 All right, so I'm Barton George. I'm from Dell EMC in the office of the CTO, and I'm going to be talking about the EdgeX Foundry. How many of you have heard of the EdgeX Foundry? Oh, cool. So I'll start from the beginning. Also, the other thing I would ask is, how many people get two choices, right? You're actually working in IoT in your company. The other one is you have some familiarity and you know about it, you just want to learn more. So who's working on IoT in their company? Okay. So good. So I'll do, and then I'm assuming the rest of you are just familiar with it and trying to get a better handle on it. So without further ado, what I'm going to do is I'm going to show you just three slides giving you a, this is IoT as we see it. Then I'll go into the EdgeX Foundry piece, which is here's the problems that it addresses, here's the goals of the project, and then here's the roadmap for going forward. This project has only been around for six months, but it did launch with the most number of member companies of any Linux Foundation project. So with that, away we go. Okay. So IoT is actually, as we've talked about here, a natural evolution of taking operational technology and merging it with information technology. So you take the stuff that has been IT goodness like networking storage servers, and you pair it with a lot of the devices and the sensors that you've had in operational technology for the last 20 or 30 years. In fact, the phrase IoT has been around for 20 years. So a lot of people will say, oh yeah, I know about IoT, it's been around for 20 years, and the reality is it has, but it's only hit a tipping point or become really interesting recently because a lot of the component parts have finally gotten to a place where they together can really make a compelling solution. So I've just picked a bunch of these here. So Cloud 20 years ago was not there. Big data, another thing, because on the back end, you need big data to analyze all of the data, obviously, and I'm going this continuum, I'll go that way. This continuum is going from the back end all the way up to the front end, and I'll show you that in another slide. Then powers become much more efficient. Wireless used to be very spotty only in different places. Now we've got wireless blanketing things. Obviously in the last 20 years, mobile phones, mobile apps have just, well they weren't here, smartphones weren't here 20 years ago, and then the cost and that's talking about the sensors all the way in the front. So RFID tags, I don't know how long they've been around, maybe 10 or more years. I remember those by themselves are going to be something that revolutionizes the world, but it's only become recently when they've dropped in fact 40 percent in cost the last year, and they're now about 10 cents. So obviously as they become cheaper, it becomes much more feasible to put them into various machines and endpoints. Then the last of the setup slides is this is showing you the architecture. So on the back end you have the Cloud, you've got the data, and this is where all the big analysis happens. Then in the middle, you've got the on-premise servers, and then you've got the gateways, which are basically an aggregation point for the device information, and the idea here is that they're actually intelligent. So they're not just routing traffic back and forth, you're actually doing analysis out at this edge of the literally at the edge where you're getting all the information from the devices which can be anything from, as you see light bulbs, delivery vans, air conditioners. There's a company AeroFarm that is actually using it for harvest and for plants. So they monitor the seed from the beginning all the way till it grows up to see how they can better grow it, how they can do different things. There's also a use case we used to talk about a lot, connected cows where you literally have RFID chips in cows to figure out how you can get better milk production out of them. So anyway, what you do is you take that information, you take it to the intelligent gateway, and there's a lot of the data here that's perishable, meaning that it's only valuable for the first couple of seconds. So if you've got a fire alarm going off, you need to be able to act on that right away to, say, close the doors automatically, shut down the elevators at the bottom floor, those types of things. So that information, as I said, needs to be processed up near the front. Now there's also the next level can be processed on-prem, and then most of the stuff for deep machine learning and analytics is back on the back end. Then as you can tell by these arrows, they're bi-directional, so it's not just the information goes there, it's as these, say, the machine learning insights come, they get pushed to the front, and then they alter how things work on the edge. So that's your quick IoT primer, and now let's switch over to EdgeX Foundry, and if you can see here, don't forget to ask me about the octopus. Remember that, see the octopus here. So with that, so what's the state of internet of things that makes something like the EdgeX Foundry needed? And there's a couple of reasons. One is it's inherently heterogeneous. So you've got on the services side, the domain expertise, you really have to know a bunch of random things to be effective in this area. Connectivity, whole bunch of different protocols that are used, and as I'll show you, there's actually, it gets worse as you get to the edge. It's not so bad back as you get closer to the cloud, but on the edge side is particularly bad. Application environments, like a lot of places, developers want to choose those languages that they feel most comfortable with, but as a result, you've got a whole mishmash of those, and then you've got the operating systems. And so even though this is being hosted by the Linux Foundation, there's more than just Linux for operating systems, right? You have Windows, you also have real-time operating systems. So, and then the last problem is fragmented ecosystem. So as you can see, just a ton of different companies in all various different spaces. So this covers both consumer IoT and industrial IoT, so it's the two together. But the reality is, it's