 Well, good morning everyone. I think we're all ready to get going. I'll introduce myself. I'm Jeffrey van Kutsem. I'm part of the open source technology center at Intel, and I'm here to present the, together with my colleague Mike Adira, the connected smart home from IoT to cloud. So let Mike introduce himself, and then we'll go through the material that we have put together for you. So hi everyone, Michael Kadera. My job is, I work with Jeffrey on IoT to cloud, but my job is actually to make the cloud easier for people to use and consume. And so I've had a lot of different careers within Intel, all for the most part within data center, but also within Intel's IT department where I worked a lot on their enterprise applications. Platform as a service, I also implemented a number of their cloud environments. And so now I get the opportunity to talk with customers and work with them on implementing their own. And so today, what we'll do is we'll show how we're actually able to bring an IoT solution to the cloud and start to look at the choices we made, why some of them were made certain ways, and how to bring together a solution architecture. With that, let's take a quick look at the agenda we will cover today. So we will first of all give you a few of the considerations that we have taken into account when we just set out to develop this prototype or demo, as we can call it. We will then go through the different building blocks that we have used. We will give you a little more background on IoTivity and the Open Connectivity Foundation, because that's kind of a very important part of the demo itself. So I want to make sure that everyone is up to speed on what that is and why we're using it and how it can be used. We'll talk about web platforms, again, a very important component of the technology we have, and we use it throughout the chain from IoT to cloud. And then we will talk about application profile and the different solutions that can be used when we're putting together a cloud solution. So the way we get organized is I'm more working on the IoT side. So I will cover the first half of the presentation, and then Mike will take over and talk to you about the cloud aspect of this. You are, of course, welcome to raise your hand if you have questions. We don't expect to use the entire time, so I think we will probably have at least 10 minutes towards the end of it for further questions and answers. But again, if you want to have questions during the session, just raise your hands and we can address those. With that said, let me just jump to the next one. So what are the different considerations? First of all, Mike and I were kind of looking, both of us, part of the open source technology center. We were looking at how can we demonstrate and maybe prototype what an end-to-end solution looks like. So when you get to talk to people, a lot of these are kind of focused on IoT or focused on cloud or focused on networking. And we were really looking, both of us, I mean, Mike coming from the cloud angle, myself coming from the IoT, we were like, okay, how could we sort of re-use and demonstrate these technologies, not individually, but also working together and sort of integrating them into an end-to-end solution that makes sense. And so our objective was kind of a, that was kind of one of the motivation we had. Now, as we set out to do that, we were considering different options and we were like, okay, being open source people, we obviously wanted to use open source solutions where we had some that we could use. Similarly, we also wanted to use open source projects and use open industry standards where possible. And what that gives us is, first of all, industry standards, we believe are very important for enabling the growth of IoT or cloud. But open source solutions is also something that we can easily, more easily control and we can tweak and tune and it's a great learning experience. So one of the things actually we got out of that is also a lot of experience and knowledge that we didn't have before as to what are the challenges that people face when they want to put together a real-life solution, something that will go in production, that will be sent to customers or sold to customers that will need to scale, that will need to be robust, et cetera. So we also looked at things like how can that be deployed easily, how can that scale once it's been deployed. The idea being, as you're a young company sometimes, you have a great idea, you develop a solution initially, and then your customer base may be 100 people. Now, as you grow that user base, hopefully to thousands of them, you don't want to have to re-implement that stuff over and over again because it didn't scale in the first place. So what we did with Mike is we looked at that and with that in mind, hoping that the stuff would scale, how do we just build it from day one so that the scalability comes very naturally. And the final aspect, which was obviously very important these days, it's considering security. So you'll see a little less of that, but you don't see that in the way we've put it together, but you see that in terms of what we selected to use in terms of technology. And all of the technology we use have a very solid security foundation. So this is an overview of the architecture of what we put together. And by the way, for those that perhaps have not been there, but we had the demo running on the Intel boot, it's the one that looks like a little dollhouse with a few sensors and a cloud solution. So this is a representation, kind of an architectural overview of what we have. So let me just walk you through some of the stuff we have put in there. So for you, it's the left-hand side. On this side, we have really the IoT part of the house. So these are the