 Good afternoon. I think the last session was so energetic that folks are over there talking about getting on the lingering conversations there. But let's start on time here. We have a very good lineup of speakers who are on the top of there in the industry. There's a lot of good work that they have done and we want to share with it. So the basic question they're trying to address in this panel is how to use cloud-based controls for building automation system. What are the benefits there? What's the cost there? Why aren't people using it? Why should we be using it? So there's a whole host of questions that we have. The opportunity is humongous, as I understand it. But the industry is not using it right now. So what are their holdbacks? How can we make this come true in the near future and make sure that we take the full benefits of these technologies? I was talking to our team here. I said, I would like to see a 10% of the building controls go to the cloud market in the next 10 years. That will be significant progress. Right now, I think it's not even a fraction of a 0.0001% kind of a thing. So it's going to be a tremendous progress in that area. And the opportunities are great and they'll talk about what those opportunities are. So we have four outstanding speakers here. He's our ECNO. He leads the R&D activities at SRI International. You have the speaker BIOS there by the program director, focused on infrastructure, security for government and commercial buildings. Very ap-topic because some of the issues that I've heard from people who are not using this technology fully is because of the security concerns. We take off from there and we'll start. All right, sure. Thank you. Yes, my name is Ulf Lindquist and I'm with SRI International, which is an independent nonprofit research center just around the corner from here and was founded in 1946 as Stanford Research Institute, but it's independent since the 70s. So at SRI, we take pride in working on important problems, not just interesting problems. And cybersecurity for anything that has to do with critical infrastructure is of course an important problem. So I'm the token cybersecurity person on this panel and this is why you may think that you have walked into the wrong panel because I'm going to preface the rest of the discussion with security and why it's a major concern and why bringing in the concept of Internet of Things distributed sensors with cloud back and can actually provide opportunities for better security, but certainly could bring risks as well. So a couple of points that I wanted to make is sort of the takeaways. Security is really a hard problem. It's an emergent issue. It's constantly changing and it's really a consequence of how we use technology. And I think it's important to keep in mind that security must be managed. It's not something that can be bought and forgotten. It's not something that can be sold. Yep, now we took care of security. Tuesday we can do something else. It's always going to be there as technology is constantly changing as our needs, our business, the threat picture changes. There's going to be changes to security. Let's see if I can bring up my notes here. And as we see this emerging Internet of Things with all these distributed sensors where everyday objects including things in building control, whether it's HVAC, occupancy sensors, all those kinds of things, structural sensors, they become networked, they get computing power, things that they haven't had before. And of course we're seeing this in lots of sectors, whether it's manufacturing, transportation, as we just talked about earlier today in the panel about the electric grid and so forth. And we're going to see many more network devices than we've ever seen before. This is connectivity on an unprecedented scale. We've never had tens of billions of devices connected to the Internet. At the same time we're seeing trends that relate to more automation of attacks. The attackers of course use the same kinds of tools that make us more productive and they use that to make themselves more organized and their attacks more scalable. We see destructive attacks. You probably all saw the news about the ransomware that hit on a global scale just a few weeks ago. Destructive attacks have been seen for example in the notorious attacks on the Ukrainian power grid. It's an excellent article in Wired just published a couple of days ago by Andy Greenberg on how that malware actually went in to destroy the boot records of computers so that it would take a lot of effort and a long time to get those computers back online again. We see cyber crime as a big business. We call organized crime organized for a reason and there's a lot of money to be made both in fraud and in extortion using cyber means. And we've also seen this proliferation of cyber weapons where malware produced by various nation states somehow makes it out onto the Internet and as opposed to conventional weapons these can be copied and used by others. And we're seeing a lot of well resourced attackers whether they're organized crime or military or intelligence services and in some parts of the world it's very hard to tell the difference between organized crime and the nation state itself. So one fundamental thing about security is that I see basically security as all about separation. We want to separate the good things from the bad things, authorized users from unauthorized users and so forth. And this notion of separation of course runs completely counter to our trend to connect everything to everything else. And the good business reasons for all this connectivity where we can gather data in ways we have never been able to do before we can remote control things and so forth. So connectivity is definitely here to stay. We're not going to go back to isolated systems and so forth. But every connectivity also brings a path for an attacker into a system and we need to actually build systems that provide security in this fully connected environment. And that's some of the things that we are working on in the computer science laboratory at SRI where we develop new security measures for the Internet of Things for example. We also see this notion of security being left behind in the rush to get new things to market and that's one of the major reasons why the state of security is so poor as it is in some cases because we rush things out there, oh yes, we'll deal with security later and it turns out that later never really occurs. And you can argue that that's sort of the case with the Internet itself. It started as a research experiment as SRI was node number two on the ARPANET and it was originally just a network between sort of institutions that trusted each other, a small group of researchers who all knew each other. But as the Internet has grown security hasn't really been been added and it wasn't part of the original design. And what we're seeing now with a lot of control of physical processes, whether it's building controls for the electric grid or self-driving cars or something else, is that the computerized systems can have a very serious impact on the physical world. It relates very closely to safety. It's one thing if your email sort of is