 Live from Times Square in New York City, it's theCUBE! Covering IBM's Change the Game, winning with AI, brought to you by IBM. Hi everybody, welcome back to theCUBE's special presentation. We're here at the Westin Hotel and the theater district covering IBM's announcements. They've got an analyst meeting today, a partner event. They've got a big event tonight, IBM.com slash win with AI. Go to that website. If you're in town, register. You can watch the webcast online. You'll see this very cool play as a Vince Lombardi, one of his famous plays. There's kind of a power sweep right, which is a great way to talk about sort of winning and with X's and O's. So anyway, Daniel Hernandez is here. He's the vice president of IBM analytics, longtime CUBE alum. It's great to see you again. Thanks for coming on. My pleasure, Dave. So we've talked a number of times. We talked earlier this year. Give us the update on momentum in your business. You guys are doing really well. We see this in the quadrants and the waves, but your perspectives. Data science and AI. So when we last talked, we were just introducing something called IBM Cloud Private for Data. The basic idea is anybody that wants to do data science, data engineering or building apps with data anywhere, we're going to give them a single integrated platform to get that done. It's going to be the most efficient, best way to do those jobs to be done. We introduced it. It's been a resounding success. Been rolling that out with clients. It's been a whole lot of fun. So we talked a little bit with Rob Thomas about some of the news that you guys have, but this is really your wheelhouse. I want to drill down into each of these. Let's say we had Rob Bearden yesterday on our program. He talked a lot about the IBM Red Hat and Hortonworks relationship. Certainly they talked about it on their earnings call and there seems to be clear momentum in the marketplace, but give us your perspective on that announcement. What exactly is it all about? I mean, it started kind of back in the ODPI days and it's really evolved into something that now customers are taking advantage of. You go back to June last year, we entered into the relationship with the Hortonworks with a basic premise. It was customers care about data and any data driven initiative is going to require data science. We had to do a better job bringing these ecosystems, one focused on kind of Hadoop, the other one on classic enterprise, analytical and operational data together. We did that last year. The other element of that was we were going to bring our data science and machine learning tools and runtimes to where the data is including Hadoop. That's been a resounding success. The next step up is how do we proliferate that single integrated stack everywhere, including private cloud or preferred clouds like OpenShift. So there was two elements of the announcement. We did the hybrid cloud architecture initiative. We're just taking the Hadoop data stack and bringing it to containers and Kubernetes. That's a big deal for people that want to run that infrastructure with cloud characteristics. And the other was, we're going to bring that whole stack onto OpenShift. So on the IBM side with IBM cloud private for data, we are driving certification of that entire stack on OpenShift. So any customer that's betting on OpenShift as their cloud infrastructure can benefit from that and the single integrated data stack. It's a pretty big deal. You know, OpenShift is really interesting because OpenShift was kind of quiet for a while. It was quiet, if you will. And then containers come on the scene and OpenShift has just exploded. What are your perspectives on that and what's IBM's angle on OpenShift? Containers and Kubernetes basically allow you to get cloud characteristics everywhere. You know, it used to be locked into, you know, kind of the public cloud or CSP providers that were offering as a service, whether it was Paz or IaaS. And Docker and Kubernetes are making the same underlying technology that enabled elasticity, pay-as-you-go models available anywhere, including your own data center. So I think it explains why OpenShift, why IBM cloud private, why IBM cloud private for data, just getting a lot of momentum. I mean, the core OS move by Red Hat was genius. They picked that up for a song and our view anyway and it's really helped explode that. And in this world, everybody's talking about Kubernetes. I mean, we're here at the big data conference all week. It used to be Hadoop World. Everybody's talking about containers, Kubernetes and multi-cloud. Those are kind of the hot trends. I presume you've seen the same thing. 100%. There's not a single client that I know of and I spend the majority of my time with clients that are running their workloads in a single stack. And so what do you do? If data is an imperative for you, you better run your data and analytics stack wherever you need to and that means multi-cloud by definition. So you've got a choice. You could say, I can port that workload to every distinct programming model and data stack or you could have a common data stack everywhere, including multiple clouds and OpenShift in this case. So thinking about the three companies. So Hortonworks, obviously Hadoop Distro, Specialist, Open Source, brings that end-to-end sort of data management from edge to clouds on-prem. Red Hat doing a lot of the sort of hardcore infrastructure layer. IBM bringing in the analytics and really empowering people to get insights out of data. Is that the right way to think about that triumvirate? 