sensors we have put in there. We have around 10 sensors today, different, they're simple sensors, but they sort of demonstrate what can be done. So we have things like a solar panel, we have temperature sensors, motion detection, CO2 detectors and that kind of stuff. For those, all of these are part of the same network running in the house, and we leverage a lot of web platforms such as Node.js and IoTVT. In the middle here, this is the gateway that we have. So the gateway is the home gateway. We run a Yocto-based OS on that one, and this is the sort of central point where it's collecting a lot of information about every sensor in the home. So that's kind of a, it's again, it runs the IoTVT stack, but that's what is then our bridge to the cloud. So the cloud is what you see on the right-hand side there. So everything in the cloud, you see things like OpenStack and Cloud Foundry. So this is what we use to bring scalability to the cloud aspect of it. So again, the idea is it's kind of easy enough to have an application that connects to a single gateway and collecting information about that gateway and providing perhaps things like extra services such as analytics or perhaps a simple cloud portal. Now, one of the key functions is as you grow your user base, you still want to be able to then scale that and have 100 portals running at the same time or perhaps a thousand portals running at the same time. So this is where we started building that on top of OpenStack and Cloud Foundry. Again, one of the selection criteria we used is to think we can use cloud providers, but we also wanted to look at how could that be done by ourselves directly. So which is why you don't see us necessarily using Amazon cloud services or Google cloud services. It's not because it cannot be done. It was really kind of a mental exercise of how would we do that if we were to do it ourselves. Kind of a part of the whole learning and also perhaps some companies actually may choose to say I'll set a private cloud versus using a public cloud. And some of the components are there. I'll go through some more of them in details. But let me talk a little bit more about IOTVT and the Open Connectivity Foundation. So how many of you in the room are familiar with what IOTVT is and OCF is? Okay. Oh, we've been good. We've changed it once. It's acceptable, I think. I've seen much worse. So I'll see a few hands. I don't need to spend too much time on the topic. But for those that don't know, the Open Connectivity Foundation is an industry consortium of companies. I think we have more than 130 companies today. So these companies got together and are defining a protocol for device-to-device communication in the IOT world. Let me just move because we're video recording that and I was just behind the pillar. So the problem they're trying to solve is the fact that today if we let companies develop sensors, develop smart devices such as smart television washing machines or even simple sensors, the problem is if there isn't a unified industry standard for defining and talking to these devices, we end up with a bunch of protocols that are basically the ones defined by the companies, respectively. So OCF is there to provide a unified solution that everyone agrees on, that you can be compliant to. And that means in your home, eventually, no matter what vendor you have, so if your fridge is coming from a specific vendor and then you change over time, you don't need to redesign and reset the rules and reapply a lot of the stuff that you've done already. The fridge will be seen as a fridge. It will be recognized as a smart connected fridge. And so that's really what enables not only it will make it easier for people to adopt the technology, but that's also how you can really make the smart home really smart. So smart today sometimes means we're starting to connect device, and that's about as much as they get. So they get connected, but still to talk to these, you still have to have dedicated applications for each. The real smart home will become available when you start having intelligent action taken based on different inputs. So things like, I don't know, nobody's home so you don't need the air conditioning system on. So the motion detection will tell you if there is someone home, and the air conditioning will be turned on enough and also based on the temperature. So you aggregate data and set up intelligent rules based on that. But for that to happen, you need to have devices that can talk to all of these sensors, to all of these connected appliances as well. So this is what OCF is doing. So OCF is very much, like I said, so it's an industry consortium and they're focusing on developing certifications and standards. Now the other project which is related to it, it's called IOTVT. So IOTVT is an open source project. It is hosted by the Linux Foundation. And their objective is also to provide a complete implementation, a reference implementation of the OCF specification. So you don't have to do it, but you can use it. It's there to be used and it's compliant. It's released under an Apache 2 license, which means it's easy to adopt for companies that want to adopt it. And then the certification aspect is something that, again, is more linked to the OCF part of it. Now onto the building blocks that we use. We have a number of hardware platforms that we use and a number of software components. So let me go through some of them for you. On the left-hand side, what you see it's a Miniboard Max, so it's an Intel-based system. We use that device as a gateway on our platform and we typically run a reducto-based OS. So you see different logos because we actually can run different OSes. We've played with different, we've played with Ubuntu Core, for example, we've played with the