unavailable for a day that could be bad for your business but no one's going to die. But if we're looking at systems that control volatile processes, controlled cars, etc., all those kinds of things that can really have an issue for safety. And in some applications and particularly, for example, in building control systems, you have privacy aspects. If you have very fine grained occupancy sensing, for example, you can figure out exactly how much time Bob spends in his office versus the break room versus somewhere else. And you can draw all kinds of conclusions around that. And one thing that's often brought up when it comes to smart homes, for example, is that it could be easy for someone who steals information from the system to see whether someone's home or not. So my concluding remark here is that we need to be at least as good at building secure systems as we are building safety critical systems. We already have systems that run airplanes, for example, and airplanes fortunately don't drop out of the sky generally. But we have systems being broken into all the time and we need to be at least as good as making them secure as making them safe because we cannot have safety without security. Thank you. He has given a very good background in terms of security requirement and which is one of the hesitancy I've seen from end users, consumers, to really use the full benefit of cloud control. So the way we have figured it out is we are going to have three more speakers. We have Kevin here, Kevin Fascinelli. He's the executive vice president at Dicen. And how many Dicen applied, I believe? And how many have you have heard about Dicen applied? He'll give you a background about that. That's one of the largest HVAC company in the world today. And then this will be followed by, and he has been, has a illustrious experience with Johnson Controls and some other areas. So he knows about the controls, which are the biggest issue when you face this one. And he's very innovative, very creative. I had some conversation with him and his thinking is always looking into five, 10, 15 years from now. So he'll share some of the ideas. If you'll be followed by presentation by Michael Franco. Michael is the CEO of a technology company called Riptide IO. And they are developing the controls architecture for some of the HVAC equipment, starting from security as the first one. And they have several big box details that they are providing today, these services. So he will talk about how they have been able to overcome and how they can address the issues of customers. And this will be followed by Mitchell Camel. He's the CTO and the founder of a company called Melrock. And they have been also doing cloud based controls. In fact, they have a recent award from the California Energy Commission to monitor and to control five buildings in Claremont Colleges, I believe. So he'll talk about his company, how they have built their software, their product from the ground up with the security in mind and how they have been able to reap advantages of these controls. Giving it to Kevin next, and I'll load your few slides they have. And after we do about 30 minutes of presentation, we'll go and have a discussion. So please keep your ready, all your questions. And all right, that's fine. I just want to move this. Okay, wonderful. Well, I'm disappointed that more hands didn't get raised when we asked the question, who is a Diken? But I'm not surprised. Diken is the world's largest HVAC company. Number one market share in Europe, Asia, Australia, and a new entrance into the market in the United States. In 2006, they acquired a company called McQuaid. McQuaid is the applied side of Diken products in North America. And then in 2012, they acquired a company called Goodman. Goodman and Amanda, the world's largest producer of residential air conditioning in the United States, literally, by far number one in installation of air conditioning in homes. So it's a it's a big company. It's making a push in the United States market. And it has a 90 year foundation really focused only on HVAC. We don't make golf carts. We don't make helicopters. We don't do anything else but HVAC. That's it from top to bottom. So my deck has a lot of information in it. I'm not going to go over all that. But I feel like when I come out, I owe it to the folks to have the ability to read through that and think about what it says and make your own interpretations. And a little bit about Japanese style, you end up putting a huge amount of information on a slide. So the stuff you're going to see today is not even close to what I have to do on a regular basis. Again, largest HVAC company in the world. We also launched the first IoT based control system, working with Intel, starting about five years ago, and have that commercially available on our products out in the market. We are focused on energy efficiency. So we came out with the first 20 sear rooftops and rooftops are the boxes that are on top of the majority of buildings in the United States. They use 80% of the energy for HVAC in the United States is rooftops from strip malls all the way up to large Walmart. This is the way in which we air conditioned buildings. We came up with a product that moved the paradigm from just a minimum sear requirements from 10 to 11 that seasonal energy efficiency to a product that came out at 20 sear, twice as efficient for the same BTUs. Great product very successful, won an award from the DOE on their energy challenge. We also have products that use magnetic bearing and oilless systems. So these are very high tech frictionless systems don't have the maintenance or the problems in terms of oil and other management. Quickly, we have a scope of just about all the products you need. So I'm not going to go through all what those are, but they're in there. Pretty much the whole envelope of what you need for your building or buildings is available from Daikon. Why to talk a little bit about this? Why are we talking about controlling buildings from the cloud? And why aren't we doing it? So to set up kind of a framework, you got to think about what are some of the things that are kind of getting in our way. So the first one is cost. So when we think about Moore's law, we all know, we probably don't all know that Moore's law says every two years, you double the amount of transistors and keep or reduce the cost of that. Now that's been going on for a long, long time. Today, you can get products that have a processor, Wi-Fi radio, multiple, call it general purpose IO for less than $10. You can get MEMS. These are sensors that can do multifunctions, miniature electromagnetic devices, technology that could do humidity, temperature, lighting, less than $10 the size of your fingernail. The technology right now is ready. The cost structure is ready. The next one is Metcalf's law. Maybe you do or don't know Metcalf's law. It just says that the value of a network is the square of the number of nodes on the network. So you can think if I had two telephones, that not too valuable. But if everyone has a cell phone, much more valuable. It's not totally logarithm. It flattens out a little bit. But in general, that problem has been solved. We have the