1 00:00:00,190 --> 00:00:04,671 How systems engineering actually fits nicely into things like cost, and 2 00:00:04,671 --> 00:00:07,280 scheduling, work breakdown structure. 3 00:00:07,280 --> 00:00:13,330 So there's a close alignment, or should be or could be, with program management. 4 00:00:13,330 --> 00:00:17,213 >> Jack, I don't know, what are the issues that [CROSSTALK] in your opinion? 5 00:00:17,213 --> 00:00:18,302 Sorry to interrupt you- >> Yeah. 6 00:00:18,302 --> 00:00:22,030 >> Is, what's the difference between systems engineering and 7 00:00:22,030 --> 00:00:23,396 program management? 8 00:00:23,396 --> 00:00:24,407 >> [LAUGH] >> I mean, 9 00:00:24,407 --> 00:00:25,570 you just showed it, know what I mean? 10 00:00:25,570 --> 00:00:27,054 Try to emphasize that. 11 00:00:27,054 --> 00:00:33,756 >> Okay, so, I teach a course on this. 12 00:00:33,756 --> 00:00:35,545 Co-author a textbook on it. 13 00:00:35,545 --> 00:00:40,339 It's just that the systems engineer really tries to focus on 14 00:00:40,339 --> 00:00:45,640 the technical risks, the technical costs, the technical side. 15 00:00:45,640 --> 00:00:50,281 The program manager is worried about costs and scheduling, 16 00:00:50,281 --> 00:00:52,746 usually, and some of the rest. 17 00:00:52,746 --> 00:00:54,090 And so he or 18 00:00:54,090 --> 00:00:57,630 she is not necessarily particularly interested in the technical details. 19 00:00:57,630 --> 00:01:01,790 The systems engineer has to do that, and how do they do that? 20 00:01:01,790 --> 00:01:03,410 They do it at the interfaces. 21 00:01:03,410 --> 00:01:07,450 In other words, at the interface to critical reviews, 22 00:01:08,610 --> 00:01:11,610 interface between hardware and software. 23 00:01:11,610 --> 00:01:13,590 There's a problem, how's it gonna be resolved? 24 00:01:13,590 --> 00:01:15,490 Well, you have to be able to talk, 25 00:01:15,490 --> 00:01:18,240 maybe point back to some modeling that you did early on. 26 00:01:18,240 --> 00:01:22,707 There has to be some way a person who understands trade studies, 27 00:01:22,707 --> 00:01:24,362 who understands risk, 28 00:01:24,362 --> 00:01:30,003 who understands the technology to talk with these different domain [INAUDIBLE]. 29 00:01:30,003 --> 00:01:34,320 There's a lot more to it than that, but hopefully, just to present the idea that 30 00:01:34,320 --> 00:01:37,685 the systems engineer has some program management skills, but 31 00:01:37,685 --> 00:01:40,493 usually there's another person who handles them. 32 00:01:42,824 --> 00:01:47,309 So, Managing at the Interface, it's essential for the hardware software 33 00:01:47,309 --> 00:01:51,947 systems, for resource allocation, technical performance measurements. 34 00:01:51,947 --> 00:01:53,957 I wish Maria was here right now. 35 00:01:53,957 --> 00:01:54,978 >> She's not here. 36 00:01:54,978 --> 00:01:56,666 >> [INAUDIBLE] Students. 37 00:01:56,666 --> 00:01:59,569 >> For a software, so if you say that to an Agile person in the software world, 38 00:01:59,569 --> 00:02:03,300 they'll go, yeah, I know what that means and they sorta do, but [INAUDIBLE]. 39 00:02:03,300 --> 00:02:07,260 Okay, need for for models in systems engineering. 40 00:02:07,260 --> 00:02:13,359 Traditionally, we would, things would be Siloed and for requirements, right? 41 00:02:13,359 --> 00:02:18,257 For behavioral analysis, functional analysis, architectural synthesis, 42 00:02:18,257 --> 00:02:20,237 what's the solution, right? 43 00:02:20,237 --> 00:02:21,430 And verification. 44 00:02:21,430 --> 00:02:27,043 And these would be in different, like spreadsheet here, and there would actually 45 00:02:27,043 --> 00:02:32,671 be some simulation packages here, and some drawing packages, CAD, [INAUDIBLE]. 46 00:02:32,671 --> 00:02:36,073 Well, you end up with all these different repositories. 47 00:02:36,073 --> 00:02:43,210 And the idea is, there's no reason now why we can't have them reside together. 48 00:02:43,210 --> 00:02:45,750 Not even necessarily in a super complex database. 49 00:02:45,750 --> 00:02:50,280 You could use things like, you could use context sensitive, 50 00:02:51,350 --> 00:02:56,110 restful APIs, and whatnot, to grab onto existing tools and 51 00:02:56,110 --> 00:02:59,768 pull out the information that you need as a systems engineer, or 52 00:02:59,768 --> 00:03:02,090 a system science to do analysis. 53 00:03:02,090 --> 00:03:08,100 So anyway, the elevator pitch for model-based systems engineering 54 00:03:08,100 --> 00:03:11,760 is that you can tie all the different views you're gonna have, the requirements. 55 00:03:11,760 --> 00:03:13,451 Well, what is that by function? 56 00:03:13,451 --> 00:03:16,448 What does the actual architecture look like? 57 00:03:16,448 --> 00:03:20,532 These are different views but they all tie together and 58 00:03:20,532 --> 00:03:24,260 they can live in a data repository somewhere which 59 00:03:24,260 --> 00:03:28,804 will make modeling of these things a lot easier, all right? 60 00:03:31,182 --> 00:03:39,031 Okay, all right, so, Where are we at time-wise here, okay. 61 00:03:39,031 --> 00:03:45,308 So, I want to suggest that the Internet 62 00:03:45,308 --> 00:03:51,725 of Things is changing the way we're moving forward in the engineering disciplines, 63 00:03:51,725 --> 00:03:57,010 in the remaining disciplines, but also systems engineering. 64 00:03:57,010 --> 00:04:01,800 I am not gonna do a discussion of what IoT is, for this for the next 30 minutes. 65 00:04:01,800 --> 00:04:04,930 IoT is everything that's here and everything more. 66 00:04:04,930 --> 00:04:06,080 So this is the Internet of Things. 67 00:04:06,080 --> 00:04:09,500 This is the idea of, thanks to Moore's law, 68 00:04:09,500 --> 00:04:12,390 we're able to have sensors on everything. 69 00:04:12,390 --> 00:04:16,172 We can add a little bit of smarts to a sensor, a little bit of processing, 70 00:04:16,172 --> 00:04:18,000 a little bit of software. 71 00:04:18,000 --> 00:04:21,378 So here things like Audrino, microcontrollers, 72 00:04:21,378 --> 00:04:23,740 Raspberry Pi, microprocessors. 73 00:04:25,200 --> 00:04:28,920 With a real time operating system, this just means that these are little tiny 74 00:04:28,920 --> 00:04:31,988 computers next to sensors that are very cheap and inexpensive. 75 00:04:31,988 --> 00:04:38,500 With very little power, energy harvesting could be going on here, solar, lighter. 76 00:04:38,500 --> 00:04:40,330 They are resource constrained, 77 00:04:40,330 --> 00:04:43,500 maybe not a lot of memory on a particular sensor array. 78 00:04:43,500 --> 00:04:47,871 I don't know, maybe a toothbrush or something, or in the car. 79 00:04:47,871 --> 00:04:52,363 But because we, A lot of 80 00:04:52,363 --> 00:04:58,030 the computations they're doing don't necessarily have to be real fast. 81 00:04:58,030 --> 00:05:01,970 And we have something called the cloud and the edge where we can push some data up. 82 00:05:01,970 --> 00:05:04,983 Have the servers there crunch numbers, push it back down, 83 00:05:04,983 --> 00:05:06,320 and then make decisions. 84 00:05:07,740 --> 00:05:11,361 This is a bit idealized, And 85 00:05:11,361 --> 00:05:16,770 everything is as the Internet of Things suggests, it's Internet accessible. 86 00:05:16,770 --> 00:05:21,323 So all the structures, all the protocols that are on the Internet we can use, 87 00:05:21,323 --> 00:05:24,624 wired, wireless, LiFi is a nice example I like to use. 88 00:05:24,624 --> 00:05:28,697 That's wireless but with light, so you have a light bulb, shines light, 89 00:05:28,697 --> 00:05:30,063 it's got sensors in it. 90 00:05:30,063 --> 00:05:33,851 You can also communicate with it, kinda cool technology. 91 00:05:33,851 --> 00:05:37,294 Cellular, 4G and 5G, 5G is just starting to emerge. 92 00:05:37,294 --> 00:05:41,768 Edge which just usually means the network gateway that you have on 93 00:05:41,768 --> 00:05:44,694 your factory floor, or in your building, 94 00:05:44,694 --> 00:05:49,516 what have you, that connects with a server somewhere in the cloud. 