 All right, good morning. Sorry again that I was running a little late. And let me get rid of this. I've got the undersized. I do not comply here with my power supply. So I'm going to go pretty quick, since I was a little late. I'm going to spend time talking about why it matters how you're developing for IoT. Open source, obviously very important period, extremely important in this space. This is going to be a little bit of a different story, I think, than you've normally heard. But I talk about the holy grail of digital, you know, get to that. But a lot of people are just hacking stuff together in this market and kind of missing the point on what the real potential is. Actually, I would argue that most people are doing it wrong. And so kind of walk through why. I'm going to go through quick if you have any interest in this story or about cats, you know, why here in a second. Follow me on Twitter, but there's a whole blog series, like one of the last posts comes out today. So you'll get a lot more if you go there, but I'll go quick. But then we'll talk about EdgeX and some other open source projects that we're working on. So as we go first buzzword of the day, okay, digital transformation. This matters. And I'll talk more why. So I know you guys are all kind of a developer crowd. How many sort of a developer versus end user developers? Yes, typically you'll find those, you know, open source and end users. Anyone? Okay, yeah, developer. Yes. Okay, so I'm going to, you know, a lot of really great talks, you know, here, but I'm going to spend a little time on sort of the business side, the why you should always ask why, because you're going to develop something that's got to be useful and it's got to goes places. This really matters. The pace of innovation is what matters in these days. It's all about outcomes. It used to be, you know, kind of the widget and whatnot. Now it's about the outcome. I mean, you look at, you know, power by the hour was coined in the early 80s. There's things like managed print services that Xerox did a while back. Nest was like the first IoT success story, you know, out there. They kind of did the outcome of improved comfort while saving power. But guess what? Nest didn't follow the first rule. They didn't stay innovative. You know, when Google bought them, they kind of fell behind the wayside and then all of a sudden a lot of different thermosets came along and they fell kind of back a little bit. There's just many examples, you know, out there now where whenever you're developing something, it's about kind of that outcome in the end. It's also about a new mindset. So it used to be when I develop a product, you know, I put it out there and that's the experience and if you want a new experience, we'll buy a new product. Okay. You know, this is like stereo system. That was your experience when you bought a stereo until you buy a new stereo many years ago. Now, of course, everything's software to find them. This is, you guys know this. Sonos is a great example of a company that kind of follows this ecosystem model. And so, you know, but even they are getting challenged by, you know, other types of products like I mentioned with Nest. But if you look at the new mindset, there's kind of a couple key things here. It's number one, when you develop any kind of software codes, product, service, whatever, it is about the cumulative value of that product or service over the lifespan of that product, not when you ship it. That's that's like kind of rule number one these days, all that stuff that you see kind of software to find everything over the air updates and whatnot. Number two, it's like being an ecosystem. Either you join one or you make one, but if you're not in one, you will die. Period. So this is the way the way it is going forward. We'll talk more about that. And finally, it's about changing, you know, capital expense into OPEX and, you know, making, you know, services like your phones have and whatnot. So this is how you kind of create that necessary stickiness and the network effect related to that and the ecosystem is extremely, extremely important. So always when you're developing stuff, kind of think about that sort of network effect and ecosystem effect. I love this story. There's a lot of stuff about this online, but maybe one of these companies doesn't look quite like the others. You know, I don't know. I mean, Domino's was named the most innovative company of 2017, one of the list and it's a pizza company and it's because of digital. I mean, there's so many different ways to order pizza online. You know, you can tweet a pizza emoji and get a pizza. There's all kinds of stuff. They realize the power of their data. So, I mean, they're doing autonomous vehicles in Detroit with Ford to deliver pizza with an oven built in or baked in is a bad pun I guess. It's crazy. First off, it makes me annoyed that I invested in these stocks, any of these stocks back then. But if you think digital transformation is a buzzword is Huey, that's a pizza company. Okay, so you guys know about cloud native. You guys know about the value of continuous delivery, all that stuff. I don't need to tell this crowd. I think about the value of loosely a couple of microservices. No, do I need to say this? Now, the big thing I will say, so cloud native architecture, it is not about where it's run. It's more about the principles of loosely a couple of microservices, platform independence and all that. I can update any single function without taking down the whole system. So, it's so important. So, come back to that, you know, where it's run. You might have engaged by the