 We're back here live at Hadoop Summit for SiliconANGLE.com's continuous coverage of Hortonworks event, which is Hadoop Summit 2012. I'm John Furrier with Jeff Kelly from Wikibon.org. We have Tasso Agarros. Agarros, great. All roads in Rome. Co-president of Teradata, Aster, Aster Data. Okay, welcome to theCUBE. Thank you, thank you, great to be here. So what do you think about Hadoop Summit? I mean, tell the folks out there who aren't here because they're watching the stream. What's it like here? What's the conversations like? What's the audience? Is it business people? Is it engineers? Yeah, I mean, the first thing I say, you guys are missing out. I'm joking. It's a great event. I mean, it was really good that it scaled up significantly from last year. You know, they're talking about 2200 people is what I tweeted early on the day. And I think it's interesting because there's definitely a lot of technology and a lot of technologies, a lot of vendors, but as big data becomes more and more mature, you get more people that talk about business use cases or how can you take Hadoop and integrate it with the rest of the enterprise infrastructure? And that's really something very important. Actually, we had a talk today about what we call the unified big data architecture, which is how do you take Hadoop and integrate it with enterprise, like the warehouses, like TheraData and discovery platforms like Aster. We have another session tomorrow about our integration with Hadoop and how do you bring together with SQL8. So, I think it's great. It has more of a business flavor than the previous years, I think, which is a great evolution. Yeah, so we hear a lot about Hadoop Enterprise ready and we also hear a lot about the integration with existing infrastructures. One of the key things that needs to happen in order for Hadoop to really be adopted mainstream in the enterprise. So, talk a little bit about how you approach that problem. Absolutely. So, I think the question of whether Hadoop is mature is not mature, it's a little bit besides the point. What's important is what are the right use cases that Hadoop can solve beyond doubt? And what are use cases that other tools like a data warehouse, for example, or a discovery platform like Aster, are much better fits. And so what's missing today, I feel, is a good understanding of how all these technologies can come together. And that is something that we're trying to pursue. We have worked with Hortonworks, we did some work to come up with a joint architecture, and that is very important. So, at the end of the day, Hadoop can have a very good spot in the enterprise, but it needs to be complimented by existing and new technologies. And that's just very important for people to realize. So, walk us through a use case where Teradata Aster is kind of complimenting Hadoop and vice versa. Absolutely. So take, for example, a use case with something we did at the customer recently where you want to capture all the interaction data of a customer across different channels. Say that you have a big bank and you want to capture web data, you want to capture call center data, you want to capture mobile data. And you want to use this data to understand what are the sequences of actions that may lead the customer to a decision that's important. For example, what are the sequences of action that may cause a customer to leave your company, to leave your business and go to a competitor. To do that, you need to capture a lot of interaction data that's really multi-structured data. You need to store it, archive it. And Hadoop is really a great tool for that, right? It's local storage, you can load quickly, and you can even preprocess data in Hadoop. And then you can take the data and move it to Teradata Aster. We can discover interesting insights, like what are the patterns that are more important to predict certain. And then once you have discovered those patterns, you can use it with Teradata to make it operational, make it actionable, distribute it to hundreds or thousands of people in the enterprise, and really make sure the right business analysts or business users have the right insight at the right time immediately available. That's an example of how you can use Teradata, data warehousing, and Teradata Aster as a discovery platform and Hadoop to collect and preprocess the data to solve an important business problem. Interesting, it sounds like you're kind of describing the big data life cycle. Exactly. Kind of from the crunching it into Hadoop, moving it onto the Aster, and then onto Teradata to actually put it into operation. Exactly, very interesting. So tell us a little bit about what's going on with Teradata Aster in terms of your integration with Teradata. It's been, I think a couple of years now, or close to that. It's actually a little bit more than a year, yeah. Just a little more than a year. But it feels more than that. It feels. So how is that going? And how are you kind of integrated into the larger organization? Absolutely. So Teradata Aster, essentially all the products of Aster have stayed independent. Teradata has made additional investments, and that has been great because we've been able to accelerate a lot of things that were in the roadmap. But on the other hand, because Teradata has an existing Salesforce that is global, we're able to take Aster on a global scale much more quickly than what we could do as an independent company. So frankly, for me personally, has been