ability to connect devices. Now the other problem is this idea of jumping over to a new technology. And that's typically called, you know, getting over the chasm. So we have early adopters that figure out how to use this technology. But at the same time, there are folks that are waiting. And that's really about being pragmatic. What is the value? What are you delivering to me? Not just the technology, but tell me why? Why do I want to do this? And what is it going to benefit me and my company? We'll talk about that. When you put all these together, right now is a very, very ripe time and has been for a number of years to implement this technology. I'm going to talk a little bit about how we've come through this journey. Not going to read through all this, but this is what our customers are telling us. This is what they want to solve. And these are the problems they see day in and day out. To break it down very quickly, there's two components to a building, you know, HVAC. One is the actual equipment. Okay, that's what's used in all your energy. It's a vapor compression cycle. It uses all the energy. So we focus on efficiency. Make the equipment as efficient as possible. So think about a car. Make the miles per gallon as efficient as you can. Now we think about controls. Controls is how you use that. How do you drive that car? How do you operate that asset? That has the ability to create efficiency as well. And it's a difficult problem. It requires domain expertise to be able to understand the protocols and the way in which this equipment can function most optimally within a building. It's not as simple as plugging it in and turning it on and off. So that's one of the reasons why I think a lot of people haven't got into this space as quickly as the domain experience has to be there to be able to operate buildings efficiently. There's just a lot of savings that can be had through the ability to understand the operation of your equipment real time. And when there's a problem, as well as be able to constantly commission or tune your equipment for the environment. Now to do that, that's difficult. You do it in your car when you're driving. You're continually monitoring and driving your car hopefully as efficient as possible. Maybe not depending upon where you're going, but you're always doing a calculation. So the cloud's very good at that is to do a principle-based calculation, constantly calculate what is the best place for me to be and what should I be doing, and then looking forward with foresight at where I'm going to go. Where do I need to get to, not where have I been? Controls today is all about what happened in the past. It has very little ability to look out what's going to go on in the future. The cloud brings in through APIs and other information, conditions and situations that I can predict in the future and then prepare for it in the most efficient way. So what we did is we started, we kind of flipped over the paradigm. This was fundamental in our thinking is before everyone was like controls are the things on the wall, the mechanical equipment's up there, and the two, one tells the thing what to do and the other one reacts to what it's being told. From my perspective, it didn't make a whole lot of sense. If I'm using all the energy in the equipment, why not control that equipment and the environment from that same space? What's held that up is really the cost of processing, the cost of networking and the cost of computers, but we were able to overcome that through some gateways and put together a comprehensive system where we extract, this unit here is a 20-series unit, we'll extract 4,000 data points from this. Now, a lot of those are static, some are calculated, but that data goes into the cloud. The vast majority of those are put in the cloud and stationary or non-updated, but about 400 points can be used within analytics to constantly tune the operation and the environment and the condition of the equipment. That's where we start using big data and analytics and algorithms. And then we present that through an HMI anywhere, anytime to a user, bi-directional, secure communications and to the operator so that if I am operating a building, I don't have to be in the building in front of the panel to make my change. I can be physically anywhere understanding that building and understanding that asset, as well as tuning and understanding the health of the asset. Again, these are some business benefits. I put them here not to go through them all, but for to give you guys some sparks and thoughts about why we're doing this and what the advantages are. We break our application really down into, probably better up here, two areas. One, I'm a technical person and I want to get the most out of my product. That's where we use 3D modeling. I can see exactly what the product is I'm working with, all the features, and I can do pretty much every function that would require me to be on the roof remotely. We're finding about 40% of problems can be solved through software alone. It does not require the role of a truck or that asset to be touched. And then the owner's view is really I am a owner, operator of an enterprise of building or a building. And I'm concerned with how am I doing in terms of Energy Star? There's a portfolio management product out there that we integrate with APIs. I'm interested in my comfort, demand response, just all the things that help me operate the building better. And this is all in the cloud and accessible. One last thing to think about is the security. So we started with security first and that's why we got engaged with some of the better people out there, Intel, around what can we do? And we said we want security not to be an issue with our clients. So we took the same security model that's in their highest performance eight-way server and put it in this product. So at the hardware level, we have a secure trusted boot, which means I can't reboot and change my boot image. We have a device extension, if you know what that is, we have white listing. We have encryption and compression, certificate authority management, buffering and flow control. And then we run algorithms in there to see if there's any time that this equipment is outside the normal boundaries of how it should be running, meaning it's being attacked or something's occurring and it will shut itself down. So we took a very proactive approach knowing that this was something we wanted to take off the table for our clients. So I'll turn it over now to Mike. Thank you. So there was a very good metric that 40% of the problems can be solved simply by this software. And if I remember correctly, the energy savings that I've been hearing about through these technologies are like 30 to 50% of the energy can be saved. So we combined these two together, this is a fantastic opportunity and he touched on some of the opportunities there. So cybersecurity now. Alright. I'm going to try to stay in my lane here. And what I mean by that is Riptide is a young company with a long history of controlling buildings remotely. And by saying that, what I mean by that is Riptide itself was founded in 2012 by a team, four of us who came out of Cisco systems. Prior to Cisco