J1 SAY 05 Gerald Crescione + Victor Valladier Blockchain based SLA Why does it matter

hello hi guys nice to have you here in this room my name is Jana this is Victor here standing next to me together will speak this morning we will speak about IT infrastructure in multi cloud environment we are going to talk about decentralized service level agreement which is very specific to our industry but I'm going to go into some details just afterwards and we are going to talk about on earth-2 Unchained data management with a specific focus on what we call state channels and more specifically multi multi lateral state state channels just before I begin are there any system administrators or maybe DevOps in the room yeah Wow three people okay I hope you guys enjoyed the show because we're going to talk about some stuff about the infrastructure okay so yeah a couple of words about us we are team of we are team of seasoned entrepreneurs we built several startups so far but not quite um Nate not crypto native though we're more coming from the DevOps system administration and startup ecosystem there's about 12 of us in in in the team where the team is spread between friends friends and Canada there's no regular office everybody is working remotely and our CEO Stefan just wrote a book about i-ight infrastructures that you can find on Amazon if you show any interest ok so a couple of words about our our ecosystem and and and and what we do so in the last years there's been a specific development in the way that people develop their apps and web services or their websites we called out the cloud native way of developing so basically it's all about work working with micro services that means that when you develop an app you do not develop it as a whole system on one whole month analytic system but more like several micro services that are interconnected but yet independent that's very useful because when when you need to move part of the micro-services or when you need to adapt it like scale it for example it's very easier to move one piece of a puzzle than they've done the whole picture you I guess you can see it a bit like a like a Tetris game maybe okay so what with what square scale is once the code is written will help then switch to we help the developer switch to production mode by by basically by giving them a one-click solution that allows them to deploy complex infrastructure in one click and in couple of minutes well I know it's not very sexy and we're going to talk about that but it's it's a major breakthrough in the industry since so far this kind of development needed like weeks of work of highly skilled people like system administrators so through through our solution we also host the the application the services websites and so far bye-bye cloud providers people like I mean regular cloud providers like Amazon AWS and Microsoft Azure people like people like that and and so far we support only only one provider in one data center but step by step we are going to add several data centers and several providers in order to move towards what we call multi cloud environment how is that interesting so basically it allows people to to host part order of their micro services by one provider part of the micro services by another provider and have the whole thing work seamlessly which is quite important because in just in case one provider goes down because of a breach then you got the other one who is able to step to step up and and you know take its place so that your whole system is is working seamlessly sajida just before talked about GDP or we also working on this on the next step of the GDP also GDP are is basically a bunch of regulations who are supposed to help people protect data privacy and we believe that as you said Salida in in the near future people related that I will have to be kept stored in their country of origin so far it's only mandatory for health data but we believe that in the near future any data will be have to be stored in the country of origin of the people and that means running several split databases this is quite complex to do and through our solution you will be able to do that in in one click which is also a part of the stuff that we could work together on with the linear linear project so we are now in in public beta mode so if any of you want to test the product just make sure to come see me at the end of the show we have vouchers for you guys to test the project free and and try and run your apps with our solution so let me sum it up for you basically in one click without any expert system administration skill you will be able to develop a dedicated scalable and complex infrastructure which will allow you to save time because you can do it in minutes instead of doing of in weeks and will also help you save some money because well it's fully automated so you don't have to pay for the team the business model of the solution is a paper use business model meaning that if in case you need to scale up your infrastructure you will only have to pay for the excess or the additional amount of power that you need before scaling down again so this is quite agile Institute ooh expensive so basically it's quite better cheaper and quicker than what's already in the market now I'm just going into a little bit of detail about scalability you you may have heard I talked about scalability quite a couple of times so far I'm not going into much details here but just for you guys to understand what I'm talking about scalability is basically adding some pepper well it's pumping up the volume on on your infrastructure so basically it's adding some either some extra computing power or some additional entry points to support to support the arrival of extra visitors let me give you an example I I guess you all guys you noticed that in December store we had some kind of massive media coverage about kryptos everybody was talking about cryptos everybody's was going crazy about Kryptos there were TV coverage about that and everybody started rushing on platform like Kraken or by Nance and stuff like that they want people wanted to open up an account and start trading you know it was like the gold fever bike in in the in the 19th century everybody thought they were going to be rich thing is those platform they were not prepared to - you know - for the sudden explosion in demand so there are servers went down one after the other scaling up these servers takes usually weeks of work it's quite tedious it's quite difficult to do through square scale or this kind of solution you could have done it in one click and in a couple of minutes so that's basically what what we do okay there's an expression that says you don't need a blockchain for that and so far you don't need a pro change for that okay let me tell you into more details the major problem or the major issue that we are facing in our industry this is a program related with trust remember when I mentioned daemon my image of a puzzle or Iran or the the Tetris game I said that the microservices were all interconnected one with each other with another well this interconnection happens through service providers those service providers sometimes they are up and running and sometimes they are down you know and sometimes when you when when you experience when you need use their service well you experience a bridge in the service but if you go check the Status page they say no it's okay we're up every all the lanterns on green and there's no problem with us well the thing is that you in front of your computer if you want your app to be running doesn't work so it's just as simple as there we are against yours the