100%. And with the Hortonworks and IBM data stacks, we've got our common services, particularly around open meta data, which means wherever your data is, you're gonna know about it and you're gonna be able to control it. Privacy, security, data discovery reasons, that's a pretty big deal. Yeah, and as the cloud, well obviously the cloud, whether it's on-prem or in the public cloud, expands now to the edge, you've also got this concept of data virtualization. We've talked about this in the past. You guys have made some announcements there, but let's for the double click on that a little bit. What's it all about? So data virtualization has been around for a long time. Its basic intent is to help you access data through whatever tools, no matter where the data is. Traditional approaches to data virtualization are pretty limiting. So they work relatively well when you've got small data sets, but when you've got highly fragmented data, which is the case in virtually every enterprise that exists, a lot of the underlying technology for data virtualization breaks down. Data coming through a single head node, for instance, ultimately that becomes the critical issue. So you can't take advantage of data virtualization technologies, largely because of that, when you've got wide-scale deployments. We've been incubating technology under this project codename query plex. It was a codename that we used internally and that we were working with beta clients on and testing it out, validating it technically. And what's pretty clear is this is a game-changing method for data virtualization that allows you to drive the benefits of accessing your data wherever it is, pushing down queries where the data is, and getting benefits of that through a highly fragmented data landscape. And so what we've done is take that extremely innovative next-generation data virtualization technology, included in our data platform called IBM Cloud Private for Data, and made it a critical feature inside of that. I like the term query plex. It reminds me of the global sys plex. I go back to the days when actually doing sort of distributed global systems was very, very challenging and IBM sort of solved that problem. But okay, so what's the secret sauce though of query plex and data virtualization? Kind of how does it all work? What's the tech behind it? So technically, instead of data coming and getting funneled through one node, if you think about your data as kind of a graph of computational and data nodes, what query plex does is take advantage of that computational mesh to do queries in analytics. So instead of bringing all the data and funneling it to one of the nodes, and depending on the computational horsepower of that node and all the data bandwidth to get to it, this just federates it out. It distributes out that workload. So it's some magic behind the scenes, but a relatively simple technique. You know, low computing in aggregate is probably gonna be higher than whatever you could put inside of a single node. And how do customers access these services? What form does it take? It would look like a standard query interface to them. So this is all magic behind the scenes. Okay, and they get this capability as part of what on IBM? IBM Cloud Private for Data. It's gonna be a feature. So this project, query plex, is introduced to this next generation data virtualization technology, which just becomes a part of IBM Cloud Private for Data. Okay, and then the other announcement that we talked to Robo, I'd love to understand a little bit more behind it. Actually, before we get there, can we talk about the business impact of query plex and data virtualization? Thinking about it, it dramatically simplifies the processes that I have to go through to get data. But more importantly, it helps me get a handle on my data so I can apply machine intelligence. It seems to us that the innovation sandwich, if you will, is data plus AI and then cloud models for scale and simplicity. And that's what's gonna drive innovation. So talk about the business impact that people are excited about with regard to query plex. Basically, better economics. So in order for you to access your data, you don't have to do ETL in this particular case. So data at rest, getting consumed because of this underlying technology. Two, performance. So because of the way this works, you're actually gonna get faster response times. Three, you're gonna be able to query more data simply because this technology allows you to access all your data in a fragmented way without having to consolidate it. Okay, so it eliminates steps, and gets you time to value. And gives you a bigger corpus of data that you can then analyze and drive inside. 100%. Okay, let's talk about Stack Overflow. Rob took us through a little bit about what's going on there. But why Stack Overflow? You're targeting developers. Talk more about that. So Stack Overflow, 50 million active developers each month on that community. If you're a developer and you wanna know something, you're gonna go to Stack Overflow. You think about data science and AI as disciplines. The idea that that is only germane to AI and data scientists is a very limiting idea. In order for you to actually apply artificial intelligence, whatever your use case is inside of a business, it's gonna require multiple individuals working together to get that particular outcome done, including developers. So instead of having a distinct community for AI that's focused on AI machine learning developers, why not bring the artificial intelligence community to where the developers already are, which is Stack Overflow. So if you go to ai.stackexchange.com, it's gonna be the place for you to go to get all your answers to any question around artificial intelligence. And of course, IBM is gonna be there in the community helping out. So it's ai.stackexchange.com. You know, it's interesting, Daniel, that when you talk about digital transformation, you're talking about data. It's John Furrier said something a while back that you