 Live from Las Vegas, Nevada, it's theCUBE. Covering IBM World of Watson 2016. Brought to you by IBM. Now, here are your hosts, John Furrier and Dave Vellante. Where are you going guys? Okay, welcome back everyone. We're here live in Las Vegas at the World of Watson event, the Mandalay Bay. I'm John Furrier with theCUBE. My co-host Dave Vellante, she found us at wikibond.org. Our next guest, Chris O'Connor, general manager of the IBM Watson IoT offering. Welcome back to theCUBE, great to see you. Thank you. So we had blockchain on earlier and you're hearing a lot about a lot of developer action. Certainly you guys have a lot going on. Certainly here you see that mojo, developer programming, building stuff. Then also you have this notion of joint development of solutions where IBM is working with customers building out solutions. So IoT is certainly red hot market. There's seven segments that we've identified through our wikibond research team. Personal, home, all the way up to IoT for industrial. And so it's a really big transformative concept. It actually has operational impact to business. And Jeff Emelt was quoted on stage last week saying at a conference, talking about General Electric, GE, we went to bed an industrial company and woke up a software company. And this is indicative of all the businesses out there as they digitize they're becoming SaaS companies basically. So thoughts on that and the challenges and the opportunities because IoT really puts a face on the opportunities and also the challenges. The DNS hack on the DDoS. Again, DDoS has been around for a while. I called the kitty scripter kind of hack. Everyone does it and there's things in place for that. But it highlights the vulnerability of IoT in the mind of potential customers. So I woke up a software company. That's the mindset of companies. The hack, this is the new stuff on the table. We'll start with the comment just off the woke up a software company. People are taking devices, they're now digital. So they've got an image that spits off it that gives it an image that you never had before. So instead of the shipping stuff in boxes comes back as a digital image. So you got to be able to connect it. You got to be able to bring that information in. You got to be able to respond and repose it. And then you got to connect it to your business. All of this is new for these physical companies, right? They've been doing it with clipboards. They've been doing it with somebody literally scanning and checking the device in and out but they haven't been done it done digitally. This presents an opportunity to connect the whole way through and you wake up as a digital company that's transformed and you can go after value then. You can go after operational costs. You can go after connecting to your clients. And you can go after building new business models like change how you do warranty and other stuff at the same time. And that's the wonder of IoT which is it gives all these businesses that were device businesses now a digital focus of being able to go after changing how they work and how they monetize the things that they do. And the security thing specifically is obviously we've talked on theCUBE all the time the last time we talked about it. The service area has increased service area for threat detection and hackers coming in has increased but still it's the same old security game. Talk about the customer's fear if you will and how you guys can counteract with product around the security. What's really happened here and if you read what's happened over the weekend and what's transformed is you're watching the evolution of a new industry really adopt what problems that we've solved before. So if you go look at what we did inside of enterprises when you had end points and you had intermediate points that's called gateways and then you had backend servers where all the data got aggregated together it's no different except you're out in the wild now those end points are driving around on the road or they're rolling up and down the street or they're flying in the air or they're in something in your house or you're wearing it around on your wrist but it's an endpoint. The information comes back to some medium point it's called a gateway and then the data comes off of that and it's aggregated all together. So along the way you have access control problems of the device you have gateway data aggregation problems you got a lockdown with security and then on the back end you got to make sure that the data when it's pulled together follows all the laws, the rules, the regulations and the things that go along with it. We solved this problem before. We solved it several times over we solved it in the client server error we solved it in the main frame error we solved it in the IT centralization error we solved it in cloud in the enterprise now we're solving it in the IoT context and the solutions are well known in terms of building patterns and being able to lock down each step along the way if you read in the press a lot of what goes along with it problems that could be solved passwords left unchanged defaults that were shipped and left out there in the wild and so this is you're referring to the devices that get shipped from the manufacturers with default passwords which is essentially just boilerplate and it goes out the door and once it's known it gets access to everything it's like you ship somebody a ploy a laptop and you told them don't change any of the passwords and just leave it sitting on your desk open for anybody to use so instead it's a device that we shipped out to somebody and everyone knows each other's password and they don't change it and that device is a compute device it's just a computer that has some other purpose out there so common sense is that play here common sense is that play on that piece of that well what's changed is scale right I mean the scale of the problem is now if we didn't we said we solved it before but we really didn't solve it we kind of look back every year and go well am I more secure than I was last year so the same prescription applies there you go it's just now but a much larger scale it's not only scale but it's a difference in players too you've got a whole new set of players that used to build physical devices that never had this ability to chat and talk and be logged onto so you've