 from San Jose in the heart of Silicon Valley. It's theCUBE, covering Big Data SV 2016. And welcome back to Big Data Silicon Valley at theCUBE. My name is Peter Burris and we are here at the Strata Hadoop conference across the street having a holding a wonderful event for theCUBE community and some of our VIPs to talk about some of the challenges associated with Big Data. Now you may not be able to see them. We've got about 120 people here to give an indication of how excited everybody is. Let me hear a yeah. There you go, see I'm not lying. All right, so this is, we've talked a little bit about the relationship between digital business and Big Data. We had a great influencer panel, some of the analysts from Wikibon, Forrester and all of them talking about a number of different issues. This is really where the rubber meets the road. We're gonna spend some time with some doers. Some folks who are actually in the front lines turning these technologies into value and solving hard problems in business today. So what I'd like to do is starting all the way over at the end, I'd like to introduce Matt Olson. Matt, who are you, what do you do? Matt Olson, principal architect at CenturyLink and my mission in life is to build well really intelligent self-tuning services for network delivery, for software-defined network function virtualization. And if there's one thing that you really wanna make sure that we talk about, what would it be? Oh, how we get from where we are now to that ultimate endpoint. Because we're not gonna jump right to the singularity but I really wanna focus on kind of an evolution and a progression to get to, well it's never an endpoint to get from here to there. Excellent, Rakesh Kant from US Bank. Thank you, it's my pleasure to be here. My name is Rakesh Kant. People also call me RK. I wanna make it easy for everybody. I am a Vice President Enterprise Data and Analytics Infrastructure Technology at US Bank. One of the things that I would like to kind of talk more about is we have a lot of workflows coming to a big data environment. I think I'm starting to see some similarities with the mainframe days as the jobs run during the night, as analysts are looking at it during the day. How can we do better workload management and dedication of the resources to the right workload at right times? John Furrier, Co-CEO of SiliconANGLE and an expert in how big data and engagement come together. Yeah, I mean I'm super excited to be up here. Thanks for inviting me up to host theCUBE. But for me, I wanna share and talk with everyone here about the role of data because we, myself and Dave Vellante, are building a media business from scratch with no outside funding, which is challenging in today's environment and we're succeeding. And one of the reasons why is we've been really focused on the role of data and open community and social media and building a data platform behind that, using signals and predictive analytics and machine learning. And we're constantly thinking about how to use a platform to create an enabler to increase the quality of content at higher speeds and put them into targeted audiences in a way that's not gonna be offensive. And so we think about this all the time and we look at engagement and we're instrumenting, trying to instrument the conversations and we have a conversation platform that we've built. Microsoft actually announced the intention for one today but we actually built one and we actually look at what people are trying to talk about and trying to create content. To create engagement for no other purpose other than to create some engagement and some interaction not to sell anything but to really create more of an intimate relationship with each other. So each of you guys have deep technology chops and a job that is normally associated with technology. How much time do you spend worrying about the technology versus how much time do you spend worrying about the business problem? Matt, why don't you start? Oh, I think I spend at least half of my time dealing with really the business side of the equation because at the end of the day, really it's about being successful as a business and also affecting transformation of organizations. And I think that if we focus on the technology and isolation, it invariably runs aground. So I'd say at least half of the time spent on the big picture and the big picture means understanding business and understanding customers certainly. Now Rakesh, US Bank has done some relatively advanced things in terms of that data supply chain, that information supply chain to a place where then you can feed it out to the various functions. How do you, how is your relationship to those functions evolving as a technology and understanding of how to use the technology evolves? Yeah, so I'm gonna concur with we have to have business on board with whatever we're trying to do. If you don't deliver business value at the end of the day, it is just a science project in a room somewhere. When we started this journey in terms of strategy, we wanted to create obviously a single provisioning point for all enterprise data in one place. Our customers have to go to different places in order to collect that information. We wanted to create a place where they can come confidently and get that data. Second, we wanted to, one of our analysts were talking earlier about reducing time to data. It was more from a real-time perspective, but I used the phrase time to data. When a business has an idea, that idea requires data, they should be able to get to it easily. And we want to create that place where that data is available for them so you can select what you want in order to drive insights for your business. So time to data was another factor. I think the third was because it's all