IoT reference OS kit, which is, again, built using the octo tools. And so we don't really have a very strict dependency on the OS. So the code that we have is very portable. But typically when we present a demo, we use a octo-based OS. It runs, I mean, it could be connected to a screen, but the way we run it, we run it headless today. So that's the gateway. That's the one, it acts in the OCF terms. We say it's the OCF client because it doesn't come, it doesn't, it's not a sensor itself, but it's connecting and collecting data about the sensor in the home. It's actually also the one bridging our network with the cloud. Now what you see there, we have other devices. Some of the sensors we have are actually based on the Intel Edison module and more and more we're actually moving to use the Arduino 101. So the Arduino 101, it's a very, very small device, low power based on the Intel curry. And we run another open source OS, which is not Linux because it's way too constrained for Linux. So we run something called a Zephyr project. There were many sessions, I think, at this conference on Zephyr, so hopefully everyone knows what it is. But for those who don't, I think the quick way of describing it, it's an open source, real-time OS, hosted by the next foundation. If I was to sort of make an easy distinction between that and Linux, it's typically you run that where you can't run Linux. So think about microcontroller type devices. Devices that have just a few megabytes of RAM or maybe a few kilobytes of RAM or Flash. So really where you can't really push the Linux kernel down to those spaces. And that's what the Arduino 101 is. It's a very tiny device. So more and more, we actually started by developing and prototyping a lot of the sensors we had using Edison because it was the easiest for us to have them up and running. And now that we have Zephyr running on an Arduino 101, we're just gradually bringing them over to that Zephyr environment. Any question on building blocks so far or on the architecture of how we've done that? No? Web platform. So let me talk a little bit about the web platform. So what we mean by web platform, it's kind of a collective term that I use because we see that technology used a different point in the demonstration or the prototype. So we use that on the sensor side, on the IoT side. We have a project called IoTVT node. These are the JavaScript bindings for IoTVT. So a lot of the development we've done is actually using these JavaScript bindings. What that gives us as people not necessarily just putting device in production, but it gives us an ability to prototype new sensors extremely quickly. It's very, very simple to have an OCF server written in JavaScript. I mean, a lot of the sensors we have and all of the code we have is published so you can go and look it for yourself, but it's about, I'd say, 30 lines of code and that's all you need. Now obviously the more complex you want the device to be, the more lines you have, but a simple sensor is about 30 lines of code and a lot of that can be reused whenever you change the device type. So very, very quick prototyping from our perspective. We initially run a lot of that stuff on a standard Linux environment because that's obviously a lot easier to develop on Linux because you don't have any kind of constraints in terms of memory. I mean, even debugging is easy because it's a full-fledged Linux environment. Now obviously in terms of hardware, running a temperature sensor on an Intel Edison is not something that anyone would really want to do in real life because you get a big beast, it's too expensive, it's too big and it doesn't make sense to run just a few sensors of that. So we started migrating those over to the Arduino 101 and what we're trying to do more and more is actually leveraging another project. Its nickname is ZGS but it's really a JavaScript runtime for Zephyr. So again, it's not exactly the same code but it's very, very similar, it's a JavaScript code and it allows us to create devices that are OCF sensors, OCF servers, that's the right terminology, and connecting them to the network. So the connectivity part is a little different. When we had the Edison board we were using either wired connections or wireless Wi-Fi. With Arduino 101 everything is done over Bluetooth, low energy and 6-low pan. Where you see other components of the web platform, you have at the middle something that says it's the IoT REST API server. So what that component is is basically an OCF client running on the gateway and it exposes all of the OCF API over HTTP session. And so that's how the cloud connects. So basically what you do from a standard OCF client, things like device discovery and then retrieving data about devices or sending commands to the device, you can do it very easily from the cloud but from the cloud instead of having to do that over co-op, we now can do that over HTTP. And in from a cloud perspective a lot of the components that exist can deal with HTTP and we have been doing that. And when I say HTTP I really mean HTTP or HTTPS that's going back to my first point about security. So all of that can be run on secure channels. So what you see on the right-hand side it's actually a snapshot of the cloud portal that we have designed. So the intention is we develop something that would be what could a user see if they were to log on into the portal or log on in their house. So that's the design we have. What you see at the top level you see a few alerts. So these are the things that should be grabbing your attention. So on the left-hand side for example you see a little alert in red that says oh I have a CO2 detector that's gone off so there must be smoke somewhere in the room. So better look at that. The second one you see is basically there