systems we were with a controls company called Richard Zeta. So we've been doing what used to be called M2M and now could some people consider that's the old label for what IoT is. We've been doing this since the early 2000s. I bring this up because I want to just give examples of what we do for cloud based controls and how it runs and why why it's a good thing. But first, so we've done this, we've put cloud based controls in probably between 4,000, 4,500 buildings. So that's, who is saying that about the percentage, right? I mean, if there are five million commercial buildings in the US and we've done 4,000 of them, it would be interesting to find out how many and see if it really is that one or two percent. But it is a small number. What our lane is small commercial buildings and it's important to make that distinction for our use case because in a large, you know, a skyscraper, one big tall shiny building, you can afford to have a building engineer sit in the basement and keep that thing running, right? Our model is for buildings that are say 50,000 square feet and smaller and are part of a geographically dispersed real state portfolio where it makes no sense to have a building engineer sit in that building. And so what we try to do is through the technology give the same capabilities that that facility guy or the building engineer can have working in the basement of a skyscraper, have that remotely from a centrally managed method, right? So and what I thought I'd do here is just, I have some slides here that go through, similar to what Kevin was saying where I've used them in other decks. But the important thing that I really want to get to is just to show you a day in the life of how cloud based building automation works. Let's just get to that. And then so what we do is we have this hub. So there's a gateway right now that goes into a building. It connects to the various systems within a building. Could be HVAC, could be lighting control, meters, sensors, refrigeration. And in our largest deployment, I mean, they went crazy, right? It was 18 submetered loads. It was, we did irrigation, we were monitoring fire and safety in addition to what you see here with HVAC lighting, refrigeration and more. And people think that small buildings, especially small buildings with national account retailers, are all cookie cutter. That they're easier to manage than that large skyscraper. But actually that's not true. I mean, there's actually just as much complexity in the systems that are in these small buildings that are in those big ones just based on the domains that you need to touch if you're trying to do some cloud based control of the whole building. So this goes in, has supervisory control over HVAC can, you know, turn, can set the temperatures to whatever the occupants want, turn them off during unoccupied hours, do the same with lighting, measure the metered loads, send those meter readings up to the cloud. We can have some logic that goes in place whenever the, you know, CO2 sensors that are measuring what the CO2 levels can, that are independent from the HVAC, can affect the HVAC to open up fresh air, fresh air access, bring in fresh air and make the, make a building more comfortable. And also increase the efficiency too. So all, that's the integration part where you put everything together. So that's the structure of what goes into our cloud based building automation, our cloud based controls. So what I have here is, I'm just going to take you through three slides. And the three slides show a day in the life of a remote centralized building manager, or facility manager, right? So, so what we have here, map of the US, selection of all the buildings that they're trying to manage. And then here are key performance indicators. So these are the analytics, you know, these are the analytics that are used that they want to run against an entire scenario. And before I get into this, let me just say, cloud based building automation isn't looking at a list of all my stores or sites or buildings, and then clicking on that and going in and doing the control on that. I mean, that's all possible, but that's really inefficient. If I have 2,000 stores, if I have 2,000 stores, how am I going to manage 2,000 stores by picking, you know, going in like that? And so that's possible, but that's not cloud based building automation. Cloud based building automation is running metrics and analytics on the aggregation of what you see and how the building is being run. So when I mentioned that middle column there of KPIs, these KPIs are customized for how a building portfolio manager wants to run the place. If you're trying to squeeze the most dollars out of your energy consumption, you're going to keep the temperature a little higher, it's going to be less comfortable than you may like, and do things with the lighting, et cetera, that maximize energy is your goal. This particular customer wants to keep their store at 72 degrees, regardless of what it's like outside, and because they're maximizing the comfort of the shoppers that go into this building. So because of that, they're looking at things like this one, for example, are there any air flow failures, are there any air valve failures, are there any VAV or fit boxes that are in a critical state? Those are, are there, where they're running heat, are they running heat or cool at night? What's the outside air flow like? So these are things that are important to them to make it run, to keep their buildings running well. So they pick one of those, run it against all their stores, and with a bunch of green dots, I picked lighting emergency override run time on this particular case. So what that meant is that they are, which stores have turned on the lighting override and left it on? I mean it's a real simple one, but it happens all the time and they have a threshold of the amount of hours that it needs to go beyond in order to trigger, whether it's a bad site or a good site. So this map comes up, shows me the sites where they need to start paying attention to, and then they zoom in on this Houston store here, where it came up with, you see here we've got this zone, zone 10, has something that was running 77.5 hours staying down here in zone 1. So the guy, the centralized manager gets on the phone, talks to the local guy and say what's going on over there. Sometimes there's a good reason for turning that override on, but lots of times there's not, and this is just how they run their day. So this is, there's a lot of noise in what we do with building automation, and by that I mean there's a lot of promise to do things that are 5 years out, 10 years out. And what we're really finding is there's a lot of low-hanging fruit with the most simple things that you can do right away. Things like just people leaving over rides on, running, not seeing the faults in equipment because it's in the roof, on the roof, and you can't see the problems as well. So what we try to do with this, in a sense, make it easy for a centralized manager to find out what those faults are. And last slide I have is just one right here. This one right here, the point is these are all common faults that happen in commercial buildings. If you look