And here is my cartoon of a platform. So here we have on the left side of this cyber physical system, we've got a plant, a physical plant here. Raw materials come in, there's a series of physical processes that take place, assembling and welding and bending sheet metal, and stamping, and dyes, and whatnot, and finished goods pop out of the other end of this. In this system, there is a high number of sensors monitoring all aspects of operation of the physical plant. These sensors are wirelessly connected into the platform. And in the opposite direction, we need a way to control what's going on in the physical plant. So the direction, in this way, actuators are controlled wirelessly. And what kind of actuators are you familiar with? How about a relay? Have you worked with relays? >> Yes. >> Yeah, so you pass current through it and the contacts close, okay? Electric motors, okay? Get something to turn, turn, I don't know, whatever, some kind of a belt or something to get something to happen. How about pneumatic, compressed air, yeah? And hydraulic, that's fluid under pressure. Various actuators, okay? So these actuators and these sensors are wirelessly connected to the platform, okay? And we have data scientist that are either on site or off site, this is very much a cartoon, but it gets the idea across. We've got logistics. Logistics folks are responsible for ordering raw material and making sure all the component show up in time so that the line doesn't go down, or this may consist of multiple lines. This company may be producing multiple products and logistics is responsible for getting orders shipped to the customer, okay? We've got management, they oversee the whole operation. And we've got operators that are responsible for monitoring and controlling at a fairly fine level what's happening in the physical plant. And if there is breakdowns or predictive analytics at work here, the machine might say, I'm going to break down, this part is going to fail. We predict or I predict this machine part is going to fail in two weeks so the operator can order a new part, get it in and do preventive maintenance on it before that machine breaks down, okay? The platform kind of can, doesn't have to but it can extend wirelessly across to, well, and it could be a hard cable also, could be a fiber optic cable or T1 or whatever. But we're rapidly cutting the wires in these systems and going to wireless, so I chose to show it as a wireless connection. So some of the platform can live over and the cloud services provider where we've got quite a bit of compute available. That's our cloud compute, cloud storage. We've got machine learning and it's a whole bunch of software applications that do analytics and machine learning, and a number of other software applications that are made available to the platform through these APIs. So the flow of data here could be, doesn't have to be. This is just an example, is sensors. This is processors running. Sensors send the data wirelessly over and it's getting spooled up in cloud storage, okay? Oftentimes, and almost always, there is a fee for how much data that goes across this connection. So what happens is at sometimes there is some computing that takes place over here on this side of this wireless connection, to squeeze the amount of data down to just the necessary data that needs to be stored and analyzed over the cloud service provider. Not always but sometimes, if this amount of data is petabytes upon petabytes of data, you probably going to want to do some pre-processing of it first, to keep your network costs down. So the data can then move up into machine learning and perform analytics on it. We can extract some insights into what's happening in this physical process. Results of that then can go to the data scientist and maybe there's some interaction here between the data scientist and the analytics process. It might be a two-way street, there may be some iteration here when we get into the machine learning and the data analytics part. We'll see that there's a lot of human intervention involved in AI and machine learning and data analytics. It isn't all just magic. You just don't feed all the data in and all these great answers pop out. There is a tremendous amount of work involved. So data scientists discover these aha moments. He's able to tease some key insights out of the data of what's happening. That can be distributed to the logistics people, the management, and to the operators. And the management may direct the operators to make changes in terms of the way the physical process works to close the cycle. So the operators are modifying the physical process and maybe the logistics people are modifying how much raw material is coming in here. Maybe they notice that too much raw material was coming in, or not enough, because their utilization of their manufacturing line wasn't high enough. And maybe they find new ways to ship products to customers through analyzing this data. The platform is the glue that pulls all of this together and makes this happen, enables it to happen. Any question about that before I go on? No, okay. So we're going to start to move into the market coverage areas. We're going to look at hardware, platforms, software, services, applications of agriculture, building automation, manufacturing, automotive and transportation, and oil and gas, and the energy sector. So this is a combined all IoT, including consumer, yes? >> So as the industrial Internet of Things grows and companies are outfitting their manufacturing lines with IoT devices, do you see companies Implementing those