 Boston, Massachusetts. It's theCUBE, covering HPE Big Data Conference 2016. Now, here are your hosts, Dave Vellante and Paul Gillan. Welcome back to Boston, everybody. Jeff Vista here, he's the man behind the event. Great to see you again. Thanks for coming back in theCUBE. So, well done. Thank you. Fourth year, you keep up in the game, great outside speakers, really motivational. Phil Black yesterday. Today, another awesome presentation from Miguel from Uber was phenomenal. The customer panel that Chris ran. So, it's good, yet you've been able to preserve the ethos of this conference, which is... Which is really important to us. Customers, right? Yeah, and I try to explain that to people, and it's one of those things where you have to kind of experience it. You can describe this as not a marketing conference. It's about substance. It's about getting real information, about making it authentic. You can say that all day long. I think when people come here, and it's happened here, people get here, you can say that to your blue in the face, and then they experience it, and they go, I had no idea. That no idea that you can go after session, after session, talk to you directly to our own engineers that will tell you the truth. You know, marketing people don't necessarily do that. So, they can get the real skinny of what's going on and get real answers, and then they can talk to real customers, and once again, get the truth about what they experienced. And that connection with those communities is just as important as what goes up on the fancy main stage. One of the things you do a lot at this conference is spend time in briefings, in meetings with customers, getting their feedback. What have they been asking for? Well, they want more of the same and a lot of new things. And what I mean by that is that, for example, with the Vertica franchise, our customers that use that advanced analytics database, they want speed and scale. And you know what, we'll be here 10 years from now. You know, maybe we'll be drinking something different, but they will want speed and scale. I can guarantee you that. So, they want to know how they can go fast, how they can bring lots of data and start applying that capability. And there's really no boundary condition. We see what we thought was fast and great and Olympic caliber before is now becoming the status quo. So, that's point one. And I think it's key for us and other leaders in this industry never take their eye off the ball. But now the pivot in this industry, and now it's gone from being a novelty or interesting to something that's becoming a little more air and water, which is that fact that you have to bring in increasingly advanced analytics into what you're doing and you have to be able to connect and work in your ecosystem. And I know that sounds obvious, but it's about being able to analyze data without moving it and being able to work within the data fabric that organizations have. The error of the data grab, I think is quickly becoming dead and buried. People want to have access to the data, they want to be able to utilize it. And any solution that says you have to put all your data with me, I think is a non-starter. So, Mike Schultz hopped on a plane from Seattle, came down, was up on the stage today. You guys announced Haven on demand for Azure. Tell us about that deal, how it went down, why Azure, how did it come about? Well, it's a huge deal for us. And it's part of a bigger strategy, which is what we call any cloud. And it's that simple that if you move from one lock-in to another lock-in, I don't think that's progress. And this industry historically has had lock-in. Maybe organically, maybe simply by design. But the old traditional data warehouses did lock-in companies and now as companies are moving to the cloud, they don't want to move from one lock-in to another. They want choice, and this is how I put it, they want choice on day one and choice on day 1,000. And we explain what I mean by that is, giving choice where you want to start up a new workload or a new instantiation to be able to do your analytics, that's fairly obvious. And that's something we've taken our unified architecture to do. But what is maybe even more important is day 100, day 1300, day 1,000, if I want to be able to change or augment, this cloud vendor has got a deal of the day. And I want to take advantage of it. Having that flexibility to be able to do that without the friction, without the risk is key. And part of that is making sure we can support in a best-in-class way any cloud. And so of course we've been running formally on Amazon Web Services, very key strategic partner, and a fair number of our install-based runs on that. And now we've added to it Azure. Now, Azure with Microsoft also brings in the strategic alliance that the Hewlett Packard Enterprise Company has formed with Microsoft to be able to deliver best-in-class hybrid computing. And so this is an element of that bigger alliance. And then I think it's going to give two excellent choices. And we're going to add more so that wherever you want to take Vertica and run in the cloud, we'll hopefully have that option for you. So explain why that's not a lock-in to people. So if I'm running Vertica in AWS, I got to go through the Amazon API. I'm running Vertica, I got my processes wired around Vertica. How is that not a lock-in? Well, first of all you need to have the ability to go other places. And then the question is if I want