one of the most exciting things. Not only figuring out what's the next big innovation we want to develop in the big data space, but how do we take a new big data platform like Aster? How do we scale it globally, all the way from Asia to Europe? And how do we scale this as quickly as possible? And that has been a great experience. So how is the, sometimes when you're bringing in an acquisition and a new product into an existing Salesforce, there's sometimes some friction there in terms of there might be some overlap with the product or whatever the case may be. How is that process? And were there any kind of bumps in the road or you had to smooth out some edges in terms of going to market and really excelling? I would probably lie if I'd said that everything was 100% smooth, right? You wouldn't believe me. But I would probably say that it has been much better than what we expected. And the reason is that Teradata is really a best of breed data management company. Teradata Salesforce, I really expert in how do you analyze data. So from all the vendors that would really acquire after and from all the vendors that we could be part of, Teradata was the one that was the most relevant to our business. They were the most relevant to big data. And that means that when we're talking to people, we're talking to people that are used to selling analytics, selling data management, we're not talking to people that are used to sell big hardware or big storage or anything else like that. And that has made the process much more frictionless than it would otherwise be. Right, it's a little different than kind of the Green Plum EMC story where they're going into a storage company essentially and the Salesforce that's used to selling storage now has to, now is also selling kind of analytics. Well, you said that, but I cannot disagree. That said, they're doing a great job from what I hear, but nevertheless. So a question I want to ask you is given your experience as an entrepreneur, you kind of lived through the rebooting of the post.com bubble, get some venture funding, a little uptick, a little 2008 recession, the Sequoia Memo who backed your company to selling the company. So you've kind of lived through the war. Yes. And the battle successfully. Congratulations. Absolutely. But in this market, I want you to share your opinion with the audience around, in this marketplace with Hadoop and the entrepreneurship that's flourishing. What's your advice to folks out there because there's a variety of different solutions. I mean, there could be some storage, there could be some infrastructure. We're seeing a lot of infrastructure activity with flash and some new things around caching and with Hadoop and high availability. And then analytics side, it's a software opportunity. So kind of two different, some hybrids. What's your advice to folks out there who are starting a company, lessons you've learned, and advice for them? Yes, so it's a great question. Yeah, and I mean, we could be talking for hours. We have time. For something like that. End of the day. Right, why not, maybe we'll get a drink to it and have a good discussion. But I mean, it's a really rich environment right now. Yes, now on your first point about the ups and downs, there were definitely ups and downs. You mentioned 2008. In fact, I remember our timing was so good that the week that we went out to get fundraising in 2008 was exactly the week that Lehman Brothers collapsed. So I can take credit for that perfect timing. But what I would say is a couple of things. First of all, if you have a young company, you have to be prepared. You know, now it feels more like boom times, which is great, but you have no guarantee that this is going to last. So you need to focus on business value, solid customer engagements, make sure you deliver as much value as you can and be prepared for everything. Right, don't just feel that this is the way it's always going to be. And that's just a basic survival instinct. But talking about the comms that come out today, I do see a lot of new companies that are coming to market. I think there's two things that I see as a pattern. The first thing is that a lot of times people underestimate the time and effort it takes to build a solid product. For example, one of the big advantages of Aster is that we have a standard SQL layer that we can layer on top of Hadoop and Hadoop data with a SQL-based technology now, et cetera. It really took us more than five years to build a solid parallel SQL engine. And so there are companies that try to do this today, but they underestimate the effort. You just cannot do it in a year. It really takes a long period of time. So being conscious about the amount of effort you need to put in to develop a new technology is very critical for your planning and for deciding what you want to do. The second thing is that a lot of companies talk about technology today, both small but even big vendors. You see talking about the use of the term big data, most of the technology description. And really you have to focus on business value. Not so much what is the big data technology that's very important, right? And we've all been focused on innovation. But what are the use cases that deliver value in big data? A company that may have a business in the financial sector or in retail or in e-commerce, what are the key big data use cases that these companies should focus on? That's something that we try to push out a lot, but most