 Okay, welcome back. We are live in Boston, Massachusetts. This is SiliconANGLE and Wikibon's the cube, our flagship program. We go out to the events, extract the signal from the noise. This is the HP Big Data 2013 event. Hashtag HP Big Data 2013. Tweet to us here in the cube. It's the HP Vertica user conference. I'm John Furrier, the founder of SiliconANGLE. I'm joined by my co-host. Hi everybody, I'm Dave Vellante of wikibon.org. As John says, we're here in Boston at the HP Vertica users conference. A lot of users, a lot of practitioners just came off the, out of the keynote where Billy Bean was giving an overview of his life, the book Money Ball was outstanding. Brian Doherty is here and he's the chief data warehouse architect at a consultancy called CMA. They're based in New York. They got offices in New York City, but they're actually headquartered in Latham, New York. Near Schenectady, John, outside of Albany, where I went to school up there at Union College. And woman-owned business. Brian, welcome to the cube. Thank you very much. Great to be here. So did you get the chance to see Billy Bean? I saw him. That was pretty good, wasn't it? Great, great example of how analytics are being used in every field now. You know what's amazing to me is how the Oakland Aces have been able to show some data that suggests that they were able to sustain that when that book came out, I read it. And big Michael Lewis fan said, Billy Bean's an idiot for letting the cat out of the bag. But they've been able to sustain in what ostensibly would be a copycat league, but really maybe MLB is not as much of a copycat league as the NFL. Are you seeing your customers use data in a similar way for whether it's competitive advantage or social gain in the public sector, et cetera? We're seeing our customers embrace using data and data analytics to drive their business. What they used to make decisions based off the cuff for intuition, they're now using data to drive those decisions. And we're seeing it across the markets that we, from healthcare to financial services across all markets. Yeah, so talk a little bit about CMA. Why don't we start there and sort of jump to head because I'm very excited about Billy Bean's keynote. Talk about your company a little bit. CMA is a one-owned business. We have about $80 million in revenue a year. We focus on really two main use cases. One is large-scale data warehouse and analytics hosting. And we also build a series of products that we OEM and sell. One's an Oracle RAC cluster computing platform and one is a big data analytics computing platform that we embed Vertica as our MPP solution. So the RAC cluster is, that's your own design? You guys do things? That's our own design. We RAC stack, wire it. We install all the software. We build the database, all the clusterware, all the storage layer. We configure all the hardware and then we ship it out. That's your own converged infrastructure. We do the whole converging infrastructure. So you saw this, you're watching the trend, right? Everybody's doing converged infrastructure. Six years ago we started this with Oracle RAC. It's been very successful because a lot of companies were coming out with reference architectures. And you can give a company a reference architecture, but then they still have to implement it. There's still a lot of risk in that. What we felt was if we could RAC stack, wire it, take that repeatable solution, take the risk out of the equation, it was going to be a very good thing for our customers that's turned out very well. What we've done with Vertica now is we've gone away from the traditional data warehousing and we moved into more of a big data analytics platform. Same concept though. RAC stack, wire, install all the software and then roll it out to the customer. So talk about that piece of it. So you were talking off camera, essentially a Hadoop solution, right? It's an integrated Vertica and Hadoop solution. So what we've done is what we try to do is in addition to delivering the hardware, we try to add value up the stack. So we deliver not only Vertica on the platform, but also Hadoop on the back end. It's kind of a landing zone for streaming data that comes in from either devices or other Oracle SQL Server databases with Vertica as the target. And that's a Hortonworks distro? It's a Hortonworks distribution that we implement. And do you do your own service? We do our own HDFS key file system, MapReduce own logic that we've built, streaming logic. Some of what we do is we take input packets from different devices and we provide our own transformations and we load that into Vertica. So it's a really unified, comprehensive solution that we roll out. So you've been a data warehouse architect. You've seen a lot of changes in the business. I often say, John, you've heard of me say it many times. It's like the Hadoop tail is wagging the dog. But maybe for good reason. A lot of data warehouse architectures are challenging, let's say. I'm going to say that some of them are downright broken. And so talk about how that whole space has evolved, made a lot of promises over the years. Can Hadoop live up to these new promises? But talk, take us through sort of the evolution. So the evolution, I'd say in the last 15 years, a couple of things have changed. First, the scale has just grown exponentially. When we first got into this business and we're talking about hundreds of gigabytes or low terabytes, now we're into the hundreds of terabytes and petabytes of storage. Brought to bytes, I heard this morning. Yes, yes. And we're also looking at the velocity of data. Traditionally, these data warehouses, we could take a weekend or a