problem is if there are they are down you can get a refund because for some of the services you pay for the services okay so if they are down if they are not delivering the quality or the service that you are expecting you are supposed to get a refund okay well the thing is it's just the very same people who are able to pay for the refund and also who declare if the service is up or down you know so you've got a major problem of trust in this in this in this field I guess you can see that well it's major problem of trust it also applies to us because we are service providers so the question is what kind of trust can our customers put into our own our own services and this is this is a an issue that we decided to tackle by by gathering a network of contributors in our ecosystem and we just asking those contributors well basically to perform the monitoring of us our service for us you know to do that in an independent way and then we are going to improper eight this monitoring through a peer-to-peer network and Victor is going to go into details about this peer-to-peer network just afterwards together they are going to reach a consensus of information about the service is it up earning or not the data was going to be to be published into the Librium blockchain and well obviously completing the few full circle this data will be fed to a smart contract and a smart contract will trigger if need be a compensation for our customers so that's just the way we we want to do that we feel it's better we feel it's better because it's transparent because it's it's well it's outsourced to the customers and and because it's also fully fully automated and we believe that this kind of system could be used by other service providers as well so maybe they can just use our solution for for their own services one of the main technical challenges that we face well there are several of them actually but one of the main technical changes that we face on the blockchain side is how are we able to scale up the system within the blockchain without exploding the cost barrier regarding the transaction cost and the storage and well obviously power the power as well and so so this is exactly what the problem that we've been trying to solve through what we call a state channel that Vito started designing just before he explains you how we we are going to do that I just want to give you one last piece of information this whole Brooklyn detector is still a work in progress meaning that we are building it right now this is not this is not fully completed okay so if any of you guys have ideas contributions stuff like that to give us feel free to interact at the end of the presentation would be very glad to to you know to build it with you this is anyway something that we want to be open to the community so that everybody can use it and customize it so this is a work that we wish to do with several people there might be some overlap or integration opportunities with people like swarm who just presented before we just had the opportunity of a little chat with him see which kind of integration opportunities we could have together so well again work in progress I hope you you will have some ideas for us okay Peter so now we are going to intern the details of the implementation of this challenge state channel so as you said Gerald the good point about hosting RSL a contract in the blockchain is that is transparency because if job each party of the contract is pretty sure the contract to be executed according to the way that it has been defined but the issue is that blockchain is blind and it can't fetch information to monitor the website of the API by itself so you have to provide that information and there we shift the problem from the contract to the information source the Tres problem so the trust issue is in fact to grant the data to be independent from square scale to be not to be vulnerable to attacks and also we have to grant data integrity from end to end from the moment when the contributor is fetching the data until the moment it is anchored in the blockchain so there is a lot of challenge and we are going to talk about that so this is a global overview of architecture with only the actors at the centre you can see the key contributors networks so this is basically a peer-to-peer network you have to subscribe to until inside this network and when you do that you start to to fetch data for us it's you have to download for example an ID an atom from your browser plug-in and this will enabled you to contribute and this is what grants in dependency of our data of monitoring data because this network is self governed on square scale has not a word to say about governance so when the data is has been fetched it is sent on both sides on the left you have square scale which provides a dashboard the live dashboard so it enables the develops on the sea side means to access the data directly from the contributors and you have also on this side the job queue where Scott's car we define which website has to be monitored on how to be monitor down and on the right you have the Unchained side of the project without smart contracts so it has to manage the relationships with our contributors and also to use the SLA itself we will uncover the data in this smart contract and process if some people needs to be compensated and if needed we the smart contract will pay these customers automatically so to process this we have several key to ensure safety data safety the first one is quite usual it to distribute the data storage so in fact we are copying the data in every contributor computer and also anchoring it on the blockchain so this enable to the data to be immutable that is to say that if you have to change the data some somebody want to attack by changing the data it cannot because it would have to change the data on every point where it is stored that's great to immutability we have also proof of contribution which grants that workers the contributors are doing good jobs and are good good willing we are going to see that just after that and state channel which grant that no party can cheat the other and that communication are safe this will be explained also though that that to last amends proof of contribution on state channel nowadays hot topics in ethereum blockchain because it enables to the blockchain to be more scalable because we are putting some tasks often so it's very interesting first to develop this this subject let's go in the details of proof of contribution so basically when you watching an information on the TV or when you're reading something what makes you believe in the fact that this information is reliable you have several ways to to believe something maybe you have several people who tell to use the same thing or you can address as the question to some people who have financial interest in the in telling you the truth and finally you can also watch the history of the your relationship with the person if somebody tells you regularly the truth you can guess that next time it will be the truth also so basically this is the proof of contribution mechanisms I won't go too much in the details of proof of contribution but if you have some questions we are there so now we are going to talk about the main contribution we have - we want to do for a theorem community which is a state channel I don't know I'm not sure that everyone here knows what is they channel so I'm going to go back to the definition so basically state a state channel is a way for two or more entities to make a state evolve of chain and didn't synchronization with the blockchain so there is a lot of key words here first