companies, they're more about talking about the technology, forgetting the business value. But if you do that, it will take forever to get good traction. If you kind of start with the technology first, you're not necessarily going to find that business problem. Exactly. And then you've wasted the... I mean, then it's coincidental, right? Then you have to be lucky to land on a good business problem. You really have to start from the business problem. How do you try to land value to an enterprise customer? And then you have to think what's the required technology to support that. And there's really no lack of problems to be solved. I mean, from healthcare to retail, manufacturing, I mean, you name it, pretty much we're seeing every industry being impacted by this. Every industry has at least a handful of very critical big data problems. The challenge is that most vendors leave the exercise to the reader to figure out what are the most important business problems. What we are trying to do, we're trying to have our own experts there working with the customer to figure out these business use cases, which I think is very important. Talk about that in terms of the services component. Absolutely. How important is that? I mean, it sounds like it's pretty important. And how do you, what's the actual approach? What's the customer engagement like? Yes, that's actually one of the great things about being part of Teradata. When I came to Teradata, I learned about some, you know, really innovation that Teradata has developed. The first is that Teradata, most vendors out there that go to sale with a salesperson and maybe a technical solution architect. But Teradata has one more function which is the industry consultant. And these are usually people that used to work at enterprises, you know, healthcare, retail, and now they came to Teradata to help the customers connect the business problems with the technology. So this industry consulting function is very critical. It removes a lot of the burden from the customer to figure out what's the best use case, what's a really big business problem, what's the ROI for a use case, and how do you utilize big data in their environments? The second thing is that professional services is definitely an important part of the story. Not so much because you need people to just wire things together. I mean, that doesn't take too much time these days with almost any system. But more about help, you know, get experts that have delivered business value with big data in other enterprises and know exactly what they need to do to deliver that value in your company as well. And that's something that if you don't have it again, if you leave the exercise to the reader, meaning to the enterprise to figure out how to get that business value, it's just, you know, you risk that the projects may fail, it may take more time, you may not get the return that you're looking for. Right, well that's, right, exactly. If you get too enamored with the technology and that's where the risk is around big data, is big data all hype and it's not all hype, but if you start with the technology and you don't, in a relatively short period of time, find the business problem and actually deliver some business value, well then that kind of says, well, maybe it was hype. Exactly, and actually that reminds me of the famous, you know, advertising motto that says, you know, an advertiser would just, a marketing person would say, I know 50% of my advertisement is crap, I just don't know which 50%. Well, in big data, 50% of the use cases are crap, but we know what 50% is crap and we can help the customers do the right things, but not everybody's thinking it that way. Well actually, I just wrote a post actually yesterday on Forbes about this whole 50% of advertising and I use that line because big data's disrupting advertising because now with social data, you can actually measure 100% of all interactions online. So what's happening is Coca-Cola just announced a, I actually broke the story of Coca-Cola running a prototype in Latin America around direct concerts, Paul McCartney, some, you know, some bands, instead of spending money on TV ads, they're spending money on direct media, cutting the middleman out using Facebook pages and YouTube analytics rolling up all on big data under a company called This Moment in San Francisco. So Coca-Cola essentially cutting the middleman out of the equation. Which actually a pretty common theme of a lot of innovations that happened today, a lot of innovations and a lot of startups, not only in the data space, but more broadly, they're all about cutting the intermediary, right? And that creates huge efficiencies for the economy, which the economy really needs efficiencies these days, right? And that's a great opportunity. Well, we're doing a software project and I would agree with your advice for entrepreneurs that's really find the use case, deliver value, but focus on the product. But to your point about efficiency, if you can create efficiencies, you'll make money in markets that are growing. So, and that are complex. So I would add to that, Thasur, thanks for coming inside theCUBE, but congratulations on your entrepreneurial success and now the next level of your journey with the big data at Teradata, Thasur, congratulations. Thank you, I appreciate it. Glad to be here. Maybe we all get the best analytics coming forward. This is theCUBE, we'll be right back with our next guest after this short break.