DAPPCON 2018: Lovelace - The Bluzelle Decentralized Database - Neeraj Murarka (Bluzelle)

Transcript

I'm gonna be talking about the blue cell database today so what is it it's a decentralized key value store if you've used Redis think of a decentralized version of that that runs on blotching technology so we don't actually run on top of a blockchain it's a question that comes up often we use decentralized trustless technologies to build swarming technology to power a database so I'm gonna be talking a little bit also about the data economy which is a slide you see here and the data economy real quick is kind of our phase two we've already released a lovely version of a database we released the first version on July 2nd what we're trying to do now is build some higher-level constructs so that people can use the database in the context of the data economy so I'll talk about that

briefly and time permitting I'll try to show you a quick demo of a solidity application that's using blues out so quick history we started the company in Canada in 2014 move the company to Singapore in 2016 and both myself and my co-founder Paulo beans are from Silicon Valley originally that's where we kind of worked in our careers but we're both based in Vancouver British Columbia and that's where all our development happens that's us sitting in an office in Singapore and we came with our backpacks and nothing else we did our I CEO in January of this year and the ICO token powers our database these are some of our clients we built a kyc product in Singapore and one of the reasons we

needed to build the databases we realize that the blockchain is actually not a very suitable database you can't change the data it's it's immutable and it's it's expensive high cost and the latency to store data on is very high so if you want to build a kyc system for a bank it just won't work this is our team myself and my co-founder and real quick or advisers we have some excellent advisers these are four of the advisers I work with the most Bryan Fox how many people the room we've heard of on Fox from the free software foundation and he's the also the Arthur of bash he loves to also hear that people know he wrote finger so if you guys use finger from back in the

day Ryan future actually is a Canadian in hero the original ripple protocol so he's one of our advisors he's helping us with consensus algorithms Prashant Malik he was at Facebook and he's one of the co-authors for Cassandra so obviously having someone with that background is fantastic on our team and then Alex Lovington one of the developers of the core developers at the etherium foundation he helps us across the board with everything that has to do on chain so the problem really right now is I people don't realize that when they're using services like Facebook or Google or Amazon or anything they're volunteering their data it's not a free service your data is actually a currency so when you're giving your data you've got these drones or Borg drones or

whatever they're assimilating all your data and this is a slide we got from The Economist just full disclosure but we love it because this is what's going on there's big Titans now that run the internet and it's not decentralize it's actually quite centralized and we're trying to push things in the back in the direction they should be so phase one was our database which is something I'm going to talk through and yeah so two of our key hires that we had recently someone from NASA Isabelle our researcher and then Andrew from Pertino so as you know in the past you had a single system model if anything went wrong in that corner you're in trouble your system went down we've got the cloud model in the middle and then the

future model is this swarming model or each of these are swarms I know it doesn't look like a swarm but you'll have to believe me and within each swarm if a node goes down it's fine because the swarm has many other nodes and your data's fine so your data basically gets chopped up and stored into these swarms so you get enhanced reliability your data is stored across the swarms and security and on-demand scalability your data is going to scale across the storms as it increases and of course there's a crypt economic model behind this the whole point of the token at least from this perspective is for people to run nodes and they get compensated with buzel tokens to be farmers we call it farming not mining because you're not actually mining anything so it's actually quite light in terms of usage of your CPU you don't have to have

a powerful CPU to become a farmer so this is what I was talking about data is now a currency so phase two is building this new data economy talking about a data wallet where your Curt your data is protected now it's stored in a wallet and that's how you protect it just like any kind of cryptocurrency so you can click on a button and it's like a kill switch it's like the right to be forgotten and this is something that's really important with laws like GD P R that have just come out in Europe and in North America ran from there's already jurisdictions that are starting to do this California just passed a law about two weeks ago rushed it through through the Senate to get privacy laws passed through at the state level so these things are

happening everywhere and it's we have to be prepared for it so storing data on the blockchain not an option and the data marketplace is a means to syndicate data so imagine that you have data that you want to sell to others you have GPS data OpenStreetMaps Yelp data whatever you can publish it now as a syndicator on to our database and it'll get on to a marketplace now whenever somebody uses that data you'll get compensated for it with buzel tokens so this is the data wallet flow I'm not really going to run through it due to time constraints but basically Alice's data goes through a data wallet and if ethers gram which is a fictitious social network wants to use that data it'll have to ask the data wallet for permission the data marketplace is very similar with the addition of the data

marketplace installed in there so now if somebody wants to access that data they have to go through the marketplace and it's like they're buying the data and again there's a cryptocurrency driving us and the decentralized web infrastructure is something that I'm very interested in because one of the key aspects of the internet right now is if you go through the World Wide Web and try to go to a website it's not too decentralized you have to deal with name servers you have to deal with centralized web servers that are storing content one of the challenges we had when we were running our ICO was we don't want somebody to hack the website where the ICO z3m address is listed so that's something that still needs to be resolved and this is something we're working on is using our database and working with other partners to

