 San Jose Convention Center, extracting the signal from the noise. It's theCUBE, covering Hadoop Summit 2015. Brought to you by headline sponsor, Hortonworks, and by EMC, Pivotal, IBM, Pentaho, Teradata, Syncsort, and by Atunituandisco, now your host, John Furrier. Okay, welcome back everyone. We are live here in Silicon Valley in San Jose for Hadoop Summit 2015. This is Silicon Angles theCUBE. This is our flagship program. We go out to the events and extract the signal from the noise. We've got two great guests here. We've got Ryan Peterson, Chief Solutions Strategy at EMC, and Mike Cookie, Senior Director of Outbound Product at Pivotal, welcome guys to theCUBE. Thanks for having us. So we were talking earlier this morning. I mean, a lot of stuff circulating here in the air. Obviously the keynote's all about the enterprise. You know, Rob Bearden's got a John Chambers like swagger up there. He's fresh, he's not sick like last year. He had the flu, I remember correctly. But things are good right now in the software business. I mean, software is eating the world. You guys are doing some great stuff there. But there's a really interesting dynamic going on between the big players, the startup ecosystems, open source ecosystems, ODP, cloud now really accelerating with software oriented tooling and platforms to make things easy. So the theme is easy. So what's your take, you know guys, what's going on Ryan, EMC? I mean EMC, complete 360 from where it was. You know, a year ago, open source. You've got cloud scaling. You guys are talking about code, DevOps? I mean code, can we? We're no longer your father's EMC, right? We're definitely changing the way we go to market. We're shifting the culture from kind of the old school mentality, the new school mentality. And we're taking our new products and we're really going out there and changing the way that people adopt these solutions. And with a great federation partner like Pivotal and of course other federation partners like VMware, we're able to really change the way that people are going to market with their, and change their strategies. Mike, so talk about Pivotal, obviously, and Pivotal was acquired by VMware or the spin out I was called that Pivotal became. It was doing Agile before Agile was Agile. It was the software, it was that generation of San Francisco, Clouderati, came out of that whole early stage DevOps movement. Now with EMC, how's that federation partnership working? What are you guys working on? And what is some of the benefits? Because we're hearing that from Pat Gelsinger, like 10 Ways from Sunday, the federation at EMC World, we had two cubes there, it was very clear, Joe Tucci, the federation is a value add for customers. Where is that value? Yeah, so first of all, Pivotal Labs was from EMC actually, it was acquired by EMC. They were the killer Agile developers for 25 years. And I think that was the beginning of EMC seeing their customer base kind of demand this next generation of competitiveness through software to your earlier point. So at the time when we started Pivotal, we had these Agile developers over here and we had these big data solutions over here and they were very split up in terms of how we went to market. What's happened over two years is this is all kind of played out the way we expected and now all of these component parts are really required to build this next generation platform. To your point though, it's not easy. There's a lot of different technologies. There's this development practice that could be disruptive as well. And I think where the federation uniquely can provide these benefits is providing that guidance, doing the pre-engineering required to kind of harden these solutions and deliver something that's really rapidly delivering value to the customer base. So that's where I think we really work together really well and it's changed. Two years ago that meant one thing. To your point about this conference, I think what's exciting here is we're hitting this inflection point where the use cases that Hadoop and Big Data are now generating are becoming application drivers. And I think that's where this buzz is creating the connection between Big Data and analytics, a development methodology and to your point the cloud. But again, so complex to put these pieces together at an enterprise class level. And I think that's a place that we're pioneering together. Let's talk about the customer journey, right? So the customer cares about one thing. I don't want to get fired. I got to produce results. Now top line and bottom line cost savings if you will, which is perfect economics. The cloud and Big Data sets the table for a lot of all that stuff. What's going on? Because again, it's not easy. DevOps is now front and center. This ecosystem has open source, which is always moving, right? SLAs are, they mean something, right? And you've got to deliver value, right? This is something that, I mean, ISLON has filled that gap in an emerging Hadoop ecosystem for a long time. I mean, the other piece to your economics point, right? You have this huge customer base leveraging ISLON as their infrastructure of choice. And then these disruptive forces come in and drive this change into the IT operation system. And something I'm proud of ISLON for is they saw that change coming and they pivoted to be that kind of foundation to guide them with a lower risk platform. So, I'm sure you'd say this, Ryan, but two years ago it was about a really stable foundation for Hadoop, for HDFS. It was about leveraging this investment they had made in ISLON in this