TOKEN2049 - Decentralising the IOT for Smarter Cities and Businesses - Vlad Trifa (Ambrosus)

so I wanted to present a bit about the IOT and how it works with blockchain because as they were just mentioned blockchain is interesting but it's always it always comes to the physical world so how do we handle how do we incorporate the physical world into into the blockchain so I'll just begin with a little bit the sad state of affairs when you read the news every single day you read something really sad about how corrupt our block chains how corrupt our supply chains are you always have people being affected by bad quality products or contaminated products and especially medicine if you take medicine there's more like than 10% of medicine which is being sold from esthetical products which are actually counterfeited so that's a big issue and are people dying every year because of that so our claim and the topic is really that if you combine the blockchain with IOT technologies Internet of Things sensors tags and you can put those on physical products you can suddenly start increasing the quality of your data increasing the traceability of products so first you will be able to trace the origin of products where they come from when the products were made how they were being processed how they were being transported and stored and the idea is really incorporate that on top of blockchain because blockchain itself doesn't really know about the real world so the key challenges that we still need solving because we're not quite there yet there are some key challenges especially in the internet-of-things space and also blockchain that needs to be solved first is the fact that crypto fees when you use tokens to pay for a system if there is volatility no big company is going to be able to use that because they don't know how much they are going to spend at the end of the month so that's a very important first prong to solve you want to make sure that a company using such a system they know exactly based on how many things they produce or move how much they are gonna pay at the end of the month the second issue is that there's no single protocol for the Internet of Things out there today they're like thousands and that's a big prawn there's very limited interoperability so when you read those crazy scenarios of smart cities devices talking to each other we're not there yet because of this first fundamental problem how can things how can devices understand each other at scale in a very decentralized and heterogeneous environment the third issue which is very much the limitations of current blockchain technology you have very limited scalability a very limited performance so you're not able to handle the 50 billions of devising sending data every few seconds out there you cannot do all of that on current blockchain technology another issue is and to end security because blockchain itself it's now it's whatever you put in is guaranteed to stay there but the problem is garbage in garbage out so you really need to work on ensuring that whatever data you're using for your decision-making comes from a trusted source and IOT systems today most of them they just use encryption which is nice but it's not sufficient with most of IOT deployments and sensors on market today you cannot mathematically prove that data really came from a sensor and hasn't been tampered with somewhere along the list and the final challenge that needs to be solved is the lack of tooling there are so many protocols therefore there are so many frameworks and pieces of software but it's it's very very limited and getting into the IOT space as the developer if you wanna take some devices start to program them you just need a lot of knowledge a lot of expertise to be able to pull that off so Adam process is the company I work with the project we're doing is called m-net and we try to combine seamlessly very efficiently IOT technologies sensor data with the blockchain so the core project we're building which is m-net is essential decentralized backbone where companies and trip runners developers can suddenly start to integrate all sorts of devices sensors ERP systems IT enterprise systems printers anything in the real world to be able to exchange data with each other to integrate data from different sources in a very very efficient way at scale for that one of the things we've been proposing is to use a universal data model so very very simple way to describe data to describe sensors what is this sensor what is the data what is the quality of the data I'm getting from this sensor can I trust it so it's really a higher level abstraction model for developers which is not just wallets transactions but ring a higher level abstraction which is assets events and I'll get to that in a second the second thing we've been trying to solve is really the ability to support both public and private data so just data encryption as I mentioned is not enough some companies are very very scared of their internal data their sensory data gain getting out there so they really want to make sure that are fully in control of their data it can be stored in a very very secure place maybe sometimes the data needs to be centralized for government reason for legal reasons you want to make sure that the fees of using such a system are stable in u.s. dollars or pegged to to any fiat currency so they know for a fact how much the system is gonna cost at the end of the month our system isn't only governed by crypto economics so all of the interactions data upload all the security checks is done on chain we have smart contracts and finally everything we're building is entirely open source so you can head out to our site start developing testing around so all of that is out there and that's really important to be in the open source element obviously because of the trust if you're using something that's proprietary in the IOT space you don't know how it works you don't know how it's there so you don't have the guarantee required so if you look at a tech stack we've put together so first we have the core blockchain that the core layer the brain the government if you want of the system which manages all the all the payments and things like that on top of that we've built a decentralized storage solution something like ipfs if you want which is fully integrated with the blockchain where you can pay different actors to store data on your behalf and they would get rewarded for it and the third level which is the most important piece in in my opinion is really all the SDKs all the tooling the api's that allows someone to build very very quickly adapt on top of that without even necessary having any blockchain