title. Ring a bell? Maybe you guys are not old enough? I don't know. AOL came along at a time when a lot of people thought that the internet was like just this magical thing. I didn't know it had been around for a long time in government universities and whatnot. You get these little CDs in the mail and you get on. They dumbed it down to make it easy to get on and start forming an ecosystem. Now, that was kind of what they did. And then all of a sudden, people started hearing about this weird word, Google. And like, you can go to this page. It's just like a logo and you type in and magical things come back. And people literally started realizing I can actually connect to the internet without going through AOL. Oh, wow. So, then all these websites come up and these tiny little trends of like e-commerce and social and mobile all kind of went on you little stuff. The thing about IoT, AI say we're in the AOL stage of IoT kind of early on. But the thing is that a lot of people are like, oh, it's cheap sensors and ubiquitous connectivity in the rise of the cloud. Okay. Sure. Right. Okay. It is as much about the maker movement as anything. In all the industries out there that are deploying IoT, whether you're doing consumer industrial, whatever, it used to be that as soon as some small company that's innovative threatens your way of life, your proprietary lock people in fat service contract way of life, you would buy them and you'd kill them. That's how you maintain status quo. With the maker movement, there's not developers like yourselves that are like innovating and coding and all that. And I mean hardware software too, or hardware like, there was not enough money in the world to buy up all the ankle biters. So either you change or you die. This is the classic innovators dilemma. If I just do more of what I've always done, I will succeed. This goes back to that pace of innovation. It's extremely important. But I still, I mean, we're getting more traction in IoT. It's early, but I like to say it's AOL stage for a number of reasons. But how many of you guys work on consumer stuff versus enterprise consumer? A little bit. Okay. Enterprise industrial. Okay. So, okay. So you guys, this is the preeminent conference diagram for IoT. You cannot go to an IoT conference without like an OT, IT, then diagram. I think you have to have it laminated on your badge. But it's important. So OT folks, operations folks, people that live in the physical world do processes. This is mission critical stuff. They care about uptime and reliability and safety and quality and all that. And IT of course cares about security, privacy, governance, scale, massive scale all over the place. The OT people historically, I mean, they've been running processes in the physical world for a long time. I mean, IoT is basically about this type of stuff hitting scale. Yes, partly because it's more accessible, partly because of the maker movement. So embedded computing hitting scale is IoT. Of course, I'm adding more sensor driven analytics. Usually it starts with kind of monitoring and then I add some analytics and then I automate and maybe do more true AI, but you always start with monitoring. The OT folks historically have been like, don't even touch it if the process is running. Like literally you could drag me away from like my process with my fingernails and the concrete from Windows 95, you know, if it's running. Security by obscurity, like you don't connect to broader networks because that introduces threats. In OT, if you have a security breach, immediate loss of life or limb or production, I mean $20,000 a minute it could be. In IT world, a breach and you will be breached as a matter of when and how you handle it. In the IT world, a breach plays out over long periods of time. It has a very long tail. This is like credit card fraud, you know, years and years of impact. This is because of the scale. It's important that these groups work together. I like to say IoT starts in OT and scales in IT, but this is a very important dynamic. Then there's the LOV, the line of business, and they actually ultimately make a lot of the decisions here. But you guys probably don't, you know, as a developer you might not kind of see some of the business side here. I don't know. But when you develop product though, code, you need to pay attention to the way that you update stuff so that you can't just push an update in the middle of a process. And imagine an OT person seeing a pop-up that says, hey, sorry, your manufacturing line will reboot in 15 minutes, save your work. You can't do that. So it's very different. So it's just something to consider. Got a lot of sayings you're going to pick up on. A lot of the market, even though we're getting traction, and because of this dynamic, we're seeing what I call pie in the sky mode. Take a Raspberry Pi class device and I hook it up to some public cloud and I just get going. A lot of the time it's the OT person on the shop floor, they know their processes, they have the domain knowledge, they're doing shadow IT, completely bypassing IT, and they just get going and, you know, innovating. But they don't really think about what they're kind of getting into. And so before they know it, once you really get going, also call a pie in the sky mode, because there's usually no business case that's just tinkering, you know, pie in the sky ideas. But once they really get going, they start realizing the cost of pumping data mindlessly to some cloud and then having to pay to get your own data back, ouch, you know, and then you have to rearchitect. So it's important as a developer to think about for end