going through one transition point, we wanted to increase the data quality. That's a perennial problem. Data is being supplied today in multiple supply chains with various varied kind of data quality. We wanted to have this one place drive the data quality and I would say conformity, right taxonomy, right harmonization of values. When somebody says what product, you mean exactly what product. And the last I would say is security and governance. Because it's going through one place and it's all in one place, we wanted to kind of have right security on top of it. So those were all the pieces of the puzzle that we wanted to achieve before we take the next steps towards kind of making it available for everybody in multiple access methods, ETL, SQL, analytics, just file access and so on and so forth. John, you have an extensive background in search and have been utilizing open source tools to do some really interesting things on a shoestring budget. You've been living by the proposition that grab that low hanging fruit. How's that going? I mean, it's interesting. When you have no real budget, you have to be resourceful. We were, we did the first solar search on HBase before Cloud Era did. It was working great. We had great HBase implementation and it was just so freaking expensive to run. I mean, for us, we had to load a server. They were $8,000 a server. I had to go host it. I was the CIS admin. I couldn't hire anybody because they wanted $150,000 a year salary to reboot it. And so it was really harsh. We moved everything to the cloud. Okay. I'm like, there's nothing to do with technology. Like here's this admin. Well, I can go to Uber and be a CIS admin to make $150,000. I'm like, go to Uber. We're gonna move to Amazon. So we moved it all into the cloud and said the cloud provides great on ramp for that initial bootstrap. And it's not about technology. Things we worry about is the partners. So we rely on external data sources like the Twitter, FireOs and other things. And that's a passive monitoring system. That has nothing to do with technology. I worry about if Twitter goes out of business or if a partner data goes out of business. We are moving quickly to control the active data that we can control. And that's what we're instrumenting with our crowd chat app, among other things in the cube video. So for us, really focusing on the process of the consumption of the user that is non-linear. And that is a great opportunity for the unstructured data but has nothing to do with technology because the software is so awesome now. You can really roll up anything. So if you think about the, we talked a little bit about data governance and making data a first citizen. That hasn't happened to most organizations. I was with a CIO not too long ago who was absolutely lamenting the fact that the assets that she was creating for the business were completely undervalued and therefore it was difficult to sustain the investment necessary because it just didn't show up the way that it should have shown up. Matt, as you think about, you're in a relatively mature business that has enormous operational expertise that is dependent on data. How is CenturyLink starting to evolve its thinking about data as an asset or what we're calling data as digital capital? Well, I think first and foremost, the key has been breaking down the silos because I think the various silos within the company have their own particular data sets. I think there is sort of in the culture, there's an appreciation for the value of that data, but the historical sort of legacy approach is I think to focus on specific limited data sets which were believed to be related to a specific problem set which the organization owns. And I think the first and foremost, the most important sort of transformation has been breaking down the barriers across the various data silos and beginning to engage users with enriched data with contextual information to join across the various silos to begin to derive really more insights and far more information from the data. It's far more than an additive process. And so I think the most important step from my perspective has been establishing not only an infrastructure but a culture which allows for the sharing and the correlation and integration of data across the various silos. And those silos are not just sort of end to end in the service delivery path, but they're also up and down the protocol stack and beginning to connect what people are seeing from a network and an infrastructure perspective with the experience that our customers are experiencing in the service delivery path, which is ultimately what's of the greatest value to the company. Well, same question, US bankers. Is data anywhere near moving into at least second citizen role? Yes, definitely. I mean, it's a journey. It's not going to be a point in time activity. The maturity comes from people getting experienced with this new technology, understanding the value as it happens over time. Obviously the sequencing of activities need to be a little bit more intelligent in terms of we initially, because it's a heavy weight item to lift from ground up. So you need to have that burning use case that is going to lift it up. And then once it's lifted up, then you can take that whatever is up and ready for other use cases. And now the overall cost for everything is kind of coming down slowly over time. So it's a collective journey type of kind of journey. I would say, I think we have talked about analytics and trying to kind of how to kind of make insights out of it. I think the basic lesson for me has been the data is still important. The quality of that data is still important. The same perennial problems that were there before, they suddenly don't disappear just because we have a new