got new players that need to be educated on really what's a lot of patterns that were part of that lockdown process you described with historically an analog mindset there you go okay now the other thing that's changed is the data it's so much more pervasive so much more data what are you seeing there I'd love your thoughts on everybody wants to instrument the windmill John always points out well before you do that you better connect to the windmill okay great so we're going to put compute out there and what's happening with analytics at the source so you're watching a classic progression of analytics take place people connect the device then they connect a bunch of them they're like hmm I went from one that was kind of cool to now I got a lot and I got a lot of data now coming in so do I want to bring all that back home do I want to bring it to an intermediate point or do I want to do analytics on the source and what you're seeing happening is that the device vendors the gateway vendors and the cloud vendors are starting to cooperate and everybody's opening up compute space to be able to put shared analytics packages on so that you can get the correct amount of data to the correct point along the way along the journey so if the device is healthy and it's fine it may choose to just send up something on a regular basis versus talking all the time if you're in a gateway and you're dealing with a million devices talking to that gateway why send it all up to the cloud if everything's fine start to send the exceptions up and put analytics on there that allow you to be able to limit what goes up and in the cloud depending on your use case you want to be doing the big data crunch and you want to be able to think about cognitive capabilities you want to be learning what the anomalies are and how to go about predicting variances and how to go about that that doesn't mean you have to have all the data that means you have to have the right data and so if you look at alliances and partnerships that are starting to get formed for example in IBM we produced a little cloud package that we can distribute out to a gateway we did a partnership with Cisco we announced it in June this year where you can put actually our analytics onto a Cisco router or switch and so if you're in a manufacturing floor without going to the IBM cloud you can run IBM analytics directly on top of that Cisco router and de-dupe and throw away a bunch of stuff and only ship what you want to up to the IBM cloud and allows us to be more local in that and we see that pattern progressing all the way down to the devices where we work with people like Intel and ARM to be able to do the same type of capability so it's use case dependent but what are you typically seeing and maybe you can help us parse the use cases in terms of is the data staying distributed for dominantly? Will most of the data stay at the edge? When you think about dominant use cases if you've got a lot of devices that are uniquely connected to the cloud they're going to bring things up so you've got washers and dryers in people's homes you may save a little bit of information on that washer or dryer but you're going to bring the information directly back to the cloud at that point in time because you don't have a gateway in your house but if you go to a manufacturing floor where there's a dense packing of IoT devices all sitting in that manufacturing floor because there's a million machines and they're making something for you you want to get that statistic and picture together before you pop it up into the cloud and you're probably going to do some reduction of data right there or if you're out on an offshore oil platform you might want to have that be a living entity that platform and reduce the data there around the models and statistics that you want to see happen be local and then only hand back up to the cloud what you need to have done so it all depends on whether you're remote it depends on how dense the number of IoT devices are and it depends on the use case whether you're dealing with consumers or enterprises on how often you talk. Connect the dots for us on the cognitive piece because you mentioned, we were talking about common sense in regard to the password debacle with the device being shipped and certainly those manufacturers need to be accountable for that, that's on them but now let's talk about conventional wisdom porting practices from the data center or information technology or IoT to IoT is not a clear move and just deploying similar practices in some cases the conventional wisdom is the opposite where you're seeing a lot of customization because every customer is different therefore you can't take a general purpose concept and roll it out to the market so could you share some thoughts and your color around how you see that new conventional wisdom or the counterculture around IoT because a lot of joint solutions and we hear that it's not a one off anymore because you say oh it's all one off solutions well everyone kind of is now tell about that dynamic of this new wisdom we sat back and we thought about this and we constructed it and we put it out in a paper called security patterns for IoT and if you think about what that means it means that you can sit back and look at the business impact of the device you're rolling out there and then decide the amount of security that you want to be able to apply across that chain of how that device talks back into the business and what you do with the data and it's a methodology that it's a classic methodology that was used in the enterprise it's now being used in IoT so to be able to help you understand what's the impact of putting this out here what does the device do what are the capabilities on it what do I need to do to secure it does it have those capabilities and then as it passes data along who's next in the line of receiving that data and how secure are they and it gives you a methodology to think about everything from the chip that's in the device all the way to the back end locking down on the data and that's what's highly repeatable now the mechanisms as you said are different they will vary because you're going to have different amounts of security that exist in the gateway or on the device some people may or may not be using the vault that's on the chip to be able to produce a secure