sensors and stuff themselves or do you see another company coming in and outfitting the sensors and managing all the communication between the devices outsourced? >> So good question. My personal opinion is always going to be outsourced. I imagine there's a handful of industrious small companies that say, whoa, let's try this on our own, and they may be successful. There are a number of big players that are willing, ready and able to come in and help companies of any size deploy these systems, design it, build it and potentially operate it, also for him. And I've got a power plant example coming up here. Yes? >> It seems like it would be the inverse of, like, the larger the company, the more headcount they can devote to managing their own infrastructure. Or I guess the smaller the company, the easier it is to try something out and then tear it down if it doesn't work. >> Yeah, I could see that. Absolutely, I don't know what the split is 90 10, 95 5, I'm not sure. There's probably a gradient there. So they had real numbers for 2015. And these numbers are based on talking to CEOs and CFOs of all of the huge big companies you're all familiar with. All the big names, Alibaba and Facebook and Amazon and Cisco and IBM and Intel, and etc., etc., etc. So all IOT in 2015 generated $130 billion in revenue. The estimate of this report came out at the end of 2016, so they had partial data for 2016. They're estimating it at 164 billion and the ones with the P's on them are projections. And so out in 2022, we see they're almost $900 billion in revenue, this is an enormous number. And this is going to transform the way products and services are delivered. And you're going to participate in this, I predict, in some fashion because it's going to be profound and it's all about this operational efficiency gain. They're going to be smarter and much more efficient business operations across the board. I was staggered when I saw that number, I mean that's a huge amount of revenue. Again, they're just looking at these numbers, the estimate for 2016 versus 2022. I put these little arrows in here, just to highlight these two things that the market growth is attributed to increasing Internet connectivity. It's almost like that other slide we saw in the history, but notice they said cloud computing [LAUGH] here. So increasing connectivity, increasing use of wireless sensors, and the mainstreaming of cloud computing. Automotive and transportation, building automation in the industrial segment are the top applications for IoT deployments. Here they're showing it, data in terms of hardware and platforms and software solution and services, and I'm not sure why they split out. Let me say it this way, I don't understand the criteria they used for segmenting what's a platform and what software solutions are. because in my mind, they're very closely tied together. You'll see, software solutions is huge, services is huge, and the hardware seems kind of low to me, because it's going to take a lot of hardware development to build all these embedded systems and collect all this data and build a gateway. We'll see what gateways are later. But this was their opinion, okay, so there's the data. Again, growing at a pretty dramatic rate. Hard to know what that curve looks like, but it's looking exponential to me. So here's where they break out software solutions services, and you see how small the platform is. That was 2015, and this was the amount in 2022, and it doesn't seem like it's growing very much when we see what IBM's platform is capable of doing. In my head, these two are together. So I see it software solutions and platforms and services. They broke down the software market into the following segments, real-time streaming analytics. So some physical plant is running and it's spitting data out across the wired or wireless connection. And then the cloud analytics are being performed and then getting, hopefully if this system is working properly, key insights into what's happening inside their operation. Network bandwidth management, managing how much data flows, how the data flows, and when the data flows. Remote monitoring, understanding what's happening exactly out in the oil fields for a hydrocarbon harvesting company. Security is enormous. Data management, how do you manage and story and analyze all this huge amount of data that is being generated. And it's only going to grow up, which is good news for storage manufacturers. When I worked at Seagate, this is just amazing, because we're shipping 200 million drives a year and, what's being stored? And I got to thinking about it, and I thought, nobody wants to throw anything away. Think about how many pictures, well, except maybe, I guess, on Snapchat, they go away but, how many pictures get uploaded to Facebook? And look, we're in Disneyland, click, and the picture goes up. It's looked at for a little bit and then it just sits on some drive at Facebook, and it's never looked at again. And we just don't want to throw anything away. And so that's just driving demand. Our management when I was at Seagate told us that, if you add up all the planetary capacity to manufacture hard drives and all the planetary capacity to manufacture flash, the demand for storage exceeds that. So they claim Network Bandwidth Management Segment is going to dominate the software solutions market during this forecast. Period. I don't know if I believe that. Certainly network bandwidth management is important. I think data management, analytics, and security in the end are going to surpass network management, bandwidth management. And I imagine