95 00:05:49,516 --> 00:05:56,820 Software, and so [LAUGH] software, do you mean firmware? 96 00:05:56,820 --> 00:05:58,870 Do you mean Arcos, do you mean the operating system? 97 00:05:58,870 --> 00:06:00,510 Do you mean the network, do you mean the middleware? 98 00:06:00,510 --> 00:06:02,390 Do you mean an application? 99 00:06:02,390 --> 00:06:05,752 So when someone says the word software, it's good to say, 100 00:06:05,752 --> 00:06:08,185 what exactly do you think you mean by that? 101 00:06:08,185 --> 00:06:10,640 Data, Big data, Little, sensor data. 102 00:06:10,640 --> 00:06:14,490 This is AI, machine learning, high-level synthesis, right? 103 00:06:14,490 --> 00:06:20,830 Which just means coming from like almost algorithms down to something 104 00:06:20,830 --> 00:06:25,318 that you might have in hardware or software or maybe some genetic structure. 105 00:06:25,318 --> 00:06:33,500 Security, this almost, almost is gonna precede all the rest of this. 106 00:06:33,500 --> 00:06:39,740 And then various models on how people use this. 107 00:06:39,740 --> 00:06:43,088 All right so, what I'm talking about is the Internet of Things. 108 00:06:43,088 --> 00:06:47,635 You can reduce that as a further example of why I think this new technology is 109 00:06:47,635 --> 00:06:51,403 going to put the engineering back into systems engineering. 110 00:06:51,403 --> 00:06:55,237 And hopefully help give us better ties with system science as well. 111 00:06:58,314 --> 00:07:03,333 These are some network things, I don't wanna bore you with this. 112 00:07:03,333 --> 00:07:07,748 The material now is pretty cheap to make anything connected to the Internet. 113 00:07:07,748 --> 00:07:12,454 You can use a Raspberry Pi, you can use Arduino, you don't even have to use these. 114 00:07:12,454 --> 00:07:14,078 You can just have a chip if you've got enough. 115 00:07:15,987 --> 00:07:22,640 Right, so systems engineering and IoT, motivation and three key issues. 116 00:07:22,640 --> 00:07:27,420 So, the motivation here, the Internet of Things is here. 117 00:07:27,420 --> 00:07:30,056 It's coming, it's huge, it's not gonna stop. 118 00:07:30,056 --> 00:07:36,124 The amount of devices that they're hoping to get by 2020 is 50 billion. 119 00:07:36,124 --> 00:07:41,587 Interestingly, they won't reach that because there are operational issues. 120 00:07:44,253 --> 00:07:48,797 Network folks are saying, you're gonna have to move slowly and 121 00:07:48,797 --> 00:07:54,870 make sure that you provision and that you put these systems together slowly. 122 00:07:54,870 --> 00:07:56,090 But we're gonna get close. 123 00:07:57,490 --> 00:08:01,399 AI Artificial Intelligence, machine learning, 124 00:08:01,399 --> 00:08:05,069 whatever you wanna call it, deep learning, high level synthesis. 125 00:08:05,069 --> 00:08:11,450 By 2025 is expected to be five-sixths of all semiconductor growth. 126 00:08:11,450 --> 00:08:14,789 So it's a huge market. 127 00:08:14,789 --> 00:08:18,520 Moore's Law slowing down, aging. 128 00:08:18,520 --> 00:08:20,824 Okay, so there's a couple other issues here that play a part. 129 00:08:20,824 --> 00:08:24,130 I don't need to dwell on them too much, and I think [INAUDIBLE]. 130 00:08:24,130 --> 00:08:28,135 We have an aging workforce in the engineering community. 131 00:08:28,135 --> 00:08:30,963 And I guess what that means is that workforce, 132 00:08:30,963 --> 00:08:34,952 it's coming has different expectations from their tool sets and 133 00:08:34,952 --> 00:08:38,450 what they want to design or how they design it than we did. 134 00:08:38,450 --> 00:08:39,107 I won't hit on that. 135 00:08:43,934 --> 00:08:47,752 So, IoT, the Internet of Things, versus embedded systems. 136 00:08:47,752 --> 00:08:50,890 The big difference is that the Internet of Things are connected. 137 00:08:50,890 --> 00:08:53,207 That means that they have some wireless capability, 138 00:08:53,207 --> 00:08:56,665 which adds some additional complexity and some additional issues to all this. 139 00:08:59,820 --> 00:09:04,690 IoT requires, yes, it's a systems of systems approach. 140 00:09:04,690 --> 00:09:08,690 And there are lots of examples of this one. 141 00:09:08,690 --> 00:09:11,661 One is there's this some kind of connection, 142 00:09:11,661 --> 00:09:16,003 this balance between Moore's Law which says we can continue to make 143 00:09:16,003 --> 00:09:19,900 electronic devices cheaper and more powerful and smaller. 144 00:09:19,900 --> 00:09:26,600 It's ending at the moment, but there's still a lot of life left in it. 145 00:09:26,600 --> 00:09:29,997 Metcalfe's Law is from the communication side, and 146 00:09:29,997 --> 00:09:34,980 it just crudely, if you'll allow me says, as you get more connected points, 147 00:09:34,980 --> 00:09:38,390 how many is there gonna be for the Internet of Things? 148 00:09:38,390 --> 00:09:39,480 Just billions.. 149 00:09:39,480 --> 00:09:43,585 When you are getting more of these, the system itself is much more complex. 150 00:09:43,585 --> 00:09:46,898 And it's gonna get complex to a point where you are not going to 151 00:09:46,898 --> 00:09:49,143 be able to connect as many as you thought. 152 00:09:49,143 --> 00:09:51,800 Now there's ways around that, but I don't think I have it here. 153 00:09:51,800 --> 00:09:57,004 Well, this occurs where these two intersect and how this will really lay 154 00:09:57,004 --> 00:10:03,841 out, which is sometimes, Interesting. 155 00:10:05,550 --> 00:10:10,539 This really is just saying it's coming, things are getting cheaper. 156 00:10:10,539 --> 00:10:12,950 There's a need for it in the medical space. 157 00:10:12,950 --> 00:10:17,984 You could have home systems that monitor and allow you to interface with a doctor. 158 00:10:17,984 --> 00:10:24,783 The global economy is built on it, on and on, okay. 159 00:10:24,783 --> 00:10:29,300 So, okay, we've kinda already covered this, it's explaining there 160 00:10:29,300 --> 00:10:33,830 are different software as different each from hardware and vice versa. 161 00:10:33,830 --> 00:10:38,532 Software, the software, I guess the only thing I would state 162 00:10:38,532 --> 00:10:43,630 here is that where's the reliability, the big concern. 163 00:10:43,630 --> 00:10:45,580 It's been replaced by something called resiliency. 164 00:10:45,580 --> 00:10:50,427 So even in the utility side there are there are things 165 00:10:50,427 --> 00:10:53,300 going on that add complexity. 166 00:10:54,580 --> 00:10:55,940 All right, let's stop here for just a moment. 167 00:10:55,940 --> 00:10:59,220 So how do you integrate hardware and software, right? 168 00:10:59,220 --> 00:11:04,190 Integration, the interfaces where the systems engineers 169 00:11:04,190 --> 00:11:08,370 is always watching and looking, keeping an eye. 170 00:11:08,370 --> 00:11:13,130 We have a prototype, a hardware system, and run software on it. 171 00:11:13,130 --> 00:11:14,305 You could emulate. 172 00:11:14,305 --> 00:11:18,579 This is usually for very complex systems like billions and 173 00:11:18,579 --> 00:11:21,289 billions of chips if you will on all. 174 00:11:22,790 --> 00:11:25,619 Chip, billions of transistors per chip, might use emulation. 175 00:11:25,619 --> 00:11:30,036 If you want more real time analysis of some program that you're running, 176 00:11:30,036 --> 00:11:32,040 you might use simulations well. 177 00:11:33,660 --> 00:11:39,450 We don't have to go into this but there's just different ways and 178 00:11:39,450 --> 00:11:44,520 what you use depends on what the systems engineering is. 179 00:11:44,520 --> 00:11:48,658 This is where the automobile vendors are really 180 00:11:48,658 --> 00:11:53,530 struggling with because they have the connected cars. 181 00:11:53,530 --> 00:11:55,301 The autonomous cars are coming out and 182 00:11:55,301 --> 00:11:57,966 there has to be a modeling of both hardware and software. 