decentralize the web experience so the token velocity many different places you know data wallet providers david marketplaces if you publish data you can earn tokens and of course if you're actually running a node as a farmer you run tokens so right now we're running seven nodes that's for our Lovelace release and you can add notes so that's the whole point is for scalability that you can add notes and these yellow ones on the sides are the notes that have been added we're using Microsoft integer and VT these are two of our partners so they provide those resources so we have quite a bit of resources to kick-start to bootstrap this whole network and that's what we're doing right now I'm going to skip these slides but here's the metaphor so even

though it looks like there's a swarm sitting here in the middle of the Indian Ocean for some reason and some nodes sitting in North America and there's one in Greenland or something these swarms are actually covering the entire world the point is for each swarm to be geographically distributed so that they act like a CDN in a way if you pull your data your data's gonna come from the swarm notes that are closest to you and of course you have the redundancy if a big tsunami comes and knocks out the west coast of North America your data is still protected it's not like there's a swarm that's only sitting there so imagine you have your key value pairs and they get partition so it's horizontal partitioning for those of us that have looked at the kind of data I'm going to show you a quick visualization

here so this is a dashboard that we built recently and what it's showing you is the health of the network so this is connected to our test net it's showing you things like the total database size total key value pairs number of commits number of peers latency the reason we have this is we want to have a public tool that people can go in and they can look and see what's going on with our database it's gonna be improving over time but one of the things I like to point out is that the key value pairs and database size can go up and down and that's because the database is immutable you can remove data from it unlike a

blockchain that you know technically keeps going without bounds this database doesn't so you can't remove data from it if you want to and what I'm gonna do now is I've got about a minute left before questions but what I want to show is a demo of using the database let's see here from and if they're names from our contract now let's see if the Internet's gonna behave with me here right so what I've got here is a smart contract and I'm gonna be presenting this in more depth in the workshop later on today but I think I already deployed a smart

contract or obstinate and what I'm gonna do is I'm gonna write a value to it so I'm just gonna put in I don't know the name the key value key of France and the value Paris and I'm gonna write it and what I'm doing now is I'm publishing I'm writing a transaction to the blockchain and it's gonna be picked up by something called Oracle eyes that's going to then transmit it to the buzel swarm so let's submit it and what we'll do is while we're waiting for this I'll start to take questions because this is gonna take a couple of minutes we have to wait for confirmations so yeah let's start with the questions and is how do you compare to ipfs so

ipfs is it's not a database storage mechanism I mean we actually want to work with IEP FS for the decentralized web aspect the reason I say that is ipfs is basically an addressing mechanism for files but there's no necessary guarantees with ipfs that whatever address you're looking at is even going to be available there's no availability guarantees and what we want to do here is actually provide that database layer in conjunction with ipfs as opposed to competition so for example imagine you go to a website you want to be able to pull you know HTML files JavaScript you want to pull video files those would come from ipfs but if you want to pull anything that's dynamic in nature like user data anything that is stored on a

database it would come from blue cells so I don't look at them as the same thing they actually complement each other next question if anybody has one from the stage you can go ahead and okay oh there's no one so what operations depend on a theorem and it's limited throughput question is what operations throughput right now is something we explicitly built to support Avenue we don't actually depend on a theorem or any blockchain so we actually explicitly

built a bridge so that you can make requests from aetherium and use oracle ice which is a bridging service to talk to our swarm but quite frankly a lot of the applications that have come out that we've run were not using it during Hornet block chain they were actually built in traditional environments like Python JavaScript even mobile applications on iOS and Android so we're not we're not constrained by that you can make direct connections with WebSockets there's another question that just came in so the question is does it involve some kind of encryption for the data wallet yes that's a key aspect of the dear moment the data wallet has to protect our data keep it away from people that are unauthorized to see it

so yes you're gonna be putting native encryption into the database it's not in there yet so the data wallet is going to be something that's coming out not in this half of the year but the next half the first half of the next year so real quick I'm just going to do a quick context switch here some of you might have observed that I wrote the key and value pairs of France and Paris and now what I did is I came to the buzel client here and I just pulled it up and there you go it says Paris here now if I put in tour in here and save it I'm going to try to pull it up now what I basically demonstrate areas I change

the value of a key value pair off the etherium and then read it back in to show you an interoperability so I'm ready for the next question yeah let me check right so there's a new question in it's asking how is a swarm different from a AWS region I don't know I don't want to be quartered on exactly eight of these regions out of the definition but a swarm does not it's not a region a swarm actually covers the entire globe when you when you join a swarm that node membership function that decides which

swarm you're joining actually determines which storms are going to benefit most from you and one of the factors is which swarms have the best most to gain from you being added so let's just say you're joining a node a swarm here in Germany in Berlin we will look for the swarm that has the most to gain baby it has weakness in this part of the world and we will add you to it so there is no regionality so to speak in a swarm swarms are supposed to be Jack as geographically spread out as possible cool all right we have one more minutes if you want to just quickly wrap up the demo there okay I apologize for the person that just asked the question yeah I know I was just thinking you

could wrap up the demo okay we have time for the last question I just submitted a read I probably will not have the time to wait for the right read response but basically the way it works with the theory before I close out is if you want to do a read you make a request it goes out to earth wise the request is made and then you wait for the call back the call back comes back with the response for what you're asking for