users, how you're developing software, and this is why we're going to be talking about how cloud native needs to be in more places and I'll explain why. You guys have heard about edge computing, it's like, you know, three, four years ago when I started working on this at Dell Technologies, it was like, I say everyone had their head in the cloud, everyone talks about edge, the cloud's not going away, the deepest of deep learning will always happen there, but you're going to see more and more stuff at the edge. And so, a lot of people say it's because of a variety of reasons, I mean, billions of, 50 billion devices, I don't really care, it's just so many devices on networks, you can't possibly send everything to the cloud. And so I like to use cat videos to explain why. So now you might understand why my shirt. So my wife and I have three cats. There's no shame in this game. We got larger capacity storage on our phones so that we can send cat videos back and forth. If you follow my Twitter, you'll see occasional cat fly by. But cat videos are used to explain the need for edge computing. So if I post one of my videos online and it starts to kind of get hits, well I have to cash it on more servers, way back in the cloud. If it goes viral, then I have to move that content as close to the subscribers that I can get it to. As a telco, the closest I can cast stuff, or Netflix or whatever, is at the bottom of, it's at the cloud edge, the bottom of my cell tower is the, you know, these points on the key points on the internet. This, this is the concept of MEc, multi-axis edge computing. So bringing content closer to subscribers. Well now if I have like billions of connected cat callers out there, I've completely flipped the paradigm and I've got instead of things trying to pull down, I've got all these devices trying to push up. That makes you have to push the compute even further down. So, you know, very, very important to kind of be thinking about these tiered models as we go. So yes, everyone says, okay you need edge computing because of latency, I don't care how faster airbag is, you do not deploy your airbag from the cloud. 5G, people are like literally saying, oh you're going to drive your car from the cloud, you're an idiot. You're not going to drive, I don't care how faster network is, you do not do that kind of stuff over wide area, you do local control. Bandwidth, it costs a lot of money to move data, you know, so you need to kind of consider what type of meaningful your data you're sending back. Security, a lot of people talk about security and IoT, it's important, but if you have smarts closer to things plus all that legacy stuff that was never connected, intended to be connected to the internet, you need smarts there for identity access management, root of trust, threat detection, all that kind of stuff. The big breach that happened a couple years ago or a year and a half ago when all those cameras took down the net, I mean first off the problem was the developers made it very easy to not change passwords. It's like password, password, you know, okay instant gratification, I want to connect up my camera. Had you had some smart, oh then the bots you took over and you took down the net, DDoS attack, had you had some smarts there, I would have been, wait, I mean this traffic is a little weird, shut it down at the source. You need intelligence, but the kicker is the total lifecycle cost of data. When you start mindlessly pumping data to the cloud and you're when you're in pie in the sky mode, you're going to quickly realize if you start doing analytics on that data, you are going to really pay for it in terms of the cost. So a lot of people out there aren't talking about that life cycle cost, they're talking about latency, security and bandwidth, which is definitely part of it. Okay, but there's many edges, there's not a single edge. Edges, lingo, bingo, whatever like, you know, to a telco, the bottom of the cell tower is the closest that I can move, you know, compute as I said for my cat videos. You might have noticed it used to be MEC used to be mobile edge computing, now it's multi-axis edge computing, they change the name with the same acronym because now it's not just mobile devices, it's things too. You can have micromodulate data centers, you can have hyperconverged infrastructure, you know, kind of gateway type nodes like on a factory floor and then to an OT person, the edge is the physical world device, the device edge as we call it. There's many edges, there's never going to be one single one, you want to be able to run across this regardless of whether you're OT, IT. The way I describe edge computing, it is moving compute as both necessary and feasible to the subscribers that need it. So if I'm a mobile user, the closest and feasible telco can move it is the bottom of the towers, you know, that might baseband units. If I'm an OT person, the most necessary place to move edge compute is on-prem, like on the same land as the stuff that's needing it. So this is how I describe it. Got to be thinking about how you dynamically, when you develop software, how you dynamically move workloads, microservices across this spectrum while balancing the needs between OT and IT. It is crazy town out there in terms of fragmentation. Are you guys actively developing FIO-T right now? Show of hands? Yes? Yep? Okay. Cool. So if you've done it and especially if you've done it in industrial, you'll know that there are thousands of protocols out there. Remember how I mentioned people would, you know, kill off companies to kind of maintain their stale business models? People would create