at the impact of the top, say 3 to 5 of those, those are easy to detect centrally. And let's not get caught up into looking at all these things in the bottom of the list. Let's focus on the easy things to detect, deal with the low-hanging fruit first, and that's the way to get efficiency from cloud-based building automation. So as we go through here, my lane is going to be having actually done this in more than 4,000 buildings. And what works and what is not, what's wasted effort. So when it comes to the discussion part, that's where I'm going to be focusing. Thank you. You have done about 4,000 of these box stores. Now, Michel is going to talk about, I think how he has done several hundreds, maybe 1,000 stores or so. So he's going to talk about his solution that he has been able to hone in and how he's going to do this project under the California Energy Commission grant. Okay, I did anyway. Give it to Microsoft, change menus every day. Do we have an expert over here who can, okay, hello, multi-slides. You go to the side. At the top. Oh, okay, it's actually... Oh. Did you reduce it? Actually... Yeah, that's fine. Yeah. Thank you. Okay, so my name is Michel. I'm with Melrock. We are a relatively young company. I started my garage about 2008, like a lot of other Silicon Valley companies except my garage was not in Silicon Valley and I stayed there about two, three years, but then in 2011 it really took off. And what we focus on is data and control messaging, basically communication between energy devices and buildings and the cloud. And I make a clear distinction between data messaging and control messaging and you'll see why as I go through the slides. But a lot of it has to do actually with security and cyber security issues because the requirements and what the applications of this are multi-fold. Sometimes customers just want to be able to monitor what's going on and sometimes they want to do ROI calculations. But there's more and more talk about demand response and about distributed energy resource management and of course what we need to talk about today, which is cloud managed devices. Unfortunately, we are in an industry that is a bit kind of, forgive the expression, it's not too sexy. If this was the car industry, for example, this would be all over the front page news, self-driving cars and you'll have a top model, not me presenting it, but you stuck with me. So we're getting to the age of self-driving buildings and in a car commercial they get nice looking guys and girls and they're driving the car, but they're driving the car. So next time you're stuck on the freeway, whether it's the 101 or the 280 or the 580, remember you're piloting your car at 5 miles an hour. So buildings, we call them energy managers and I did some calculation after I sat with Mukesh last night and you told me 3 miles per kilowatt hour, where's Mukesh? So you said 3 miles per kilowatt hour, right? So you have that and I sell it on number of 2 to 3 miles per kilowatt hour and if you drive 12,000 miles a year, that's about 4 megawatt hour per car, electric vehicle car that you use every year. Well, your 30,000 square foot building uses 5 to 600 megawatt hour per car, per year. That's about up to 150 cars. We don't have 150 occupants in a 30,000 square foot building. The other point is the building energy manager versus the car pilot is actually managing a lot more energy than the car pilot is, right? So we should pay attention to it and we should enable that energy or the building pilot with as much resources and capabilities and computer power that we can and hence we got to go to the cloud, right? So I'm going to talk about our first deployment of a building or cloud-managed building at Pomona College. This is the project that was funded by the California Energy Commission and we're going to take 5 buildings on campus and literally disable the local energy management system and put a gateway in there and totally manage every device from the cloud automatically with no energy manager there. And it's going to be exciting and we are on the record with 20% energy, I think we're going to save a lot more. But if you think about it, if we can save 20-30% energy by going to the cloud we may not need the Paris Climate Accord and all the headaches that come with it. So in summary I'm going to talk about this and I'm going to talk about the platform, focusing on the platform because not all cloud solutions are born equal. A lot of people will take something and now you can see it on the screen. So the reason I'm going to talk about this and going to the cloud why there's a lot of reluctance is because there are no real standards on going to the cloud. What is it involved? What are the security standards? What about the open interfaces and what not? And there's also another reason that I want to talk about why we need to go to the cloud and it's not only for the 20-30% savings that come with it that building belongs to the old grid. The new grid, if we want to support 20-30%, 50% and I think there's a bill now in California for 100% renewable grid we cannot have buildings that react the next day or a month later. We have to have buildings that react every second basically to the grid and every second interact with the buildings around it and with the grid and you can't do this from an isolated islanded energy management system. So when you talk about a cloud platform you really need to start laying the standards for, you know, I like to call it a CSI model, the cloud system interconnection, you know, it's kind of a spin-off of the OSI model that makes the internet possible. But you do need to think about multiple layers. First you have the interconnectivity layer where you actually talk to the different energy assets and that's a tower of Babel. If you've ever been in a building and if you've ever looked at the energy devices in there you've got at the very least half a dozen protocols that are being supported and potentially even half a dozen different physical interfaces, whether they're wired, wireless or pulse counters arm waving and so on. So you have all these different interconnectivity challenges to deal with these assets that exist and then we like to separate upload processes because if somebody hacks into your data upload process okay, there's some maybe proprietary issues to deal with and other things, but the damage is limited but if someone hacks into your control, your message delivery process, now that's a big thing, right? They can shut down your building, they can take down the LA grid just by hacking into the inverters, solar inverters. So we got to separate this plus the bandwidth requirements are different. You need a lot more bandwidth in the upload because you're monitoring 4,000 points, right? But in the download you may be sending one or two control messages. So we separate these and we have different requirements and standards for upload and download and of course you have to serve it. And all this has to be done on a secure platform, right? So security has to be the foundation on which you build these things and another thing that's big is the open services. We cannot build a cloud platform today with the legacy traditional culture of