 At Hadoop Summit in San Jose, keep on saying Santa Clara, San Jose. Last year's, yeah. Yes, I'm here with my co-host, Abhi Mehta. Thank you, just got my promotion to a co-host. Doing a bang up job. Thank you, I appreciate it. And we're joined now by Andre Dragomir from Adobe. Thanks for having me. Thanks for coming on, really appreciate it. Thank you. You know, one of the big things we're trying to do here on theCUBE is bring in some practitioners, some big data practitioners, to really talk about working in the trenches and how they're really making use and deriving value from big data. So, why don't you tell the audience a little bit about yourself and then maybe we can go into how Adobe is kind of working in the big data space. Sure, definitely. So, I'm a, as you say, a practitioner, a software engineer. We've been working with, me and my team, we've been working with these technologies, Hadoop, HBase, and France since 2008, actually. So we're really early practitioners with this stuff. We've been running it in production since late 2009. So again, we're pretty advanced for that time. And what can I tell you? We've had a lot of experience, me personally, working with other distributed system, building some internal stuff at Adobe, and then we looked at this, these open source projects and it made a lot of sense to start looking at them and using them. And basically, we placed a big bet on Hadoop and HBase and it's starting to really pay off. What drove it for a technology company to pick this thing? I'm sure you have incredibly smart developers, your products are incredible. What drove this move to going open source and being such an early adopter? What was the driving factor? Well, I guess it was more like a necessity as the mother of all, you know. Basically, you know, Adobe was traditionally, they had a lot of experience and some hugely smart people working with image technologies, video technologies, but you know, nobody, us included, we didn't have much idea about analytics, big data, distributed systems. So in that sense, it made a lot of sense because in the beginning, what we were on about was earlier was fast prototyping. So we did a whole bunch of prototypes for a whole bunch of clients and products for video analytics, we did machine learning. We did a project that was really interesting of analyzing flesh content on the internet so we can detect stuff like flash player crashes and what functions are people using. Proactively. Proactively, yeah. So we knew there was a use case. It's like we were growing into this technology while the company was growing into this need. And very interesting. So it was kind of a serendipity with us. And now, Adobe is moving more and more towards services, what with the launch of CS6 and the subscription model and the Creative Cloud. And we're going to be seeing much more and more user-generated content, user-generated data like somebody mentioned in a presentation, observational data. So we need to have the tools on hand to be able to derive meaning and business value from this data. Very interesting. We spoke recently with Peter Goldmacher, who's a Wall Street analyst account. And he mentioned, his thesis was that you're going to see the majority of value generated from big data from practitioners. And he specifically pointed to Adobe. So I was particularly excited to talk to you today. So why don't we dive into a little bit about, as you kind of go into that services business a little bit more, the role of big data and how you're turning that to your advantage to actually create value for your customers and actually value for Adobe. Yeah, so I guess the first thing that we're looking at is helping, matching business value basically there's a couple of things here. The first of is what do we need to do in our software offerings to make it more useful for users? One of the best way of doing it is actually observing what users are doing with the product, right? That's one thing. On the other hand, it's something totally unrelated like we are creating new software products, which that's their reason to be, to analyze data and create value for customers. And again, I'm referring here to the digital marketing initiatives and what we're doing with big data in those spaces. And finally, again, there's a lot of stuff that's related to internal metrics and the question is quality of service data, how well does a product work for my users? And that's a huge amount of data that traditionally and in most companies today still stays in its raw form because the tools don't exist to query it, basically process it, query it, and get meaning from it. Slightly change the topic and we're making it more around is this as much, it's where you said, necessity is a matter of invention, is it as much the fact that you're moving to a as a services model, which seems interesting and logical, you hear that a lot, Jeff, a lot of people coming here, seems like this collision between the cloud and big data is almost forcing both for the customers and on B2B side, but also for companies and as a service model. What about something a little bit more separate on the technical side? This whole HTML5 versus Adobe's core technologies and some of your larger customers, the larger users saying, well HTML5 is going to be the way. I mean, how do you guys view this shift in technology? We're talking about the analytics side, but what about shift in technology on as new technological waves come? Because HTML5, I can make the same logical argument around Hadoop, open sourcing a key part of the stack. Proprietary technologies have to look in and ask them a tough question. It's very reassuring to see you guys picking up Hadoop on the analytics side. What about on the core business? How does HTML5 change your perspectives on proprietary technologies? Well, again, that's not my area