 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 everyone. We are live in Boston, Massachusetts for HP's Big Data Conference. This is theCUBE, our flagship program throughout the events to extract the signals from the noise. I'm John Furrier, the co-host with Jeff Kelly in this segment, Wikibon analyst and big data analyst here inside theCUBE. Our next guest is Brad Hooper, VP of Technology, Office of the CTO of TIPCO, and I came from Spotfire, welcome to theCUBE. It's great to be here. TIPCO's been around the block for a long time and everyone in the web business knows the success they've had in web services and beyond. Notifications, events, large scale complex stuff. Pulling it together. But now we're in a different world. Give us the update, I'm obviously still relevant. What's the new thing at TIPCO for the folks who want an update? Give them a quick update on where you guys are at, position-wise, what's your key markets? Sure, so TIPCO, long standing player, key player in the integration business, but TIPCO's been expanding our technology and use cases pretty dramatically over the past 10 years to include in particular event processing, which we were one of the innovators in that area, but we recently, with our business events product, we recently added stream-based acquisition coming from Michael Stonebreaker, kind of heritage. Also analytics, the spot-fire team that I hail from before becoming part of TIPCO proper, and recently added Jaspersoft as well for embedded analytics. So we've been pretty focused on this, the use cases that involve understanding data and then taking action on data in a kind of closed loop, full circle. So you must like the segment we did with Tom Davenport we just had around business competitive advantage. That's your shtick right now, using data for dollars, both top-line revenue generation, but also for competitive strategy. Absolutely, spot-fire actually was featured in his competing on analytics book back in the day, and we're still doing that. One of our key concepts is that automation on its own is impossible. You need to have what are the rules actually that are gonna run and what are the decisions and the actions that you're gonna take. So this idea of combining human-based decision-making with automated kind of events, response and action-taking, and that loop is really critical. So let's get Jeff Kelly to kind of, that's the real deal here. Let's find out, does that make sense? Do you agree with what Brad was saying? I mean, obviously he's a little bit biased, but incredible, he's a VP and techie. So what's your take? Well, I certainly agree that you need a combination of both the automation to make those real-time decisions where you actually have a system-making intelligent decisions based on real-time analytics. You've got the human aspect which you can't take out of the equation, both in terms of kind of validating that the real-time automated decisions and actions are correct, and doing some of that deeper analysis, historical analysis that really feeds the models that you're building as they move into automation. So it sounds like, well, what I'd like to ask about is the kind of the marriage of both kind of the traditional TIBCO side with the both Spotfire and now Jaspersoft. Maybe I do have some more questions around Jasper in particular, right? Did that go into Spotfire? You guys put that together as a still separate question. It is together actually at this point. We combine those into an analytics group which is run by Brian Gentile, the former leader of Jaspersoft. So basically how are you integrating those two parts, the kind of the TIBCO proper and the more the analytics side in a way that closes that loop, if you use that term, I think that's a good one. We heard yesterday in the keynote about not from Colin Mahoney, not just about closing the loop but tightening that loop. Continually tweaking it and fine tuning it. How do you guys go about marrying those to the automation and the analytics? Right, well, in the era of big data, it becomes even more important that we tease out the nuggets using good, powerful analytic techniques and also the business expertise of end users who are skilled in that particular area. And so we focused on being able to transition from that human-based understanding into building a model and taking that model of business behavior and pushing it into the real-time system. So it's not just about statistical modeling and by the way we have cutting edge, closed source execution engine for R and the only one in the marketplace, 10 to 50 times faster than the native open source R engine. Not only about building the right models but about or about the automation but about how you smooth that path from human-based insight development through to execution against the right opportunities and then of course taking the action whatever that may be, making a contextual offer or redirecting traffic in the supply chain and there's a hundred different use cases to which that model is lost. Right, we're gonna have some fun here. We're gonna have some fun here. Oh, is that in supply, Jeff? No, I was just gonna point out, I mean, that's one of the biggest challenges in this space there's a huge, obviously the market for data visualization is a big market in BI where you've got a, you're a human being, you're an analyst, you're an executive and you're looking at some data and you're visualizing it so you can better understand it but ultimately you've got to translate that into action and increasingly with the speed of business you don't have time for a human to sit there and say, well, okay, we're going to make these decisions. I mean, the old, whether it's LinkedIn or Amazon is probably the example people understand the most is people you may know when on LinkedIn you log in there, they recommend some people based on all that kind of analytics in the background and that's an