proprietary protocols to lock you in so your switching costs are really high. So many protocols, of course, in the IT world, there's tens that matter in the OT world. There's literally a thousand plus protocols. I mean, you count for proprietary ones. We'll never have one standard. The old standards show because we'll fix the standards problem with one more standard. No, you need to bring them together. Domain expertise, it takes a village. You cannot tell developers that you must only program in Java. You know, it doesn't work. You need polyglot if you want to build an ecosystem. And then in embedded, the closer you get to the edge, the more and more fragmented choices get. Not only hardware, more and more custom, the closer you get to the physical world, but also software. Software gets more and more complex. Now, the software gets a little, the curve of the complexity of software increasing and the closer you get to the device edge is a little flatter. And then it goes up steeply when you hit controllers. Hardware starts going up and then eventually you're at parity. So software, you can software define things more readily closer and closer to the edge while also recognizing you still need embedded hard real-time control systems. But the closer you can get your flexibility to the physical world, the better. Crazy town platform world, running joke was 150, three years ago. Now it's a 450 platform. It's not all the same. It's just not sustainable. There's a lot of reinvention happening right now. I talk with a lot of companies out there. The last thing I would do right now is start up a generic IoT platform. I would focus on value and we'll kind of get into that a bit more. How do I get to advanced class? That's just like getting out of AOL stage. To get to advanced class, I need to do a handle. I need to intelligently handle streaming data. So make decisions in memory in the moment. Is that useful? Is that useful? No. Scrap it. Scrap it. Scrap it. Send that back. Analyze this on the spot and act. This intelligence and streaming data at the edge is important. Otherwise, you end up with like pumping all this data back in. It's like a bad episode of that TV show Hoarders, just stockpiling data that you'll never actually touch. Scale. Once you have some success, that pace of innovation, it breeds more interest for more data and then I've got a scale problem. I've got a lot of people that we work with that are like doing pie in the sky. Command line interface is cool. Managing party of one. Go try to manage a billion managed objects in a connected car scenario. Because every car has got about 108 ECUs and then you multiply it out by a fleet. Try to do that. This is why management really matters. A lot of people don't think about that up front because they're just kind of doing basic stuff and getting started. Then of course, security matters. We'll talk a little bit about that here in a bit. The true potential that was a system of systems. How do I get to this true potential of IoT instead of a bunch of intranets, everything interconnected across an ecosystem. This is the real power of stuff. Energy, manufacturing, farming, all that. Well, put some blockchain on it. Have you ever seen the movie My Big Fract Greek Wedding? The dad sprays blockchain or Windex on everything. I call blockchain one of the Windexes of technology. 5G is also one. A little bit of AI. There's a lot of people saying, oh, I do AI and they're actually doing it. If this then that rolls in, they're like, no, you're not. Anyway, so blockchain, very, very important. It'll change society as we know it. Way, way over hyped. There's a lot of challenges. Read the blog for it. I can't get into all the details. I like to talk about technology in a practical way. It will matter and I'll come back to this. It matters in many ways, but also be careful with the hype. You have to start separating the application plane on top. The insights, the domain knowledge from the infrastructure. Way too many IOT platforms are reinventing the same ingredients over and over and over again for security management tied in with the underlying or with the overarching applications. When's the last time your ERP system managed PCs? You don't do that. It just doesn't make sense. And so you want to have platform independent data ingestion, security and management that feeds into your choice of applications. Most people are doing it completely backwards, but there's more. Okay, so I want to extend cloud-native principles to the edge, so now I can update individual functions without taking down the whole system. This matters to an OT person. Now, continuous software delivery freaks out the typical operations person. You know, they're like, I don't even touch it. But as the innovation happens around you, you're going to have to be ready to start innovating yourself or you will die. So architect today for flexibility, even if you're just going to hard code to one cloud now, if that's the right thing to do, so be developing software today. You've got to be thinking about where it goes. Open source matters. Open source is a period of good way to minimize undifferentiated heavy lifting. You guys know this. You're passionate about this community and whatnot. It's also a great way to build ecosystems because you have tangible code that you can develop, you know, ecosystems around. So lots of good stuff out there. This is a mix of standards and this is not even all of them, but in some open source projects. I'm going to talk about edX. There's, I think there were some things. Have you guys heard about edX? Did you see any of the sessions? Definitely