the industrial IoT that is taking off first. Commercial became a lot more interesting early on because it's something that we as individuals can understand. Milk in your refrigerator gets low. The RFID tags sends something that tells you to go and buy it, then the store knows how many to stock, et cetera. So there are those types of operations, but where the real value will be, and where they're furthest along, is on the operational side. So what does this all lead to? Well, this is an actual customer sticky note. So when we went and we got a bunch of folks together, both customers and partners, to try and understand where the issues lay, this is what somebody wrote up there, what they want. So provide a platform that will cure the paralysis of companies that are not deploying IoT due to the fear of making the wrong choice. So basically, nobody wants to get in when it's such a mess. So they're gonna wait until things sort of settle down, and as a result, the market is not taking off the way that it could be, even though the technologies are now in place. So with that, oh, and then one more thing, specifically when we talk about where is the problems, where are they the most acute, it's out at the edge. So if you look at this and kind of think of this as the United States in, say the 1800s, back here, the cloud is sort of the established East Coast where people came first and things are pretty much in control. And just like every cowboy movie you've ever seen, as you get out here, it gets pretty lawless. Things are pretty, the wild, wild West, right? So on the back end, you've got standard protocols on the front end, hundreds of them. You've got just IP based on the cloud side, but you've got IP, non-IP based. You have APIs, consistent APIs on the back. You've got them very spotly used. So they're not, they're used here, they're used there, but not uniformly. And then also, as you'd imagine, you could have centralized servers and insecure areas on the back end, whereas a lot of these places out on the edge are often in secure areas. So if you think of things like oil rigs would be an example which are exposed to a lot of weather and the elements and cars that are, or say vans that are driving, they're not, they're places that need to not only withstand the nature, but at the same time, need to be able to send that information back. So now I've laid out the challenge. So where does the EdgeX Foundry come in? So the whole idea was, okay, if it is out on the edge where this is the worst, what can we do to help this? So what we did, and so we at Dell about two years ago started this thing called Project Fuse, idea of fusing things together. We wrote a bunch of code and then realized if we just do it, it's not gonna, it's not gonna be a big deal in the market. So we got together a whole bunch of customers and we got a whole bunch of partners to help out with it. And so as you see here, the vision is to create a common interoperability framework that enables an ecosystem, a plug and play, EdgeX certified components. And I'll talk about the different bits of this, of this vision as we go forward. But the idea is it's vendor neutral open source hosted by the Linux Foundation. As I mentioned, it's only six months old. I talked about the lines of code over 50 members. As I mentioned in the beginning, it was the largest number of customers, sorry, partners that any of the Linux Foundation projects started with and it's based on industry and customer feedback. So if that's the vision and why it was done, then here are the goals of the Foundry. So you're taking to build and promote EdgeX as the unifying IoT Edge via common open platform. So I've got an architectural schematic that's coming up in a couple of slides, but it will show you how this is all put together in this as a gateway operating system, middleware that works to aggregate all of these different protocols and inputs and outputs in a common understandable framework. So that's where these plug and play components come in. Once again, I can show you that in the architectural diagram. These components will then be certified that says the basic idea there would be, it will say that this component doesn't break the APIs. That's in essence what that will say so that people know if they get it, it will be able to plug and play into the overall framework. Tools will have to help work this and then not doing this by ourselves. Even the 50 companies, we wanna work with other projects, other industry standards group, and believe me in the IoT area, there's a whole bunch of different industry consortiums and standards. So here is the architecture diagram. So first of all, and the one thing I should have mentioned is when we, the community were putting this together, we looked a lot to Cloud Foundry to try and take lessons. So if I talk about the name, Cloud Foundry, sorry, EdgeX Foundry is a nod to Cloud Foundry. The idea, we also took the principles in the idea of these loosely coupled microservices connected by a common API. Then the, and then just the X on it, it came because we merged with another project called IoTX, so you get that. So anyway, it's loosely coupled microservices, common API. These microservices are written in any language and then you can deploy them either by containers or VMs and just like it's polyglot, it's also hardware and OS agnostic. So if you look at the diagram, maybe I'll turn sideways here and try to use the pointer. Okay, does that work? Okay, stand over here. So basically the part that everybody is working on is this purple part here. And so this is where the core is, this is where you've got your microservices. As I say, you can run them in containers, you can run them in VMs. And it's the part that acts as the middle or the translation between the southbound and southbound is all the devices and the northbound, which is all of the, where the analysis takes place on the back end. And as I mentioned before, it's this area that becomes the wild, wild west. This is a little bit more organized. So you've got the core services, you've got the device services, which takes the data from the endpoints, puts it in a common data structure and then can take it back to the