these will start to shift up and this one will shift down here, where it settles by 2022, it's anyone's guess. Yes? >> [INAUDIBLE] CAGR? >> Pardon? Compound Annualized Growth Rate. I guess if they're using that number to project and they don't have it flatline ever, then it is exponential. So I wonder in the market research, do they have a forecast for when all the needs are met and the industry is saturated. No and if history is a teacher, it will never be satisfied. In the last lecture did I tell you the story about when the ENIAC was first came out and they were just three machine? Yeah, okay. [LAUGH] Three machines. Yeah, no. [LAUGH] We have historically been wrong over and over and over, 640k, who would need more than that, right? And on, and on, and on, and on. This one's a key slide, I've stared at this often. And this I think I agree with, although I think oil and gas might move up. But look at automotive and transportation's projection terms of growth rate and revenue. 30% growth rate, industrialists is a manufacturing sector, 32 compound annual growth rate. There's a lot of opportunity here for you to deploy your embedded system skills. As I see it, I see tremendous opportunity for you guys. Gals, I'm using guys in the androgynous sense. This is fabulous, and we're going to, later in the semester, I plan to, I haven't finished the slide set for it yet. But we're going to do a deep dive into automotive and transportation and look at all the amazing things that bleeding edge companies are doing in terms of, we all know about the sexy one is self driving cars, right? We hear it constantly, self driving cars, self driving cars, okay. Yeah, maybe we'll figure out self driving cars eventually. But there's a bunch of other applications in this space that are, Actually achievable now, and there's companies working on some of this stuff, and so later in the semester we'll look at that. Here's some data, I don't know why they made it look like a I don't know, a jester's hat or [LAUGH] something from [LAUGH] medieval times. This is real data from 2015 and you'll see the automotive and transportation segment accounted for 19.2% of the market share by application. So here's consumer electronics over here at 7%, and that's the one we hear about all the time, right? But there's all these other segments and tremendous opportunity, just absolutely tremendous opportunity for you guys. There's a quote sort of everywhere I took text out of the market report, you'll see little quotes around it. And I'm just going to read these, hopefully it's not too boring for you. Strategies such as agreements, collaborations, joint ventures, and partnerships collectively accounted for 36% of the total strategic developments. And this strategy was adopted by all major players in the market. In the years 2015 and 2016, as companies look beyond their core strengths trying to figure out how to grow at a rapid rate in the IoT technology market. So they looked at their core competencies, and they said, well, we're good at this and we're good at this, and we're not good at that, and so we need to go find a partner to partner with. So there were a lot of partnerships. So what's the trouble with partnerships? >> I recently saw a video where Micron and Intel, they collectively launched their Jet 3D Crosspoint solid stabilizers, so if both of them are partnering together, how do they make time to compete with one another on the same platform? >> [LAUGH] I don't know. I've worked at Micron for nine years, nine months. No that's not right. How long have I been there? Certainly not years. I joined the company last July, so how many months has that been? Seven? Six, six, yeah. And since I work in engineering I'm not privy to what the senior management decides that their going to do. They don't really tell us. So I don't know. So, they're a partner and a competitor. And it was right on, because have you heard the term frienemies before? [LAUGH] I have lived through this over the course of my career. I've been involved in situations where we would do a joint development agreement with another company and everything seems great, and you get a little bit down the road. And all of a sudden things aren't so great. And your partner becomes your competition, and that's exactly what you were talking about. Yeah, so they can become outright competitors in the long-run. So in the short-term there may be some enablement and mutual benefit to both companies. In my experience, many of them fell apart and yeah, just didn't work out as planned. Yep. Something to be aware of when you get out into the working world. Question, or okay. Here's a geographical market breakdown. Markets and Markets estimated that the Asia Pacific region is expected to grow at the highest compound annual growth rate of 38% from 16 to 22 followed by the Middle East and Africa. I'm a little surprised by the Middle East and Africa, that they would say that, but the data doesn't seem to support that one. I look at this, I get a different take away. >> But I see the CAGR is at the bottom. I was getting that confused with the market share pie chart at the top, although it looks like the blue compound annual growth rates are all in order. Although that doesn't match the 38. >> Point is there's tremendous opportunity. I certainly believe Asia Pacific region is expected to grow. We all know about the companies that operate in this space or in this geographical location, right? So that makes sense. And China is doing a lot of investment providing a lot of financial incentive within their country for IoT development.
