 It's okay. Okay. So, this is going to be the last bit of the presentation this year. Are you ready? Just start. You're late. Okay. Hello. My name is Eric Schruell. This is the issue of the repairman. We're going to take you through some open source edge. She says at scale. She says if you would be able to click her. What are we going to do today? We're going to show you a little bit of a back story about how we, in the last four years, bringing to you insight, use of what's going on with customers at scale and the various. So, this is a little bit of a background of why this became an issue. We've all been in IT for a while. Let me start you with a little bit. Anybody here build a house? We've all run into architects. Okay. Looks really pretty. We have these really specific lines with the designing software. What he thinks when they design these things, this isn't going to come out exactly. I measured it to the center view or to the inch. What I'm going to do, right? It's very easy to send it to somebody and build it. But what you end up getting is different interpretations. And I know this is really crazy. Think about this. Someone built this. They did this. I'm not pointing fingers. But I mean, it's just crazy, right? You're like, how did you get this out of that? But I mean, how many people have done software architecture? I've done functional designs that I hand off. Sub-day or sub-part of a project to somebody. I'm like, how can you best this up? Friday they come back with something that I didn't ask for. Who hasn't seen that? So what we'd like to do is we'd like to take people that have done things really, really well and be able to expose it to them so that we can all build stuff like this. If you map this out correctly or at least give you some kind of insights, we're all smart enough. We've been long enough that we're able to come up with some kind of architectural designs and plans. But I'm just looking for the hints, the tips, and the tricks that people have added us. Tie this together and scale. It's just a big problem. How can we get this done? So what we did is we stepped back and we thought there are several problems in this space in the earth and organizations trying to share. Public references. They're hard to get along with. Often it's very light, fluffy inside. Really detailed stuff that we do when we go. We do a customer. Those are the things that we're interested in. That's sort of an information flood that you don't do on that. What we decided to do was to step back to customer research. Two to three is a certain use case or a topic. We can sit down and try to find three of these repositories that are open. That's the information that we have. I do not want to map exactly. What I want to do is find the common similarities. To narrify that, present that in a form that's edible for us to do something in our own. So, came up with a diagram tool. You can see a slide in the middle. There's going to be three different levels of diagramming plus architecture came later. It doesn't even have a diagram tool. It's just how do I talk to people that have no understanding of the stories that we put together behind this. It's so interesting. We used to enable the presentations around this. We tied together a short public-facing story. That's going on inside. The original targets were our special architects. Often looking for something like this that we would like to talk to someone. This is as close as I can get you to. That's kind of the point. As we do solution brings, we'll be together with the publications. So, solution bring is just a type of form of that. So, you get one or two pages. If we have any internal-facing demos, we'll tie those in. You can imagine these use cases are rather broad until they bring software. So, one demo is kind of hard to feed. You'll see stuff attached like that. The architecture, which is plenty of examples. See on the left there, a logical view. First thing you notice, this doesn't look like sequence diagrams. Toe gaffes. That is just a flood of information that requires an understanding of my side of the world. It's very complex. These things we decided to step back and say, hey, let's take three, four colors, see some branded icons, and some templated size and setups. Logically grouped. What I found in this article. Listen. You can complicate the story or simplify the story, and you should be able to demonstrate. These are much more specific. This is a schematic, a physical diagram. This is pretty much what we look at from, like, suggesting something on the front side. You're lucky like this, I actually numbered this one because it has to do a good one. That's a really fun one where you get these on the floor. Say I want to zoom down in on this thing. That's what this is. So imagine you're talking to some audience or something. Really want to zoom down in on anything. Sort of talking in detail of what's happening. In area high level stuff, who I talk to or talks to me. I'm trying to force you into my way of doing things. It's telling you how to do things. You start taking through a few of these. Any idea? Okay. So we're going to be talking about, you know, edge use cases and their number. I want to make sure that we are setting that up. Why? Just based on what's driving them is some of the challenges they have. This is a change from anything and thanks to our work. And then if we summarize computing, all the hypotheticals come in. It's another iteration where data is going from edge side to edge side. So something really off in the oil and gas. Play. A very little thing of some sort. Cellular in the autonomous field. Versus future. That's not great for the legality of sending the data back. We're talking about a lot of data. There are engines. Engine can generate autonomous data and data. So all this data cannot be sent to the center. Some way of seeing what's important, what's not important in some processing of the edge, then there should be some things wise and right. A very good requirement. So the large number of euro of data in the global side can be sent to the out-of-country, out-of-Indian department. A lot of these requirements are financial, electrical, government data, proprietary data. And then finally is distribution. You can store and all of a sudden connect to the cloud. You still want the ecosystem that's functioning in the cloud. A lot of these needs are driving people to move to the edge. And people have, it's not any concept, right? People have been doing this. I think what's happening now is there is data computing on the edge and even for businesses are saying that they, I think that's what is important is where the edge should be. So edge is no simple edge. I mean the idea is that there are different tiers of edge. So for example, starting from the right-hand side and that's, you know, four of this circle of money right now is the center. As you roll out towards left, for telecommunications providers, their edge is instead of my COVID-19 and trying to put it in the city. And what's on that is the edge of that. It could be really hard, right? So we don't talk about them over here. You know, ours is over everywhere. That's the last one. And for its customers, their edge starts with factory or office. And then changes on what kind of business you want. The edge of the extreme left is devices of sensors that we spread all over. And just to make a point here, a lot of examples we see are for, you know, hopefully for, you know, our cell phones, our smart projects, our