is a motion detection so that could be something at your front door telling you that there is someone walking up the aisle and others that we have and we've started building up and adding to that and that's what Mike will talk about a lot of the analytics. Because collecting data is good if you want to take a snapshot if you want to check what temperature you have in your home. The smart home again being really really smart will become when you can do some sort of clever analysis on that data. Perhaps telling you that it looks like you have too much electricity going consumption going on for the temperature that's outside so perhaps you have a faulty heating system or the air conditioning perhaps is about to there is something wrong with it because it's consuming too much. So that kind of data is the kind of or that kind of information is stuff we can retrieve if we have powerful data analytics engines running in the cloud. Another aspect of the demo that we have as well and you show it we can see there at the bottom on the right hand side we have also developed a smart phone application it's an Android app and you can run that directly so it again runs the same IoTVT stack in the single app that can detect and really allows you to see all of these devices. So similar to what you have in the cloud but from your smart phone and with that we'll just go over the application profile and I'll pass that on to Mike. Alright, thanks Jeffrey. Okay, so I've really enjoyed working on all the IoT projects and there are a lot of there are a lot of fun because for me one of the challenges I've had in working with all the customers that I meet even when I was in Intel's IT department is helping them understand the benefits of cloud applications and IoT really lends itself really well to that because sometimes they're connected, they scale really rapidly, people start to really understand the concepts and why you need to have failover and those types of things available. But what I want to do is just go into a little bit more of what's different about IoT applications because if you've come from an IoT environment you know that the first thing when a customer comes to you when they want to land an application or a series of applications is you lay out the application profile. So you really need to start looking at those things that are different, there are many things that are similar but let's take a look at some of them. So starting with the behavior initially you're going to have some of those projects that are predictable, those are great to have, from an IT perspective you know what to expect, you can go ahead and plan it but not everything performs that way. So some of them of course high growth how are you going to respond to that? So if you have to scale rapidly are you going to have infrastructure in place? If you're moving from purchasing to get to new hardware to landing it, power, cabling I mean I know when I was in our IT shop it could take anywhere from two to three months on average just to land infrastructure from the purchasing point all the way through to having it powered up and landing customers. You need to plan ahead if that's part of your constraints or think about bursting into other data centers public cloud advantages. So additionally IoT applications are not always connected sometimes they're on or off. You need to know how to transfer that data if you have a truck that's moving through the mountains and it loses its connectivity is it going to have the information it needs to maintain its route are you going to be able to update information on your inventory and that it's safe how are those connectivity how is it going to be managed. And then lastly kind of the heart one of the harder ones to manage is that random or periodic bursting. So really setting things up so you can grow your applications and scale them to other data centers or other resources because sometimes your customers may be where you don't have data centers that can reach them. It might make an advantage to either set up a data center there or leverage public clouds continuing down the list micro services have you heard of has anyone here heard of chaos monkey testing. So a few people have yeah so this is something that Netflix really started a couple years ago I think it was like five or six years ago and the whole idea is that imagine you have a monkey just running crazy throughout your data center and it's going to go through and it's going to start chewing on cables and beaten up systems with the whole point is you don't know how your application is going to perform under that kind of circumstance so what you we've developed is they developed and others have implemented is writing your applications in ways that it's resilient to services that you aren't expecting failing and that's how micro services kind of come into play here. You want to really move away from those gigantic do it all applications and start pulling out those individual pieces so for example you have a login service and a catalog service and all these things need to work together as well as operate independently should something happen and so what you need to do when you're working on micro services is think of those logical ways of dividing these applications and you got to be careful because sometimes I've seen people go a little bit crazy with the micro services and every little thing is in a micro service but think about bundling it together and bringing those solutions together but what happens is that you can then eliminate single points of failure in the past where you have this one gigantic server that houses everything you can now have a farm of servers and VMs or containers that would allow you to do that failover and scaling that you can implement. Now