I'm going to name names, so please forgive me. I'm Honeywell and I do this and I'm Schneider and I do this and I'm JCI and I do this, right? We can't, we got to open the data, make it available because that's how advancements happens in today's age. So three simple steps. Connectivity secure big data pipelines and then cloud analytics. I love to show this because interconnectivity is really a bigger challenge than people think. I like this cartoon by Randall Monroe. They get together, hey there are 14 standards around this issue. Let's create another one that's universal and now we have 15 standards around this issue and this is not only typical to the energy industry but really every other industry. So the last thing I want to mention is security cannot be like Ulf said, an afterthought. It's got to be by design and 100% right, it's got to be continuously managed. So we're always thinking about adding defense layers like anything else, you always need to add defense layers to your platform. One of the things we've invested in is putting actually FM radios. So on your way here you may have listened to quite FM or some other radio station. We actually go and piggyback on those exact same commercial stations where there's like one or two massive antennas in every city for every station and we piggyback on it, we hitch a ride and we send our own control signals to our own devices. Now try hacking that from China or India or anywhere else right, it's practically impossible. And now when you couple that with the other communication channel whether it be 4G or anything else then you have a very, very robust system with kind of dual or multiple path of security and layers of security. And our routers by the way I don't have a picture of our router here, it's that little diamond thing on the bottom right but our routers do get hacked. They do get hacked almost, I'm an attempted hack, like whenever we put one at a customer site we literally get few times a second continuously right, that we actually have our own firewalls inside our router even though we're behind the customer's firewalls. So security has to be embedded in the design from day one because we're not going to have cloud management and we're not going to have big data engines without the data. And so you need the secure and robust data pipeline to the cloud. Thank you. I'd like to open up for discussion. Any comments you have, any questions you have we'll have for the audience. You can pick up any one person or ask in general we'll be happy to answer it. Yes, sir? Jerry? I know that the security is coming all this is great stuff. Do you folks who are deploying things, do you have penetration teams that test them out and go do on-site studies to look for human error? The reason I ask this is I'm a real ancient Bell Labs unit guy and I used to do security products. And we hear about a lab that said they have a great new password scheme and we do the odd, we'd say, yeah, your password scheme is good but there's two problems. You've got yellow stickies everywhere with the root password and the lab gets lost so you leave the door open. So I've always thought human laziness was, you know, probably the worst security problem. So talk some about the practical stuff that you do to try to look for problems like that. I'll take the scar of this. Very good question and Ashley, we're really lucky enough because we're at the stage where you're not working with mom and pop shops, you're working with large retailers, we have from Fortune 1 companies as our customers to, you know, Fortune 100 companies and we do our own testing, we have some third parties but they do the testing too. So before we put any gateway in any one of these customer sites, it has to go through the IT department and we have our own exchanges with them and there's two things, it improves the robustness of our system but also helps us to quickly evaluate who's the smart, you know, who are the good people who are not doing the testing. So it's a surprise, there will be some really large companies but the IT department is not the best and then you'll have another company and they're doing everything beyond and beyond to do this. So to answer your question is yes, we do some of it ourselves but we also rely a lot on the partnership with our customers where we send them the unit ahead of time, anytime there's a new firmware upgrade or anything like this and then they really put it through and they usually have more money to do. One of the things we do is third party testing on all of our security models and stress it and look for any issues. When we selected our topology we have in our gateway 3G, 4G, plus Wi-Fi plus Ethernet. So literally we can run off corporate network directly into the cloud and stay isolated from the IT department entirely which is very attractive if you've ever walked into an IT department and asked them to push building data around their network. They're very risk adverse and they tell you the first thing is no and then the second answer is no and that's really the problem with a lot of these implementations is the IT department does not want to take the risk, the reward is not explained great enough for them to take that risk. So we take that kind of out of the equation. We're also working with another company that just received the first gateway device to UL 29 102-2 which is the UL cybersecurity standard first IoT gateway to get that approval through the UL for their security model. So yeah, we're very cautious as we're a $20 billion company we do not want to be the problem for our large enterprise customers in terms of their security systems or any of their data. So we try to isolate as much as possible from those networks. So let me reinforce what you said about the IT department's hesitancy to let the building data flow on their network. One of my previous job, I started a project in 2007 and by the time I left in 2014 we had gone through hundreds of meetings and it was very difficult for me to identify the IT department. And this was one company that was saying everybody else you should be moving your applications to my cloud. So you can think about what goes on there. Question please. So what's the penetration rate this is actually for Kevin. What's the penetration rate in the United States for IOT for buildings for what you do with IOT? The second applied was really kind of the spearhead within the diken industries for IOT cloud based control system. So we are starting to see more and more of the orders coming in with this feature enabled. Our next generation really is about putting this feature into equipment at a zero cost position. So there is no upsell. The gateway built in and one doesn't you know you're not going to pay more for that. The expectation is it's there you may use it or not and that's our philosophy is to have it on 100% of everything at least the capability to be activated if the customer desires. What you're saying is that you're abandoning the way you used to do it. Is that true or not true? Are you going to IOT? Yeah, we're primarily a hardware company. We do have some controls capability but when we look at our controls, building controls portion of the business, it is