183 00:12:00,720 --> 00:12:07,727 Siemens, I think just came out with this system modeling work. 184 00:12:07,727 --> 00:12:13,928 And I call it your attention only because we have hardware and 185 00:12:13,928 --> 00:12:21,241 software within a systems engineering context that's been modeled. 186 00:12:21,241 --> 00:12:25,654 And this one is XML based, SysML based. 187 00:12:25,654 --> 00:12:29,490 And that brings up another question with this modeling. 188 00:12:29,490 --> 00:12:37,086 Do we try to force everything to have a common language that's XML, right? 189 00:12:37,086 --> 00:12:40,986 Do an XML based sort of thing, or 190 00:12:40,986 --> 00:12:46,770 do we use more of, say a contextual mechanism? 191 00:12:47,850 --> 00:12:52,273 Well, I would argue that there's, aside from our world of systems there are folks 192 00:12:52,273 --> 00:12:56,440 out there in the hardware and software space, the EDA space, the word space, 193 00:12:56,440 --> 00:12:59,310 packaging space, that have their own sets of tools. 194 00:12:59,310 --> 00:13:01,450 And they work fine. 195 00:13:01,450 --> 00:13:03,337 They need to integrate them together, they're not gonna go. 196 00:13:03,337 --> 00:13:05,860 [LAUGH] They're not going to. 197 00:13:05,860 --> 00:13:10,623 We need a way of getting that information out without causing them to redo 198 00:13:10,623 --> 00:13:12,900 their entire ecosystem. 199 00:13:12,900 --> 00:13:13,928 So for this to happen, 200 00:13:13,928 --> 00:13:16,963 we're gonna have to come up with some creative ways to grab data and 201 00:13:16,963 --> 00:13:20,660 of course model or simulate it, what have you, and that's what's happening. 202 00:13:23,680 --> 00:13:26,042 Actually in COSI and IEEE are both. 203 00:13:26,042 --> 00:13:29,320 You were involved in the modeling or requirements. 204 00:13:29,320 --> 00:13:30,623 Well anyway, there's whole modeling activity- 205 00:13:30,623 --> 00:13:32,130 >> Yeah, no, requirements. 206 00:13:32,130 --> 00:13:33,556 You're looking at the wrong guy. 207 00:13:33,556 --> 00:13:34,395 >> Sorry, sorry. 208 00:13:34,395 --> 00:13:39,008 [LAUGH] >> Yeah, this manner [INAUDIBLE]. 209 00:13:39,008 --> 00:13:39,508 >> Yeah. 210 00:13:42,704 --> 00:13:48,049 >> In terms of an avenue for systems engineering to potentially work closely 211 00:13:48,049 --> 00:13:53,651 with system science, I think what John's point is we got these new devices, 212 00:13:53,651 --> 00:13:56,510 though it's on them and so on so forth. 213 00:13:56,510 --> 00:14:00,590 So it's just gonna change our whole way of looking at thing. 214 00:14:01,730 --> 00:14:05,336 What tickles my fancy and what interest is the International Council on 215 00:14:05,336 --> 00:14:09,678 Systems Engineering, they have a big effort in model-based systems engineering. 216 00:14:09,678 --> 00:14:14,429 And John just pointed out kind of the level that some of these 217 00:14:14,429 --> 00:14:16,347 activities are asked. 218 00:14:16,347 --> 00:14:20,497 The SysML, that would be more of a management level, would it not? 219 00:14:20,497 --> 00:14:25,646 I mean, if you take a look at a system of systems, IoT or an aircraft carrier or 220 00:14:25,646 --> 00:14:30,164 whatever, I mean the domains are going to be not necessarily used, 221 00:14:30,164 --> 00:14:32,490 they don't know a thing about it. 222 00:14:32,490 --> 00:14:35,850 Unless they're a software engineer. 223 00:14:35,850 --> 00:14:38,394 >> Right, but so, I mean, you're right, 224 00:14:38,394 --> 00:14:42,323 that's high level as to how maybe the system architecture, but 225 00:14:42,323 --> 00:14:46,750 really the process of systems engineering would take place. 226 00:14:46,750 --> 00:14:48,740 That has to marry down with the domain here. 227 00:14:48,740 --> 00:14:49,406 >> Yeah. 228 00:14:49,406 --> 00:14:54,588 >> Remember the engineer, it has to marry up with the existing billions of dollars, 229 00:14:54,588 --> 00:14:58,970 trillions worth of tools in these different spaces that are already there. 230 00:14:58,970 --> 00:15:01,500 And it happens, and it can happen. 231 00:15:01,500 --> 00:15:05,287 And with the Internet of Things, this Internet part means, HTML and 232 00:15:05,287 --> 00:15:09,792 some other things, that there's a common way to interface grab data out of here, 233 00:15:09,792 --> 00:15:11,583 use it, and even put it back in. 234 00:15:11,583 --> 00:15:14,207 So all I'm saying is that the system, 235 00:15:14,207 --> 00:15:18,230 there's a systems engineering piece that we focused on. 236 00:15:18,230 --> 00:15:22,625 But there's a whole 'nother piece that's there and we need to bring them both 237 00:15:22,625 --> 00:15:27,240 together for digital continuity, for digital prints for this digital printer. 238 00:15:27,240 --> 00:15:32,070 And it will happen, it will be painful but it will happen and it is happening. 239 00:15:32,070 --> 00:15:36,234 >> So my question would be what would you say is the maturity of model-based systems 240 00:15:36,234 --> 00:15:37,960 engineering at the SysML level? 241 00:15:41,282 --> 00:15:44,411 >> [LAUGH] I guess it depends who you talk to, 242 00:15:44,411 --> 00:15:49,842 there are some people it's very maturely, but there's other than, 243 00:15:49,842 --> 00:15:56,880 say, there are other tools, as I say that are saying, well, that's great. 244 00:15:56,880 --> 00:16:01,550 And we can use it, but we have other ways of grabbing data without these 245 00:16:01,550 --> 00:16:05,885 huge databases that the different, IBM, 246 00:16:05,885 --> 00:16:09,360 [INAUDIBLE] and Oracle. 247 00:16:09,360 --> 00:16:11,280 So we don't need to buy this huge package. 248 00:16:11,280 --> 00:16:14,557 And a matter of fact, the IoT folks doing Internet of things, 249 00:16:14,557 --> 00:16:17,470 they're not gonna buy expensive, costly programs. 250 00:16:17,470 --> 00:16:21,644 This is a type [INAUDIBLE] kind of a martial area, martial in terms of profit, 251 00:16:21,644 --> 00:16:25,940 right, which is why keeps jumping in and out of it, as an example. 252 00:16:25,940 --> 00:16:29,480 But I don't know what's gonna win out. 253 00:16:29,480 --> 00:16:33,156 I think some combination of the two, right, I mean XML, system L, 254 00:16:33,156 --> 00:16:36,930 UML certainly have their place and they're fine at a high level. 255 00:16:36,930 --> 00:16:41,613 But how is that gonna interface with these other tools, and even tools like MATHLAB, 256 00:16:41,613 --> 00:16:45,260 right or SPICE or for example, high level synthesis tools? 257 00:16:45,260 --> 00:16:47,580 Well, there are ways and it'll get better. 258 00:16:47,580 --> 00:16:50,153 But the thing that's fortunate is this Internet of Things, 259 00:16:50,153 --> 00:16:54,400 if you'll allow me to use that ASY, a lot of things wrapped together, right? 260 00:16:54,400 --> 00:16:56,450 So that's what's gonna force us to relook at all this. 261 00:16:56,450 --> 00:16:57,640 We look at the process. 262 00:16:57,640 --> 00:17:00,020 Systems engineers are gonna have to look at things a little bit differently. 263 00:17:00,020 --> 00:17:03,420 Because the system you have, 264 00:17:03,420 --> 00:17:06,980 hardware folks are gonna say, yeah, okay, but the software guys are gonna say, 265 00:17:06,980 --> 00:17:11,580 no, the network guys are gonna say, okay, so how do we address all those problems? 266 00:17:11,580 --> 00:17:13,000 We'll have to find some way of doing it. 267 00:17:13,000 --> 00:17:16,329 Sorry, it's not a very straight answer, but I'm trying to just share that 268 00:17:16,329 --> 00:17:19,080 the new elements here are- >> I think the point I'm trying to make is 269 00:17:19,080 --> 00:17:20,280 there is no straight answer. 270 00:17:20,280 --> 00:17:21,660 I mean, this is what we're facing. 271 00:17:21,660 --> 00:17:23,580 There's the technical connections. 272 00:17:24,640 --> 00:17:29,365 There's the the legacies, the demands of what they like to use, and try 273 00:17:29,365 --> 00:17:32,950 to put all this together, which in theory should be the systems engineer's job. 274 00:17:33,980 --> 00:17:34,480 >> Right. 