 Live from Boston, Massachusetts. It's theCUBE at the HP Vertica Big Data Conference 2014. Brought to you by HP. With your hosts, John Furrier and Dave Vellante. Okay, welcome back. We're here live in Boston, Massachusetts. This is theCUBE, our flagship program. We go out to the events and extract the signal noise. I'm John Furrier, the founder of Silicon Angle. Our next guest is Shilpa Luwande, VP of Engineering and Aiman O'Neill, Director of Product Management here at HP Vertica. Folks, thanks for coming on theCUBE. Really appreciate it. Good morning. Thank you for having us. We're gonna talk about the product, product management, the product roadmap. Obviously, Colin Mahoney at a great keynote. This conference is all about your customers and it's great. A lot of practitioners on. We had the U.S. Postal Service on, which was just an earlier, fantastic interview. I mean, that represents everything in big data. Internet of Things, legacy, you name it. Great customer testimony. You guys have had great success. Congratulations. But one of the things that Colin Mahoney said in my interview this morning with him was, is that I asked him if HP Vertica's solution was a Ferrari. And I was kind of teasing out the question of, hey, you know, it's a high end, high performance. He said, yes, it's a Ferrari. That's really the flagship. You guys are known, had been known for high performance. But as you guys have gotten with Haven, as HP starting to broaden their portfolio, what does the product roadmap look like? Because not everyone can drive a Ferrari and no one wants a Ferrari in their driveway because they're going to be embarrassed and can't drive it. And if friends see that they can't drive it, they're going to say, no, just for that reason. So how does a customer engage, that just wants to drive with big data? Not necessarily. And then maybe grow into being a Ferrari driver. Well, we've made the product much more accessible very recently because we launched a cloud version of it. So this is for a business that doesn't necessarily want to maintain their own hardware or set up a software cluster themselves. They can now access it, host the BIOS. We also shipped an appliance. We have a few customers now who like the fact that they get up and running much quicker. So they're able to do analytics much faster because they're not spending so much time on set up. So I think it's become more accessible than a Ferrari while it's still as fast as one. And we've always had the community edition, which is free for you to use for unlimited amount of time, one terabyte on three nodes. That also is, we have a big focus on community, so on building out the community of Vertica. So it's a Ferrari that anybody can test drive, and it's a Ferrari that... It's funny, I mean, I brought that up. I kind of baited him on that question, but the reality is that looking at your wins and successes in the marketplace, it has been really high performance data environments. And that's not going to be the one off anymore. It's going to be the rule. Pretty quickly, as people start realizing they can get their data pretty quickly, that's going to be a requirement. So I kind of wanted to get that up, but not everyone's ready for the Ferrari. I was talking to someone last night when I was walking back from Legal Seafood, and he's here, huge data warehouse, huge enterprise, and he's here to learn about columnar store and what you guys have, but he's coming from a legacy data warehouse environment. What's your product do for him? He's new to Vertica, so how does he engage and how does he view your product? So the way to look at Vertica is that the legacy products for data warehousing, they basically stop at that. They stop at the data warehouse use case, but now with all of the machine data that is being generated every day and the volumes that are growing and the diversity of data, those products don't handle that. Vertica is probably the only product in the market that can handle that entire gamut from the traditional business EDW type of use case all the way to the big data type of use cases. And so I think that's where, we're seeing a lot of traction in these really large data sizes in enterprises who are looking to go from just the EDW to the broad platform. Hey, man, how about integration? How does someone integrate Vertica? And let's just say that, oh, I love the vibe, love the product, love the roadmap. How does someone get involved with the product? What's the entry point? Well, I think with Differency, it's the huge number of certified partners. Most of them are here, but from ETL to BI, there are a lot of products out there that showcase the fact that they run well on Vertica. And so there's, in addition, though, there are plenty of APIs that are on our community where developers are very active in supporting each other. And our developers answer the questions live, but also many of the developers from our customers do. So I think that's one of the best ways to integrate is to get involved with the Vertica community. Got a great tweet from someone from another interview just said, I asked some examples of how people are pushing things, not forcing things into a relational database, but looking at the data perspective. I have to give some examples in this one person, Terry, said, a bank could not load ETE transactions fast enough in Hadoop, use Vertica as stupid fast ETL. Now they can identify profitable customers and transactions. What does he mean by that? Well, upstairs right now, the HPIT is giving a presentation about just that. They're showing you how you can analyze your sales pipeline. And they're using Vertica to examine, for example, how often salesperson changes the predicted close day, changes the size of the deal. And so they're showing people how to make, not be surprised at the end of the quarter and really understand the future of their business. So we've got a question here from the folks here on CrowdChat, which is better for getting things done? A Ferrari or a pickup truck? Guess it depends on what done means. I guess it's not a question, it's more of a statement. So engineering-wise, that kind of implies a road exam. What road