on top of that because you're going to have these distributed services you need to look at collection orchestration and what that really means is how are you going to organize the deployment of your application you need to have that database first you need to then land caching services and web services that will connect into that in the right order that doesn't stop at the cloud you also need that with your individual IOT devices your gateway in place how are you going to do that handshake of new devices that come on and how are you going to grow and scale them as well all needs to be thought of because you know they need to grow together you know one of the things we found in many of the device or applications that we deployed is that our database usually tended to be our constraining point and so what we would do is we would have one database and then or a database with its replication and then caching probably five to six on some of these applications and web heads ten to twelve depending on what it was and so that was basically our collection and we would be called it a pod and we'd scale by these pods and that's how we would grow and shrink back that's the key piece is you also need to know when to shrink back and you don't want to be especially if you're paying a premium dollar for that compute you want to make sure you're scaling back down and releasing those resources okay the life cycle I work quite a bit with open stack and in the community one of the things that they've really been focusing on really over the last couple of years is upgrade ability specifically where things are really challenging is when you have multiple services all working together you need to make sure that their APIs all work as one how do you do that when you're going through and doing this upgrading and that's why you have to have sometimes maintaining multiple APIs and do that rolling upgrade to allow critical services to continue forward that really becomes very important in the cloud but you have to think about your gateway and those services that are going to be tying into that as well continuous integration and deployment you know this really gets into how you're going to be managing those individual devices if I have a device that's in the wall of my house and I don't want to have to rip something out of the wall to have to go in and flash it I have to think about how I'm going to be pushing code to it pushing updates keeping it secure that becomes a big part of that as well so you need to plan for those things in your environment security you know when I last presented on this it was back in the fall you may remember the web camera that was hacked and created the denial of service attacks so it couldn't have come out at a more perfect time as to when I was talking about security not that we're looking for an advantage with that but it actually helped me make my point there's so much with security and not only making sure you're patching those devices and keeping them up to date but looking at authentication user authentication as well as device authentication making sure you know what's on your environment and you know what to allow data encryption whether it's at rest or in transit it becomes really important I really started getting into the IoT tinkering because my kids kept on leaving the garage door open so I didn't want my bikes and my tools going walking off and so I put a sensor on there so I knew when it had been left open for us so depending on where I was I'd always get the alert and I can close the garage door fortunately today I had a little bit of a problem true story my garage door fell back down on my car I was packing out so I need an alert that warns me if it's not safe to back up it's one thing to add so we talked a little bit about patching but then on the far end with the network intrusion detection making sure that if something is an anomaly you're actually picking that up so of course security goes much further than this lastly data that's why we're here and why we're looking at IoT is to bring all the great things that data analytics and knowledge of it can bring you but location comes into mind and what you can do there's so many different options because the cloud they can seem like endless amounts of resources that you can use and scale you can have historical information a lot more data library available for you to use but you can do that detailed analysis at the cloud but now start thinking about as we start moving towards the edge and this is what there's a great presentation that Tom Bradich from HP gave about I think it was about six to nine months ago and what he introduced was the spectrum of insight and I thought it was a really good way to help give everyone an idea of how things are changing as far as compute available from the cloud to the edge as you start moving to the left that's where all the devices are and you get into the tier one and tier two that's where you're getting into the more intelligence that's being added with those gateways and what he's saying here is that there's options and trade-offs that you give at each level so in the cloud as I mentioned you have endless scalability you have endless power cooling all of those things available to you and so you can do that detailed analysis and it doesn't come at a high cost but if you move to the edge and you start thinking about if I have a device that I need failover for or if I need detailed analytics at the edge so for example if I'm driving my car and a truck goes out in front of it I should say it's an autonomous car and a truck goes out in front of it I want to make sure that I'm not sending that data off to the cloud for processing I want to process that locally and not barrel right into that car or that truck so