month or daily load. And now we're really ingesting data at the speed of thought. We're ingesting data every second, pulses every second. So the adaptability, the velocity and the scale have all changed significantly over the course of 15 years. And because of that, an adaptable, flexible platform and some of the key characteristics of Hadoop, we can take advantage of our key to our solution at this point. Okay, so where do you see this going? Do you see the Hadoop-like architectures, you talk about your solution. Do you see that becoming the sort of center of the universe for data? Or do you see the traditional data warehouse sort of morphing to accommodate that world? Yeah, we think what's going to happen is a couple things. We think we're going to see a slow morphing of the traditional enterprise data warehouses where we have an ETL platform, a metadata platform, a BI platform, a data warehouse platform. It's going to morph and change somewhat. We don't really have the opportunity to take our time and to have kind of a structured, rigid platform. We think Hadoop is going to take on more of that traditional ETL platform. So something like a Oracle Warehouse Builder, Informatica, some of the more expensive and costly products are going to be displaced. Hadoop is disruptive and it's going to provide more of a fluid streaming ETL platform. Working in conjunction with topologies and databases like Vertica, they're going to apply that infinite scalability here. So the guys are all, the folks out there watching, the stream, this Billy Bean keynote is getting out. You hear a little background noise and we're going to up our mics a little bit here so you can hear us. Great keynote by Billy Bean. Brian, talk about this purpose-built environment. You guys actually have a turnkey solution. The hardening of some infrastructures, what we're seeing at infrastructures of service, certainly at the cloud levels as well, independent of what you guys are doing, it's creating an innovation for software. So the enablement of say Vertica or whatever solutions changes the software model. Could you talk about what you're seeing on top of you? What are you guys enabling? Well really what we're trying to do is to come out and provide what we would call a grid, kind of a large-scale clusters, a private cloud if you will, that we roll out and that we can very adaptively cut up into clusters. So we can roll out 128 nodes, 64 nodes and we can cut those nodes up into very flexible clusters, the Duke clusters and Vertica clusters and then we can drop software on top of that that gives these companies the flexibility at a low cost, a much lower cost, to be able to cut, slice and dice them up and deliver a heterogeneous set of functionality. How much lower cost? Can you quantify that? Significantly lower cost. We found when we moved from kind of our traditional big iron, large mainframe or large Oracle Rack clusters, we're cutting our cost, hardware and software costs by a factor of four or 500%, significant cost. So what used to cost us 10 million in hardware, 20, 30 million in software, we now are trying to target down into less than a million dollars. So for hundreds of thousands of dollars, we're providing this comprehensive platform. That's astounding, economically. What do you give up when you go there? What's that little 20% that maybe not everybody needs, but some people do, what's that, Jim? You always give up some degree of functionality, I think. I think when you look at what we're offering with our big data solution and you compare and contrast that to Oracle Rack solution, there's certainly some functionality within Oracle Rack solution that we do not deliver within our big data platform, but there's also an enormous amount of functionality in the big data platform that we deliver. So this solution, this migration to this big data platform in Hadoop, doesn't have some of the mature technologies that have been around 15, 20 years that have really embedded a lot of rich functionality, but that tends to kind of mitigate and erase itself very quickly. So we expect within a couple of years, and as the economics of the hardware come down and the commoditization of the software is just, it's going to be harder and harder for the large vendors to command that extreme premium for their software functionality. So the business value of the new stuff is, in your view, overwhelming or going to overwhelm the traditional? We think it is. We think the economics, when you've taken the consideration, the commoditization of the hardware and the software, it's going to be very difficult for traditional companies to continue to command the premiums when you have these disruptive technologies that are basically leveling the playing field. And we've looked at this pretty closely at Wikibon and Silicon Angle and where the money, this is not a race to zero because where the money needs to be made is services. Yes, that's correct. And integration. And that's a key part. There are a lot of companies offering converged infrastructures out there, kind of the hardware layer. What we're trying to do is really move up that stack to a full functioning platform that delivers not only integrated hardware, but software and then retail software on top of it. Right. Okay, so talk about what you're doing here at this event. Maybe the interactions that you're having with customers, what are those like? We're doing a couple different things. We're seeing a lot about Hadoop. We're kind of seeing here, and I'm talking to a lot of customers about what they're doing with Hadoop, what we're doing with Hadoop. So for example, one of the kind of the technical things we're