about like how the Internet is made and how some things are made maybe won't even use anymore like what the thinking was like but way back with like just recently for some like things like two days ago it was really interesting news about like the most protocol and actually I thought that that was insane because the thinking that went into this was like I wasn't really aware of it but the cool thing is it was defined in it in the door see like years ago and I think I can still read about that I could actually I was like oh yeah they have some interesting issues and like these issues are relevant and like some of the system that they built there for example really similar to what people are building today was basically about that indexing identity information you know like it was was just had like a bit of a flashback I was like you know what why do we have something like that like there is no sort of like repository before like all the people you know condition for example and this is just about basic techniques you know like thinks that you have to like everyone who goes in the PDP has to learn about all that stuff over and over again and then they like eventually it was covering there was this this is super cool software project that you know I didn't even know about and they have they have this awesome way of like handing that or you know I think these kinds of things are just like I feel like the space right now is because we're not being in academia we're just you know like just people was different development I guess is is that this very product eyes the whole thing like you're always talking about that what's my product what is my product - how's my product constructed even the product of the source this is so I mean who's gonna fund that I mean I people to figure that out [Music] and we look to kid since the beginning two years ago and we realized they were like a lot of design yes assumption that they didn't work for us so in our I guess problem we had to do like a lot of designs bags trade our research back as you know you mentioned and there's a lot of I think like maybe like content to be that you like we like comment both and community knowledge sharing and big healthy product because that will provide monetary view by the same time like if you think how many graphical each key goes into a other or Society or ecosystem measure wasting more money on that side of like you know with the whole life which [Music] depending on how much public energy mass pretty medical protocol communications anyway attention to the project but not to actually lead into other projects that may give us a different approach and worked through that and it also today to compare them I think you just like know that he came in funding to make the kind of comparison were possible then I person because I think it's worth - I wasn't keeping up to date any way to do it all so basically that will help between probably informants all these different cues about a different decision that be more aware of why he said or he said because there's no organization that is coordinating everything or that goes around interviewing and collects data it keeps like a comparison there with no software 25 minutes on like challenges and it seems like we're proud of all we know this one particular travel engine plane who's experienced in like the latest poll this alert also if you want to like you know change topics or proposals we can still go through relaxed last few people yeah but be you know 20 minutes left we call so the top pivot in the solution solutions problem yeah that's a major yeah we just want to look conversation not only did you get us continue this ain't like we could like five minutes say yeah you know I gotta say the there's too many unsold problems throughout be gentle I think that's that's part of what back and I think respect as it was mentioned the specs are definitely important link right now an art scene you can see now what was proposed to like disappear whatever but the web brief updates on the team after there was a browser I think we're still good for downstate we're still trying to figure out the waters I don't think we're ever got a point you know just like hey let's just you know put it all in a nice document that way and because we don't have to say were concerned there's just too many things we haven't yet figured out properly I'm swishing from facilitator yeah like on that like there's a teepee sssh recipe with my tools and I imagine you what right animators and there's probably gonna be peace there's probably gonna be a bunch of stuff that are gonna be abandoned and forgotten we are now super excited about but there [Music] yeah so like okay so we find like a funding entity that wants to help me to be the right role or if I'm an engineer coming out right down while people join the space portal and maybe don't ever answer whatever like where do I go for this list of the problems he's the for me to head problems because I think you're right I just need like you to be took you time to arrive at this conclusion so how do we speed this out for that one space monoxide desert island so like before judge my love I should build lose a bunch of friends or pergola Delta we want to make every c1 and the idea was like neutral zone it's not because it's there's a lot of innovation that we just need to just make love equally you're like really deserve thinking them to a bit unlucky a dating to see other growth data problem directly to that diluted in the back burner because we have a business to run but we're thinkin T to that would you take a synergist yep and she's a feral nation where we would elect or is it like to work so that it's going to teach you crazy from us I started to them [Music] we should have probably an aggravator and then at the same time a forum with all of this miners like you're not seeing like fast venture toolbox that Eric we've been talking to love teams and only been in touch opinions don't need feel about it there's innovation lucky about particular says they're still in today's where it's been a product built their custom that I only have 20 minutes I know I wanted to say is that means before you leave the room please take that that you believe in got behind with you so I wanted to transition to this discussion that receipt may be continuing on the discussion and thinking about burdens both what sort of like this is kind of like peer-to-peer organization to centralized organization space so world people enjoy freedom is disgusting though people enter they're just discussing like the challenges of the peer-to-peer spaces and one of the challenges we kind of founded on was how you choose a toolset to build your your house and then kind of linens to like well everyone's kind of this trying to build apps and building on the product because they're after other things an open-source reading documentation is already a huge challenge alone for your own product compared to other products that make it easy for developers sketball so it's become all sort of a difficult space adventure for a lot of folks and so you're trying to think about those kind of problem you're assessing and we're going to go over the next 20 minutes into solutions just kind of like popping out solutions making it cool to like maybe go around and see think about this and then we can have a discussion later or maybe we can have a happy hour I think that the main problem is that we all will be separate and we're not collaborating that we're not like integrating what other people are doing with our the thing that I'm currently