take a look at this project. So this is something that we've been working on with a bunch of great companies because it's basically extending cloud native principles to the edge to create an open ecosystem. Think of edX doing for IoT what Android did for mobile. It's the easiest way to put it even though it's not no less. So it's a loosely coupled microservices architecture. It's a Linux foundation project just like cloud founder. You can't plop cloud foundry down on a Raspberry Pi. It's a little heavy for that. Kubernetes we're working on extending it down and whatnot in the community, but think of it as like you take your pick of protocols because you'll never have one standard protocol. Use this common SDK, pass through these APIs from brief moment in time and then be as custom as you want on the other end too. So it's architected to bring together commercial value add around just enough of an API layer that bridges together this stuff. So it's an OS agnostic, hardware agnostic, protocol agnostic. It's built for cloud native edge. We believe it's really the only way that we're going to get to a true ecosystem to scale this out and then of course be as commercially proprietary as you want. This was a post-it note during our early session when we were collaborating on it says provide a platform that will cure the paralysis of companies not deploying IoT something like that with a fear of making the wrong choice. That's what's happening out there. A lot of customers are like I don't know what to pick. I don't want to get locked into one cloud so I don't want to take a leap of faith off a cliff so I'm going to do nothing. This is not good. So more on why we helped get this started. I'm not going to get into all the technical details. There's a lot of great information online. There's people around. There's actually the TSC technical steering meetings that are completely open. It's open source are running this week so if you want to stop by you can just register go stop by. But imagine this whole thing running on something like a gateway. It could run on anything. The purple bar is that API layer around these core services. If you've ever seen Big Lebowski, I call it the dude. It's like the rug that ties the room together. You can be as proprietary as you want out around it. In fact you could replace every bit of code with proprietary code and still be edge x and point. That's why it's the x. It's going to be a certification program that you follow the baseline APIs. Anyway so more on that later. It's built to be distributed. You can run it on all across distributor architectures. Lots of people coming on board. There are a lot of larger names coming in now. You know a lot of the startups that we've been working with for a while joined early on for a four mention reasons. The big names are like ah I got this all figured out. I'm going to own everything. I'm going to lock people in when now they're realizing man this is hard. There's a lot of mess here. We need to band together on the commonality. We're following and meeting our commitment so the deli release comes out in a couple of weeks. You're going to see I mean the code was initially we prototyped it when we had contributed it initially from Dell. It was like two and a half gigs of memory footprint because it was more about the architecture and cat herding. Cat herding less about the code up front. Now it's gone from two and a half gigs and kind of a Java base. We've redone it as a community in Go and it's 128 megs and half of that as the reference database. So it's compressed quite a bit. It's really starting to take off a lot of stuff coming. So certification program coming more and more security management as we go. Imagine a world where every microservice out there at the edge could advertise how you manage it in a common way. The facto standard way start me stop me reset me here's my quality of service needs. Really cool stuff that can happen when you when you have this kind of capability. You can granularly manage every function in a common way even though they're proprietary individually. Good for distributed deployments. There's even proprietary versions coming that are compressing the entire code base into this is from from from commercial vendors into a C binary. But I still use the API's around the wheel and get value from you know cloud connectors that are plugged in or sensor connectors that are plugged in the bottom those API's. It grows an ecosystem but also allows room for proprietary differentiation. We're also working with Acreno a new project which is telco focused. Acreno is another Linux foundation project. Edge X makes no decisions on the interoperability or like the infrastructure layer because if you do you're going to be wrong to somebody so we're completely protocol agnostic OS agnostic we're agnostic to Kubernetes Docker Mesos whatever. Infrastructure in Acreno is being kind of selected so these projects actually go really well together. So imagine where APIs developed by Acreno in that community are coordinated very tightly with API's and Edge X that's in the application plane. I could literally even though I would like to prioritize the CATS a healthcare app would get more priority over the CAT app in a de facto standard way even though you're building commercial software around it very very cool stuff can happen. Okay so the new world this is about floating boats for scale collaborating and then make sure your boat is really good and really fast. This is how it works period. I mean no more lock-in stuff this is about pace of innovation as I said up front. Three rules for IoT scale you