 from San Jose, California, in the heart of Silicon Valley. It's theCUBE, covering Hadoop Summit 2016, brought to you by Hortonworks. Now, here are your hosts, John Furrier and George Gilbert. Okay, welcome back everyone. We are here live in Silicon Valley in San Jose for Hadoop Summit 2016, soon to be called DataWork Summit, they announced on stage today the name of the conference will be changing. I'm John Furrier, my co-host, George Gilbert, analyst at Silicon Angle, Wikibon, theCUBE. Our next guest is entrepreneur, founder and CEO of cask.io, cast.co, not cast.com.co, Jonathan Gray. Welcome back. Thank you very much for having me back. Great to see you. Good to be here. I've been following your career, you've been a really great entrepreneur. You have worked at Facebook in the early days, you understand data, you understand Hadoop, you understand distributed systems, you understand data. Yes. What's going on in the ecosystem? Some are saying it's a mess. It's changing, obviously the naming the event, DataWorks. This is kind of like a transitional moment for Hadoop 10 years now, it's new clothing, new name. It's getting there. What's your thoughts? Because you've been involved in this at a root level from the beginning. Yeah, I think it's really interesting. I mean, I think there's no doubt this has taken a lot longer than we all expected it to. You know, I think in one token, I look at, I joined Facebook in kind of 09, 2010, and by 2011 we had multi-pedabyte user-facing applications being served off of Hadoop and H-Base. And we're talking five or six years later from that, and that's way beyond what most enterprises are still doing today. So in some respect, you still have this gap where the internet companies took the exact same technology and did amazing things really fast with it, and then it comes out to the enterprise and a whole different environment, a whole different set of constraints, security and governance matter already. Not everyone is a complete rock star developer and you have a very different reality. And I think it's taken a little bit longer for it to mature and grow up in the ecosystem, but I think what you're seeing now is architectures and patterns emerge. I think, you know, the data lake as the kind of next data warehouse, and that's kind of coming to be an architecture and not just a technology, and an approach that organizations are actually organizing around. And I think that's one of the things that is changing in big data, is you're seeing the organizations now adapt to how the new software is being used. So now they get comfortable with data lake, they can put their toe in the water, so to speak, and build the data lakes, now they got to put some value put to work. Exactly. And is that a? I think that's the next big gap we have here and that's what's really going to help the ecosystem to explode is the business value, the applications and the use cases to exist on top, extending all this great tech into the hands of people who understand the business more. You know, we always say, Dave Vellante, I always talk on theCUBE about, you know, Hadoop really is about industry standard hardware, and you know, on-premise, but now cloud's right around the corner. So it has always been kind of a horizontally scalable concept, distributed networking. But now you've got the vertical apps, and we had customers on from US banks saying, look, I got analytics driving business value, that's an app-facing thing. You've got the uberization of the world going on, I just used that word as a reference point, but that's what app developers are looking at. So, you know, where is this kind of tying together? I mean, you have DockerCon kind of showing that the containers are bringing some goodness there, but that's an ops issue, but yet customers that want analytics, is there a horizontally vertical connection today? Yeah, I mean, I think what you're seeing is, and Docker is an example of it, Hadoop is an example of it, there's all kinds of examples of this shift from scale up to scale out, you know, and I think that's also about what cloud is and all this kind of stuff. And now we have to rebuild all of the layers again. You know, the app servers and data integration products and the operations tools and the governance solutions that worked for relational databases and data warehouses and three tier architectures don't really work in this new world. And I think you're seeing now in the whole development operation side, you're seeing Docker just completely grab that entire community because they're saying, look, you had this really easy way of providing consistency and packaging and deployment and life cycle management when you did traditional apps. Now you have these multi-tier apps with all this open source and all that's gone away. So ops is a disaster in a lot of cases and Docker all of a sudden this is breath of fresh air. Here's something that is consistent, repeatable, has allows a developer to prescribe something and the operators to much easier transition to that. And I think what you're seeing now is a similar type of thing happening in big data, which is essentially this crossing of a different kind of gap, which is the gap between the infrastructure teams, these horizontal Hadoop systems data development engineer type skill sets with the line of business, the product teams, the even data scientist, business analysts who understand the domain, maybe even understand some algorithms, but don't necessarily understand how to put Hive and Parquet files together with Scoop and HDFS to connect them to your BI. That's an interesting perspective. Gaps, container gap solves the gap of ops. The gap that you mentioned is a more of an app business value gap between technologists here. Is that where cast.co is solving? Because you guys talk about containers on Hadoop, focusing on your website, focusing on application insights, not infrastructure integration. Exactly. What does that mean? Is that an abstraction layer? Is it like a new explain? Yeah, we think of it as kind of abstraction and integration and more and more self-service. And I think that's really what is the big, big shift here. I think abstraction and integration is almost an obvious value proposition. I think people understand and it's recognized that Hadoop is hard, that it's low-level, it's open source, it requires a lot of low-level skills and you really have to understand the technology to get value out of it. And so the need for abstraction, the need for higher-level layers is really, really valuable to broaden access to that platform. The easier you make it, the more people you can touch. So what problem specifically do you guys solve to your customers? I mean, what's the number one thing that you guys target and saying? Time