looking at is, is it better for us to deliver Vertica cluster and a Hadoop cluster on different clusters and then kind of wrap them together in one 42-year rack and roll that out? Or is it better to deliver a shared cluster node? So run Vertica and Hadoop on the same cluster, kind of manage the resource within the cluster, reduce the cluster set and then deliver that. So we're talking a lot about Hadoop, a lot about Hadoop integration with Vertica, a lot about packaging. How do we package Hadoop and Vertica together to get the lowest cost and highest value for a customer? And we're just talking a lot about streaming data. One of the big topics here at this conference is how can we kind of capitalize on all of this streaming data that's flowing in and deliver really two things, immediate response back out to the device and then on the back end, take it in for deep analytics going forward. Brian, what's changing? I want to ask you the question. What's changed, obviously, in the past couple of years, specifically in the database area and also in the data warehousing area that's changed? Obviously, it's the legacy industry, both on the BI and data warehousing, ITL and BI. And stuff was built differently, right? It's like horses had roads and cars have roads and sports cars have tracks. Now you have performance levels going on and compute that people have never seen or dreamed or creatively thought of before. So those old constraints kind of need to go away, but yet they're still bloated in some of the old software models. Well, this is not paving the cart path, is it, John? Yeah, I mean, the old, you know, whip and buggies build data warehousing. The cow path. Yeah, so, you know, the cow pasture put the old stuff out the pasture, but what's changed? What's the new software look like? What's falling down? What needs to be put out the pasture? The biggest change, I think, is this whole concept of speed, speed to market, and responsiveness of the enterprise platform. You know, we don't have the opportunity to take, you know, two to three years to deliver a very large scale enterprise data warehouse and then have a very structured and rigid infrastructure where we can't really adapt and change. It's a much more fluid environment. There's lower cost hardware, lower cost software. Something that needs to be much more adaptive can take in data from a heterogeneous data sources, not tied to databases only. So it's a much more fluid environment, much more heterogeneous environment. You know, when you think back 15 years ago, no one ever thought about ingesting device information into your analytics or your enterprise data warehouse. Now we see the device information as a really high growth area going forward. So. And now you have the Internet of Things. Cisco calls it the Internet of Everything and then GE calls it the Industrial Internet. But at the end of the day, it's the Edge Device. It's the Edge Device. That's really where the, really the growth, the high growth area is, and that's where we start to focus. John wants the instrument genes. I don't know, it's interesting. I could have gone there, but if we had a female guest on. But at the time, you know, having sensors on your genes would be an interesting. We had guests on, really. Yeah, guest genes, you know, kind of a joke. Very predicted, folks. You heard it here first. But you have to be able to use, you have to be able to use that information. Hey, video games, use it, you know? You see Tiger Woods swing. How does that happen, you know? I mean, we're all going to be in one big video, first person shooter game in the future. We're trying to make sure that we have a good window of use of that data coming in, because I think what we're going to see is, I think we see a lot of companies overwhelmed. Sure, you can open up all this windows to this device information, but that's also an opportunity for companies to drown. So we try to provide kind of a window and ability to drive those hundreds of terabytes of data and companies to get use out of that information. All right, well, hey, love to hear more. We got to take a break now. We're coming down to the end of the day. This is the day one coverage of HP Vertica's end user conference. I'll give you the final word. Obviously, it's an event among customers, people who are sharing stories. So let's end this segment by sharing what you're hearing in the hallways. What are some of the coolest things you've heard from? Other folks and peers and maybe not peers or competitors or just HP, what have you heard that's been exciting in the hallways? Yeah, I think that one of the most exciting things I'm hearing is just the hype and the interest around the conversions of the database industry and the Hadoop industry, and kind of the value add of Vertica. A lot of people are excited about Vertica. We think it's a great platform for us and a lot of excitement about Hadoop, combination of Hadoop, Vertica, and being able to handle that streaming data. Okay, Brian Dardy, thanks for coming on theCUBE. We really appreciate it. Thanks for sharing your knowledge. A lot of people are excited who want to get their teeth into this new modern architecture and a new modern infrastructure. The old stuff's going to be put out to pasture, right, Dave? So we talk about that all the time. Data Warehouse is a dinosaur, some people say. The horses are not on the track. It's the car's NASCAR. We'll stay with the NASCAR analogy here and again, it's just the beginning. So it was exciting. We'll be right back with our next guest after this short break. This is the HP Vertica users conference. It's theCUBE. John Furrier with Dave and lots of right back.