got to decouple the infrastructure from the insight plane you cannot have over and over again different vertically focused platforms you know with the same you know with the different security management features because you'll end up with wildly different things for like multiple different use cases in the same in the same organization. It will never scale it's bad news. You want to decouple the edge from the cloud as close to the physical world as possible. The moment data is created all data is created in the physical world. You want to decouple it so now you know and Edge X is a great way to do this in a cloud native fashion right above the controllers and all that the hard real-time stuff because now you can control your data destiny if you decouple it you know you can send it anywhere on-prem in cloud multi-cloud multi-tenant whatever it's going to be multi-edge multi-cloud world. The clouds are trying to produce what I call IoT gateway drugs literally giving you some dev kit and make it all too easy to get locked into their cloud because they want to get your data and then charge you to get it back. So the clouds are doing great things I'm not picking on them too much but at the same time as a developer think about how for your end users you're giving them that flexibility because we've only scratching the surface but wait there's more oh the last thing domain knowledge you need to have the domain knowledge about industries you know manufacturing energy transportation retail healthcare whatever separate from the technology and then you can pick the best technologies with the right domain knowledge right now the vertically focused platforms are the ones that are actually doing something but then you have like 10 platforms for 10 use cases that's bad separate it out that's kind of a catch all rule but wait you have to get to the real grail and I'll tell you what that is I need trust everywhere and I've got a couple more minutes so I'm going to run through this quick but again you'll read the blogs and whatnot but I want to cross over between public and private domains web and mobile brick and mortar you know retail auto home industrial this is the true potential I cannot have any single entity own the trust and consumer you build trust with particular brands and sometimes the trust is violated but in the in the business world or B2B to home health retail usage based insurance you cannot have one company like an Amazon own the trust it does not work to get there you need technology is help you need silicon based root of trust down at the silicon layer where data is created you need trusted zero touch provisioning so we're working with a bunch of big companies a little small smaller companies that are doing all kinds of cool things intel and arm just announced last week that they're collaborating to help with a trusted provisioning across the supply chain you arm and X86 and arm devices open API is a crane or edge X things open source projects that create that sort of transparency yes you need ledger you know blockchain stuff and a little AI for context awareness and I can get to the Grail and I've had hundreds of conversations with very smart people and nobody has said well that doesn't make sense that this is the Grail I want to sell data resources and services to total strangers I want to create data in the physical world set some terms on it put it off into the ether and clip checks from complete strangers I mean there's people doing this resources is like networking storage compute energy cars any kind of consumable thing now service my domain knowledge you've never met me before but you it's like Angie's list on steroids that's my computer ledger the other stuff you need these open technologies you have to collaborate on open plumbing where you will never have the trust and transparency you guys are probably too young for this reference or overall have you ever seen the movie Mr. Mom it's a great 80s movie go watch it there's a clip in the deck I don't know if this deck will be published whatever but this lady comes out he's trying to drop his kids off for the first time and she's like hi hi Jack I'm Annette you're doing it wrong most of the clouds are are are doing it wrong a lot of people are thinking about the problem wrong in the movie there's a a scene in there that's basically some guys like trying to or you talk with them and put them down and he's trying to Michael Keaton's trying to press the guide and the guy asked him about his old he's rewiring the house what are you going to do you can put in two twenties again two twenty two twenty one whatever it takes it's a classic line is a lot of people that quote on the Google that are like I guess my age I literally check into my room last night my blog on this this topic posted this morning I was in room two twenty one and look at the end of the hallway two twenty two twenty one whatever I kid you not so so I know I went quick you know read read my blogs if you want to kind of get the full story it's a lot of funny moments I got my little sayings or whatever maybe it makes you like cringe but either way if you're interested stop by the edgex technical steering meetings they're literally co-located here type edgex TSE Edinburgh and you'll find it just register maze in the back with some questions from Linux Foundation it's not just about edgex we think it's very important just collaborate and let's go figure this out holy grail three to five years out but right now it's like get people started when they're in pineapple sky mode without getting locked in you need to keep your options open so that's it thank you