to value, essentially. I mean, I think, I said the highest level of value props that we provide are really time and talent. It's really, really hard to hire people who know how to get value out of these types of systems quickly. And it takes even talented people a very long time to do it. And a vast majority of your time gets spent on undifferentiated, low-level kind of infrastructure, not things that are adding compelling business value. They're not differentiated things up on top. And so the more we can help companies find, you know, broaden access to that platform. Are your customers a developer or what's the customer for you? Yeah, and then I think the ultimate end game there is also to enable more services, more service area so that you expand more self-service and give people who are not just developers access to the platform. And basically the way that we sell today, you know, I think what you see in the market today is essentially people who are selling down into IT, they're selling infrastructure software, the Hadoop distros and no SQL vendors and people like that. And then you have data prep companies and BI companies and people who are selling directly into the end users, these more self-service types of tools. And we're taking this kind of, I think a little bit of a hybrid approach, really targeted on this gap and saying the owner of Hadoop today is an IT team on the platform side. That's who buys Fortinworks, that's who buys Clodera. And they are responsible essentially for delivering potentially use cases, but most often delivering integrations to the business analysts tableau. Or they're responsible for getting a bunch of ETL data somewhere so someone can generate this report every day. Or these developers really need this kind of golden data set of pre ETL, pre-secured and governed and encrypted and masked, whatever it is, data. That's that IT team's job. And we really are kind of trying to enable them. And part of that is building a UI on top that they can give to their user. Because if that platform team can expose some point- So you're freeing the IT guy up. You're unshackling him from doing all this one off. We like to say self-service with guardrails. So we want to enable them to provide self-service to their users, but we want to have the IT team and the security team and the governance teams be able to actually kind of provide some guarantees. So for example, when you use our UIs to configure pipelines, we track, audit every single thing. We have complete lineage over everything that happens. You can apply security and governance on top of all that. And so it's really trying to empower more and more users to access the platform, but to do it in a way that allows the platform team to ensure governance and compliance and security and things like that. And I think that's one of the huge thing that's different from when you leave the palace of Facebook and you go into an enterprise. I think you think a lot when you're there that it has to do with kind of like talent, but it has a lot more to do with constraints. Security and data governance and things that Facebook is a lot different than it is at a bank or a regulated industry. If you're a healthcare company and you have all kinds of data on people or you're a telco company or it's very, very different how you have to treat that data and who can touch it and what the policies and security and privacy and things that privacy at Facebook, it's kind of on or off. Like once someone can see any of the data you don't have any privacy anymore. So it's almost easier to deal with. Yeah, and the app is the app. The company is the app. The company is the app. And you don't have a whole bunch of different lines of business and a whole bunch of different things on top. The gap between the products and the horizontal and vertical is very small in internet companies. But taking that, the gap being larger at your average commercial shop, you would seem to be, or CASP seems like it would be the perfect module that translates what the IT shop knows and can enable then business value. So I would think all the Hadoop distro vendors would want to take something like CASP or CASP itself, make it an Apache project, pay you off, send you to Bahamas for the rest of your life and then essentially ship a platform where it is zero time to value. Yes. And I guess my question is, what's constraining them from doing that? From building it themselves or buying us? Buying, yeah. We have big dreams at CASP. Okay. So that would be one thing. Okay, got it. You know, we see ourselves as an independent company, partnering very, very close with all of the infrastructure providers. Part of our value proposition is our independence. So one of our big value props, not so much to individual enterprises, but some of the big users, if you go look at Otelco, who's adopted Hadoop for five, six years, they probably have all three major Hadoop distributions and cloud. How are they supposed to do the same security and governance across those systems? It's really challenging. We provide portability against it. We recently announced, Ericsson made a strategic investment in the company and you know, part of the reason. Your company. In our company and CASP. And part of the reasons behind, you know, why our value aligns with a company like Ericsson is, Ericsson is primarily a, you know, a services company, a global SI to the telecom industry and other related industries. And they do a lot of big data stuff. But it turns out that, you know, Ericsson has been using MapR a lot. Other telcos use Hortonworks a lot. Other ones are using Cladera. Other ones have gone into the cloud. The environment is very mixed. So if I'm a product and services company selling into a vertical and I'm building Hadoop based solutions, I got a lot of work to do in this ecosystem, which is essentially a broadening ecosystem. Maybe not forking, but a broadening. No, there's a different, you have standards. You don't have standards and you have an intentional differentiation occurring. Right. So before it was kind of everyone with the same and that was a value. But now as everyone needs to differentiate in the market, they purposefully need to create different solutions for governance and security and other things. The customer problems. Now the customers have an issue. And I think that's going to get doubly complex as this big move to the cloud happens. Yeah, you mentioned a lot of Hadoop stuff. How does the cloud impact your business at Cast? Because Hadoop on the cloud is not really working in what we hear in the hallways here. S3 seems to be work well. If you want to do batch, I even have Hadoop. So that's kind of like stalled right now. Yeah. It's a very interesting transition. I mean, it's ultimately, you