275 00:17:36,740 --> 00:17:39,758 Well, it certainly will enable him to do his job a lot better [INAUDIBLE], 276 00:17:39,758 --> 00:17:41,160 [LAUGH] at the very least, right? 277 00:17:44,140 --> 00:17:46,230 The second issue, if you'll allow me here and again, 278 00:17:46,230 --> 00:17:50,090 the idea is systems engineering, and the Internet of Things effects. 279 00:17:50,090 --> 00:17:54,070 So there's the Internet of Things is different from just the embedded world. 280 00:17:54,070 --> 00:17:58,608 The other issue is the engineering expertise. 281 00:17:58,608 --> 00:18:01,234 How will the semiconductor and embedded folks and 282 00:18:01,234 --> 00:18:04,120 if you don't understand a lot of this, that's fine. 283 00:18:04,120 --> 00:18:06,720 You probably do, but I mean, I'm really just saying, 284 00:18:06,720 --> 00:18:11,830 there's a lot of intellectual property that goes on here. 285 00:18:11,830 --> 00:18:14,340 It's a global activity. 286 00:18:14,340 --> 00:18:19,828 And the folks that created this, the old guys [LAUGH] and 287 00:18:19,828 --> 00:18:26,520 gals are comfortable using tools that require a lot of attention. 288 00:18:27,660 --> 00:18:31,560 The new generation of engineer, it's been argued, are used to Facebook and 289 00:18:31,560 --> 00:18:33,065 iTunes and things like that. 290 00:18:33,065 --> 00:18:38,036 I don't mean that it's gonna be dumbed down to where just any Joe Blow can 291 00:18:38,036 --> 00:18:41,703 use it, if you will, but that it's going to have to, 292 00:18:41,703 --> 00:18:45,720 we're gonna have a lot of inexperienced folks out there. 293 00:18:46,860 --> 00:18:49,840 Maybe that's okay, because the hardware is being commoditized, and 294 00:18:49,840 --> 00:18:52,780 we're getting lots of intellectual property on the software side. 295 00:18:52,780 --> 00:18:55,320 So we'll be able to have packages of things. 296 00:18:56,610 --> 00:19:01,441 This is diverting a little bit, I guess, but it's just trying to point to the fact 297 00:19:01,441 --> 00:19:05,712 that we're gonna have to do things different from just the domain side, 298 00:19:05,712 --> 00:19:07,062 if you will, as well. 299 00:19:07,062 --> 00:19:12,451 Predicted 2 billion devices in less than a year or 300 00:19:12,451 --> 00:19:16,570 two, no, 50 billion by 2020. 301 00:19:16,570 --> 00:19:21,060 IoT is driven by data and subscription business models. 302 00:19:21,060 --> 00:19:25,170 The point I want to make here is that there will be tens 303 00:19:25,170 --> 00:19:29,710 of thousands of new IoT businesses and devices. 304 00:19:29,710 --> 00:19:33,630 Which, if you look at the number of engineers, hardware and software means 305 00:19:33,630 --> 00:19:40,740 that each individual is gonna have to designing far more systems than he can. 306 00:19:40,740 --> 00:19:45,394 And so it has to be a different approach in the tool side and 307 00:19:45,394 --> 00:19:47,680 on the engineering domain side as well. 308 00:19:49,440 --> 00:19:54,830 Well, okay, I mean, this is just down in the details of it. 309 00:19:54,830 --> 00:19:58,681 It's just saying there's gonna be a mix of engineers even more so 310 00:19:58,681 --> 00:20:00,650 that are coming into this space. 311 00:20:03,360 --> 00:20:07,356 A majority won't necessarily, I'm sorry, this is kind of focused on the chip side, 312 00:20:07,356 --> 00:20:08,630 so it's a little bit much. 313 00:20:08,630 --> 00:20:10,532 But the idea is that there's gonna be new challenges. 314 00:20:10,532 --> 00:20:11,970 There's gonna be a lot of them and 315 00:20:11,970 --> 00:20:15,120 there won't be enough engineers to handle it without some sort of automation, 316 00:20:15,120 --> 00:20:19,590 without some sort of getting smarter early on, as with with models. 317 00:20:21,560 --> 00:20:23,520 And complexity and chaos, okay, the third point. 318 00:20:26,580 --> 00:20:31,220 The overwhelming number of life-cycle tools that is changing 319 00:20:31,220 --> 00:20:35,190 engineering skill sets, life-cycle costs. 320 00:20:35,190 --> 00:20:40,060 The acquisition within any number of industries that you want to point to, 321 00:20:40,060 --> 00:20:47,260 are going to make this system, this Internet of Things more challenging. 322 00:20:48,650 --> 00:20:51,550 Well, what are we gonna do about it? 323 00:20:51,550 --> 00:20:52,790 There's lots of things. 324 00:20:52,790 --> 00:20:54,296 Again, this is kind of a chippy part, 325 00:20:54,296 --> 00:20:57,739 which I would have hidden if I was half decent at PowerPoint [LAUGH]. 326 00:20:57,739 --> 00:21:02,667 What's gonna come out of this is that we're going to 327 00:21:02,667 --> 00:21:07,595 have more systems that are kind of platform based and 328 00:21:07,595 --> 00:21:10,520 that have reference designs. 329 00:21:10,520 --> 00:21:14,876 [INAUDIBLE], there's gonna be things more complex than that, but 330 00:21:14,876 --> 00:21:17,072 these are easy examples of that. 331 00:21:17,072 --> 00:21:19,922 I don't think you particularly care about what it looks like. 332 00:21:19,922 --> 00:21:26,059 Okay, so in summary, and this is still kind of, 333 00:21:26,059 --> 00:21:33,220 I'll let you decide, but these points are relevant. 334 00:21:33,220 --> 00:21:36,820 So with all of the things that are happening, 335 00:21:36,820 --> 00:21:40,380 one of the ways we can deal with it is how we've always dealt with things, 336 00:21:40,380 --> 00:21:43,540 we raise the level of abstraction, right? 337 00:21:43,540 --> 00:21:50,170 More modular architectures specific to in this case, the Internet of Things. 338 00:21:50,170 --> 00:21:51,652 There's going to be more intellectual property. 339 00:21:51,652 --> 00:21:55,010 There's gonna be more hardware and software that's been tested and 340 00:21:55,010 --> 00:21:59,300 verified and that you can just put into your design. 341 00:21:59,300 --> 00:22:03,718 But you're still gonna need a systems integrator to go between systems and 342 00:22:03,718 --> 00:22:04,496 platforms. 343 00:22:04,496 --> 00:22:06,725 Analog models, yes, digital systems, and 344 00:22:06,725 --> 00:22:09,494 there's gonna be more software abstractions as well. 345 00:22:09,494 --> 00:22:13,677 And we begin to see that sort of, I mean, API's, yes, 346 00:22:13,677 --> 00:22:18,750 hardware interfaces but I mean, Python and other things that allow 347 00:22:18,750 --> 00:22:23,915 us to put more of these things together at a higher level, all right. 348 00:22:23,915 --> 00:22:26,224 So systems engineers, how do we improve? 349 00:22:26,224 --> 00:22:28,500 Systems engineers must be more agile. 350 00:22:28,500 --> 00:22:33,330 Not necessarily agile in the software sense, but that's fine. 351 00:22:33,330 --> 00:22:36,096 There's things to learn there, but certainly more flexible. 352 00:22:36,096 --> 00:22:41,217 Lean techniques, which what I mean there is just what we've always done, 353 00:22:41,217 --> 00:22:46,304 tailor the systems engineering process to the thing we're trying to do. 354 00:22:46,304 --> 00:22:53,760 Ilities, reliability, the automotive industry right ISO 26 2016. 355 00:22:53,760 --> 00:22:59,331 You will not be welcomed into the new age without 356 00:22:59,331 --> 00:23:06,023 some guarantee of reliability, autonomous vehicles. 357 00:23:06,023 --> 00:23:09,840 Security, a big issue in hardware and software, in another discussion. 358 00:23:09,840 --> 00:23:16,750 So going back to kind of my lead in slide here and my end. 359 00:23:19,660 --> 00:23:24,220 Hopefully, I argued that developing the Internet of Things systems 360 00:23:24,220 --> 00:23:29,170 are gonna require systems engineers 361 00:23:29,170 --> 00:23:33,670 to kinda of rethink how they do what they do. 362 00:23:33,670 --> 00:23:36,990 Requirements are critical, but they're gonna have to be more than just 363 00:23:36,990 --> 00:23:43,840 requirements jockeys, process experts and requirement jockeys, if you will, experts. 