of competence, but if you're looking at what we're doing right now in the space, we're actually, we're becoming a very short time and that's related strictly to the HTML5 questions. We're actually becoming really big players in the WebKit ecosystem. Ah, okay. So we actually have a team in Romania, we have WebKit committers, we've implemented, and the way it integrates with our products is basically we have a lot of knowledge with stuff related to video, audio, image, topography. One of the things that we did, and again this is unrelated to big data, just this, we implement stuff that does CSS regions basically allowing you, with HTML5 and open web technology to have really, really advanced layouting techniques for web pages. So we're embracing that with, and the reason is why, and again, this is, I guess it's pretty much common sense, everybody benefits, right? We benefit because we're- And it's almost like you're making the distinction between the IP, the intellectual property, is not the core technology, the software. It's decoupling it. It's a very interesting point of the fact that it's your knowledge, it's just semantic knowledge of processing images and videos and documents that is probably more important than the technology that you wouldn't do it. And if you can apply that to HTML5 or any new technologies, open source or not, you can retain your advantage. Is that the concept? I think so. Interesting. I think so, yeah. So, you know, back to this conference here, I mean, what are you seeing here in terms of the vibe and in terms of the attendees here? Is it heavily on a technology and the developer side? And kind of, you know, you're an early adopter, you've been using this for a while. Are you seeing a lot of your brethren early adopters? Are you seeing some new faces and what's your take? I'm seeing a lot of new faces and I think it's like the free lunch is over, so to speak. And this is good. There's nothing like a free lunch, huh? Yeah, right, right. Up to a point, Hadoop was this technology where you had to invest a lot in creating your own team that could install it, operate it, and develop on it. Correct. A lot of these functions now are being replaced by either vendor solutions or they're being rolled up into the core product itself. And for me, this is something that's really, really awesome to see because typically, that's the way that you want to bring an open source project forward. Take care of the underlying infrastructure so you can concentrate on creating business value with it. So how do you see the larger players who have now come and embrace the HDFS ecosystem in the last 12 months? IBM, EMC, HP, VMware versus Hortonworks and Clara. I mean, what is your perspective? Have you used a lot of them? What do you think of what each of them offers? So we did look at some commercial technologies that's building around this ecosystem. We've evaluated some of them. Well, I think it depends. I can't speak for the motivations of each party, but what I want to say is that competition is a really, really fine way of driving stuff forward. I agree. So let 1,000 flowers bloom and let users have choices and let enterprises invest more in Hadoop and the existing ecosystem. So related to the vendors, that's pretty much my idea. What was really interesting coming, and this was the first time at this year's Hadoop Summit, was seeing a lot of open source projects in the Hadoop ecosystem, but which are not tightly related to the core. So we've seen presentations on, well, each base is now pretty part of the core, but we've seen presentations on ZooKeeper, on Storm, and all kind of technologies that really, really complement Hadoop. Completely agree. And moving forward, I think that's, I mean, that's not just empty talk. That's the Hadoop ecosystem. It's a big one. It's a big one, exactly. We at Jeff's does, I do, I call it the HDFS ecosystem. It's actually not about Hadoop anymore. It's not just not produced in HDFS. And the HDFS ecosystem is very vibrant. We see it with all these conversations, all the projects, Storm is, there was a standing room only crowd for Nathan Braz for Storm today morning, which was very good to see. It's interesting in real time at Hadoop. We get the question often. I agree with you. This is a very vibrant ecosystem. It almost compels you to, when Jeff, what do you say? Right, that open source is the secret sauce. Exactly. And it's, but it's open, right? It's the secret is open. The emperor has clothes or not. We know it. You know exactly what it is. Absolutely. But you know, we're also seeing, if you look at some of the vendors here, I mean, we're seeing, there's some of the more traditional data management vendors that are now trying to kind of position themselves in a big data world. That's a good point. You know, so versus some of the Hadoop focused upstarts in the open source world. And it's really interesting to see how they're playing together. And, you know, there's- Are they playing together, Jeff? Some are. Some are. You're seeing some proprietary vendors working with some of the distribution vendors, for instance. That's, you know, we're seeing some of that. You know, but we, I think we coined the term big data washing on this, on the cube a while back. You know, not everything is, not everyone who kind of slaps that term under their product under the box is really big data. So, you know, it's definitely, still an evolving ecosystem right now. Definitely, definitely. I think also from what I'm seeing, most vendors are also testing the waters. Right. And, you know, implementing some stuff. There's, again, there are a lot of vendors which are selling stuff that helps Hadoop easier to use. I think that's- I agree. That's still like a motif that comes up in a lot of what the vendors are doing. No, right, absolutely. I mean, one of the keys of