expertise so just to jump a little bit more in in the tech of that the technological aspects how it works is that first we have this Pentagon here which is Hermus that's basically the name we gave to that service it's entirely open sourced and essentially that is a gateway on one side it provides a very simple REST API for all sorts of developers to send data from devices or build an app that reads data that is a centralized component which can be run and operated by any company in their own data center or it can be used in the cloud something like infra if you want now what happens is you're storing all those sensor readings locally but what happens is the the hermit's gateway takes a snapshot of every single modification every single entry you're doing there and puts that in a bundle so first this is very very fundamental because it's like an audit log of every single operation in your database that gets put into a bundle a bundle is essentially a huge five hundred hundred fifty megabytes that contains the metadata of every single transaction you've done in in your her mesquite way and you upload that on the blockchain so what happens when you do this is you pay a fee which is stable you paid in tokens but if the token price increases you need less and less to store a bundle the price of uploading a bundle is fixed it's 12 to US dollars and you can put around 15,000 16,000 sing sensor readings or any type of data you want to record in a bundle then the third step is you have this network this distributed data storage network which are Atlas nodes and then essentially their only role in the network is to download the actual bundle the full file only a proof of the file goes stored on the blockchain in the transaction but the atlases download the file they can verify the file they downloaded is actually matching the hash of the bottle on blockchain and for every single bundle you pay the twelve dollar fees and that will be paid out to every Atlas not holding your data and what's really interesting here is that it's incentivized on one hand but on the other Atlas nodes actively watch each other to make sure no one is modifying any data because there is a reward for any node to say hey at last 24 he actually changed some data and they earn as well which means we as a company or as a project we don't need to store with only to monitor what what the nodes are storing all of that is handled inside the smart contracts so just to finish a very simple use case it's one of the type of application that you can build very very quickly on top of m-net so first this is a vaccine company who wants to send some vaccines they want to transport it so the first thing they will do at the factory at the packaging line for every single box of vaccine they create they generate an asset and an asset is essentially a crypto crypto secure globally unique identifier think of it like a wallet address essentially for every single product which has a serial number and then the physical product itself is sealed with a tamper proof rfid tag so you can actually know if someone tried to open the bag or if someone is trying to create counterfeit medicine they won't have the digital identity that was creating the first place on the blockchain so all those years all those vaccine packs get packaged together into a cold box it gets closed at the factory and in the coal blocks you can put a sensor like this one which is something we've built in-house it's a very simple Bluetooth sensor which measures temperature humidity motion light it has a battery and what's really interesting in the reason we had to build our own sensors for that is of the Shelf sensors you buy very few of them have actual crypto chips on it so this works like a hardware wallet there is a chip which contains a private key that is only available and generated in Hardware on the device so you have a way of measuring the box if someone opens the box if someone if the temperature increases above a certain level you have all of that that is stored and those are built in such a way that if you open the device so if you try to tamper with it then you will get detected and what's really interesting once you put that whole box into into a trap that truck moves along it can also have other devices inside the truck that read data from this sensor forwarded to the blockchain and when it arrives at a pharmacy or at a doctor they can pick the medicine they can open the cold pack and they are able to verify not just that the product itself is authentic seeing thanks to the asset the unique that digital identity but it can also measure the entire conditions of transportation which they can share with their patients or with any doctor who is buying one single pack they can verify all of that so you know just to conclude I think the the key main points I was trying to conveys first current blockchain approaches and and the idea that every single node must have all the data obviously is not gonna scale and especially not when you talk about IOT scenarios where you have billions of devices all over the world not all the data of every single sensor needs to be shared but only a proof of that data needs to be visible and publicly accessible so the best trade of where you get both the advantages of centralized systems and the centralized one is where you have one single public chain basically like the the presentation before where you have one single root chain if you want that it's public visible by anyone on top of which you can plug a variety of either private or semi-private change where the data full of that is centralized so this combination of this centralized and centralized system is essential if you want to get the performance the second aspect is really interpret ease so you need we need to find a single way to define to describe data to describe sensors what they do if you really want to build a global scale autonomous IOT systems that can be used in an ad hoc manner the third aspect is the security level that is used in blockchain is essential very few IOT solutions hardware devices use that today and it's fundamental we start adopting the best practices that are in been using in blockchain directly on devices as well and finally the most important bit is really the developer tooling the easier you make for someone to build a hello world to build a prototype and then be able to scale that that's going to be really a trigger point if you can if you really want to build a truly global autonomous ecosystem so thanks a lot for for your time happy to take any questions if any otherwise I'll be around for next today so very happy to discuss ok any question no good go to lunch at zero thank you