 Good morning, good afternoon, or good evening, wherever you are. So welcome to the session at KubeCon. I am Garmer Creos. I am the HSE lead. The HSE is the health service executive in Ireland that the organisation tasked with both fighting COVID-19, but also delivering health services to every citizen in Ireland. And I'm the lead for the COVID tracker app, which is our digital contact tracing solution. I'm also chair of the Linux Foundation Public Health Advisory Committee. So this is a story that happened that started over a year ago for me. It's a story about COVID-19. I think we all know what this is. At that time, so back in March of last year, we could see, so it was like glimpsing into the future, we could see what was happening in China, we could see what was happening in, in Italy, we could, and we're now seeing what's happening still. And it's, it's awful to watch what's happening in India at the moment. But we're seeing health systems come under huge intense demand and dealing with the capacity. So you've got to find a capacity when you're faced with something like COVID-19. And a challenge like that, it's how do you deal with something like that? So at that time, I was working in the health service executive. So I work as a digital advisor for the CIO, working on eHealth projects and arranged different things like that across the board. And but my, but my background is I'm an architect. So we got thrown into this, it was myself, another guy called Tim Willoughby and another guy called Owen Harris. So we're kind of at the heart of this. But it was developing, or trying to understand where we choose technology to, to help the fight against COVID-19, in particular with the area of contact tracing. So cast your minds back to this time last year, we didn't know about apps, we didn't know about Bluetooth based detection of contacts, no, that stuff existed. It's all new, but the technology exists and the ability to do it in the proof concept, but certainly nothing that was kind of shovel ready to go. So the graph that you're seeing at the moment is the graph of cases from March of last year up to July, which is kind of the period that we were working in. So it was a very intense, pretty short period. And I think the reason I use it is because it sets the context for just the intensity, because you could see this mounting, mounting, mounting. And we actually thought that if we didn't get something done really quickly, that this would be over. And it would be pointless doing the things that we were doing. Now it turns out that we were wrong, that this has gone on an awful lot longer than most people expected. I think it's, it's met my expectations anyway. So the challenge that we were set ultimately was once we defined what the problem was, was trying to find two people who were too close for too long, where one of them had tested positive for COVID-19. So we're subject to guidelines from the European Center for Disease Control. They had to find two closest to meters and too long as 15 minutes. This is obviously all being thrown out the window with new variants, B117, which seems to have no respect for these guidelines that were given out by the European Center for Disease Control and seems to want to go at further distances and shorter timeframes. But at that time, that was what we were working towards. So Neil and Fred, if Neil has COVID and Fred is too close, we want to identify Fred and then get Fred tested to figure out what's going on. So the search begins. And again, as I said, we didn't know that Google app exposure notification system, none of that existed at this point in time. So it began a search around what technologies could we use. And this was, wasn't just kind of just a desktop thing. So we were actively experimenting, gathering data, going out and testing things, trying it. So the first incarnation of the app was a big data-based kind of geo-timebox piece where we were trying to put two people in the same timebox, in the same place, and trying to identify if that worked. Obviously, there's limitations with GPS in terms of what you can do with that. We looked at agent fires. We looked at kind of other location-based technologies. We looked at cold data records from telcos. We looked at social media data. We looked at beacons. We looked at getting people to scan in with QR codes, et cetera, et cetera, et cetera. So it was a, and it was a big team. And we're all running in parallel, trying to coordinate and trying to run as fast as we could. So this is the team. And it was a multi-agency. It wasn't just the HSE. So the team included people from the Department of Health, people from the HSE, people from, so central governments, the guy in the middle of the top there is the CIO for Irish government. We also had Tim Willoughby from the police force, Andrew Sullivan from the police force. We had the Defense Forces involved. We had Central Statistics Office, Organ and Survey. You name it. Anybody who could help did help in this effort. And it all came together very quickly. Most of us had never worked together before. So for anybody who's been in that situation, the usual forming, storming, norming happened at an incredible rate as people were trying to figure out how do they help each other and how do they work together. So also what was happening is around the world, everybody else was trying to solve a similar problem. And Singapore were ahead of the game because they could see this happening. They'd been through situations like this before. They'd activated their, their digital government team and a guy called Jason Bay led a team that released Trace Together, which was the first, and I guess it's the first Bluetooth proximity based contact detection system is born. So it comes up and it's using Bluetooth as an estimation for distance. And then I'm recording time. So have I seen this phone? Have I seen this phone again? Have I seen this phone enough times within a 15 minute period to kind of look to call it a valid close contact? And for anybody who's familiar with the Bluetooth