 from New York, extracting the signal from the noise. It's theCUBE, covering Spark Summit East, brought to you by Spark Summit. Now your hosts, Dave Vellante and George Kuehlberg. Everybody, this is theCUBE. theCUBE goes out to the events. We extract the signal from the noise. Ashish Tussou is here. He's the CEO of Kuehl. Kuehl's been on theCUBE before. Welcome to theCUBE. Thank you. Thank you for having me here. Actually, personally? Actually, this is the first time. Oh, great. So you are Kuehl, but Kuehl has been on. I know, at one of the O'Reilly shows, we had you guys on. So it's when I first met you all. It was a while ago, pre-Spark. So what do you think, Spark Summit? Kind of a geeky crowd, sort of fun. It's a fun crowd. It's a different crowd. So I think over the years, Hadoop obviously started with the developer revolution and then started becoming more business-like. And here we see a lot of grassroots people, machine learning crowd, data science crowd, developer crowd. So it's fun to be here and be a part of this movement. Yeah, we love it too. theCUBE loves the early days. We were there at the Hadoop World early days and then the suits came in. I should talk and so they took over. But I'm always in a suit in these things. But so, Cubol, tell us kind of what it is, how it works, why it matters. So Cubol essentially provides big data as a service to a cloud-based platform. We run a SaaS platform on the three primary clouds, whether it's AWS or Azure or Google Compute. And the big value proposition that we bring to our clients is that once they get to Cubol, they can find all the technologies that are needed for a modern-day data team, whether it's Spark or Hadoop or Hive or SQL, for different workloads. Whether it's machine learning and streaming workloads at Spark or analyst workloads, SQL analyst workloads at Presto or Hive, they can do all of that in Cubol in a turnkey manner through a SaaS platform. And what the Cubol platform does is, provides full automation of the infrastructure at the backend and provides a unified access panel for access to all these technologies, thereby making it very easy for a data team to enable the adoption of these technologies inside their companies. So the idea is they don't have to worry about all that heavy lifting. You guys provide that as a service, and then they can spend their time doing more value-added stuff. That is correct. We essentially make big data self-service within the enterprises, and that has huge implications. That has implications around opening up a lot more use cases around big data within the enterprise. And that has implications in terms of also the support that the data team can provide to their enterprise users. And before Cubol, we were at Facebook, where we had built a very similar technology internally, and we saw it very close hand as to what implications that model has, as opposed to a model where there is data infrastructure and there's a data team sitting between the infrastructure and the users. So through a self-service platform like Cubol, essentially you take away that bottleneck, you make big data self-service, whether it's the analyst in the company, whether it's a data engineer in the company, or a data scientist, they can go to the same platform and essentially run their workloads, different types of workloads, and all of that is provided through a SaaS platform based on the cloud. So your mission is your why, if you will, similar to Databricks, but in a different context. That's correct. Simplifying big data. That's correct. I remember one of the early Hadoop worlds we were talking to Jeff Hammabacher, and he said my mission when I was at Facebook was to obliterate the storage container, lower the cost of storage. That was kind of the early days. That's right. And then it became, okay, George talks all the time about the data lake and how it became a lower cost alternative, but it still seems a really long way to go. So as you simplify big data, what are the business problems that you're now seeing people attack beyond reduction of investment, attacking that storage container and dealing with massive amounts of information, eliminating sampling? What's the real business problem that they're trying to solve? So I think you're right. Time has changed. Initially it was, hey, store all your data. Now I think a lot of people are asking about, I have all this data, what do I do with this data? What kind of business problems are you going to solve? We see a lot of those. Some of the verticals that we see in Cuba is around marketing and ad tech. So the real world applications that are analyzing clickstream data, analyzing impression data to do user modeling to figure out what kind of campaigns work well, what kind of campaigns don't work well, what kind of users have a better propensity for certain campaigns and so on and so forth. We see use cases in product analytics where companies like Pinterest and of the world are essentially analyzing a lot of what their users do with their product and how they can improve the product. We have started seeing applications in the IoT space, essentially the companies who are gathering data from set-top boxes and doing analysis on that. Again, the common theme here is either it's user-generated data or machine-generated data. The common theme here is that they are using that data for optimization of existing business processes as well as using that data to create recommendations and newer products out of that. So I wonder if I can ask you Ashish, we've been working on this premise which is that over the last couple of decades, information has become so widely available to consumers. And the access to that information has leveled the playing fields with the brands. The brands used to have proprietary information that nobody else knew, they had pricing power and that's gone now. The consumer has all this pricing power. It seems like one of the big opportunities with big data analytics is to kind of regain that asymmetry in pricing information or information about the consumer. Get a competitive advantage. A lot of times brands know more about me than I know but you know when I run out of paper towel is my latest example. They can predict. They can predict it. I don't like that, I don't log this stuff. I just know it's gone, okay great. And so are you seeing that as driving competitiveness? We're certainly seeing it with Google, Facebook, maybe Uber, Netflix, some of the big leaders but is that hitting mainstream? Is that a valid premise on our part, that asymmetry sort of trying to get that back? I think it's become a very competitive world. So essentially what most companies are trying to do is use their data assets to gain that competitive edge as you mentioned. So that is certainly happening. Has it happened at a very wide scale? No, I think it's, we are still at the very beginnings, beginnings of that revolution. But people are thinking in those terms. People are, in certain cases we have seen deployments similar to that, not exactly that. We're trying to target some of those things. What happens with big data technologies is that now for the first time, well not for the first time for the last few years you have access to technology that can help in enabling those use cases. I think people have been talking about, brands have been talking about these use cases for ages but now you have access to technology at a price point where you can, those use cases can be enabled. In addition to that, with the advent of the cloud, not just the access of technology is there but also the access of the operational ability to operationalize the technology is there. Also there's now with the cloud you also have access to a self-service platform like that. So I think when you combine some of these trends