ID and other people will be and so you got to decide when you make your device what level of security do I want to meet in this and we think that by applying a business pattern look to what you're going to be doing and the exposures you take at each chain along the way gives you that ability to lock the pieces down so this tool methodology is kind of a way to get the architecture out there to deploy now you said it's out in the wild so it is out in the wild IoT is out in the public it's analog means digital so new patterns are going to emerge this is where I want to connect the dots on cognitive so once you have essentially used that methodology to deploy now stuff's happening you need the notifications you need the alerts you need to have some intelligence in the device to understand the landscape of the device itself what does that mean for customers and how should they think about that what's your point of view there so when you bring back the information you want to use things such as machine to machine learning and you want to be able to use cognitive engines to be able to look for variance because in the end you're going to have a model that's going to prove out what you want it to be and then you're going to see anomalies that are going to take place as that device either freezes or thaws or as it moves around or goes to places you didn't expect it to go or gets taken apart and put back together and you're going to want to look for anomalies and in those anomalies you're going to be able to understand what's an acceptable pattern and what's not an acceptable pattern and this is where while you might be reducing data along the way bringing back the variances the whole way back and then comparing that with something that looks like a machine learning algorithm similar to what Watson has it gives you the ability to see the variances and be able to see the differentiation that's in there at the same time. So baseline deployment and then using the new variance modeling understand the new data does it reiterate through that? I mean does the model iterate and get smarter? The model's continuously running and it gets smarter each step of the way and the longer you're on the model the more devices you put into the model and the more type of statistics you give it the better off it is and IoT is this interesting combination of we used to teach you in the IT data center to reduce the number of different types of data with IoT we actually think your accuracy gets better with multiplying more different types of data so if you bring in weather and you bring in other types of data feeds to compare to it you get that variance down to the environment of the device as well as the device itself and we think that's critical in IoT to be environmental around the device not just about the device but also the data and sharing data among systems is a key, key tenet for this, right? That's right, that's why we also include a weather company feed in our IoT platform because we think as you bring that data in you're going to want to combine it with real time what's happening with the weather and use that as a way to show variance and understanding what's taking place out there in the wild, so to speak. So Chris what are the outcome discussions like with customers? I want to start with the outcome and then work back to the solutions. I mean are they all over the place are you seeing patterns emerge? There's kind of three outcomes that customers are driving for. First outcome is just operational impacts. I can run it better, I can understand it better, I understand its health, it tells me if it's happy, it tells me if it's unhappy and instead of just changing the oil so to speak every 3,000 miles the device can actually tell me what's my status and what do I need and what kind of parts do I have. So the outcomes that customers get inventory of parts on demand. Better maintenance, better scheduling. Just in time maintenance and better positioning of things to be able to take things in and out at the same time. Huge savings that clients get. You can drive a couple of percentage points off of somebody's bottom line. We drove eight or nine percentage points off of the Porta Cartagena's use of cranes, for example. Millions to their bottom line in terms of what that means to have that crane always moving a box because the boxes sit, it's cost some money. So that was the utilization. Utilization, right, asset utilization. Second thing is talk to your clients. Go out and be able to work with your clients, establish a loop with your clients in a way that you can't before. It lets you disintermediate other industries. It also lets you talk directly in a way that you couldn't before. You take somebody like an appliance maker. They're typically not in touch with the person that uses the appliance because it's sold through some department store, some change store, some distributor. If you start to connect, you can actually start to sell services. You can actually get services back. You can actually work with your clients. You can go into your engineering team with that data that's coming directly from your clients. You can re-energize the engineering team to go make product improvements, software loads or hardware changes that make real changes in the product. So that whole connected client piece is something that we see customers racing for because if they can talk to their client, they get direct feedback versus this pass-back effect that's existed for the last 100 years. Then the last thing is new business models. Think about things you can do with warranty. You can be driving down the road and at 35,990 miles, they can decide, hey, we think you're a good driver. How'd you like a new warranty on that? All right, in fact, push the button right now and we'll initiate a payment from the device and we'll give you six months more warranty. And you get this instantaniousness versus the letter in the mail that says, we think that you're going to run out and it's already happened. It's already done. By the time you look at the piece of mail, you're out of there. So the device participates in new business models and warranties. That's a live device with a consumer. Put it back into supply chain and to being able to understand what you do and into distribution chains in terms of how things get distributed and how accurate