364 00:23:43,840 --> 00:23:48,555 We need some way of, which is, thanks to Herm and others, 365 00:23:48,555 --> 00:23:53,340 which is really why the PSU program was created. 366 00:23:53,340 --> 00:23:56,940 Was to reach down to the domains and to teach systems engineering, 367 00:23:56,940 --> 00:24:00,500 the hardware guys, the software guys, chemical engineers, the business or 368 00:24:00,500 --> 00:24:03,870 what have you, because it's necessary skills for 369 00:24:03,870 --> 00:24:08,430 them to move forward, to move into the things that we're doing now. 370 00:24:08,430 --> 00:24:12,510 Model-based systems engineering is coming, it's here. 371 00:24:13,530 --> 00:24:18,480 It's going to be an asset because the rest of the world is moving towards 372 00:24:18,480 --> 00:24:23,080 this digital manufacturing, digital design, this digital twin world. 373 00:24:23,080 --> 00:24:28,590 Which will allow us to model all sorts of things, reduce time, improve reliability, 374 00:24:28,590 --> 00:24:30,710 do all the good things that you can think of, at least in theory. 375 00:24:32,650 --> 00:24:36,645 And hopefully that model-based approach will also allow the systems engineering 376 00:24:36,645 --> 00:24:41,569 folks to talk more to the system science folks. 377 00:24:41,569 --> 00:24:46,093 And that's it, so are there any questions? 378 00:24:48,152 --> 00:24:51,436 Or just anything at all. 379 00:24:56,127 --> 00:25:00,878 >> So there's so much stuff. 380 00:25:00,878 --> 00:25:05,530 >> I know, there's too much probably. 381 00:25:07,680 --> 00:25:09,050 >> So I could see a little bit of how, 382 00:25:09,050 --> 00:25:12,470 I mean, I see connection between system scientist and system engineering. 383 00:25:12,470 --> 00:25:15,948 I see how we could work together in general or 384 00:25:15,948 --> 00:25:19,529 how we're doing things kind of the same way. 385 00:25:19,529 --> 00:25:25,804 So someone's pointed in the comment here, Andy says maybe this has been 386 00:25:25,804 --> 00:25:33,050 needs to partner with a cognitive scientist to understand social layers and. 387 00:25:33,050 --> 00:25:39,075 >> Is the idea that, so the system engineer learn all of this stuff and 388 00:25:39,075 --> 00:25:43,702 then engineering people, or is the system 389 00:25:43,702 --> 00:25:49,094 engineer partnering with all the non-engineers as well? 390 00:25:49,094 --> 00:25:52,046 Are they integrating between the rest of the world and engineers or 391 00:25:52,046 --> 00:25:54,010 just between all the engineers? 392 00:25:54,010 --> 00:25:58,240 >> Well, excellent, great question from. 393 00:25:58,240 --> 00:26:03,356 So I would almost say that in that case the systems software engineers, 394 00:26:03,356 --> 00:26:06,512 the one developing the system interfaces, 395 00:26:06,512 --> 00:26:10,279 developing more with the social and the cognitive. 396 00:26:10,279 --> 00:26:12,311 It doesn't have to be that, but just as a- >> Mm-hm. 397 00:26:12,311 --> 00:26:14,940 >> To answer your question, though, what does that mean? 398 00:26:14,940 --> 00:26:17,968 Does that mean we have a systems engineeringing degree now the last six 399 00:26:17,968 --> 00:26:20,550 years because you gotta do hardware and software? 400 00:26:20,550 --> 00:26:24,501 No, I think what it means, and this is just from our experience, 401 00:26:24,501 --> 00:26:30,440 is that it means you try to introduce more systems concept in an undergraduate. 402 00:26:30,440 --> 00:26:32,480 But really what's gonna happen is you'll come out, 403 00:26:32,480 --> 00:26:34,850 people work in the hardware space, in the software space. 404 00:26:35,990 --> 00:26:38,410 You do a good job in one or the other and the boss will say, geez, 405 00:26:38,410 --> 00:26:39,380 you're a great hardware guy. 406 00:26:39,380 --> 00:26:42,348 I'm gonna give you some software folks and you're gonna go out and 407 00:26:42,348 --> 00:26:44,008 do this this new Internet of Things. 408 00:26:44,008 --> 00:26:46,310 Okay, great, let's see, what do I remember about software? 409 00:26:46,310 --> 00:26:51,090 And hardware guys know a lot, or vice versa, right, software, hardware. 410 00:26:51,090 --> 00:26:56,350 So then there'll be the opportunities through master's programs or 411 00:26:56,350 --> 00:26:59,930 certificates to where you can come back and get refreshers on and 412 00:26:59,930 --> 00:27:03,390 expand upon systems engineering. 413 00:27:03,390 --> 00:27:06,280 And a lot of that will be much more modeling, I think, 414 00:27:06,280 --> 00:27:08,060 in the future than it is now. 415 00:27:08,060 --> 00:27:11,700 I mean, just in a practical sense, I can't imagine how else you would do it. 416 00:27:11,700 --> 00:27:17,111 And that seems to be the way most of our- From a system science side though, 417 00:27:17,111 --> 00:27:21,199 I mean, you look at this and you're going well, okay, but we really don't, I mean, 418 00:27:21,199 --> 00:27:22,600 what do you think of all this? 419 00:27:22,600 --> 00:27:28,560 I mean, you're not really so concerned about the and all that. 420 00:27:28,560 --> 00:27:32,420 You're obviously more in a bigger space. 421 00:27:34,420 --> 00:27:36,530 So before this talk I was doing a little more research, and 422 00:27:36,530 --> 00:27:40,660 there's a nice chart up there on your wall actually, about how technology through 423 00:27:40,660 --> 00:27:44,002 the years has changed system science and changed systems engineering as well. 424 00:27:44,002 --> 00:27:50,980 I mean, do you see that what's happening now will change things? 425 00:27:50,980 --> 00:27:51,830 Is it just another 426 00:27:53,630 --> 00:27:58,412 more advanced technology that won't really affect how you view system science? 427 00:27:58,412 --> 00:28:04,830 >> I think, Let's see how to say it. 428 00:28:04,830 --> 00:28:08,370 I think we have, so maybe there's a general feeling and 429 00:28:08,370 --> 00:28:11,900 some different ways that, like with what you lay out, 430 00:28:11,900 --> 00:28:18,360 that there's an increasing need for the system's viewpoint on things. 431 00:28:18,360 --> 00:28:23,314 And whether or not people will read about it, whether or 432 00:28:23,314 --> 00:28:29,080 not that's taken to heart on projects is I don't know, but tere's a need for it. 433 00:28:29,080 --> 00:28:35,070 And the way we kind of go out to see them, we see it in different domains or 434 00:28:35,070 --> 00:28:38,080 we see we have different interest in different problems. 435 00:28:38,080 --> 00:28:44,585 And so maybe our idea on how we are making system scientists is that they 436 00:28:44,585 --> 00:28:50,332 get a lot of breadth and then they have depth in one area maybe. 437 00:28:50,332 --> 00:28:53,490 Start to have some expertise and they in other things. 438 00:28:53,490 --> 00:28:58,400 And we send them out to be kind of an integrator, so they have to 439 00:28:58,400 --> 00:29:03,800 understand a lot of different perspectives and domains or be able to pick them up. 440 00:29:03,800 --> 00:29:08,690 And they're doing sometimes formal technical modeling and sometimes mental 441 00:29:08,690 --> 00:29:11,790 modeling of how the different things work to bring them together. 442 00:29:11,790 --> 00:29:15,940 So I think we have a lot in common in that sense, right? 443 00:29:15,940 --> 00:29:19,746 Then the question is, so if a system scientist goes and 444 00:29:19,746 --> 00:29:23,553 works on a project that gets to the engineering point, 445 00:29:23,553 --> 00:29:28,960 then is there a system engineer also that work together or do they hand it off? 446 00:29:28,960 --> 00:29:33,468 We're doing all these things that are very parallel. 447 00:29:33,468 --> 00:29:38,030 It's not quite clear to me how and when we link up. 448 00:29:38,030 --> 00:29:43,729 >> I would think that the development of, so this development of neural networks and 449 00:29:43,729 --> 00:29:46,631 artificial intelligence is pretty new. 450 00:29:46,631 --> 00:29:51,667 I think that would probably be of more interest to system science 451 00:29:51,667 --> 00:29:56,703 folks because you're, I don't know, you're modeling could 452 00:29:56,703 --> 00:30:02,300 be behavior of people, behavior of different groups and systems. 