are you on? Because a road, handle the car, pickup truck for loading things. What is a pickup truck in the big data world? If Vertica is a Ferrari, what's the pickup truck? So we do see this, people moving a lot of data. First of all, you've got to store the data. And so I think we see HDFS or the Hadoop file system becoming sort of that place where people collect all the data. And then getting the value out of that data is where I think Vertica shines. So whether you're just exploring that data, Vertica's broad SQL capabilities, as well as our performance, helps you sort of get value out of the data to understand what's in it. And then when you're ready to put that data in production, is when you use the high performance, the Ferrari part of Vertica, right? So Vertica I think has these- Is Hadoop the dump truck? Is it the pickup truck? It's certainly the place where people start. So it's certainly where people start. But John, I'll actually finally exhaust this car and truck metaphor to answer your question about the data warehouse. So we have a new feature that came out a couple of weeks ago that we call the zipper truck feature. Because the biggest challenge for these people with enterprise data warehouses is how do you keep all the traffic in different lanes? Because they have a lot of demands from them, from people who want to run batch reports to do payroll or comp plans. And then executives who want immediate response on their ad hoc queries on their, say, Tableau or ClickView dashboard. And typically these compete and fight with each other because the same queries, these different queries are going against the enterprise warehouse. So we have, for a couple of years now, we've had a workload management feature in Vertica where we can put these different kinds of workloads in different traffic lanes so that they don't step on each other. And we took it further in this last release where we can dynamically change the lanes. So you've probably seen these zipper trucks early in the morning when you're commuting that pick up the barrier and create four lanes inbound to the city one time a day and then four lanes outbound at another time a day. So that's the feature we did. We can now dynamically adjust the resource pools for different workloads depending on the time of day or the business priority. So I got to ask the roadmap question. How do you guys, what's on the roadmap for you guys? Because obviously you've mentioned Tableau, visualization, you guys are a database, right? At the end of the day, there's a big database. Now you got Haven. As these things come together, where do you guys stop and other people pick off? Because we saw news around Hortonworks as a new partner for HP. You have a relationship with MapR, Tableau's here, all your customers here. On the roadmap, where is the innovation going to stay core on the engineering side, product positioning side, and where do you leave off and others pick off? Yeah. So certainly that Hadoop ecosystem is something that we are trying to become a part of more and more because our customers are using it that way. And so the H in Haven is Hadoop. So what we are going to continue to innovate on and expand on is providing that Hadoop user base ability to do advanced analytics. Trying to leverage that data through the SQL paradigm and trying to get value out of that data. We are going to rely on the Hadoop distributions to really provide the core of Hadoop itself, right? So we are not intending to create our own Hadoop distribution. We would try to integrate with these. So we were working with MapR for a while, working with Hortonworks on things like Yarn integration. So we're going to play seamlessly with the other components of the Hadoop ecosystem. Same thing with the BI vendors. So we're leveraging, you know, Logi analytics, Tableau, MicroStrategy. We have a lot of partners who help us with the visualizations in how you consume the data that Wharticka now enables you to get at really fast. So that's how we see it. Any comments on that too? Yeah, well, we're going to focus on this scale. So we've gotten into some peripheral technologies recently but we focused on handling the massive amounts of data. For example, we have started to do some text and sentiment analysis of less structured data. But we still pull in autonomy libraries to actually break open that data and find the structure in it. But then we feed it into Wharticka to do analysis of massive amounts of data. So you'll see that again and again when we partner with, you know, when we get into peripheral technologies, we're going to partner with various people to understand the formats. But then once we extract data from those, we're going to be the scale answer. I got to ask you a product question, Amon, because this has been debated. Certainly Silicon Valley has been talked about. Product management, the best CEO of startups and product managers is a tough position because you're always saying no on one hand. And you also got to incent people to get things done on time. So you got to be kind of, you know, kind of played both roles, like a helicopter, get high level, get low level. What do you say no to a lot right now? Because big data is one of those markets where there's so much demand. I want this feature. You're simply smiling. I'm going to see just what your engineering team like. No more, we need higher, more engineers. There's so much demand. Where do you, what are you saying no to right now in the product roadmap? And where are you guys focusing on? Because that's really, I mean, everyone wants this bell and whistle in the software. So it seems to be unlimited and out of features that could be in products. So how do you handle that product question of the no? What's the no? Well, I am, I am blessed in that we have a- Oh, the ship will say no. But what makes it easier for me is the tight integration that we have with other parts of HB software like ArcSide, Opera Analytics and Autonomy because when I'm asked to add bells and whistles to say our tech searching or to our interface for, you know, showing how queries are happening, I don't have to put those into verdict and necessarily because I can