that's why you're starting to see vehicles that are becoming basically data centers on wheels but it comes at a cost there's a lot to go into that safety design the compute and cooling that's needed at those levels you have that trade-off but you're seeing more and more capabilities coming to you and being able to do that analytics at the edge when I pull my car into the garage I don't want to actually track every single moment that my car was out driving that my tire pressure was 35 psi I know that, I want to know the anomaly so I don't want to have to have a data array in my garage just to track that track the stuff that's important filter the stuff that's important send that to the cloud for the analytics you need to track so there is so much we can talk about with the cloud so I'm going to give a high level overview if you have any questions about these areas in detail just come over and talk to me but really looking at the applications in the cloud there are lots of choices available now looking at it from the cloud perspective you have lots of great solutions software as a service that many of the bigger manufacturers have out today Samsung has their things so you can just plug in a sensor have your gateway recognize it boom you know how to use it and it integrates with your environment Philips has stuff like that too but if you're developing your own you'll need to think of in bringing that to the table so as Jeffrey had mentioned we're using cloud foundry and that's a platform as a service solution now what this does is it gives you an API that you can just push your application to it would recognize that you're using the lamp stack or various components and it'll deploy and provision that in locations that make sense for you and what it does is it allows you to scale rapidly from just a couple of instances deployed and it'll load balance that as you scale it on up it makes it really easy in allowing your developers to go from code to production rapidly but additionally there's big data analytics we're using a python solution in our environment and what we're showing is collection of data based on some we found some house analytics based on power and temperature we overlaid that if you stop by our booth and check out our demo we overlaid that with Portland weather and so you can see how the house would perform at that but what we're getting the value out of it is when we start adding additional simulations of data at scale and seeing if you're for example managing a grid and what you can do with all of that detailed data analytics so within the environments here there's lots of tools to allow you to scale and deploy big data as well as regular databases to help you with your applications now VMs containers that comes down to choice and how you like to deploy your applications VMs traditionally a little bit more secure but a little bit more heavy weight with speed of deployment and ease of provisioning but so each have their tradeoffs and then containers you know it takes a lot more to secure them definitely you can there's lots of solutions available and their use case is wonderful real rapid start up as well as things like doctor hub you can create images download them deploy them it makes upgrading and keeping them going forward very easily maintained and so we also have we're doing a lot of work within Intel and in the community to help secure containers a lot more there's lots of different options from trusted attestation to I'm wearing a clear linux shirt they have something called clear containers that help to leverage that technology at the hardware level at the physical level to secure those more like a VM but I really those are choices that you'll have to make and I really like now that I've used both platform as a service and containers they really make development a lot easier now underneath all of that I mentioned of course I use open stack but there's lots of things available Kubernetes to orchestrate containers VMware has their solutions Microsoft has Azure and of course there's multiple public solutions with Amazon rack space and bliss goes on so the important point is that you have the API just like you would with a platform as a service solution to allow you to scale your compute storage and networking and lastly with physical layer the important point there is you have your resources where necessary failover of course to keep in mind but making sure you're taking advantage of the capabilities at your physical layer so making sure you're passing through things like if you're doing heavy duty encryption workloads you can actually use things like AES and I maybe an FPGA if you have to want to offload that capability to to the hardware it really makes the performance of your applications that much better so really kind of summarizing this and where we've been with a lot of this project as well as the experience that Jeffrey has had in working with multiple customers and IOT and mine and the data center is really looking at how you're going to manage that data because that's a big part of how you're going to construct your overall IOT solution at the edge in the cloud little bit of both you can even have intelligent network devices there's lots of options available for you and the balance of analytics with that location compute power is available of course in the cloud but the heavy duty compute power at the edge could become costly if you're operating under extreme conditions lots of vibration cool or hot temperatures things like that scaling always make sure that you build these applications to grow opportunities like cloud foundry open stack make it a lot easier to design that scaling into the cloud but also look at it from the gateway how much your gateways can handle how you're going to scale