the state which is in the world in the expression state channel what is it in the spot contract States it's data you store you walk risk so it's quite important in your smart contract so in the data and the state generally neighbors the members to make some updates on the state but without contacting the blockchain and this is very interesting because you save money you don't have fees to pay gas to pay you save some time because everyone blockchain is getting slow nowadays and you permit the blockchain to do something else and so to be more scalable and regularly you have to synchronize your blockchain in order to eat to stay as safe as a smart contract so we are going to take daily life example in fact bakers they use a channel every day for example the morning let's say the Joe Baker is withdrawing some money from its ATM he puts it in the cash register on all day long he would have some transactions with the customers he will receive some money give back some some change and his balance will go up but the bank is not informed of this balance going up and it's only when the banker will go back to the bank to deposit money that the bank is aware of the the actual accounts balance so this is kind of a state channel you have a lot of transaction that the blockchain is not aware of and only sometimes once a day for example you synchronize the blockchain so to be more specific about what we do at square scale what we work on it's not only basic state Channel that we could use in square scale we'd like you have the ambition to do an open generic multi-party state channel the lot of key words again when we say open say channel it is to say that this project will be open source we believe that the community will have better trust in a tool which will have been developed by the community and we could be challenged we could receive some help some suggestions and then this will make our tool more safer in fact and the Dinaric part is because when talking about state channels you can cover what variety of use cases it can be hard to hunt transaction it can be money transfer or information providing like Foursquare scale so the tool have to be very customizable to fit your needs and the needs are very wide so to do that we purpose to develop kind of a framework with some holes some function that you can overload and this will emerge to customize everything in the state channel for example you could make use of state configuration defining your own format for the state something like a database format and you couldn't define the way you define the jobs that you distribute them adding maybe some scheduling and validation of the results all these functions has to be very precise and fits your needs particularly another way you have to customize your state channel is the levels of security you you want for example there are some stage and L wire that you use we - what remember of your family or with some very well-known partners and there you can be sure that this members will cooperate with you but if you open your state channel to some thousands of contributors like for square square you can't be sure that there will be always online collaborative collaborating with you and you have to be resilient to their failures for example next the multi party keyword so it is just to say that you can have several members more than two who work on the same stage to update it though this tool doesn't really exist now you have Wyden who makes a channel but it's only one to one relationship so this customizable aspect will be maybe our strengths and the open source also so now we are going to do to explain how do we grant the safety of the data even if we are off chain so basically it's the use the use of cryptographic signature so quick telegraphic signature it's to say that you can prove at any time that a member agreed with a version of the state at a given time and you can exercise your rights on the state who has been approved you have version comparisons because each member of the state channel has to be to agree on the same state is very important and that way you can grant that no one can be can lose too much in fact you can lose more than one drop value and to take back the example of the Baker if the Baker don't give you your bread you just stop paying it pay him in the same way if you have a you have been cheated by a member of the state channel you just quit the channel so this is a draft of what could be a job addiction process so as we said it's a work in progress but it could it gives a good example I have got requesters a creator of the state channel the owner is a tech channel so as a process start with a contributor who asked for job to do it can be compared compared to the get work function in the mining world and here is every science job specification a mission to do and when receiving this the results the requester can integrate the job result to the states and he will send back written contributor a signed version of the state's updated and this signed version of the state is the key for the contributor to one verify that its contribution has been taken into account and to exercise its rights on its later if needed when synchronizing with the blockchain so this was in technical details about output a mission about contribution to daily safety of chain and I will now let Jared conclude with the next steps of the project Thank You Victor I think we were running out of time so I just let's let's just move to the to the questions if you guys have some questions here's our roadmap but you can we're going to publish this anyway on on our paper in the coming days just if you guys have any questions please feel free feel free to ask just before you so I'm going to answer the first question that you still haven't asked but that's going to save us some time we talked a lot about infrastructure and we talked about a lot about multi cloud well I know it's not fully decentralized I know we are working with again major player players like AWS and so on and so forth thing is we are focusing on heavy-duty production environment okay heavy-duty production environment means that we have to be able again to provide a quality standard that needs to - well - we need to make sure that what we provide as a quality standard is is is enough for big applications with massive number of customers to run smoothly and and so far the state of the art or the of the of the blockchain doesn't allow allow us to do that we hope that in the coming years it's going to come and we are going go step by steps towards less and less centralization and and just you know by adding some touches like the state channel ID the fact that you can split between server providers and so on and so forth you are trying to make our best effort okay questions no it's not quite that it's not quite out I mean you mean well I guess I guess there's no there's no whole solution there alternate if there are some solution that will allow you to use for exactly example distributed computing with people like I exact you know we're talking with them because we've known each other for a couple of months we're neighbors by the way and so we're looking into the opportunities of using part of their decentralized computing power to pump it up into our own infrastructure but the so far there's no global solution allowing you to have both storage computing what we call also what we call also sorry about that I'm missing the word well part of the part of the scalability process that we need for the micro services there's no global solution allowing people to do that so far okay thank you very much guys for your attention [Applause]