together you are essentially reaching an inflection point where it becomes much easier for applications of those sorts to become more and more widespread and I think that's what we'll see going forward in the future. A question related to that, this inflection point. We had Ken Sai of SAP whose VP of cloud platform and data management, he's responsible for the Vora product that runs in conjunction with HANA, anyway it's for accessing the non HANA operational data and he used this example of predictive maintenance which we've all heard as this great big data application but he said he took it a step further which was to tie it into the whole MRO process that's already in SAP so that you know how to coordinate all the resources related to that business process. Are you seeing any of your customers tying these these functions into something that's part of a larger process or are their own sort of applications sort of not repeatable enough? Like is everything still custom built? So I think in certain areas you have started to see the emergence of repeatable templates. But in broader areas I think things are still very custom built. And so where are the repeatable ones and where is custom? So again I think the segments that have been using big data for the longer, you have started to see repeatable templates and I would say marketing is one big segment there. And in fact there's a plethora of marketing applications that have come up. All of them at the back end are using big data for various different reasons. For better scoring of leads that go off to sales. For prediction of whether a certain lead will be a better sales prospect and stuff like that. So there's a plethora of these applications that have come up and all of them fundamentally at the back end are using big data technologies to power that. Now as you go towards newer cases around big data that you've started to see in more emergent IOTs is a big use case. There I think it's still those, some of the architectures have started to emerge. Certain template architectures started to emerge. But I think template applications are still emergent. There you would start to see a lot more custom built things. Similarly in some of the other business processes that have not typical in a standard enterprise that have not typically used big data. I think that there's also an emergent pool there but template applications as such haven't emerged. But I think in the next five to 10 years essentially that would be a strong area. And that would be one of the reasons, one of the ways in which big data power gets more and more democratized to end applications in themselves. You know it has to happen, right? I mean when we first started covering the big data market, we're the first to quantify it. And we noticed that the predominant revenue source was from services. So I was like, oh, because it's so hard. Okay, and it sort of dawned on us, well that's not sustainable. So either humans have to get smarter, well it can only get so smart that data is growing faster than we can get smart. Or we have to train a gazillion data scientists, which everybody's trying to do. We heard Databricks today, started a MOOC last year, trained I think 20,000 was the number. Okay, but it's still a brute force approach. Software seems to be the only way. Yeah, to get the scale. We're going to be able to get that scale. So we see it coming, maybe not as fast as Mike Olson predicted five years ago, but you're opportunistic. Optimistic I should say. And opportunistic I hope. But optimistic about applications, actually packaged applications versus custom pieces. Look, I think we have seen that happen already in certain fields. Like if you look at- Some examples. In marketing you have a lot of new age companies that have come up that are putting big data to use. There's certain innovative companies who are actually crawling the web to look at job sites and stuff to predict that certain accounts might be more favorable for a certain solution. Are you talking about spider- Spider book is one. Yeah. So there are a bunch of these things which have come up. So it's already happening. Already a customer? No, we use them, they're not a customer. But these are, in fact, actually they're a customer as well. They also use us. But these are emergent trends, which you're already seeing that happen in verticals that have been using big data for the longest. For a long time, big data was very used for data collected from consumers. So there you're starting to see emergent. So I'm very optimistic about it and I think it's happening. And they're kind of stealthy apps in a way. I'm stealthy in the sense that Peter Burst talks about how he predicted years ago that there would be more SaaS platforms developed from non-tech companies than tech vendors. And that's kind of what's happening here. And they're all, they're all database digitization, digits are data. Yeah. And so you're starting to see these apps emerging maybe in ways we didn't expect. Yeah. And you know this has happened in the past also, right? In the past, the 90s was all about creating templates around business processes. Transactional, transactional systems came in place. And on top of that, why did databases take off? Because a lot of business process applications were built on top of that. And that was all the focus of the 90s. Then now we are seeing the systems which do interaction data very well, which is these big data systems. Those systems have reached a level where now it makes a lot more sense of these analytical applications to emerge, which are essentially, these are different from the transactional applications of the 90s. But I think the trend is the same. So I don't think there is, it's an emergent trend. It's going to happen. And that I think would be a very powerful driver of the adoption of big data technologies in your day-to-day needs, not just some data science person or some advanced user doing things. So we only have about a minute, so we'll give you the final word on Q-Ball, the company, what we should be looking for, what's going on in the valley, you guys are in the valley, we're not, but people want to know, say we hear B rounds are tough, we're hearing about flat rounds, but so give us the update on sort of what's happening company-wise. So company-wise, Q-Ball is essentially charging to our vision of using cloud, the power of the cloud, true clouds, dynamic infrastructure and so on and so forth to simplify big data, both from access perspective as well as operations perspective. Company-wise, we just talking about rounds, we just closed as you see in December, so now we're putting that money into use in order to make that vision more and more successful. We are also heavily focused on making sure that our platform is a complete platform and not just a single technology platform. So we are, you will see in future more engines being part of Q-Ball, so they can be a single data platform which a data team can rely on for multiple different use cases. And then of course, as cloud becomes more and more mature, as cloud adoption becomes increases, and as cloud vendors, my prediction is that the cloud vendors will also diversify. Q-Ball will be provided as an option in all those, in all those clouds. So that's what we are running towards. Excellent. All right, well good luck with that and thank you so much for coming on theCUBE, we appreciate it. Thanks for having me, glad to meet you. All right, you're welcome. All right, keep right there, everybody. George and I will be back with our next guest right after this. This afternoon, George will be releasing his first ever Spark forecast, first look, all right? Not the final one, but keep right there, everybody. This is theCUBE, we'll be back. We're live from Spark Summit East.