microscopic compared to Honeywell or to Johnson Controls who I used to work for. So it's really kind of instead of investing in DDC layered control for a building, we said let's invest in future topology based upon a cloud in a much narrow IP stack from sensor to cloud. So we take out multiple field controllers layers and that allows for a much more affordable system and if it's architected appropriately, you don't necessarily lose any of the redundancy or control capability. If anything you gain some of that based upon where you place that if it's on the edge or in the middle device layer. I think field controllers are passe once we go to IP there's just no reason for that. You still could have supervisory and that's really could be placed in the gateway itself so that supervisory control can be done locally as well as advanced functions done in the cloud so if you lose the cloud you don't lose control of your building. Two months ago in April at the American Physical Society we held a high level workshop two days in Monterey and the title was the actualization of all the other things and it really got down to the grizzly details. I can give you all the talks that I'm on and you're pretty interested in following up. Now we had DOD we had several of the MAT Labs California universities GE, IBM, Bowen and a number of other organizations that I'm interested in right now but what I wanted to bring up there was a lot of concern about security and communications particularly between sensors and data collection devices and two topics were brought up to which there is not yet quite a solution. Two factor authorization and quantum encryption. Now is this community looking all the way? Yes. We're actually working on both of those issues how to do secure authentication especially between devices because one of the speakers mentioned what used to be called machine to machine is now called IoT but it's really a lot about devices communicating with each other with little interaction from people and the other issue about quantum safe crypto for example. So this notion that the encryption mechanisms that today really form the basis for the protection of our data in all e-commerce and so forth could easily be broken supposedly through a quantum computer when such computer actually materializes because they're built on the fact that some problems like factoring are very large integers is a difficult problem but it's much easier with quantum computers so the research community is working on post quantum encryption technologies and our team at SRI for example is very involved in developing that so there are other theoretical constructs other than factoring of large numbers that can be used but that's being developed actively right now the NIST the government's standards agency is running a competition for post quantum crypto functions that we're participating in and so forth so yes we're looking at that but there are other threats to our crypto mechanisms than just quantum computers it could be that any time now a big mathematical breakthrough could happen that makes it that has the same effect so that's something that always needs to be considered that we can't hard code in mechanisms we have to be ready to swap in what new security mechanisms is needed I think for the depending upon what the information is that you're transmitting and the importance there's different techniques obviously I mean if you look at blockchain as one technique that is extremely difficult because of the way in which it's architected it's distributed and it's decentralized that type of technology may be appropriate for certain functions within this architecture going from a MEMS to a an HVAC device directly I'm probably not going to use blockchain on that but there are definitely a lot going on in this space they said the new UL standards and I think it's a very appropriate question for anyone getting in this is what are you doing to secure your customers data the sort of bigger aspect of the question is here given the type of systems that are put in place in buildings and so forth as we all know they tend to have very long lifetimes they tend to be in place for decades perhaps so we need to make sure that we're forward looking when it comes to not just the functionality but also security again that it's possible to upgrade security of systems etc and of course moving a lot of the security functions to cloud based solutions makes those kind of updates easier one fundamental question we haven't asked as we talk this is all about the cloud so what is the cloud really one can see it as your data resides in someone else's computer the fundamentals of it a computer that seems to have unlimited capacity for storage and computation and there's all those kinds of issues that comes with actually having your data on someone else's computer and you mentioned for example making sure that the buildings can still be controlled even when connectivity goes down so things of that nature it needs to be considered as well thank you thank you excellent discussion I'm Imanath from APRI one of Mukesh's colleagues very good discussions but you kind of touched upon it but can the whole group talk about kind of the big picture in the sense that if Mukesh initially said you know it's a very very small percentage of buildings are controlled by the cloud so if that expands what it would do to efficiency on a large scale for the whole country and even expanding beyond what it would do to climate change carbon dioxide emissions to power plants how much reduction they would be in kilowatt hours so can you talk a little bit about the big picture on where you think this is going this is to every one of you and the second item is I didn't hear anybody talk about customer comfort productivity improvement with doing these things with new technology of controlling buildings so two questions one is big picture on carbon emissions the second one is on customer I'll go ahead and start with your second one when I was showing those screens they were all for customer comfort and it's a great point because what Michelle said about us not being in a sexy industry is really true and the closest we get to sexy is to talk about energy savings but that's not fundamentally what drives our customers what fundamentally drives the reason they buy our system is comfort and maintenance and repair so over the life of equipment the amount they would save on energy over the life of the equipment eight times that amount on repairs and maintenance so if we can have an impact on repair and maintenance that has a much more important factor there and then the second thing is comfort absolutely we can save all kinds of energy by turning off the heat or off the cool but we're in retail and that's going to chase all the customers away it sounds silly but take it to another level of comfort it's not just temperature it's humidity, it's CO2 and other factors like that these kind of systems can bring that in run the analytics on which places have high CO2 levels where there are humidity issues and other things like that I can maybe break it down to some really tangible examples for you all education today people are looking at how well kids are