453 00:30:02,300 --> 00:30:06,229 I would think at that point where it moved from that 454 00:30:06,229 --> 00:30:11,530 to actually implementing something, I mean, that might. 455 00:30:11,530 --> 00:30:16,702 So again, maybe that movement, the AI movement might lend to 456 00:30:16,702 --> 00:30:22,499 more of that interaction between systems scientists and systems engineers. 457 00:30:22,499 --> 00:30:27,385 >> I've sat in on the systems science seminars for more than a year, and 458 00:30:27,385 --> 00:30:32,030 last spring we had that dynamite data science sequence of talks. 459 00:30:32,030 --> 00:30:40,484 So as an example of what I see the connections, I'll say it again. 460 00:30:40,484 --> 00:30:45,777 An example of connections between what we're doing in systems engineering and 461 00:30:45,777 --> 00:30:50,526 what you folks are doing, the data science sticks up like a sore thumb. 462 00:30:50,526 --> 00:30:56,335 [LAUGH] I don't know, because you folks are doing amazing things with 463 00:30:56,335 --> 00:31:02,610 a bunch of sketchy data saying, yeah, here's the valid part of it. 464 00:31:02,610 --> 00:31:04,340 Here's not the valid part of it. 465 00:31:04,340 --> 00:31:06,700 I'm gonna draw some patterns from this. 466 00:31:06,700 --> 00:31:10,810 I got these patterns, we now can make some decisions in the future. 467 00:31:10,810 --> 00:31:15,710 That way of going about doing things has to be in a systems of juniors toolkit. 468 00:31:15,710 --> 00:31:17,045 If we're oblivious to it, 469 00:31:17,045 --> 00:31:21,289 it'd almost be like being oblivious to fundamental equations. 470 00:31:21,289 --> 00:31:24,784 In, say, the electrical engineering sense, Kirchhoff's Law, 471 00:31:24,784 --> 00:31:26,512 mechanical engineering sense. 472 00:31:26,512 --> 00:31:28,284 If you're oblivious to those, 473 00:31:28,284 --> 00:31:32,520 you really don't have a full understanding of the fundamentals. 474 00:31:32,520 --> 00:31:37,240 And I would see that as a great example of the intersection. 475 00:31:37,240 --> 00:31:40,526 And it's really up to us as system engineers to reach out to you, 476 00:31:40,526 --> 00:31:42,680 if you ask me. 477 00:31:42,680 --> 00:31:48,010 And we've got one master student here who is in system engineering, who is 478 00:31:48,010 --> 00:31:54,720 in the data science here The other area that the system scientists are, 479 00:31:54,720 --> 00:31:59,238 in my opinion, way better at is the human side, is the cognitive side. 480 00:31:59,238 --> 00:32:03,530 Cuz you've got a huge number of folks in this program and 481 00:32:03,530 --> 00:32:07,400 that you bump into, I mean, they have a humanities background. 482 00:32:07,400 --> 00:32:08,691 They're interested in psychology. 483 00:32:08,691 --> 00:32:10,920 They are interested in and so on and so forth. 484 00:32:10,920 --> 00:32:13,653 Our guys who tend to, and gals, who tend to be more, 485 00:32:13,653 --> 00:32:17,402 comes from a hardware software end of things and product development. 486 00:32:17,402 --> 00:32:21,460 Not totally but, I mean, that's kind of the big pushes. 487 00:32:21,460 --> 00:32:24,655 And those are just examples that I see. 488 00:32:27,223 --> 00:32:29,592 >> Yeah, I would agree with that. 489 00:32:29,592 --> 00:32:32,550 I was just thinking about this, where do where do they meet up? 490 00:32:32,550 --> 00:32:36,704 It seems to me like the data side is so huge, these Internet of Things, 491 00:32:36,704 --> 00:32:41,510 is there are a bunch of sensors going out there, they're collecting data. 492 00:32:41,510 --> 00:32:45,980 The huge network that they create that you can when there's a problem in that you 493 00:32:45,980 --> 00:32:50,690 need a lot of data about latency, about all the actual topology of the network. 494 00:32:50,690 --> 00:32:54,295 All these things that then needs to be analyzed and look for relationships 495 00:32:54,295 --> 00:32:58,035 between variables, which is very new to the whole system science thing myself. 496 00:32:58,035 --> 00:33:03,462 But I'm taking DMIT and learning about that finding models for 497 00:33:03,462 --> 00:33:06,725 possible, how are these things related. 498 00:33:06,725 --> 00:33:10,814 Especially if so much of, as they make you to believe, AI, and 499 00:33:10,814 --> 00:33:14,364 especially these black box AI, they put neural nets, 500 00:33:14,364 --> 00:33:19,640 when things are happening and you wanna understand how is it making decisions. 501 00:33:19,640 --> 00:33:24,660 Sometimes you just need to collect lots of data about what went into that and 502 00:33:24,660 --> 00:33:27,014 what came out and didn't understand, I guess. 503 00:33:27,014 --> 00:33:31,110 That analysis side is where I think the system science and 504 00:33:31,110 --> 00:33:32,460 engineering kind of comes together. 505 00:33:34,080 --> 00:33:39,780 >> The predictive maintenance for aircraft and cars and everything else. 506 00:33:39,780 --> 00:33:42,580 I mean, that's just one example, but 507 00:33:42,580 --> 00:33:47,220 that's big, right, if you can predict when an engine's gonna go out. 508 00:33:47,220 --> 00:33:51,540 So you have some algorithms, you done some, again, it's more technology, 509 00:33:51,540 --> 00:33:53,020 I guess, than interfacing with people. 510 00:33:53,020 --> 00:33:58,525 But that's a huge market but companies can realize significant gains. 511 00:33:58,525 --> 00:34:00,820 It's AI, it's mainly AI, right? 512 00:34:00,820 --> 00:34:04,730 The sensors, you may have the sensors there, so 513 00:34:04,730 --> 00:34:07,830 you got into legacy systems or new systems. 514 00:34:07,830 --> 00:34:11,350 But, yeah, it's really a data piece. 515 00:34:11,350 --> 00:34:14,930 >> Yeah, getting away from system science per se, and 516 00:34:14,930 --> 00:34:20,530 maybe addressing the question that was asked. 517 00:34:20,530 --> 00:34:24,430 If you take a look at the product development environments in Portland 518 00:34:24,430 --> 00:34:27,835 area versus Seattle or San Francisco or something like that, 519 00:34:27,835 --> 00:34:33,420 there's different size companies, the different maturity of companies. 520 00:34:33,420 --> 00:34:36,745 And the way a system engineer would go about doing their business, 521 00:34:36,745 --> 00:34:40,610 who they would interact with, would differ from company to company. 522 00:34:40,610 --> 00:34:45,118 A smaller company you may have, there is no title systems engineer. 523 00:34:45,118 --> 00:34:48,560 They're a program manager that knows a little bit about software and 524 00:34:48,560 --> 00:34:50,780 they make a product that has hardware. 525 00:34:50,780 --> 00:34:55,160 Well, and they were chosen to put stuff together. 526 00:34:55,160 --> 00:34:59,610 But kinda getting back to our examples of very clearly data with 527 00:34:59,610 --> 00:35:05,280 the system science folks, the data science, the human side of things. 528 00:35:05,280 --> 00:35:07,030 And here's two good examples. 529 00:35:07,030 --> 00:35:09,159 Let's talk about our relationship with software people. 530 00:35:10,230 --> 00:35:12,760 They're the ones that developed, what was it UML? 531 00:35:12,760 --> 00:35:15,506 What was before? 532 00:35:15,506 --> 00:35:20,591 UML, it's a graphical language for them to organize what 533 00:35:20,591 --> 00:35:25,896 the activities are and what the objects are in the software. 534 00:35:25,896 --> 00:35:29,830 I mean, they got so sophisticated that they can actually generate code from them. 535 00:35:29,830 --> 00:35:33,680 Well, we, if you're in an environment where you've got a software guy that's 536 00:35:33,680 --> 00:35:36,820 doing that and your boss is saying, well, why can't you get this product out in 537 00:35:36,820 --> 00:35:41,420 six weeks instead of six months, ou want to talk to the Agile folks. 