technology. Absolutely. It enables certain aspects of it that you can scale, you can process it volume and so on and so forth. But the basic requirements still remains the same. Absolutely. All right, so audience, questions for our panel. We have one. Joey Bedash from Sama Technologies. Earlier I asked a question from Tony, but I'm more curious now asking the customers. There is a lot of buzz about systems of insights. And we were collaborating. I said, if it's not actionable or it doesn't have an impact, it really means nothing. So what's your thoughts revolving around systems of insight or actionable insights? Well, my take is I think there's definitely a problem with getting action out of the insights. The insights is easy, I think. And certainly I think it's early days. Here's the problem with action. What does that mean? So the context of action is, one, defining what you want to target insight to. And then having a destination endpoint actually be compatible with receiving that insight. So we have a huge problem in our business in the media business called content management systems. They were built for polling databases, not push notifications. So every single media company on the planet pretty much has a flawed infrastructure on receiving an endpoint, which is essentially a JSON feed, for instance. Real-time updates require precision and targeting. So the endpoint actually isn't ready for the insight. So there's no action that can be taken because the apps are not mature. So I think there is an insight there. So that's my observation. Even if you could produce the actionable insight, getting it right is really, really hard. And then two, having an endpoint that can receive it. So that's a huge problem right now that I think will be overcome pretty quickly. And actionable insights, we talk as if there's only one insight at a point in time. But I think when you look at a customer, there might be multiple insights that are applicable to that customer. The question is, which one we should move with? You're eligible for a loan and you're eligible for a card. Which one is better, even though both are actionable insights? So sometimes it needs to be which one is the right? And sometimes it is, even though those are valid things, you don't need to move on it because we have called them multiple times or we don't need to kind of upset the cut. So actionable insights and the relative importance of those is also important in action. And I could speak to this, I think in probably a very different context, which is the network itself. And in my world, I think the key is sort of joining the data that's generated from the underlying network with the data that's generated by the customer's interaction with that network and the services it's delivering. And at the end of the day, it's the customer interaction, the quality of the experience, which tells us where we have opportunity or where we have challenges, what really matters to the business. And it's the data flowing from the underlying network and infrastructure that tells us what actions need to be taken. And from my perspective, the advent of software defined networking and network function virtualization is really exciting in this regard because it provides an infrastructure, which is by definition ready to accept the actions, the insights via API calls and such. So I think really for the first time in my world, we're seeing the opportunity to build really intelligent feedback loops and build intelligence into the service delivery, which was almost unimaginable. Yeah, I mean, the only thing that's actual right now, just to add in my comment is that you see that this is a clear indication of the maturity of the insights being generated in the consumer data. You see smiley face or sad face, okay? That means there's not there. And then, you know, what animal are you? You see that on Facebook all the time. So, or what does your Twitter feed say about you? So that is so early. They can't actually generate the insights. So there's a lot of unknown spots. So that's an indication that the predictive analytics only predict extremes. Are you happy or sad? And that's just baggage from natural language processing technology from the 80s essentially. Yeah, I was, I wrote a paper in the early 1990s on action support systems that I found out that that name wouldn't play well. The acronym wouldn't play well in Long Island. Another question? Who are you? Thank you, George Simons, I.C. Ventures. Not that John can answer this, but I'm curious with the other two. I've always believed you need corporate buy-in to make big data successful within your organizations. I mean, is the CEO understanding, involved, committed? What level of corporate buy-in do you have? I'll take the first. I think, yes, we do have corporate buy-in. The chain of management has been behind this effort for some time. We have been working on it for two and a half years and it has been critical and crucial to support our financial crime and compliance initiatives. So it has been very, very visible, the burning use case, and visibility across the organization. And I think from my perspective, we unquestionably have buy-in at the highest levels. The issue, of course, is defining exactly buy-in to what, in practical terms. And it's very much, in my experience, a situation in which there's iterative development and demonstration of value and an evolving strategy. But I think at a high level, in the telecom industry as a whole, it's very clear that we're in the early stages of a massive transformation. And there's full buy-in to embarking on that transformation. And there's, I think, a full buy-in in understanding the fact that that transformation is inherently dependent on intelligent use of the data. So in big