automated process that happened based on a lot of deep data science and that's one of the challenges of going from that deep data science to productionizing that into essentially an application that actually does something. So that's an area I think, John, where there's ripe for opportunity. Yeah, and let me just respond to that and then also respond to your early query about Jaspersoft, because I think those two kind of fit nicely together because we've been focusing for a long time at Spotfire and at Tipco on creating applications that are very easy for a business user to interact with the statistical model. So they don't need to be a statistician in order to take advantage of that intellectual property that's developed. So, but that's within the analytics sense kind of business intelligence kind of framework in the context of delivering an application to the user that's dealing with the data and so on. Now, there's been a kind of longstanding rule that only about 70 or 75% of people who need or who would benefit from business intelligence are actually using it. And one of the hypotheses of the Jaspersoft organization is that the reason for that is not because they're not smart or they don't need it, but they're actually, their typical business use case is in an application that does a particular job, whether that's a supply chain application or a CRM application or so on. It's not a BI application, it's a business application. So the Jaspersoft commitment and now the TIPCO commitment is to deliver on what we call embedded analytics. So we realize that people can always spin around in their ergonomic chair, from one application to a business intelligence tool to a statistics tool, we wanna bring the analytics to the user who's making the decision in the context of that business process. And so that's where they fit. Spotify is more focused on an analytic use case, whereas Jaspersoft is focused on an in-process decision-making use case. So they're very distinct. So embedding was, as Colin says, put smarts on the data. That's a Jaspersoft, where you're more analyst with the Spotify. Yeah, I think Jaspersoft is, put the smarts in the app. If you're already building an application, a SaaS-based application, or just a desktop application that's gonna be used in your organization, you still need decision-making content. And so let's allow that to be possible as well. So I gotta get your take on something. We'll have some fun here. I wrote a tweet earlier. What's the difference between a statistician and a data scientist? The answer, salary. That's a great question, yeah. So that was my snarky creepish, which Sunil Rawat just said, hey, that's highly irresponsible. Tweeting back, highly irresponsible statements. Statistics are an infinitely small portion of the tools data science is used. So it was overheard common that we overheard in theCUBE. I suppose that was kind of a joke. But this data science bubble we're in right now. There's a lot of demand. Wall Street Journal highlighted this week. Statistician, analyst, or business user, versus the hardcore data scientist. Obviously, the salary is going to weigh on the data science side. So it actually is kind of true. But the point is, the question about that Sunil brought up was, we're now comparing disciplines. Your job is to make analytics easy for people. So what is going on with the data science bubble? Well, every statistician that I know has changed their title to data scientist now. So I mean, in some part, your statement is kind of true. And actually, practically speaking, statisticians who work in a business, like I came from the semiconductor yield management business, those folks know a lot about that business. So they start to actually take on some of the characteristics of the hypothetical data scientist. Who knows about stats? Who knows about the business? And who knows about programming? But so I think they're kind of merging into one. But to respond to your other point, absolutely one of our primary goals is to make easy things faster to do. But also to support the more complex use cases. And maybe a more global statement would be to facilitate each player's role in the process of constructing and using an application that spans all of those disciplines. So as a statistician, I should be able to take that IP and broadcast it through the app instead of having office hours. If you do your job properly, then there'll be a level playing field, in my opinion, on the roles and salary, except for the corner cases of the ninjas who are like just super gurus worth the worth of money. But I want to bring that to another level. So we're in this notification economy, I'm calling it. My notifications are going off on my phone. Twitter updates, all these apps sending me notifications. You guys basically invented and created this stable platform for complex events. It's not getting any easier with today, with the mobile apps. So how do you look at this new API economy, notification economy? I mean, Tom Davenport wrote the attention economy in 2001. This is all happening right now at a whole another level. So what are you guys doing around this? Well, there's kind of two things to say. One thing is what's kind of state of the art and that we support, which is a little bit along the lines of what we were speaking about earlier. We have the technology so that the data scientists or statisticians of the world can easily construct the right models to do better targeting of whatever it is, whether it's an offer or a campaign or what have you. And then the infrastructure for deploying that into a high speed kind of event driven methodology. So using our technology, you can be much better targeted and get the message where you want it to be. And we really do believe we're at the cutting edge of that whole soup to nuts closed loop process. Having said that, you talk about this kind of signal to noise ratio. We