know. Possibly or negatively. It's playing into our hands in a good way. I think we have tech to build to really benefit from all of the changes. However, architecturally, we are an agnostic kind of platform, right? So we treat something that is S3 or something that's HGFS or something that is EMC Icelon or something that's, you know, Microsoft Azure Data Lake Service. For us, we treat it exactly the same. So as we do our metadata governance, security, all these types of capabilities, our platform is agnostic to it. And so I think we have a great story when it comes into the hybrid world. And I think customers are going to be really, really challenged as they kind of go from one collection of loosely coupled services to another collection of loosely coupled services and try and figure out how to manage both of those. Jonathan, when should a customer call you guys up? Give us some insight into the kind of environmental kind of pain points where they bang in their head against the wall when the ideal scenario should, when they should call you guys up. What's the configuration of their problem or environment? Sure, I'd say there's really, there's been one historical way and now we've gotten a lot earlier. So we are now engaging with people at the beginning of projects. And we actually have customers who don't yet have Hadoop distributions purchased. They're buying us first because essentially we help shortcut a fast path to success, you know? So when we engage with customers who are early, we say, hey, what's the project? How do we deliver success in two weeks? So that's architecturally in the front end. How do we do two weeks to get you into production? Do that as the fast path to success. And then they'll go and say, okay, what Hadoop distribution are we going to bring in here? Do we want to do it on-prem? Do we want to do it on cloud? But let's start with a win and then use that to back up our big data initiatives. So front end, in the front end process. Quick wins to success. And I think as people come to the Hadoop market today, customers, they don't want to do the same way that the banks and telcos and healthcare media companies did when they went in early, which was 18 months of a lot of pain. Yeah, and so is the other scenario that they have sprawl of Hadoop? Yeah, the other is app three, we like to say, or cluster three, right? Which is you did app one and it took you a lot longer than you thought it would, but you're like, we got through it. And you go to do app two and then by the end of app two, you realize it was exactly the same process. There was no repeatability between app one and app two. I had to rebuild basically all the same thing because the nature of the ecosystem is you kind of integrate the projects while you're building your application logic. And so you get these- There's no leverage at all. There's no leverage, there's no abstractions, there's no layers of reusability. In traditional apps, you would not write SQL queries in line with your code. You would use an ORM, you'd use Hibernate, you'd use C systems that were built. But now people are putting HBase code directly in their applications. Well, what kind of business or app developer even is supposed to understand the HBase API? So we still- You're getting really efficient too, that's efficiency. So it's all about repeatability efficiency. I think one of the common patterns is companies are able to hire one or two people who really understand the ecosystem, really understand the technologies. And there's a much larger set of people around them who are not as familiar. So as much as you can also empower the few people who really understand it to say, hey, I want to really understand how to build an anomaly detection model on network data, or I really want to understand how to model CDR data inside of HDFS and HBase. Let those people, those individuals who actually really get it, build out some libraries or build out a REST API or build out a connector and then everyone else drag and drop it or let everyone else use it as an API. It's just good development practice. Exactly, it's not a big surprise, right? We already did this and in a lot of ways, I think. And that's why I think my background is why this company is why I started it is I started out building apps on HBase, on Hadoop as just trying to build products and services and get them to market and essentially the patterns emerge. Well, I think you're in a sweet spot. I think this is this abstraction layer on application development and having a data platform. It brings that best practice to use leverageability efficiency. That's what developers want. They don't want to have to recode everything and hard code other disparate parts of their code base. Thanks for coming on. Just take a quick second to just talk about what you guys are doing with CAS, hiring, number of people, what you guys are doing as a company. Yeah, so we did our Series B at the end of last year. We've got great strategic investors. We've got great venture investors. We're growing rapidly. We're over 50 people now. Really, really good customers. Lots and lots of traction and attention. And essentially, I think the fun part for us now is that we've built great products and user interfaces and stuff around really great tech. And I think that's been something that's really helped to blow the door open for the company. Is we've really built great technology that solves hard problems. But if you can't service it in a way that's really understandable and if you can't show time to value in a couple of minutes for a customer, it's a lot harder to sell. And what would you say it's an interface with guardrails? What do you guys call it? It's self-service with guardrails. I love that. I think the ability to show somebody a demo and in 10 minutes walk them through a use case, start to finish. I love that. That would take them a week or two to do themselves. And you can just show the stark contrast of what the experience is like. And then you say, by the way, I just audited and had lineage against this and this will push into your existing MDM for free. It's all open source, by the way. It becomes a totally different way of engaging. And so I think time to value and trying to get up the stack and delivering more apps and things like that is what the ecosystem needs and then is what really helps companies like ours get to market. Well, congratulations. DevOps makes IT ops more efficient with CAS. Thanks for John for coming on. Jonathan Gray, founder and CEO of CAS.co. Check it out. We'll be right back with more live coverage of Silicon Valley's theCUBE. I'm John Furrier with George Gilbert. We'll be back with more live coverage from Hadoop Summit 2016. Soon to be called DataWorks Summit. We'll be right back.