leverage how they exist in autonomy and ops analytics and in our security product. So as these four or five HB software assets come tighter together, it actually makes it easier for me to concentrate on the core. So I'm going to ask the question I asked Jeff Kelly. It's a Hadoop world. I want you guys both to answer if you can't get on the same answer. You get the same answer, you don't need to answer. It's a Hadoop world. How does Vertica fit into it? So interestingly enough, the world of Hadoop and databases seems to be sort of converging quite a bit right now. The Hadoop world started out with MapReduce and, you know, I think as everybody has come to the conclusion that MapReduce is pretty much dead, you know, it's on its way out for sure. And the Hadoop world is moving now into a sequel world, right? So this is where I think Vertica has a lot to offer because one, we have a very mature sequel engine and over the course of working with so many customers, we've sort of grown up with the big data community in terms of expanding the nature of our product. You know, when I first started with Vertica, it'll be almost 10 years now. A terabyte felt big to us, you know, to do with a sequel analytic database doing a terabyte. We were very excited when we got our first benchmark on a terabyte, but now a petabyte doesn't feel very big to us. So I think that's where all of the learnings that Vertica has over the past years of what it takes to do a sequel on large datasets, that's what we have to offer to the Hadoop community. And your thesis there is MapReduce needs other tools are too complex to program. They're too complex and I think they're limiting. So sequel I think is a language that is, people realize that there is a lot of skill set out there that people already know how to use it. And then it's very simple to express a lot of things in it, right? So it's not something that can do everything, but it can do a whole lot of things very easily. And so if you can figure out the technology under the covers to make sequel be the language and leverage the whole ecosystem that's already built around sequel, then you can get a lot more out of your data lake basically. I think it's an HDFS world. When I ask customers over the last couple of days why they love Hadoop, the part of the stack that they love is HDFS. And so what Vertica can do for you is in our recent release a couple of weeks ago, we changed our storage layer so that now you can plug in HDFS to be your storage layer. This means you can ask us so much more data with Vertica. Previously there was maybe data that you didn't think was very valuable. You didn't query it very often and you didn't wanna put it into a database because you thought that was gonna be expensive or take your time to structure it. Now you don't have to trade off because we can query everything that's in your data lake because HDFS is our storage layer. So true or false, the future's about thinking about the data and not the database. Do you see that as what customers should be thinking about the data itself versus the database they're storing it in? Or is it still a database decision then, the data? What comes first? Thinking about the data or thinking about the database? I think, oh sorry, go ahead. I was just gonna say, if you don't have to think about the database, just pick Vertica, so it is all about the data. That's a good answer. I think the database is still important because what people loved about SQL is that it abstracts their business questions from the way it's physically stored and that's what a database does for you and you don't have to care so much about the details of the data's location. So final question, we're getting cut on time here, Aiman, I want you to help with the product perspective and Shilpa, talk about the engineering perspective. The question is, looking at where Vertica is today inside of HP, HP software, what you guys have done with Haven and across the portfolio, explain to the customers in your own words out there, people watching, what's the update? What is it all about? What is the product all about and what's the technology all about, Haven and all the big data activities? I think it's about being able to consume many kinds of data, so we'll take any format. Also, don't throw any data away. You can bring many, many petabytes now and also it's no longer about just BI, with the new ways of deploying Vertica, we've gotten it out into the line of businesses, so it's not just for the specialists anymore, it's become more accessible to a lot of more people in your enterprise. Shilpa, from a technology perspective, what's it all about? So HP has this unique combination of assets, so there's Vertica, of course, but there's autonomy which allows you to do human data and get value out of that. We have some leading enterprise security products and Hadoop, of course. And Haven is really all about putting those assets together and building an ecosystem around them that enables people to use these assets to really solve problems. Ultimately, HP's mission is to make things matter, make it matter. So we believe that with this Haven suite of products, you can actually make a difference to your business, to lives, and so on, as you heard today, the Conservation International Project. That's a great example of what you can do by putting some of these technology pieces. And making it reliable too, making it have a reliable product. And making it reliable, making it available in different form factors, so making it available through the cloud, through appliances, different ways to consume that data, but it is really all about trying to get the most value out of all your data, not just structured or unstructured. I've been called the Big Data Fanboy because I love anything to do with big data, but I think the conservation example this morning shows the value of across the entire world, industries from non-profit to for-profit, how transformative data can be. The engineering and product perspective here at Vertica, this is theCUBE at HP Vertica's Big Data Conference live in Boston, this is theCUBE. We'll be right back with our next guest at this short break. Thank you. Thank you.