next generation use case. Today, we've kind of solved that. ODP, to your point, the open data platform has created kind of a very consistent and predictable foundation to support. Now it's about advanced analytics on that data set. And this is, again, the place I see ISLON. We've recently just made some advances on their platform. They made advances, actually, that enable advanced analytics executing right against their storage array, which I think is a sign. The customers moved from landing data to processing data. Now they actually want to operationalize that data. To Michael's point, I mean, we've had customers, two years ago, we're talking about MapReduce, MapReduce, MapReduce, and we were really focused on how do we make MapReduce really efficient and work well against ISLON. But the conversations today are, how do we make Hock work, how do we make Hive work, how do we get into more of the upper layer stack, and then more importantly, how do we actually drive results and new value from the data? My favorite stories now are the ones where people actually tell me what they've done to change the business. And I tell you, I met with a telco, for example, and it dropped calls, how do we don't know? We don't know how many drop calls are because of the person's handset being bad. We don't know if it's because they disconnect from the tower, or if the tower was a problem. But if it's a handset, we can fix that. We can go out to the handset, tell the customer, hey, replace your handset. But the problem is, in order to get to that point, people had to start driving data into a facility, storing that content with no understanding of what they might get out of it. And that I think is to some of the analyst reports that are coming out to the point. People are not understanding what the value is. They're looking for a case study, a use case to actually start working on. And actually it's almost backwards. We've got a lot of our customers are saying, hey, I just got to go collect all my Lego pieces from around the house and bring it all to one place so I can figure out what I can build. Because I don't even know what I've got. Well, that's step one, right? You got to have that brute force energy of having the mindset and cultural shifts and hey, we're going to think differently. This isn't a, have a meeting, assign an action item, some department goes build something. Is it like this stuff that has to get done first? But it's a massive shortcut to get that done, right? You mentioned SLAs. If you're doing this with proven, mature infrastructure solutions, right? Iceland's already embedded. The cluster's already there, maintaining an SLA enterprise class, storage infrastructure for the customer. So if you're able to adopt that existing infrastructure and those existing SLAs and proven experience, then all of a sudden you've got this new use case with a lot less risk. Well, this is a great point, Mike. This is the future proof we've been hearing and as a word that's been around since I was, broken in the business 30 years ago, which is always code word for charge more money and promises will come out later. But in a market that's moving so fast, this is what people need. They want to have a starting point. They want a bridge and a partner to cross over that bridge to the future with. So it's very relevant right now. So I got to ask you, there is an SLA requirement, but also there's also training is also, I need to knock down some quick wins first. I need to operationalize it. Now customers I've heard said, I have no problem spending money on gear and tech. They've used to, they write checks all day long, EMC, IBM, everyone gets checks written to them, but they don't want to get stuck with irrelevant stuff and a cul-de-sac with no outcomes. Right, so again, that's why everyone's using the word business outcomes. So I got to ask you, what is, if we believe that outcomes is what they want and they'll write checks for stuff, whether it's cloud, gear, how do you get a customer to walk the first few steps and then what is that bridge? What is that future proof message? What do they need to have in place? So I have a conversation about this actually tomorrow at 8 p.m. It's a great question. The discussion is, where do I move my data to make it so it's future proof and accessible by anything in the future? And this is where I really believe that our story shines because if you move your data into a tool set, then you've locked your data into that tool set. And it could be something as simple as an Oracle database and you put that and you've locked it into a DVF file and all of a sudden that data's no longer accessible. You have to access it, move it, translate it, convert it, whatever. And I think that the future of the data, the data lake is all about keeping that data in a raw format, making it accessible to any particular tool set and ultimately abstracting that value up into Cloud Foundry, where I can then build any kind of application against that data, whether it's an analytical application or a business driving application. Yeah, I used to love Joe Tutti's comments around resource pools and it was generic. He didn't say EMC, it was implied EMC because they had a market share, but now with commodity hardware or cloud, you can put it anywhere. So as long as it's not locked in, it's accessible. Yeah, you can move stuff around, but putting data into a new schema and to new ETL systems, these are the, that's the lock-in, right? That's what you're saying. You're locking it into the tool then. As soon as you lock into the tool set, well