being an enterprise-ready platform is to, it has to be able, you know, easy to use. So, to abstract away some of the complexity with some of these drag and drop tools and GUI, you know, graphically user interfaces and things like that, so you don't have to write the raw code. Exactly. So, yeah, so, and I think that's one way that we're seeing the more traditional proprietary type vendors. That's right. That's what kind of what they're trying to do is apply, oh, their data integration methods. Right. You know, to Hadoop and we'll bring our, kind of our user interface to that. Yeah, I agree. And I think that's important. I mean, that's one way, you know, we're going to reach that level of adoption to make the tools easy to use. What's exciting for me, Jeff, is to see, kind of like Adobe, you know, I've been a practitioner, I'm now a developer, a vendor, trying to change the world with my little initiative. But it's very exciting to see companies like Adobe, but while we're not classified, it's almost like what you described to us is the evolution of your company. Fundamentally built around big data. People may not think of it that way, but you are a big data company yourself. Well, I wouldn't know if I could go so far, but there's, I mean, there seems to be an emergent pattern in what we're doing right now. But you can use, the one I was trying to make was you actually are leveraging the same tools that a Google has leveraged for search and Facebook for social. And banks can leverage on my platform for financial data to actually look at building an Azure service model for Adobe, which is very, very exciting to see because, Jeff, in my mind, that opens up a whole new world of existing technology companies can redo their own infrastructures and their own business models around analytics and predictive work. As you mentioned, it's very exciting. It's very exciting, but it is also scary, I think, to some businesses. To change my whole business model, it's like, well, we spent years and years building, so there are, but it's going to be those companies that jump in early and really are committed to it as Adobe is, that you're going to see flourishing. Again, back to our friend Peter Goldmacher at Cowan, he said back in the late 80s or the 90s, if you're trying to pick which vendor was going to be the ERP vendor, you never would have bet on SAP. But if you had bet on the companies that were using ERP, those companies, their value exploded. So the practitioners, that's really where a lot of the value is. And it's hard, we are here on theCUBE, we love to try to pick who's going to win the distribution game of the war or whatever you want to call it. And that's important, but really, it's the practitioners that are developing either new lines of business or completely new business models or startups building off this platform where the value is. And another thing is that we try to look at everything that's happening, but again, we're trying to be pragmatic. We're building some of this stuff on our own right now. If there's a solution as a service for somebody else and if it makes sense from all standpoints, technology, performance, price, then why not? We're open to any kind of suggestions. The most important thing, again, is being a practitioner and actually getting inside the mind of the big data ecosystem. I really agree. Yeah, in terms of, so what advice would you have to IT pros out there, they know a bit about, maybe they're DBAs or they've got experience in data management, but not in big data. And they're wondering how they're going to transition in their careers to really take advantage of this. What advice would you have for them to kind of get moving down that road? I think just go ahead and do it. I've always been a fan of just doing it basically. And then there's- Nike is the big data company, just do it. They did it. I hope they won't sue me for this. I think that was a good question. I mean the point is that, as opposed to 2008, 2009, the tools are much, much more finished. Everything is much more easier. You've got Amazon, you've got Hadoop infrastructure in Amazon, you can get started- Hadoop is enterprise ready. It's about time to- Exactly, you can get started in more than one way. And you can just start doing it right now. Just learn the concepts, learn how to think like a big data practitioner or scientist. And the community is pretty good. The community embraces you. The 2,000 people here are testament to the fact that if you are willing to your point, chef, if you have an idea and you want to implement it and you want to reach out and figure out, if you have a question, because I agree with you completely, it is a much stronger and readyer ecosystem. If you have a question, reach out to the community, there's enough people now with expertise and experience and open to share it. Yeah, and we're bringing them on theCUBE. And they're bringing them on theCUBE. Yeah, I absolutely agree. I mean, I've been covering, as a reporter before joining Wikibon as an analyst, covering more traditional business intelligence and data warehousing. And as I've kind of, as we've all started to learn about big data, what I've really been struck by is exactly what you're talking about. The community, just the ability, the willingness to share data. I mean, Abhi, you've been coming on theCUBE a lot. You've helped me kind of understand the market. So yeah, I agree. It's a great community of people. You know, there's a little bit of that co-opetition on the vendor landscape, but that's good, you know? Frenemies are good, there's a perfectly healthy phenomenon. Exactly, frenemies is a very good thing to have. Absolutely. All right, well, thanks so much. We really appreciate you coming on theCUBE. Thank you so much. Thanks for having me. We'll be right back in just a few minutes with our next guest.