to run on Azure or I want to go back on-premise, and it might not be moving your whole job, you may be spawning up other workloads and you want to be able somehow to cash the data you need and run on that so you don't impact your mainstream job. So it's not always switching out willy-nilly, but it may be augmenting what you want to do because of a need to have that capability. By having one analytical engine where all your queries, everything that you've written, all the projections that you've written will run unchanged no matter what deployment model you're in. That makes it viable. And that's something that's very rare, unique to us that we can go from on-premise, on open source, and in the cloud, without any changes to the query jobs and the models that you've put together. The other point we did by the way is we've always had support for Hadoop. And that came in two flavors. Flavor one was ingest in from Hadoop and we've had that for a while. And our competition for the most part has that too. So you can bring data into Vertica from a data lake. So check that box. What makes us more unique is what we had announced a year ago, which is running Vertica directly on a Hadoop node. So if you have those servers and those resources, you can bring the analytics to the data. Check box two and do that at Hadoop Economics, which is part of the Hadoop value proposition. Now of course you're in a Hadoop format. That's not always going to be optimized. So there's going to be trade-offs. But for a lot of companies, certain use cases that may make sense. The trifecta, if we can talk, course racing here. It's over this weekend. There you go. Is that we just added support for Hadoop analysis in place. And let me emphasize what that is. That allows you that if you have a traditional Vertica deployment, let's say it's running on premise in your data center. And you have your mainstream financial data in Vertica in its optimized format. And then you want to join in some data that is sitting out in your data lake. Maybe some customer support information. Now you can go and connect and query and analyze that data without moving it at all. And be able to bring that supplemental data into the query model. You don't have to bring it into Vertica. You can if you want, but you don't have to. And you don't have to bring Vertica to it. So it is a third offering and the whole idea is we want you to increasingly analyze the data you want to analyze and do it as effortlessly as possible. And to our knowledge, we are the first vendor that offers this capability and the only vendor that offers these three flavors. So it's about embracing open source and that's the other element. It's not just speed and scale. It's not just getting to the cloud. It's being able to bring analytics in a best fit way. Do you do that through connectors? I mean, how do you get at that data? Well, the product or the capability is called a reader. And that doesn't necessarily sound very sexy, but it's optimized. So we have to understand, we've worked closely in this case with Cloudera and Hortonworks to optimize our reader technology so that we can really, really intelligently understand the formats that they're in and be able to have high performance to do that. So we not only connect to it, but we have to, in an optimally way, be able to access that technology and then be able to run those queries and analyze it. And that is what we have facilitated. And that's why you don't have to move the data and you're going to get high performance. And performance means concurrency, not just throughput, right? And that's a big challenge with a lot of the Hadoop technologies. They're still fairly early day and you take up concurrency into lots of jobs running. You'll see other vendors have challenges, challenges that we've surmounted years ago and that's one reason you can take a proven technology like Vertica. So it's best of both worlds. You get the data lake of your choice, but you're going to get proven high performance. It's talking about cloud portability. I mean, isn't data sort of the new lock-in though? If you have 10 terabytes of data on Amazon that you're doing your analytics on, you can't easily move today over to Azure and work on that data. So how do you address that problem and giving customers that choice? Well, there's no magic wand, but I think one thing is that point of bring the analytics to the data. So if I don't have to force you to move the data to accomplish any task, drive a business process application, do analytics like we're talking about, perhaps we'll talk about the machine learning APIs that we came out with that can interrogate that data. That gets you half the battle done, then you don't really care where the data is. Now, where the data does reside and the format's it in and is it on a slow network in Prague, or is it in a high performance flash server sitting two feet away from your server? There's laws of physics that imply there, but if I'm able to have a technology that has an I don't care view about where the data happens to be stored, you're halfway there and you're not initially locked in, you haven't locked in your analytics with your data sources, you're able to separate. And that allows you to do two things. It allows you to use the right tool for the right job and it allows you to supplement your data sets. That's the big thing today is these organizations have these diverse silo data and they're dying to be able to bring that