stack, that's not a great, it's, it's not specifically designed for something like that. So a lot of the work that's been going on has really been in terms of trying to understand, understand what it can do, what it can't do, its abilities in different phones, if you've got different antenna rays in different phones, there's different signal strengths coming off it. So the transmission power comes into play. So there's a bunch of different things to do with Bluetooth that we, which I wasn't very familiar with beforehand, but I kind of speed pretty quickly to try and solve some of these problems. Another piece that happened at that time as well was this debate was raging around centralization versus decentralization. So this is the idea that am I collecting information about the people you've met? Or am I collecting information about you and making that available to the people you've met should, should one or either of you test positive? And it became this good versus evil debate. But actually, if you look at how public health works and how contact tracing works, contact tracing works by developing graphs and depth of people. So it's knowing who the index patient is, who's the person who's tested positive, and tracing it through to other people in that chain and trying to follow that around. If you try and have a conversation with heads of public health, which I did, and explain to them how we're going to do this and not have an index patient, there is this baffled look as they've never considered a world without an index patient, because it's not the way they think. However, there are concerns around if I'm developing this massive graph of every connection with every person, then what are the impacts of that? And what does that mean? Because obviously, it has privacy implications as well, and a breach in a system like that would be correct. So this whole thing, this whole episode was a series of trade offs, a series of proofs. And again, from from operating inside of government, a lot of things, some things take time, like government is actually quite good in situations in a crisis in terms of responding quickly. But usually the default is we'll take our time, we'll do the analysis, we'll work through this. And seeing is believing. So our first working prototypes were within a couple of days of us starting. So we were showing people the whole way through and just making sure and bringing them on that journey and building confidence that we could actually do this, because it was a novel technology in a novel space under incredible pressure. So how do we do something like that? And the critics define the meme space. So the meme space is kind of the stories that were kind of going around in the media at that time, the screamers, the superstars, the concerned idealists, the uber experts, they were all there. And this was going to be the end of days, as far as some people were concerned. The key thing that came out was the right demands for transparency. And we had committed very early on to open source this. At the start, it, I guess, it, it was in our heads to do it. But we were kind of working so fast that it was kind of an it basically became an afterthought. So we were, we were kind of said, we'll, we'll get to that piece, but actually open sourcing the technology and making it available and making it visible and sharing the data protection impact assessment, doing all of these things led to an increased transparency and that drove trust. And I think trust is one of the underpinning things in terms of success of these apps. So we hit a point where we could not get Bluetooth working in the background. It just didn't work, didn't work reliably on Apple. There was challenges on Android as well. And you can kind of either face into it and keep fighting and keep fighting, fighting, fighting, or you can decide you're going to take a different path and see if you can do it in a different way. And we were constantly, so all of the options that we were exploring earlier, we kept all of those options on the board. It wasn't like we doubled down. We were kind of committed to building like a Bluetooth based conflict tracing app, but we were also keeping other options open and making sure that we had that optionality as part of our landscape for us. So Google and Apple then stepped in. This was kind of April of last year and announced that they were releasing this Google, Google, Google, Google, Apple exposure notification system. Now, that had two things for us. One is it solved our problem in that they would solve the technology problems, which meant we had ultimately had to rip out about 10,000 lines of code. The other side of this was that we actually went from centralized to decentralized, which pleased the privacy people. But at the same time, we then had to trade off with the public health people to explain to them that you're losing an index patient. You will no longer have that connection, but hard, strong connection between the two people, one who infected another person. Then understanding how that works inside of a contact tracing system that's designed around index patients, that becomes a challenge as well. So we were reeling this really fast. Amazon were helping us. We had a dev team called NearForm who were like building the tech. They're a major contributor into Node. They work in these kind of crazy and sane projects all the time. So we were kind of working with them testing from Xplio. We had security and privacy values from a company called ISAS plus all this government help working in different directions. But Amazon were there as well. So we were working with Amazon to help define this and get there as quickly as we could. So we committed early on to use native services. And that was a commitment at that point in time. And it was really driven by the time constraints and just making sure that we should make this thing work day one out of the box. And we subsequently built an interoperability architecture which allowed us to so we were the first country to deliver interoperability