picture, conceptual terms, I think we have solid support. The key is now to iterate rapidly and demonstrate value and build momentum. Another question from the audience? Ms. Bill Warrior. Hi, I'm Sandy Holder. I'm CEO of Insight Business Advisors. I actually have helped fund quite a few companies that are growing and developing in the business analytics environment. And the biggest question I think that we all have is that at what point in time do you have a CEO or CFO that today could capture information on his mouse, on his dashboard, slide it across, slice and dice the business, and be able to capture information and then, more importantly, be able to have this, as they said before, advanced intelligence to be able to kind of predict or have some measure of probability about the kinds of things that they could or should be done. Either there's gaps in the business, there's challenges with whatever part of the business there is, but in business visibility, because I think that's at an affordable cost. So the question really is time and cost, how long do you think it's going to be before that happens? So I think it's a journey. If there is a use case, it can be done today. The technologies are available. The question is, what is that kind of questions that we want to answer? A predictive and I would say more interactive kind of analysis requires a little bit of thought. I mean, we've seen this, you need to be kind of aware of what the data is, what it means, and so on. So it can be done today, I would say realistically, when self-driving models and those kinds of things that can start kind of creating those kind of intelligence, we are, I would say, three to five years away. Yeah, I would just say I concur completely, and I think the key is not so much the general availability of technology which supports this in theory. It's the continued development and refinement of the models and integration, both of the predictive capabilities and the underlying algorithms with an environment that allows for user interaction with that information, with the data. And I think we're starting to sort of take baby steps at this very early stage. I think we'll just see continued development over the coming years. Well, we run all of our business on Google spreadsheets and stuff, so we don't really have a system for that. But all the CUBE interviews I've done, I've heard a lot of great insights there, which is I just think it's so far-fetched right now because I think it's not a technology problem, it's most people are trying to figure out what their business is. And for the first time, I hear in the CUBE all the time that this is the first time in the history of the world that everything can be measured for the first time. So that's mind-blowing in and of itself. So that is kind of causing people to say, oh my God, look at my baby, it's not as pretty as I thought it was, or I'm in the wrong business, so you have all these kind of moments happening right now and it's either gonna scare someone to death or transform them. So I see it as the problem of what to look at, what data do I need to be slicing and dicing, and that to me seems to be the macro problem. And I'd say August 19th, 2019. So I'll have one more parting thought before we close and open it up for moving out into the hallway and talking to each other. And I think one of the most interesting tests of a lot of the things that we're talking about is there are a lot of companies out there talking about how they can apply big data to the challenges of marketing and revenue. And I have a question. If that's true, why aren't those companies growing faster than any other company in history? And I think that for what John just said, one of the most interesting tests here is that this marketplace should be using itself and its own technology to evolve and develop. And I think as a test of whether or not we're really that close yet, it hasn't happened yet. There's a lot of marketing analytic technology companies out there that aren't blowing the doors off. There's a lot of big data companies out there that talk about what they can do for the customers but aren't blowing the doors off. So it's gonna require a lot of work. It's gonna require a lot of commitment and it's gonna require an organizational effort without question. All right, so with that, let me thank everybody here for joining us in our sessions early. I think we're closing out Q Broadcasting tonight. Is that correct? Just to lie, we have a giveaway raffle. Okay, we have a giveaway. When are we gonna do that, John? After we break here and... Okay, we'll break here. We'll go to the audience. I wanna thank everybody. Thank you very much. We can do it now. Greg can put you out. Oh, we're gonna do it now? Okay. Yeah, pull it now. Live on the air. Instant... Woo! You must be pressing to win. Oh, we are giving away a smoky titanium. Oh, I mean, galaxy Tab A. So we're giving away a galaxy Tab A and am I pulling it? No, how about the customers? Yeah, pull it. All right. You all right here? All right. And the winner is... Uh-oh. Uh-oh. TK Akbar. Chief growth officer of Zooming. Hey. All right, come on up. You're gonna be on the cube. All right. We're still live. Okay. Excellent. Ah, I congratulate you. What does your company do? Awesome, fantastic. Awesome, congratulations. Thank you. Thank you. All right. All right, so with that, I guess we're right here for John Furrier, Silicon Angle and Rakesh Kant of US Bank and Matt Olson of CenturyLink. Guys, fantastic. Great insight, great analysis. Very helpful. And Matt, you've now got an hour and a half to talk to everybody about how we're gonna get from point A to point B. Thank you very much from the cube at Strata Hadoop. All right, thank you. Thank you. Thank you. Great job.