hope that the best possible targeting and alignment of the offer and the customer will get their attention. But the fact of the matter is there's a whole bunch of other stuff happening at the same time. So the tip code bus was a term that everyone would talk about the service bus. What is that now? I mean, what is the new bus? Is it a cloud bus, mobile bus, data bus? Is it now a new architecture? How is that? That's a question. Well, the bus as a concept persists in exactly the way that Vivek envisioned it. We shouldn't have to be doing point to point routing from multiple technologies. It's a multiplicative problem instead of a linear problem. So if you can have one conduit into which all of the different sources and sinks of information and action can flow, that's a good idea. So we've been doing that on a kind of enterprise on-premises basis, if you will. But we do have a cloud bus offering, which is, as you think about organizations who are on their journey to the cloud, or maybe we'll get part of the way there only, there's gonna be naturally a requirement for integration that spans internal premises systems and then true SAS or cloud-based systems. But you see the bus relevant in the social omnidirectional. It becomes even more important because now there's even more sources. Now you have multiple clouds to connect to each other and also to the on-premises data and systems. Well, we heard from the Royal, Phyllis Royal, the international issue is huge. Now you have buses for parking data in Germany versus... That's a whole other set of constraints, right? What do the regulations say about where the data needs to be? But fundamentally, if it's not illegal, it might be a good idea to have a kind of single source of flow for that information so you're not point to point wiring. Okay, so now let's bring it back. What is TIPCO? I mean, just explain to the folks out there who may not be familiar with the company, share with your own words. What is the TIPCO? What's the company? What's the DNA? What's the core secret sauce? What does the brand stand for? What's the brand stand for? So TIPCO was invented, it stands for, if you want, the information bus company. So then that's still true today, just as the topic that we were on, we want to reduce the kind of spaghetti code that needs to be done in order to create a holistic organization in which all of its parts are functioning smoothly. So toward that end, we've added on top of that bus, we've added componentry that is useful for, so the bus you might call integration, right? The bus is the concept that you build with that integration toolkit. Then we have analytics for decision-making against the data that's flowing through or about to flow through or has finished flowing through the bus. And we have event processing, which is kind of high-grading that deluge of information. By the way, every big data source started as a real-time stream. So our mantra is to glean the information that you can from the big data, but then put that model upstream so you're taking the action before it's too late. So we've got analytics, we've got event processing, we've got a number of other technologies, social and mobile, we've got technologies for social and mobile. So really it's just a few pillars and then the idea is that we're very, very focused on putting those pieces together in a smooth way rather than just saying, hey, here's a bunch of stuff that we have to sell. So where are you looking, East Coast, West Coast? I personally am here in the Boston office, just a few blocks down, but tipcoast headquartered in Palo Alto. Great. So quickly, obviously, we're here at the HP Vertica Big Data Conference. What's your relationship with Vertica? Obviously, it's a source of data, it's a processing engine. I imagine the relationship is, from the analytics point of view, kind of sitting on top of Vertica, but in your own words, what's the role and the relationship between the two? Yeah, well, Vertica is very much in the data management business, both in terms of what they're doing literally with the Vertica database, but also in the way that they manage and interact with and utilize no SQL sources. And Spotify and tipco are not in the data management business. Instead, we're in the decision making and action taking. So as a matter of fact, it's pretty much a perfect fit and we're actually working with Vertica and HP at many, many different levels. HP, of course, is a big deliver of service and does implementations of tipco technology, but also we're working on a number of active projects right now in the area of this kind of closed loop, event-driven use case that we've been talking about. And Ken, actually, is speaking at the moment and saying a little bit about one of the projects that we're working on. So we got some commentary from Sunia, we got some virtual participation here on the CrowdChat. He says, I said every status system changes their title to data scientists. Chiropractors claim to be doctors too, sometimes helps, loves status of the seams that anyone can pass for data science these days. And he says, I've routinely seen big data projects fail even at Fortune 10 companies because people can't tell the difference between data scientists, stats, coders, quote, pet peeve. So his pet peeve is there is a little bit of title inflation going on with the... Yeah, you could say the reverse though too. You could say that a lot of data scientists are pretending to be statisticians. I mean, so there's definitely a broad set of skills that you need to be proficient at in order to be a true data scientist if you take the McKinsey definition, if you want. Brad, thanks for coming on theCUBE. Really appreciate the commentary. We went geek, we went high level. Thanks so much for your content. This is theCUBE live in Boston here at the HP Big Data Conference. We'll be right back with our next guest after this short break.