core. And here's a whole bunch of the protocols that will then be able to feed into it. And when we talk about for the device drivers, sorry, devices, we get it so that they have a common driver format so that you can plug into this. And then on top, supporting services is actually where the analysis takes place. And then the export services, as the name would imply, is where you take that information and learnings and you push it back up through the, back to the next tier. And then you've got, of course, device management, system management here and then security. So when you get this, and this is, I'll show you the roadmap because this isn't all available today. This is the end state vision. What will happen is you can get this whole shebang as one, but if you want to pull out different bits and pieces, so for different partners, different customers, you're gonna wanna take out, say, your secret sauces and the support services. So you're gonna yank that out and you're gonna put in your own proprietary support services or your own open source support services. So you can decide which you wanna do. But as I said, this is where then those components will get certified so that they are indeed plug and play and the certification, as I said, says basically we will not break the common APIs. And then, so you can mix and match or you can use this whole thing right out of the box. And then this architecture, if you look at it, how it works across the continuum from the edge back to the cloud, is you can actually have this on various, on the gateways, on the servers itself. And how many microservices you have running on this depends on what it's doing, the hardware it runs on. Once again, because it's loosely coupled, it allows you to run it across. And then the benefits, as you would guess, is by having a place that isn't the Wild West, you can actually sell more hardware and use a customer can buy hardware that you know will work with the various components. ISVs can write software packages that plug into one place, as opposed to a mess. Sensor and device drivers, sorry, device manufacturers, as I said, if they write the device drivers, then they can plug into this and then system integrators, their life becomes better. And of course, most importantly, the customers then have something that they can use and they don't have that paralysis that we talked about on the sticky note. So the governance structure, if you're familiar with various open source projects, is pretty standard. So you've got up top, you've got the member companies and then you've got the three pillars underneath it are the technical board, sorry, governing board technical steering committee and the certification committee. Certification committee is coming the second half of next year. Technical steering committee, as the name implies, sets the technical direction. It also coordinates and creates the working groups. And then the governing board is on the business side. And then all of this, because it's a Linux Foundation project, can leverage the support team there. And so I talked about 50 at launch. Six months later, we're at over 60. And so on here, you'll see both big companies that are familiar, also a lot of startups and companies that are in this space, specifically IoT, ones that will become household names, ones that will become acquired to be part of the IoT portfolios of larger companies. So here, last but not least, before I talk to you about the octopus, here's the actual release roadmap. And the idea is to do this bi-annually. The first one is Barcelona and that is coming out this month. And what it does is it basically provides the first instance of it that is tire kickable, right? So this is where customers can try POCs. This is where devs can get their hands on it and start working on it. And then the next one in February, it's not a release, but what it will be is a preview in Golang. So the idea is at this point, it's written in Java and C++, which makes it pretty big, has a big footprint. We will now be doing, or the community will be doing a Golang, the both Golang and C-based microservices replacements, which of course makes it much more performant and much smaller. California release in June of next year, so next summer, that's where they're gunning for the first release that's commercially available. And then last on this roadmap, next December is the certification program that we talked about. Once again, pretty standard in the industry. If you have things that you want to be interoperable, plug-in playable, if you don't have a certification program, then quality goes out the window and then it doesn't really work. So let me actually give you some contacts and further information. I'm not expecting you to write these all down, but my presentation is uploaded to the, which has the description of this talk. Also, the version there has a lot more words on it, so it will be easier to follow and have more information than I took off a lot of the stuff for these slides. So with that, for an end, back to the octopus. Does anybody know how many brains an octopus has? It has three hearts, but that's not what I'm talking about. How many brains does an octopus have? And the only reason why the guy who leads this effort knows it, because he was watching his kid's science show. So this is something I guess you learn when you're in sixth grade or something. But anyway, so it has nine brains, and so it's a distributed animal. So at each of the end points, so each of the tentacles, it has a brain, and then it has a large central brain where things aggregate. So very appropriate for the IoT space. So with that, if there's any questions, let me know. Questions? You can also just come up, I'm gonna hang out after this if you like. So that was your fun trivia for today. Very easy to join the foundation, sorry, the Foundry, as I said, to use the contacts there, and you just, you don't have to be a member or company to work with the code. So if you're interested, please get involved. You can use those contacts. If you come up, I can give you my contact and I can get you to the folks that are actually leading the project. So with that, thank you very much and have a great afternoon.