thought, this is like five or six years ago. He said, data is the new development kit. And now you guys are essentially targeting developers around AI, obviously a data-centric, people trying to put data at the core of the organization. We see that that's a winning strategy. What do you think about that? 100%. I mean, we're the data company instead of IBM, so you're asking the wrong guy. If you think I'm... You're biased? Yeah, possibly. But I'm self, I'm acknowledged to be biased. I, the data over opinions. Right. All right, tell us about tonight, what we can expect. I was referencing the Vince Lombardi play here. You know, what's behind that? What are we gonna see tonight? We were joking a little bit about the old school Power Eye formation, but that obviously works for your... You're a New England fan, aren't you? I am, actually. If you saw the games this weekend, Pats were in a Power Eye for quite a bit of the game, which I know upset a lot of people, but it works. Yeah, I don't know. Maybe we should have used it as a Dallas Cowboy team. But in any case, it's gonna be an amazing night. So we're gonna have a bunch of clients talking about what they're doing with AI. And so if you're interested in learning what's happening in the industry, it's kind of the perfect event to get it. We're gonna do some expert analysis. It'll be a little bit of fun breaking down what those customers did to be successful, and maybe some tips and tricks that'll help you along your way. Great, it's right up the street on the West Side Highway, probably about a mile from the Javits Center, people that are at Estrada. We've been running programs all week. One of the themes that we talked about, we had an event Tuesday night, we had a bunch of people coming in, there was people from financial services, we had folks from New York State, the city of New York. It was a great meetup, and we had the whole conversation got going. And one of the things that we talked about, and I'd love to get your thoughts, and I kind of know where you're headed here, but big data to dupe all that talk, and people ask, is that, now AI, the conversation has moved to AI, is it same wine, new bottle, or is there something substantive here? The consensus was there's substantive innovation going on. Your thoughts about where that innovation is coming from and what the potential is for clients. So if you're gonna implement AI for say customer care, for instance there's gonna be three raw ingredients. You need data, you need algorithms, you need compute. With a lot of the infrastructure that we laid down to capture data, it wasn't captured inside the traditional data systems, really anchored by Hadoop and the big data movement. We landed, we created a data and computational grid for that data today. With all the advancements going on in algorithms, particularly in open source, you now have, you can build neural networks, you can do statistical machine learning in any language that you want, and bringing those together are exactly the combination that you need to implement any AI system. You already have data and computational grids here. You've got algorithms, bringing them together, solving some problem that matters to a customer is like a natural next step. And despite the skills gap, the skill gaps that we talked about, you're seeing a lot of knowledge transfer from a lot of expertise getting out there into the wild when you follow people like Kirk Bourne on Twitter, you see he'll post like the 20 different models for deep learning and people are starting to share that information and then that skills gap is closing. You know, maybe it's not as fast as some people like, but it seems like the industry is paying attention to this and really driving hard toward it because it's real. Yeah, I agree. You're going to have Seth Dobrin. I think it's Niagara, one of our clients. What I like about them is the, in general, there's a two skill issues. There's one, where does data science and AI help us solve problems that matter inside of my business? That's really a, trying to build a treasure map of potential problems you can solve with the stack. And to Seth and Niagara, I think they're going to give you a really good basis for the kinds of problems that we can solve. I don't think there's enough of that going on. There's a lot of commentary, communication actually work underway in the technical skill problem. You know, how do I actually build these models to do? But there's not enough, and how do I, now that we solve that problem, how do we marry it to problems that matter? So the skills gap, we're doing our part with our data science lead team, which Seth opens, which is telling a customer, pick a hard problem, give us some data, give us some domain experts, we're going to be in the AI and ML experts, and we're going to see what happens. So the skill problem is very serious, but I don't think it's, most people are not having the right conversation about it necessarily. They understand intuitively there's a tech problem, but that tech not linked to a business problem matters nothing. Yeah, it's not insurmountable. I'm glad you mentioned that. We're going to be talking to Niagara Bottling and how they use the data science lead team as an accelerant to kind of close that gap. And I'm really interested in the knowledge transfer that occurred. Of course, the one thing about IBM and companies like IBM is you get not only technical skills, but you got deep industry expertise as well. Daniel, it's always great to see you love talking about the offerings and going deep. So good luck tonight. We'll see you there. And thanks so much for coming on theCUBE. My pleasure. Thank you for right there everybody. This is Dave Vellante. We'll be back right after this short break. You're watching theCUBE.