then it's, If you can do a lock-in, lock-in to a platform at least, right, I mean, or have no lock-in. Ideally no lock-in, right? Yeah. I'm a big proponent and I think EMC'd be okay with me saying this, is I'm a big proponent for keeping it completely flexible at the storage level, at the compute layer, at the software layer, and ultimately up in the application layer. And I think that by using open source commodity components and having an open and flexible data lake architecture, you can do that, you can do that really well. Mike, talk about the conflict between open source and SLA, because this is something that's come up last year. Certainly a big data SV, we talk about this with the ODP, with the folks from Pivotal as well. We kind of hit them over, we're hard with this, but they answered it really well. They, you know, it was, okay, open source is moving really, really fast, but the customers aren't, right? So the customers need hardened something, it's something that's hardened with support. So we were speculating and discussing candidly, it's not a bad thing to be lag a little bit, because if someone's on a migration path for say a year or two, they don't really, they want a trusted partner to keep an eye on what's going on in the public, open source. I think, Move the innovations in. Is that ODP's mission or is that just, or generically more of a guiding principle? No, that's exactly the mission. I think also what Ryan was just pointing out, those stages he just said, there's foundations with absolutely open communities around them for both those stages he was mentioning. So that's exactly right. I would say, you know, the open source momentum has been amazing. It's resulted in this phenomenal development curve that we really haven't seen before. It's amazing innovation. But you pointed this out earlier, you were talking about lock-in and what, you know, one thing the customers really invest in is their software stack for analytics, right? So you have these very mature software analytics packages and a relatively mature existing analytical warehouse, data warehouse ecosystem underneath it. And when the customer goes to mature towards a Hadoop use case or an advanced use case on top of Hadoop, they need to know that that software stack that they're using can be migrated. The skill set that knows how to use that software stack has to be able to be migrated. So I think ODP Open Data Platform gives us the ability to say, look, all of you amazingly powerful ISVs, your software packages have to migrate with your customers so they can leverage this next generation. And the ODP enables us to effectively do what you said, except the idea is to really shrink the amount of time that they have to wait between an Apache innovation and maybe an enterprise implementation of that Apache implementation. So yes, the ODP kind of rides above the Apache ecosystem and tries to collect a very particular and predictable set of the Apache projects so that enterprises, but also the ecosystem around it can kind of start to adopt it rapidly. Yeah, Ryan, I want to get your thoughts. We had Donna Prilich on earlier. She's a geek. She's awesome. She's been involved in Hadoops since the beginning. We love talking to her. She wrote up a good point, which is no one does 200 node clusters out of the box. So this is the reality of customers. We're beyond POCs now, right? Okay, so POCs, they're happening all day long and has its own agenda. But when people put their toe in the water, they want to get their feet wet. They're going to come in 1925 node cluster, somewhat running the engine, redlining it, they're going to test it, and then move into some sort of reference architecture. So I want you to take me through the state of the union on reference architectures. How important are they? Do you guys have them? You see that as critical success for the customers. Can you elaborate on that part of the customer interaction? Let me just back you up one second because I think there's a little more variety in terms of the ways that enterprises are adopting this. There's some enterprises that are under such critical threat because their entire industry is being toppled over. And you better believe it's not about 20 nodes for them. It's about a complete replatforming. Like, how do I immediately change what I'm doing because the trajectory that I'm developing at and leveraging data at is going to fail. So there's a lot of companies, a lot of enterprises just being completely toppled. And in that case, there's a lot more urgency and a lot more interest in building a next generation platform potentially in parallel. But I think to your point, most people- Well, let's take those two cases. Well, those are two good use cases. So the one is just like, okay, we're going to hobble along. We're going to put some mile markers out on the milestones and we're going to go through this transformation. The next use case that didn't fit the old picture, let's build that new thing out. We're not under, we don't have a gun to our head. Now, the second one, which I love, which is like, I'm a financial services company or someone like that, it's like, I better, I got to move, the game is on, the clock is ticking, I got to transform as fast as possible. I don't look like Amazon. Something's wrong, right? Yeah, so take me through that. Culture, tech, roadmap, Bill Schmarzo walking, or you guys send the team in, parachuting in, how many people on the customer side, give us a taste of what that flavor is. Questions, where does the decision being made? If the decision