doing so I think that's one of the main thing and that's why I said that we should probably have a diamond grader or the need to be networks and at the same time rate them and that we put some kind of rating system there I don't know bro or something like this could be useful our systems and often times there might be reasons why certain project us delivered differently it's not easy for somebody who's moved to a figure out why but it's very hard to figure out all these aspects my heart you differently and then being able to think about this but maybe resolution well I don't know I have a feeling that it's a funding project problem because if there is more money that means you can explain things more you can communicate better if it will better documentation with an experiment better and also like when you are trying to use peer-to-peer technology a lot of times if you want to get funding like we are not creating we are using that for example yeah but then you have to like people want you to use the technology that is more popular because then it's easier to get funding so this is also but yeah what we want to do wrong yeah we'll just focus on solutions here and I'll start with signal on and they'll just start reading that's we can see yeah let's go around what maybe it's because we don't have we don't have a group or something that who received that money we don't have data organism you know maybe we have a foundation for this or I don't know I don't know what's the best book but every location or something here kind of knows me out standards also very helpful results like some of these issues would be if there be some place to letting publish these things and I know that like these places exist you know you could just you could just go through my testing for example it's just that for a lot of people in this community also very like mutilating internet in general and might not actually you know like you know I'll get into these kinds of things so it is the forever like it has a bit of a traffic problem so you could either just for your own standards organization which is chrome at an existing one but then somehow make it so that I can what about doing at our implement sensitive document to get some key with the except you're starting with JavaScript or something like that you're avoiding my sister didn't take it difficulty music go okay geillis listen I know we've done a lot of problems but I think in this section we're gonna train for every one into Center we're talking about solutions to the issue of like the fragmentation and the preference base and how to make it easy for new people to get involved as well as like solves for the questions around funding and standards like that sensible the solution is not very simple example for tech that you get kind of made that will be coming across every action cool we're gonna have 20 20 more minutes I think first of all I think it's fine if we have some level of fragmentation and expiry it's true we're trying out different things in but it's also useful to have any companies eventually the most successful the solution is going to be about that but I think we're at the stage where okay so we and the solution is is to have an activator please look at organization yeah where we can stand there's funding for what come here I mean I'm imagining maybe for me the quickest way that I think matches the depleting projects they may be taken all the they talk to each other thank you their funding to put it together and say that some rusty vigil or maybe small teeth posed interviews or research is reduced in comparison if the team's agreed in just yet for all their documentation that particular github repo or whatever it is just a quick way to start so this is comparison I have no idea maybe they could pull in funds and figure out what will be all need and then one thing that is offering that across all just universe thanks for ditching way protectively who's here for the next floor treatments like on my part I don't think a lot of Fundy was that may issue wipe these ecosystem oversharing we have been like and also I think like you very surprised at how little cost for each project to have just like community people even wanted to like study issues and I mean there's another problem for which like the each project I went through that my voice being like we were implementing the best process for our project we'll get a doctor so you know like this like no I'm not incentive to share right away we do want to share later after the fact because then you get a selfie on the credit but before the fact is but a similar thing like it does and the Sun biggest issue I saw is not what the design aspect they were sort of eight or more like what are the options they saw that you refused no you need it and then why I think actually that's them learning like them it's like this Beach like that's so we don't want them to say bye before like everybody because no one explained why they didn't choose I so yes I if we just come up with like initial for all those cars and how do we spread among or the project thing like everybody we they surprised us more is there to share learnings yeah that's that's the biggest part that is a lot of projects point yeah but I'm gonna say I sort of did it the title Asians in my head you make something out likely to people for six months pretty good fifteen more minutes our ambush I would just go in London [Music] it's just that I think the Dow is late potentially actually very or say we should probably sort more generally just defining what maybe find a lotta pieces here what is up here those are things that are my priorities if I had been declining yeah I was gonna say well had something to listen said it was something but yeah I really think having an initial scan of the committee and funding that through the communities like the projects that have a lot more money to spend like maybe they convention like 30k to pay for some of the work of the specialist so I would say like an initial report and maybe also like something's like an investigative report that like comes out with like some Doc's maybe and then like an event you like get everybody sharing roadmaps around sharing the road map I mean road maps is this something opening up hiring a debate the commitment just means like I'm gonna start streaming r5b please and those are you know it's pretty I think advertising Curren don't worry the technologists already have in this field will help you find the real problems under this list first and then the funding so everyone is the first time you should just over finish what you have advertisement unit yeah more like yes just to go back to them just say no myself problem section under protect pens like I think what a lot of people of Reyes was that a project can't necessarily or doesn't necessarily a time or one prioritize often that sort of work because they're already getting clients or something so 30 getting to do other things that aren't that so it's kind of like sometimes it's hard to do some doctor or you always be prioritized so the 70s they these generate foxes and assertions about you have programs so he respiration like monthly dues that tourists I think that vehicles that particular solution assaults actual horizontal fish I think air yes like questionnaire definition ice or this resolution actually sorta start something [Music] story what everyone likes to write though right - hackathon so you don't plug the Ducks oh I have a prospect I think after listening to this provision you might be wonderful it is I mean what's time today [Applause] you