anywhere globally. And that was between us and Northern Ireland. So we put a shared border on the island and we want to make sure that people can travel over and back across that border with ease. And that was in terms of opening up movements and travel. And then subsequently we were part of the first group, the first wave as we were opening up and interoperating with Europe. So that then became another program as we went through it. But meanwhile, in other news, so we weren't the only show in town. So I think we've seen probably more transformation and more, more incredible work gone on inside of government in this country and governments around the world in the past 12 months. Then it's just been incredible in terms of the pace of change. But COVID-19 was just impacting all over the place. And that's where Kubernetes came into the rescue. So it's underpinning clinical information systems. It's underpinning and kind of future, future pieces around how we do healthcare integration across the sector. So there's a lot of work there that it's really been pivotal in terms of building reliable, robust infrastructures underneath that can cater with the load. And it's a lot of systems that really weren't designed to kind of work in this way or scale to the to the demands that they were kind of facing. And COVID really beat the crop out of a lot of things. And so technologies like Kubernetes and some of the other technologies that we're looking at from CNCF are going to play a pivotal part of how we kind of rebuild and reinvest in infrastructure across the board. And I don't think it's not just health. I think it's more broadly across government. So 110 days later. So we had the app in the app store after 100 days, like ready to go in the app store and the play store. We were then without a government because we had an election around this year. We're only around last year. It all the time gets compressed during COVID. So we had this, we had an election that had been run COVID hit. We got stuck. It was determined that this record cabinet decision to go live because it was seen as putting health in everybody's pocket. So it was a big decision. So this was the first decision that this new government that we have now made was to go live with this app. So we went live on the 7th of July. It's a fully integrated content tracing solution. So it integrates into the core manual contact tracing piece that's there. And so it's an end to end. So it's seamless. It's not just a piece of technology that sits on its own and does its own thing. It's integrated into the entire system. So 7th of July, we went live today. So we've got right now we got 1.3 million people on the platform. We've had 15,000 people who've tested positive and uploaded their diagnosis keys. We've notified over 30,000 people. Most of them are contactable. But we have a symptom checking feature in the app and we've had 60 million check-ins on the symptom checker. We're interoperable with Europe. We're interoperable with Northern Ireland. 10 plus releases at this point. But the code, we contributed the code into the Linux Foundation Public Health project. And it's now being used in whole or in part in 10, 10 implementations and there's more coming. So it's working in that it's doing its job. It's doing what it's meant to do. Could it work better? Absolutely. But as we learn about this technology, we're constantly tweaking and tuning and understanding as our Google and Apple. So we're kind of working on this to try and understand how it works better. So I think you're never going to please anybody. But one of the things that happened was that the approach we had taken the open sourcing and the transparency around what we were doing actually won over a lot of privacy. People who would typically be critical of these kind of moves were actually won over in terms of this is a model for how we do things like this into the future. And next steps, well it's a global impacting. So as I said, it's in 10 different implementations. We gave it into the Linux Foundation Public Health project as a contribution in there. We're actively participating in that group with other public health authorities trying to figure out different ways, not just to solve problems of COVID-19, but also to look at preparing for a future. So what happens post pandemic response toolkits? What do they look like? And then other broader public health issues that we face. So I guess for many of you and I know this is an audience from around the world, this will mean nothing but there's a satirical newspaper in Ireland called Water For Whispers that is usually skewering anything that's done by government, anything that's done by public figures, anything that's done by celebrities usually get skewered in the front of it. When this headline came out, and that was August of last year, or yeah, it was actually August when it came out, but it was to do with July. And there had been so much chatter on social media that it'd been so much chatter around the world about how this whole thing would work. The relief that we had, and I know it seems like a very small win, but when you're not being skewered on the front of it, when you're not the subject of the joke then and somebody else is but you're being used as part of that piece. There was a certain sense of relief, a little bit of Shab and Freud, if I'm honest. So this is this is me. If you want to find out a bit of green, the link to the to the open source repository is there. Welcome anybody who wants to ask me questions. Welcome anybody who wants to come in and join community as well and help because I think what we're trying to do with the Linux Foundation Public Health Project is really bring some of this stuff together and bring some of the the enthusiasm, the minds and that that open source community thing, which is for Ireland, it was one of the first real experiences in terms of how this can actually help us solve some of the major problems that we've got in the world. Thank you all for that. And I hope that was worthwhile. Take care.