is at the executive level, they see the threats and they understand it. Then we take somebody like a Bill Schmarzo, we say, hey, let's go do a vision session, let's talk about how you're a threat. How can we help you change your business completely, revamp it, and come with a whole big data strategy behind it and let's execute against the particular vision to make that happen. But that's a smaller majority. Really, the group of people that you're talking about, I think, is those people that are from the bottom up. The business hasn't quite. Get ahead of it. Yeah, the business hasn't quite gotten there yet. So IT's saying, hey, I see all this data. I think I could probably find some really unique value from it. And we've got great example use cases of customers where they went to executives and said, did you know we could provide you this information? Wow. Well, I think the majority of people here at this event are in that camp, but I think what Mike's pointing out is there are named accounts that are huge, that you have that relationship to. So I want to kind of tease out the order of magnitude of the seismic shift because compliance server back in the 90s, when I broke into the business early 90s was, that was a mandate. Okay, let's move from mainframe to client server. And SAP, HP sold gears, sun was around then, or we all the guys we know. That was a three year process to load a CRM system, ERP system, to get printed checks. Like, okay, yeah, what do we get? For the zillions of dollars we spend in three years later, but we have data coming in and we have nice fancy checks and other cool automation. That's three years. What's the timetable now? What's the equivalent shift in acceleration? Again, I think it has to be done in a matter of months, six months, you've got to be up and running. I mean, your first project, I think you get a line of business request maybe that hits the operations team and they say, this doesn't look like it's going to fit in any of the holes that we have shaped, right? So they say we got to do something differently. Again, lowering that bar of entry and saying, hey, we've got this Isilon system here, let's carve this up for this new use case, immediately turn on HDFS on that system and all of a sudden you can leverage your existing infrastructure. You start landing those technologies in. I think it's really about just defining that next platform for that single use case, right? Because the economics are different and the return is drastically different. There's an app driving this, right? So they're predicating the apps doing all the plowing. The other is the piece that Ryan was mentioning, which is if operations teams start to leverage these technologies themselves, there's amazing insights that can be had at the operational level as well, right? So IT is becoming an innovation engine. I think we're seeing, if you think about the last 10 years, IT has been a service industry constantly a cost center for the company. Underfunded, a bottleneck, right? No options as an operation, note everything. All right, so if we take an operations team that becomes a software is eating the world operations team, all of a sudden they're actually driving insight into the business about behavior or application. DCIT is a bottleneck right now. Do you think they've come to grips with the fact that they need to be driving the bus or getting out of the way? Each company's different, right? Some of them, they're really innovating and they get it. I like to say CIOs are either chief information officers or they're chief innovation officers. And it really splits between two camps and every single time you meet the customer you can tell exactly how their CIO thinks. And the rest of the organization thinks that same line. Yeah, hello, Cloud, Shadow IT. Right, so again, let's get to the Cloud. Sure. To wrap up the segment. Cloud is, I mean, OpenStack was an eye-opener. I mean, so Brian Gallagher ran billion dollar P&L at EMC's now. Building from the crowned up, EMC's open source, open presence, open the open EMC, EMC Open, or whatever you guys call it. You have Azure kick and butt. You have Google out there. You got Amazon. You got VMware. Trying to go fast as they can. HP, I mean, IBM. This is interesting, right? So what's interesting is that Cloud Foundry, a pivotal spin out, or they've created a foundation, is the center point of that. So if you look at ODP and Cloud Foundry, I'm smelling some synergy there, guys. Connect the dots between that. And how is that going to accelerate what I've seen as a slowdown in this ecosystem? Sure, sir. Or am I, did I? No, no, no. I'll say something quickly and then you can take it from there. I think that's where we actually stand out, really uniquely, right? We launched something called the Federation Business Data Lake and that is literally the collection of pivotal technologies with EMC technologies and VMware technologies that is a reference architecture and engineered solution to deliver to enterprise customers. But I think that's where you find the Federation really has a unique stack. I mean, the ability to combine where we're heading with our PAS and our data services being combined into that platform as a service and then how the EMC Federation behind it can supply that kind of predictable infrastructure and rapid scale out. And then the other thing I was going to pass to you is you were saying, how do we achieve these initial projects? Their service is heavy today. And you were saying earlier, customers love to write that check to the infrastructure team or the IT