these devices and systems with them in the end with the micro architecture is just understanding how these will all break apart and just know your application requirements that's really what in the end that's what it comes down to is you have to know what your customers want and how you're going to service that and lastly IOT applications they lose connectivity that's not I wouldn't necessarily call it an issue it's more of a feature design your application to really withstand that and have the failover necessary you know you don't want to have a factory that's for example like a bread manufacturing and just because of a cheap little temperature sensor breaks you want to make sure you shut down the factory you want to make sure you have redundancy within that as well just kind of bring it all back together as I mentioned know your application requirements so remember that and that's one of the things you can take from that and plan for scalability so let's expect those services to drop that's one of the things with micro services you can have failover you can also do that with devices as well our demo that we have in our booth is actually available to help jump start some of your projects fully open source go to 01.org slash smart home all the directions are there plus link to our get route repo that qr code will link you there as well and lastly iotivity this is what makes all of the sensor connectivity and the ease of use within that environment much easier to use so just iotivity.org and with that I will open it up to questions yes questions yeah so the first question was about privacy and protecting like for example web cameras and streaming images of your family or things like that and making sure you have privacy there you know what I found with applications in general across the board is that security comes with a cost and it really comes down to how secure do you want to make it not everything is 100% secure but you really need to do the I think do diligence and looking at the companies and providing how well are they patching their systems how well are they securing their environments have they done the penetration testing necessary to make you feel at least comfortable that it's going to be maintained it's a choice hopefully everyone makes the right one yeah well you know in the end we all support our companies with what we purchase right yeah now the the next one was really cost and that cost model you know really if you look at that trade off as well there's great solutions out there that provide that automatically for you can pay a license for someone who's done a great solution and carries that burden cost that start up cost of a large center to support those applications however if you look at things like for example just getting started you know there's low cost startups that you can use like for cloud services that to get started on that side now on the design side I probably turn that over to Jeffrey if he wants to talk about start up costs with IOT devices if you had any recommendations there on how that would scale so yeah I think the question was also how that would scale and who would pay for that eventually we've not looked at the cost model that deeply I think that the way we look at that is as we enable the smart home I think a lot of the value in this IOT domain we actually come and that boils down to having all of these sensors being able to talk to each other that's why we started choosing things like analytics I think anyone I mean throughout the chain would be more willing to chip in and provide I mean put money on the table if they're really getting value out of it and the value just cannot be only I can read the temperature sensor remotely so that has to be a combination of things so it's providing a service that I perceive as adding value to my I mean if it's you as an individual it's going to be like in just the example of smart home it's probably you as a family you see value so for example like Mike saying Mike knowing that the garage door will be closed if it should be closed because there is no one in the house and then so that that's a value to him so it's probably willing more willing to put in a few dollars or whatever is needed to get that up and running and so the whole IOT explosion I think it will be unleashed whenever you have these interactions and when the systems really become smart that's for the smart home and I think the dynamic would probably be similar but a little different in other segments so things like the industrial the industrial guys will probably have similar things when you can automate and manufacturing plants for example they'll probably sustain a lot of the cost but only if they get really good return on their investment so yes boy well in my opinion I look at the manufacturer to stand behind with a warranty yeah and then the ownership is also you know it's my so for example it's my due diligence to make sure that my home network is secure you know I don't I don't rely on Comcast to do that for me or others I mean they do I expect when things go out beyond that I think that that we all own a portion of making our own environment secure and knowing how to use those tools I think the when I look at that one of the dynamic we've seen so far is a lot of very often the gateway in the house is the stuff that is I mean it is provided with service so if you get a Comcast and they probably have a gateway that they sell to you so I think I'm not excluding that things like home gateways maybe boxes that will be subsidized by the companies providing the service I think we look at that other devices in the home will most likely be the stuff that you want to have like an alarm system or you'll have I mean you buy your fridge because you want that fridge of