538 00:35:41,420 --> 00:35:44,460 And one of our other students in the program is is in that area. 539 00:35:44,460 --> 00:35:45,912 And these are just examples. 540 00:35:45,912 --> 00:35:50,500 I can't give you a matrix of here's when we would do what. 541 00:35:50,500 --> 00:35:54,570 There's a couple examples of what we would do. 542 00:35:54,570 --> 00:35:57,297 >> And then throw into that Internet of Things and now you've got, 543 00:35:57,297 --> 00:35:59,340 you throwing in some wireless thing, right? 544 00:35:59,340 --> 00:36:01,955 >> Yeah. So it's the software folks, they're great. 545 00:36:01,955 --> 00:36:02,955 The hardware guys are great. 546 00:36:02,955 --> 00:36:06,475 And now suddenly I need to put someone on- >> Radio insurance. 547 00:36:06,475 --> 00:36:08,311 >> Those are easy to slap in. 548 00:36:08,311 --> 00:36:11,825 I can just slap them in the software called MAC layer and 549 00:36:11,825 --> 00:36:13,915 yeah, maybe it doesn't test out. 550 00:36:13,915 --> 00:36:15,295 Maybe something else interferes. 551 00:36:15,295 --> 00:36:19,970 So again, I mean, again, they won't be called systems engineers but 552 00:36:19,970 --> 00:36:24,800 there's going to be a need for that kind of, to use the tools of 553 00:36:24,800 --> 00:36:28,540 systems engineering in terms of knowing how to watch the interface. 554 00:36:28,540 --> 00:36:31,440 Knowing what to watch for on an interface to tell you if you have problems. 555 00:36:31,440 --> 00:36:34,650 Is there I have a certain power allocated, I have a certain performance, 556 00:36:34,650 --> 00:36:36,360 I have a certain? 557 00:36:36,360 --> 00:36:41,170 So that's the special sauce hopefully that a systems engineer will help with. 558 00:36:42,450 --> 00:36:47,529 Of course, we can model that on early on, which you probably can, 559 00:36:47,529 --> 00:36:52,170 and you can start to add another parts to it as well, anyway. 560 00:36:52,170 --> 00:36:53,160 Okay, well, thank you. 561 00:36:54,550 --> 00:36:55,375 Any other question? 562 00:36:55,375 --> 00:36:57,498 >> I was gonna say, I think your, 563 00:36:57,498 --> 00:37:01,667 we talked about your point that security is maybe the first or 564 00:37:01,667 --> 00:37:06,212 the most interesting or highest thing and- >> He glossed over that, 565 00:37:06,212 --> 00:37:07,482 I might add [LAUGH]. 566 00:37:07,482 --> 00:37:11,322 >> He said it was important, but- >> Now it's the most important. 567 00:37:11,322 --> 00:37:19,975 >> But very broadly, part of the issue is maybe that almost anyone now could make, 568 00:37:19,975 --> 00:37:25,400 I could make a product with a cheap Wi-Fi sensor, 569 00:37:25,400 --> 00:37:28,900 router thing in it and on my own. 570 00:37:28,900 --> 00:37:32,802 I could design this and order 10,000 of them and start shipping them. 571 00:37:32,802 --> 00:37:37,762 And it's easiest and cheapest if you do a really bad job at security, and 572 00:37:37,762 --> 00:37:40,135 that's what everyone's doing. 573 00:37:40,135 --> 00:37:43,587 And it's happening all over the place and 574 00:37:43,587 --> 00:37:49,010 people who just don't know better or don't care are doing this, 575 00:37:49,010 --> 00:37:53,685 all these Internet of very poorly secured things. 576 00:37:53,685 --> 00:37:58,530 And that's the thing where there's technical security issues, 577 00:37:58,530 --> 00:38:00,040 but there are these social, 578 00:38:00,040 --> 00:38:03,270 there are very high level security issues about what it means that these 579 00:38:03,270 --> 00:38:07,340 things are insecure and what you can do with the data while manipulating them. 580 00:38:07,340 --> 00:38:11,350 >> And what if you have one of these that running your infrastructure, 581 00:38:11,350 --> 00:38:13,805 your power plants, your airlines? 582 00:38:15,730 --> 00:38:20,365 It's amazing how poorly all companies spend on security until, 583 00:38:20,365 --> 00:38:24,720 so I'm afraid it's gonna have to be a governmental thing where it says 584 00:38:24,720 --> 00:38:27,270 you will have this level of security. 585 00:38:27,270 --> 00:38:28,400 Cuz otherwise, why would you do it? 586 00:38:28,400 --> 00:38:29,250 I mean, we're capitalists. 587 00:38:29,250 --> 00:38:31,710 So the product did what it was supposed to, 588 00:38:31,710 --> 00:38:35,830 yeah, it's hackable and now we're vulnerable to attack. 589 00:38:37,970 --> 00:38:40,012 >> You see it on the news every day. 590 00:38:40,012 --> 00:38:43,371 >> It's easier to pay the fine than to make the initial investment. 591 00:38:43,371 --> 00:38:43,939 >> Yeah. 592 00:38:43,939 --> 00:38:45,289 >> One time fine. 593 00:38:45,289 --> 00:38:46,012 >> For now it is. 594 00:38:46,012 --> 00:38:48,272 >> Yeah. [CROSSTALK] 595 00:38:48,272 --> 00:38:49,848 >> Especially if you're a startup and 596 00:38:49,848 --> 00:38:53,943 you're just out there and you only have so much money, you only have so much runway, 597 00:38:53,943 --> 00:38:55,440 and you need to create a product. 598 00:38:55,440 --> 00:38:56,570 What's the thing to cut? 599 00:38:56,570 --> 00:38:59,190 Well, what are the odds that something's gonna be? 600 00:38:59,190 --> 00:39:02,730 Hell, we put a product out there, no one even knows it exists at first. 601 00:39:02,730 --> 00:39:04,718 We all have time to put in security later. 602 00:39:04,718 --> 00:39:07,260 I mean, I work at a company that's technical IO2, 603 00:39:07,260 --> 00:39:09,600 that's exactly what we've done. 604 00:39:09,600 --> 00:39:13,104 We don't have a huge lot of market penetration, 605 00:39:13,104 --> 00:39:17,462 it's a very specific product we put into retail stores, but 606 00:39:17,462 --> 00:39:22,700 there's definitely vulnerabilities there that it's a problem. 607 00:39:22,700 --> 00:39:25,661 >> And that's where the modeling side would come in because a lot of it is 608 00:39:25,661 --> 00:39:26,795 the human error, right? 609 00:39:26,795 --> 00:39:30,261 It's the you never change your password, 610 00:39:30,261 --> 00:39:35,181 you don't use two way, or two, yes, thank you. 611 00:39:35,181 --> 00:39:37,898 So there's lots of thing, so how do you change sort of behavior? 612 00:39:37,898 --> 00:39:38,968 How do you? 613 00:39:41,216 --> 00:39:44,943 >> So to get back to the first questions, and 614 00:39:44,943 --> 00:39:49,650 we were talking about it, is, okay, I'm a system engineer. 615 00:39:50,650 --> 00:39:52,750 Who am I really at this company? 616 00:39:52,750 --> 00:39:54,250 Who I interact with. 617 00:39:54,250 --> 00:39:57,750 More importantly, what the relationship is to system science, and 618 00:39:57,750 --> 00:39:59,176 that's what today's talk is about. 619 00:39:59,176 --> 00:40:04,905 I know this is gonna sound crazy, but here's a topic, systems thinking. 620 00:40:04,905 --> 00:40:07,990 You guys live, breathe it. 621 00:40:07,990 --> 00:40:11,317 I mean, if you don't do it, you shouldn't even be in this building, right? 622 00:40:11,317 --> 00:40:16,918 Now, in the system engineering world the same would be true. 623 00:40:16,918 --> 00:40:21,060 But I don't think we've treated as formally as you folks do. 624 00:40:21,060 --> 00:40:23,900 And which get us back to, for example, security. 625 00:40:23,900 --> 00:40:28,357 I mean, it would just seem to me that if you're system engineer, 626 00:40:28,357 --> 00:40:31,843 you're involved with the Internet of Things, and 627 00:40:31,843 --> 00:40:37,533 security is not at the forefront of your mind, you are not thinking systemically. 628 00:40:37,533 --> 00:40:39,700 You're leaving out a big piece. 629 00:40:41,422 --> 00:40:46,360 And so I see that as a connection and one of the challenges for us. 630 00:40:47,630 --> 00:40:51,670 >> And your comment was right because we should think about that. 631 00:40:51,670 --> 00:40:53,464 But as a systems engineer, 632 00:40:53,464 --> 00:40:59,243 you're watching which also program manager's going, you're over budget. 