in. And sometimes the data's not in their four walls. Sometimes it's public data, it's data from their partners. It could be weather data. And right now you don't necessarily want to bring that all into your data center, but you want to be able to tap it and utilize it. And we haven't even talked about IoT and the data wave that's coming in from the edge. So it's not about one huge data store. Those days are gone. It's about let the data sit where it naturally is, use economics and performance judgments and make the right trade-offs. And those can be dynamic and then have technologies and platforms that can bring analytics that can deal with it. And that's the target we're going to have. And I think the answer to the locking question is choice. I mean, open has, the definition of open used to be UNIX, right? And now it's become either open source, which is okay, that's one spectrum. And I think the market's proven, it's two-edged sword or it's choice. I can run an Amazon, Azure, on-prem, et cetera. So I have choice of deployment now. If I go and put processes around Amazon, well, that's a formal lock-in, is what it is. I don't think that's ever going to go away. There's going to be in different ecosystems, different capabilities. And if you take advantage of them, then they're unique to that. But with Kafka and these messaging buses that are very intelligent, the ability to pull out from your main data center, get a task done and then return back, that's the kind of orchestration we're talking about as opposed to one-stop shopping and everything happens. It's not one shopping mall anywhere, right? I mean, it used to be, you went to where the Macy's and Nordstrom's were and you could shop at the stores in between and that was it. And now, you want to be able to go to any store anywhere, any time. And that is where I think computing is going. It's that anchor, the anchor concept of start in one place and everything's got to be there, it doesn't make sense anymore with the modern hybrid systems. Well, Tim Crawford's coming on too. I want to get his perspective on this issue, which is from a CIO perspective, the trade-off between not having lock-in and getting value and how a CIO thinks about that because it's not a no-brainer. I would argue that actually most people would accept some degree of lock-in, maybe even a lot of degree of lock-in if it delivers business value. And that's a personal decision that every business has to make. Well, you're hitting on an important point, which is we've seen it with these various open standards that drive a lot of innovation, right? And that we embrace, but that doesn't make that maybe early-stage product magically the best. And so the term we're using much more often is about best fit. It's like, you want a great solution. You don't want lock-in. You want to be able to have choice. You want to be able to have flexibility, but at the end of the day, it's got to perform. There's no free lunch. And I think sometimes when the new wave of, quote, whatever the flavor de jure is comes where you see the fact that it creates this open standard, it doesn't magically make that technology performing. And we're seeing that repeat in waves and now I think people understand it's a balance. All right, we got to wrap. Give you the last word. The conference, so tonight you got another customer event, right? Half day tomorrow. Give us the summary of the event. Well, this event I think is built on what we've done before and really taken what separates this particular business from others, which is we're about making it real and very authentic as we connect and service our customers. We have pushed the envelope with some new announcements, whether taking advanced analytics, columnar databases to the next level, or the amazing, exciting announcement we made in applied machine learning, now with this offering called combinations. That I think the big takeaway there, if you check out our information, is it takes this topic of machine learning, putting intelligence in for the mainstream software developer. And I think that's the miss of the whole industry. Everybody's either doing, building a self landing rocket, or they're trapped in kind of the geeky side of the data science, which can be very compelling and is suitable for some solutions. But there's the other 90% of software developers that are just trying to make their app better. They're trying to get that new consumer mobile app out. They want to take their CRM application and create an advantage with it. And they just want to get a solution that's going to help them. I don't necessarily want to get a PhD in data science. Haven on demand and idle, those two franchises, which we coined this term applied machine learning, they're about pragmatism. And that separates us from, I think a far majority of the industry, we've got a lot of interest in that and that's worth checking out, because that's a big bet for us. Well, Jeff, thanks for coming on theCUBE. Thanks for having us here. This is, I can say, our fourth year in a row. You guys have been very supportive and also respectful of our editorial model. So we really appreciate that. You guys are an institution and you keep us on our heels and you never give us the free pass and we appreciate that. So thank you very much. All right, our pleasure. All right, keep it right there. We'll be back with our next guest. This is theCUBE. We're live from HPE's big data conference in Boston. Right back.