that color that model and so you want that to be able to interact with the rest of the systems but I'm not necessarily saying and I may be wrong but I'm not necessarily saying a model where these devices become all part of a one just big giant package that you get from a provider yeah yes I did hear about that yeah yeah okay yeah that's a good guess I think there's a lot going on in the courts with legal that are you know it's evolving and people don't know how these use cases are going to eventually be used good or bad you know I don't have an answer for you because I I'm not sitting on the court and I don't yeah but I think there's a lot going on yes the second question really is that if you look at today's devices that I have they are being driven in their architectures not by all this open protocol but driven by the cloud providers because the cloud providers that you decide they have an SDK that you download that's the first thing you do on your platform you download the SDK that provides different activities of platform and Azure has it you know Microsoft AWS Samsung IBM and they all have and then you write your application next to that okay which basically is reading a sensor and putting it in some JSON and then you're done that seems to be economy happening this is my question people designing a network in the home that needs to be very complicated and then just the cloud providers saying this is the way to do it quick time to market and I don't know where it's going to win so what's your view on this whole because I see a big that economy of things happening like this I don't know where it's going to go but it's nobody's just talking about it and I see many devices that we have supported in our company that do it from the cloud provider side that's your first choice where's the SDK and who owns the data well that's what it is the data is it goes to the cloud because the sensor costs 20 bucks and that's the end of it I think that's the reality today is that the cloud providers are driving this and I just want to what cloud providers are supporting this stuff because that's what I do I help people download the SDK and they'll update the SDK every now and then with MQTT whatever they're doing whatever communication they have and then you're locked into that there's a lot I know from the cloud side that's going on to help with as far as data ownership and making sure that your data is yours for example from when the point that the data leaves you transit and it stays encrypted and then you maintain those keys elsewhere that's been a lot of work I know that has been going on in that space multiple companies working on that this is just you're right there's no easy answer here there's a lot that's going on to really implement and understand security I think that we're seeing more announcements like the web interface the webcam that I had mentioned I think that those types of things are making the issues in security more relevant that people I need to pay attention this is a big deal we have one more minute one more question alternatives to IOTivity and also the question is if there are different similar solutions in the industry for okay the way I compare them typically is with something similar to what Apple is doing with HomeKit for example, or Google Weave in terms of functionality now they're different in the approach because obviously Google Weave and Apple HomeKit they're driven by a company so you're sort of tied by what the company decides to do and how they want to take that the approach that OCF is taking is more kind of a compromise in terms of let's get the industry player together all agreeing on a on a platform I think from what we've seen is Google Weave for example HomeKit they're very focused on the smart home IOTVT's by nature they've had a lot of focus on the smart home but they're really also looking at that technology something that could be used in different segments so not only just the smart home I think this is there were back in the days I think there are two standards that were really going to hide on that was mostly well what was known as OIC back in the days and all join all seen now all of the players backing up one or the other have actually come together and now OCF is kind of the merging of these in many ways so I think this what we're seeing now and this is going back to your point earlier I think a lot of the dynamic to see is it's not really it's driven by opportunities by companies they see a gap that they can fill by providing a service that doesn't really exist and so it's a little chaotic chaotic at the moment and I think what we will see eventually is the market mature and as the these technologies penetrate and make their way into their homes and we will see things settling down and in general when you have a more mature market that's when you start seeing industry standards set while really getting a strong hold of those so now that will have to be paired with compiling services too so I think to go back to your question I think that this from an industry standard there isn't really a second industry standard iotivity is already the result of different trends going in slightly different directions and sort of looking at what they were doing and thinking okay we just get we have to get together on that it doesn't make sense to have two or three yeah I mean it will never converge to a single solution that doesn't exist but I think it will consider I mean things will consolidate over a handful of solutions is something that we'd expect so I think with that we're out of time we'll be happy to take more questions outside but we'll have to make room for the next session and if you want to have more questions we'll be outside or you can grab us whenever you see us around thanks