633 00:40:59,243 --> 00:41:00,509 Your okay, I'll deal with it. 634 00:41:00,509 --> 00:41:03,220 >> It's a game theory thing really when I think about it 635 00:41:03,220 --> 00:41:03,740 as- >> But 636 00:41:03,740 --> 00:41:06,270 the things are rolling out now, right? 637 00:41:06,270 --> 00:41:10,180 And they're starting to, so these Internet devices that you hook up to, 638 00:41:10,180 --> 00:41:14,400 maybe on a factory floor, and they're just rolling them out and 639 00:41:14,400 --> 00:41:18,930 they're just letting the network find them, right? 640 00:41:18,930 --> 00:41:21,000 Which is a terrible thing, doing security. 641 00:41:21,000 --> 00:41:23,700 And so the IT guys are saying, well, this is bullshit, I can't secure this. 642 00:41:23,700 --> 00:41:27,390 You're not gonna put anything else out until you find a fix. 643 00:41:27,390 --> 00:41:31,166 So Intel and a bunch of other companies came together and said, well, 644 00:41:31,166 --> 00:41:35,460 we have this onboarding where we put some ID in the chip maybe or an. 645 00:41:35,460 --> 00:41:38,420 So then they address the security issue, but it's always after the fact, 646 00:41:38,420 --> 00:41:41,110 after the dollars appear. 647 00:41:41,110 --> 00:41:43,736 So I don't know how we're gonna do that. 648 00:41:43,736 --> 00:41:45,920 >> But isn't that real business? 649 00:41:45,920 --> 00:41:50,180 I mean, there are several tasks that you can do without initially, 650 00:41:50,180 --> 00:41:53,670 and then once you start really understanding the market and once you 651 00:41:53,670 --> 00:41:58,700 start making some penetration, you go back and you re-engineer certain things. 652 00:41:58,700 --> 00:42:00,710 >> But except the risk now is too great, isn't it? 653 00:42:01,810 --> 00:42:06,400 I mean, the risk is national security, it's Bitcoin, I suppose, right? 654 00:42:06,400 --> 00:42:08,164 I mean, it's financial. 655 00:42:08,164 --> 00:42:14,990 Our system isn't set up to deal with it, I guess, aside from us folks, I mean. 656 00:42:16,070 --> 00:42:18,890 So it's something we know we have to do, and yet 657 00:42:18,890 --> 00:42:21,728 we can't do it sometimes because of business. 658 00:42:21,728 --> 00:42:26,515 >> Well, I mean, it's got to become, something needs to sort of enforce it, 659 00:42:26,515 --> 00:42:28,460 whether it's the government saying, all right, look, 660 00:42:28,460 --> 00:42:32,350 now all IoT devices need to have this level of security. 661 00:42:32,350 --> 00:42:35,713 But then, of course, for small businesses, for startups, well that's 662 00:42:35,713 --> 00:42:40,536 hindering innovation, gonna have to spend so much resource on just securing it. 663 00:42:40,536 --> 00:42:45,450 For us, my company, I'm amazed when we go to 664 00:42:45,450 --> 00:42:49,246 put it in a retail space that there isn't more netting on their side to say, well, 665 00:42:49,246 --> 00:42:52,280 we're not gonna put this in here unless you have the security. 666 00:42:52,280 --> 00:42:53,780 Now suddenly it becomes, well, 667 00:42:53,780 --> 00:42:57,080 we're not gonna get any sales until the security is in here. 668 00:42:57,080 --> 00:42:59,270 Well, now the security's obviously gonna go in there. 669 00:42:59,270 --> 00:43:01,620 But because they say, no, it's fine. 670 00:43:01,620 --> 00:43:03,570 It's just a little thing, we can put it in there. 671 00:43:03,570 --> 00:43:08,000 We're like, okay, we've got all these other features to build, the security is 672 00:43:08,000 --> 00:43:13,790 gonna get knocked down to the priority less until someone says it's a priority. 673 00:43:13,790 --> 00:43:17,020 >> Well, they trust you, they trust you. 674 00:43:17,020 --> 00:43:20,599 >> [LAUGH] >> I'm always very upfront about 675 00:43:20,599 --> 00:43:25,004 the vulnerability that are in the system that we're putting together. 676 00:43:25,004 --> 00:43:25,711 Look, this is the situation. 677 00:43:25,711 --> 00:43:27,436 We need to look at this and this and this. 678 00:43:27,436 --> 00:43:30,164 Okay, how big a deal is it? 679 00:43:30,164 --> 00:43:32,648 Well, probably not that big a deal, only in one place. 680 00:43:32,648 --> 00:43:38,210 But then- >> It's at a point now where it's like 681 00:43:38,210 --> 00:43:42,638 if you make a product that blows up or you sell someone else a product that blows up, 682 00:43:42,638 --> 00:43:47,720 if you're that you're liable in some way, potentially. 683 00:43:47,720 --> 00:43:49,500 You know about that, 684 00:43:49,500 --> 00:43:53,110 your lawyers, everybody involved knows to check for that. 685 00:43:53,110 --> 00:43:56,790 It needs to get to the point where there's that structure around security and 686 00:43:56,790 --> 00:43:59,360 people understand that it's just not a liable thing. 687 00:43:59,360 --> 00:44:03,620 There's way too much potential cost there with 688 00:44:03,620 --> 00:44:05,392 shipping something that has a security problem. 689 00:44:05,392 --> 00:44:08,749 It's starting your own kind of fire, and 690 00:44:08,749 --> 00:44:13,455 we need to get to a point where everybody knows that's a really bad idea. 691 00:44:13,455 --> 00:44:15,142 >> [LAUGH] >> I'm not gonna sell the product if you 692 00:44:15,142 --> 00:44:18,000 can't assure me that it's not gonna start a security fire. 693 00:44:18,000 --> 00:44:21,367 >> Well, getting back to another example, the intersection 694 00:44:21,367 --> 00:44:26,370 of systems engineering and system science, I was thinking about systems thinking. 695 00:44:26,370 --> 00:44:32,720 But, I mean, security is kind of a subset of things like trust. 696 00:44:32,720 --> 00:44:34,380 What do you guys mean by risk? 697 00:44:34,380 --> 00:44:35,760 What we mean by risk? 698 00:44:35,760 --> 00:44:37,180 What does the customer mean by risk? 699 00:44:39,030 --> 00:44:43,140 And so, I mean, there's just a ton of higher level of 700 00:44:44,540 --> 00:44:50,160 higher level thinking that we should be thinking about. 701 00:44:50,160 --> 00:44:54,260 You guys, I mean, it's just part of your DNA, and 702 00:44:54,260 --> 00:44:56,176 I see that as being very important. 703 00:44:56,176 --> 00:44:59,788 >> I mean, we're being forced, so I guess with us, with systems engineering, 704 00:44:59,788 --> 00:45:01,860 it's more of a reaction, right? 705 00:45:01,860 --> 00:45:05,590 So more and more you're finding systems engineers happy to talk to the legal folks 706 00:45:05,590 --> 00:45:09,570 because they're doing that business deal with the global supply chain. 707 00:45:09,570 --> 00:45:12,240 And they're saying, well, what happens if this screws up? 708 00:45:12,240 --> 00:45:15,166 Well, here's the points where it's gonna screw up. 709 00:45:15,166 --> 00:45:19,264 I mean, I'm an integrator so I understand where the weaknesses are. 710 00:45:19,264 --> 00:45:24,478 So we should probably have a little side course in how to deal with intellectual 711 00:45:24,478 --> 00:45:29,300 property and also what lawyers are gonna want to know and due diligence. 712 00:45:30,900 --> 00:45:32,306 Not everyone but, I mean, 713 00:45:32,306 --> 00:45:35,549 so I have another part of this talk was kind of the educational, 714 00:45:35,549 --> 00:45:38,932 how can we offer something that'll be of value across the board. 715 00:45:41,383 --> 00:45:46,450 You can add in things, but the basic premise is still [INAUDIBLE]. 716 00:45:46,450 --> 00:45:48,912 >> Keeps John and me in business. 717 00:45:48,912 --> 00:45:51,680 >> [LAUGH] >> Should we adjourn? 718 00:45:51,680 --> 00:45:53,071 >> Yeah, let's wrap up. 719 00:45:53,071 --> 00:45:54,242 Thanks for the John. 720 00:45:54,242 --> 00:45:59,037 >> Thank you. >> [APPLAUSE] 721 00:45:59,037 --> 00:45:59,754 >> Thanks very much for coming. 722 00:46:02,463 --> 00:46:03,157 >> Thanks, John. 723 00:46:03,157 --> 00:46:03,717 Okay. 724 00:46:03,717 --> 00:46:07,176 >> There is too much there really [LAUGH].