that is. And the device itself can now tell you what's going on in terms of the monetization of your business. And then business model. Two and three are life changing. So you're saying that today when my ice maker breaks, I'm going to sit around for three hours waiting for the guy to show up. I can't really tell, right? So I mean, that's a major productivity impact. And how many times when the person shows up, do they have the part of the truck the first time? One percent. So if the device can tell you, here's what's broken and here's what model I am. Right. Huge efficiency. And how about you mentioned weather before, the acquisition of the weather company. We learned was actually about IoT as much as it was about data science. And so how has that played into your offerings? So if you think about what we're doing about assets, traditionally we've taken assets that they've come in through the fiscal systems and we've been able to put them in tables and we help the depreciation life cycle and it's there, right? But if you think about what we're doing with IoT, we're now connecting the assets in real time. So we bring them in through the platform and we bring a weather company feed in with it at the same time. And when you think about scheduling work on that asset, you can start to understand what's the conditions going to be that day? Do I have the right parts for the right conditions that are going to be operating? And do I have the right environment? Do they have the hanger reserved or do I have to do this in the open? Right, and is the device in a place where I can service it in a native format and what's its exposure? An aircraft that's been exposed to a hundred hours of turbulence is a different aircraft than one that has been exposed to clear skies and sailing the whole way. The amount of parts, the inventory you bring to bear, what you expect to change, is a completely different list from that perspective. So the weather company provides this enormous benefit of making your IoT environmental, not just about the IoT device itself. It's predictive and it learns also from the past when a car slips on the road, it can tell the other cars for the same conditions, what to watch out for. Chris, talk about your business. What's the strategy right now? What's the IBM straight for the customers watching? Because obviously they're interested in IoT. It is the hottest area, has the most impact. I mean, you're talking about millions of dollars. That's right. What's the strategy for your group? What plans do you have? And if the customers want to learn more besides going to IBM.com, do you have any big events coming on? What can they expect? How do they engage with IBM? So plans for IoT. At the enterprise level, looking at industrial components, we've got offerings for each major enterprise that's out there, or each major industry that's out there that have little quick entry points that you can sample and see how you want to be able to get started with IoT. If you approach it from more of a consumer or developer point of view, go out and take some online courses. We've developed a set of courses with a company called Coursera. And there are IoT courses out there. It's 15 hours of education you're going to take. You get a Raspberry Pi, and you can sit at home and you can play and take a course, get certified in IoT. And you can learn a little bit about the IBM platform, but more importantly, it's an IoT oriented course. It teaches you what a sensor is, how to connect it, how to get value out of it, uses real life chip sets and data, stuff that you'd use in an industrial set. I think a culture kind of thing. But it gives you the ability to play with it at home almost in a Heath kit oriented fashion. So I would tell you, go out and play and go to Coursera and download our IoT course. And for clients who want to bring IoT into the enterprise, whether it's turning the industrial into or it's a small medium sized enterprise, these kinds of environments, how should they engage with IBM? So a couple of ways you can engage with IBM. You can start again at the courseware. We have clients that we walk into, they already have Bluemix IDs, they've got the IBM course, they've got chips up and running, and they've done that experiment as a team building exercise among the engineering team. And that's kind of an engineering platform statement. At the enterprise level, we go in and we meet with chief digital officers and chief security officers and chief executive officers and we talk about business transformation. And we engage at that level also with these quick sound bites of what you can get started to be able to realize the value. And World of Watson, it's a big show for you guys. There you go. Is there any like subshows, regional things you got going on? We've got 25 different sessions here at the show. We've got clients that are out giving sessions everywhere from Cisco to our friends at Cone Aid, other companies that are here talking about what they've done for values. Go sit in at one of the shows or one of the sessions and pick up on what other customers has done for us. All right, Chris, final question. What's the one thing that customers should know about or people watching about IoT or IBM's IoT that they may not know about that they should know about? It's one of the most consumable things on the planet. Go out, download the course, play with IoT and have an experience. That's what I'd want customers to know. Go out and sit down and start your IoT journey today. It doesn't have to be an enterprise-wide transformation project, it can be something you can start inside of your development or your engineering team to learn. And there's a viable path to digitize something. And it helps you start this whole path down, the digitalization of your devices. All right, Chris O'Connor, general manager of the IBM and Watson IoT Group here at IBM, we're live at the Mandalay Bay with the CUBE coverage of the World of Watson. The new show branded from IBM Insight and catches everything from big data across the board. Obviously, we've got wall-to-wall coverage two days. Go to ibmgo.com for all the on-demand live action. Got the keynotes, got the CUBE, tons of content. Check it out and also join the community there as well. I'm John Furrier, Dave Vellante. We'll be right back with more after this short break.