learning and schools are rated on how well kids are learning and the environment which they're learning in has a direct impact on their ability to absorb information and process it so when we're doing our best job, we're invisible you don't hear us, you don't feel us we're not distracting your senses away from what you're focusing on so that would be in the education market in the healthcare market I guarantee you within a week you'll get a survey of 40 questions because the healthcare industry also gets paid on your satisfaction and a good portion of your satisfaction is your doctor and care but also when you're hospitalized in that building there's tangible things that you recognize as being a good environment or a bad environment for example in the hospital administration you talk to someone that runs a hospital they're number one concern number one infectious disease control so the environment the environment has a big impact on infectious disease propagation so Mike said humidity temperature this is bottom line dollar and cents for so many industries you just have to step back from it and see how you big picture wise I'm an efficiency believer I don't think the cloud there's so many tangibles but really it's about more efficient equipment so people aren't talking about training you to drive your car more efficiently you're pretty much saying hey get a more efficient car that's going to be the bigger impact because you may be able to squeeze a couple miles per hour out on how you drive it but you're never going to get a 20 mile an hour gallon car to do 50 miles per gallon this is possible thank you let me add two things so first I want to reinforce what Mike said I think actually the pendulum will swing the other way in terms of customer comfort will probably increase you see over on one of our customers I sat down for a few days with their energy managers about half a dozen people in a couple of containers and I'm looking at they have like five screens in front of them with every possible data point and I quote trend and I say that's impressive I go how do you know something went wrong and he looks at the phone he goes the phone rings at the end of the day the phone rings and I think with the predictive analytics in the cloud and other things what's going to happen is the phone is going to ring less now I do agree that the equipment has to be efficient however you can have the most efficient chiller if you turn it on all day long you're not helping and there's an ectodal thing which is actually not an ectodal a lot of customers when they benefit from the lighting retrofits incentives with Southern California Edison SCE has stats on how many customers actually their lighting went up even though they put more efficient lighting and that's because they replaced the light bulbs and have them were burnt and they didn't really need them right so the point is yes you need the efficient equipment but you also need to use it properly and use it efficiently the worst thing you can do is pay a lot of money for efficient equipment and you have your economizer stock open on a hot day right and it happens all the time right last question thank you the most important question so in the construction industry today one of the most exciting trends that I've seen revolutionary trends is prefabrication panelization effectively and panelization allows us to add sensors and walls build mesh networks you can put them in the roof and you can design the ultimate energy efficiency system because you're not doing a custom job every time you go to a built site so I'd like to ask you guys how does that affect your respective businesses for mine I totally understand where you're coming from and we have an approach for our systems are factory mounted installed control systems are terminated like VV boxes it's just you plug the wire in and it's we discover it and it's based on the cloud and we're not we're not doing artwork on every building and that's what you're really getting at is there a best model for running this type of building for this type of application and if there is why are we allowing every individual contract or an architect to go a different route on how they're going to control that that's just how the industry has grown up but all of our stuff is let's just say ready to start up so to do a building control job typically the expense is binding of the points and then setting up your sequences that you've chosen as the contractor or installer that you think are the best that takes on an average building weeks our stuff you plug it in you run the application and there is it's all bind it already it's all together and now you're tuning for different case scenarios like Mike said you know I'm tuning for comfort or I'm tuning for energy savings and you know working on that we've also built in a comfort index so we look at not just 72 degrees we look at humidity we look at the time of year we look at the amount of activity in the space of humidity as well and we run analytics on that that's based on an AII standard that will spit out where you are in terms of your comfort index based upon six different variables so that we are not using too much energy or potentially ignoring the environment not using enough I'll say one thing about panelization I think that the installations we do are a mix of new construction and retrofit and when you can so new construction gives you all the flexibility and everybody in our industry loves new construction because you come in right away but the majority of the business is retrofit and in my lane of retail stores do refreshes all the time and panelization of the partitioning walls makes it easy to create new spaces and new sales floor over here or something like that but do you know like the dirt system do you know that one? so dirt is a great system for panelizing interior spaces makes it tough on retrofit for HVAC if you've got a duck system and it becomes real expensive because now you've got to redo everything there so I know there's a lot of duckless systems that make it easier but so the zoning aspect to get that maximized comfort if you're going to repartition it and be really flexible there there is a factor if I wish I could snap my fingers and do that rezoning together with the partition space that's a tough one thank you very much so I think we are going to continue this conversation there is a session after this one that's called conversation for action that will be in the main hall there are a couple of tables where most of the people will be available we can continue the conversation and you will hear something very surprising I heard last night and that is trying to maintain a building besides 72 degrees uses less energy less maintenance cost then a system like you will see in dollar general that is not maintaining the temperature very correctly uses more energy and more maintenance cost these are some of the observations that I found out in my conversation with this team here they will be available there so we can continue the conversation after a short break first I want to thank the whole team there they came from far off distances here they did a very good job so thank you very much appreciate it