system, which are looking for the edge, but they are more important. Just because they can be able to decide not on the edge. So, this is how the example for edge we talked about when the edge would be different for different industries. And here you see that, you know, this is not something that will happen. They have been doing that. They are exaggerating why they are talking first. They did talk about, you know, how our cell phones are vendors and office agents. They are doing it by G. They are doing it against the platforms. They are doing it on an industrial scale because it's manufacturing. So, they will be talking about how do I do victimizing so that instead of being seen coming out, then my cell is coming out and I figure out something is going to happen in the window of maintenance and fix it. So, this is the same thing on energy side. So, oil and gas. There is a stationary of gas coming from on how oil and gas are meant to be. It's really for upstream to be able to manage and really decide if it's going to be here to remember as we look at different cell phones. It's up here to be just in the manufacturing, oil gas. One thing we want to remember is that how do we, like any other government want to extract out so that every implementation out. So, the common elements will cross over the season. So, in this case, we want common elements to be able to standard that from that extent irrespective of whether it's a health care or it's a health care. So, the idea here is very simple to, I think, sort of at the bottom of the data center is your which conference to be able to talk about that this will be of the oil and gas and the patient will be able to manage it in the same way you are managing it. So, these are the cooling, but we want a similar development in the talk of the tax reform. We have the same template that Eric talked about starting with an architecture slightly new, very high-level view. So, here we have a physical example of a task model that is a one of the biggest gas pipeline that gas-friendly stations from the distributed flow, one of the things that was They had to modify it, they had to do their own operations. They wanted to think, I have a service as an architecture, for real time, screen space, even legal automation, they wanted to do the same. So what they did is this was, they deployed a new edge, they were still not open to it, the work of edge location was still, it usually came from, and this was either using a cluster or through the building, they were doing, they were micro-services, deploying the cluster, they were doing it on the other side. That was an example, they were doing management, data, and one of their things was, if it needs more energy, that's not an element of their demand, I should be able to control the supply from where the supply used to be. It's a reminder that now I should be able to change my supply based on the virtual value of that. That is what we talked about. So on the right, you have your centralized practices, your data center, public cloud location, they have their spread across. And all of this is done on a common virtualization background. And they are running on a... On the story behind this, what you're looking at here is a very simple intro. This is a text here, something they got added about it, they said, why are the text aren't so important if we've got enough feedback? Let me jump up to the next example. This one is industry-med, and I think it's good in the audience and Japanese, it's a very different one. It's the same, and we have factory site, processing of data, we have a dashboard that shows what's happening on the factory. They have a data application that they put normally in the production data, and they have a computer, and then they have the data actually coming from sensor on the assembly floor. Data is coming in, and the data is sent back to the center processing, all of the events, what they're trying to drive with this, obviously. Data, real time. Inside of this data, you can pass, you know, do it in an automated manner, particularly in the state, and doing this, there's any update to the application from the center side, what is the update, this is the product, the factory, it's not a typical edge location, but it has a factory data, so it's fairly easy for this. We have line data server, that's connected to sensors, you have firewalls, connected data, all of the data, and then on the center side, you have all the document storage, data modeling. So when you see this, you think, oh my, what am I going to do with this? Everything he's talking about, so we're going to have an architectural center, you have to find it. Actually, there is, so if you're interested in how the object, how what's connected, there's many kinds of it, showing that the data is coming in, and there are a couple of things. How do you explain from edge side to the center side? The model is being updated, how does it get pushed to the repository, and then to the generator, how does it get pushed to the repository, and all of that stuff. It shows the end result. This one is medical diagnosis, what can I say, so there are many, so you have image processing happening on the edge side, like this case is possible, and you are taking more x-ray, and then you are taking more x-ray to the next thing, and then you are taking more x-ray, and you are taking more x-ray to the next thing, or if it's scheduled, it may not be possible. Thousands of these X-ray will be taken, there's time lag, even if the data is not in, the actual... And the use case behind this was a pneumonia detection, so one of those. This is where the developers are seeing. This is where the... What I've been noticing is that all of these sites, I mean, I've been talking to you so far, very common way of doing things, you have the satellite site that has common elements, then you have the X-site that is just that what kind of data you're purchasing varies from the X-site to the X-site. Then you have a satellite phone and you try to make a look for something to the browser. If this doesn't happen for the test provider, is that you are ready to make a name, they are upgrade, having a couple of viewers, that is both of them. It's a great message stand while they are going to get a new company, and you are running down an application, what wants to offer you? Sub-script 5G, they want to make a party strategy. Same with off-the-shelf server, have a layer of abstraction of software and then software applications running off in different environments. In this case, all these capabilities are software, so in the other product, they want to make a new company, they want to create a part, right? This is a program that will distribute your name, and then again, this is the all this traffic stadium, right? They serve these capabilities while they are operating these other software applications. This is the breakdown before you get any idea. It's actually getting the greatest experience, sometimes you may have to search for them, they are basically getting components from one of these software. For example, Verizon is an example there, they have to do a starting product, but still they are not able to do it, so they have got these three hours on the any changes and that's why the factory is down in three hours, so now what's happening is that they're great, but it's not very quick, it's on hold, right? The risk of algorithm is still bad, but they are not able to find this at that time moment. So it is the beginning of three hours.