technology team. They're starting to wonder why they're writing those big service checks. And I think our job is to create a situation where they don't need to do that anymore and they can actually achieve it on their own. You've got to deliver a relevant solution so they're not stuck in some cul-de-sac with a product and or app that's not delivering. Because they're already out of the limb. So, and this is going to come out on our keynote tomorrow with CJ, so it's still a thunder. But, you know, we're releasing a whole series of big data products to handle exactly where the customer wants to put it, how fast they want to run it. So ECS, for example, or Elastic Cloud Storage is for those service providers, for the ability to have service providers start building those solutions, whether it's with a big service provider or with VMware or whoever it might be, we're going to start providing that ability to transition data, whether it's on-prem, Icelon type of implementation, or in the cloud, whether it's the new DSSD technologies we're bringing out, all these new things are really about making it so you have an open and flexible opportunity to put any data anywhere. So, Ryan, you mentioned CJ, CJ Desai, who's the SVP runs the new emerging technology group at EMC. Which, by the way, Icelon is not in because they're in the core technologies. No, we're over at ETD. We're under CJ, yeah. Icelon is. Icelon, oh, no, Icelon is, I'm sorry, Icelon is, but ExtremeIO is not in. So one's emerged, but no, it's a cash cow, I get why. I explain that on my EMC views, I don't want to waste my time on that now. But tell me what it's like at EMC right now, because CJ, he's a geek, he's a nerd, he's awesome. You got a lot of cool mandates there, you got a lot of flexibility, you got a lot of stuff going on in the flash. What's going on? You got the big data, you got the cloud, Icelon anchoring the whole thing. I got to say, I always tell people, they ask me, you work for a storage company, it's not boring, no, this is an awesome opportunity. We're right in the middle of innovating the next generation of what EMC will be. It's incredible, and having people like CJ, the awesome leader, Sam Groek, I mean, all these guys are incredibly great leaders. We're really building something new, innovative, flexible, and I think you'll see some things coming out in the next year or so that's going to just be around. So we were speculating on theCUBE last week, can a storage company stand alone and win? The answer is no. Right. And NetApp, obviously the CEO got, there was a thin boardroom thing that went down, still trying to get to the bottom of it. NetApp, darling in storage. You guys are shifting away to software, right? An emerging cloud, pure still out there. So it's interesting, this storage market, if you're not in an integrated stack, this is my personal opinion, if you're not an integrated stack, software, you're going to be not positioned well. It's about thinking differently. So CJ will tell us, don't worry necessarily about innovating at the hardware level, we're going to do that DSSD example of that. But really we're abstracting the storage value from physical storage and going to the software. So whether it's I salon, it's going to get abstracted out of the hardware. You got all of our software defined storage systems. That's the future, software will eat the world. I love chatting with you guys, we're getting a hook here. I want to press on, I can do another 30 minutes. But I want to give you guys the final word for the folks out there watching. What's going on at this show? What are you guys doing? What's happening in the Hadoop big data ecosystem? Talk about some of the dynamics quickly. What's happening here? What should they know about what's going on here and what are you guys doing here? I'll start, the business has shifted from collecting all your data to building applications. We're going directly from here's all the information Hadoop was being played with, application providers are coming out of the woodworks and creating new capabilities that you can just buy as opposed to having to build manually. I think that shift to application development and shift to automated applications, that is the future in my perspective. And I hate to say that I feel like we were right, but in February we said that this year was going to be about open, being agile, and being cloud ready. And we basically talked about all three here today. You know, from an agile perspective for us, you mentioned this, but we have to support fully virtualized stacks across all of our database products, all of our analytics products, and all of our application stack. The merging of our application stack with our data analytics stack in the big data suite has been picked up really rapidly. And I think the cloud and agile piece basically feed each other, but there has to be this IT operational flexibility without killing the business. I think you guys just summarized my argument against Paul Martino, my friend now venture capitalist, who went on Bloomberg, no, went on CNBC and said, there's an enterprise bubble. There is a consumer bubble, maybe some outliers in the enterprise like Box and whatnot, but in Dropbox. But what you just said is an innovation dream, right? Cloud is really coming on strong, huge opportunities for startups and companies. So guys, thanks so much for coming on theCUBE. We'll be right back with more after this short break. This is Silicon Angles theCUBE, we'll be right back after this short break. Thanks for having us. Thanks for having us.