 But for all the best content on the planet, that's free. No banner ads, banner free, and always a fun place to get all the data and information. And I'm here with Mark Manson, industry analyst, elusive, omnipresent, guru, Twitter influencer. Welcome to theCUBE. Thank you. So tell us what your impressions are, I'm only kidding about it, you are influential. But we were chatting last night about big data. What are you seeing right now at the end of the show day two, or day one, officially the big day, but really day two of the actual conference we've been broadcasting. What are you seeing, what are your thoughts? What's the big trends that are popping out at you? Well, I think the big trends that are popping out are probably the shift from the early stage startup scene to the much broader enterprise IT market. I think that the era of startups selling technology to other startups is ending and the runways run out on the free cash from the VCs. Now it's a, oh, enterprise is wanted by this stuff and they have the money and the startup scene is run out of money, we better move. So that's my nutshell. Okay, let me just, we have some time to talk about that, so we don't want, okay, that's the end of our interview. Thank you. So let's talk about that. So we just had John's dog deal on, and he said, and big data is coming to the enterprise through the front door, and it's coming fast. So let's unpack that a little bit. So, you know, the rubber's meeting the road, okay? The VCs are funding quality and business value. So you got to have some tech to enter the market, so you got to have some chops. Not a lot of, hey, I launched a company, I'm done. You know, celebrate, get my VC funding, get some VC funding, some scrutiny right now on a lot of things, product market fit and the go-to-market path. So with respect to that, where are you seeing that business value connection on the product market fit and the go-to-market for these companies? To do the business value equation, because no one's going to buy the hype that's just too much noise right now. Yeah, there's a boatload of noise. So I would think that what you have a lot of right now is technology searching for a solution, or, you know, it's like, I have this technology, what can I solve with it? As opposed to people actually asking for it, and that's a big problem if you're a tech company, you know, it's a push. And in some areas, it's a pull from departments because they're frustrated with IT. So a lot of the big data messages been around get us in your data warehouse, get rid of databases, do this, do that, it's faster, it's agile, it's easier, which it may or may not be. Or it's a part of the underserved. And when you look at things like transactional data, databases do it well. Time series, databases don't do it well. Networks, databases don't do it well. So what are you going to do? Well, you're going to have to turn to something else. And so we're seeing a lot of big data as applications sneaking into other parts of the organization. And what's that called? What kind of effect is that causing when you're seeing? And obviously some confusion, but also some impact. It's causing some change, right? Yeah, it's causing change because you've got the IT people being disrupted. You've got a part of the IT organization that never dealt with data, developers dealing with data. And you have the data people being left out because they don't have the technology to support what these guys want. And so we're in this one world's collide period where developers on one side, data guys on the other, each one trying to talk the other's language. So what are we people doing? I mean, what's your sense of the, obviously people can get stuck, right? They can break through it. What are you seeing that the action right now on today's marketplace? I mean, is it just, okay, everyone's got the pause button, is it evaluation time? I mean, there is confusion. I mean, I was just talking to someone out there that's like, I don't know who to call. I walk into the exhibit hall and I see all this noise. I don't know who to call. I saw EMC's in the news, maybe I'll call them. So there's a lot of, I don't know who to talk to. Yeah. What are you seeing there? Because of the confusion in the market, I think we've got a situation of disambiguation that has to happen and this is probably the year that it does because people say it's an analytic solution. Well, maybe, maybe SAS is good enough. It's a database solution. Well, maybe database. And there, I think that you're seeing right now the strawman being thrown up by the vendors who are threatened, right? The Oracle or the IBM or the Informatica are your choice in the sort of, in the EMC as, yeah, definitely in the store. The green plume announcement yesterday. Terra data. Everybody's being threatened. IBM. So let me ask you a question. Let me ask you a question. Terra data versus EMC, green plume. Who has, who's going to, what's going to happen there on the BI side or data warehouses side? Because the green plume came out with a very focused, we're going after the data warehousing. SQL on Hadoop, Hadoop is a nice little Hadoop wash there. Come in, ride that. You talking about Hawk? Yeah, Hawk, yeah. And you've got Terra data that's already out there. You think they'll put a dent in Terra data or? At this point, probably a very small dent, but EMC has a lot of feet on the street. So maybe, but you really got to roll back a little bit. Yeah, you have to roll back a little bit and look at what does Terra data do? What does EMC do in this Hawk thing? And how are they different in terms of both implementation and what you can accomplish with them? And it comes down to worldview. If you believe the world is going to gravitate towards the database wrapping things like it has for the past 20 years, then Terra data might be the way. If you believe that bringing things like SQL to Hadoop and building out that world and migrating it closer to the enterprise data world, then it's the green plum side. And so, I don't really know, to be honest, because they're both potentially viable futures right now. Well roll back a little bit more if you would. So where I was going with that was, you had Vertica, Natesa, Green Plum attacking, essentially Terra data when they came out, the MPP guys. And they made a little bit of traction. They got good exits, but Terra data did fine. Okay, now the big data thing comes on, it's like wave two of attack on Terra data, which brings us forward to your last comments. But I want to flip it now, because you consult to both the vendor community and the IT community, the practitioners. And so my question, you've already answered it, but I'm going to ask it anyway, is my data, I'm an IT practitioner, is my data warehouse a dinosaur? It's not a dinosaur, but what you really have to do is look at the uses of information in an organization. And data warehousing is designed around assumptions of stability. Things don't change a lot. Core transactional data, essential financial pieces of information about the business. Customer data, behavioral data, external signal, never part of that inside the firewall transactional apps. And so you've got the monitoring, see exceptions, and maybe do some analysis of those exceptions. Figure out the root cause of what's going on. Choose amongst three different choices, like I could make a decision to do X, Y, or Z, which one is going to lead to the best outcome, and what are the downstream implications of that? Or forecast or predict this outcome? Data visualization stuff can't really run effectively off a data warehouse, because it's in human interactive time and data warehousing, the eye tools are not. And so how do I convince you that you need to change this thing and we need to take this to the CEO and get it approved as a business project? And so convincing, educating, elucidating, laying out problems and explaining, all of that is part of a spectrum and the data warehousing market drew a little tiny box around the first piece which was monitor and identify exceptions and it's a very stable thing that very rarely changes. That's where we are. Yeah, you described it very well. It's like this hardened piece of cement. Every month a little bit changes. Add some new data, whereas this new world is, hey, let's redo everything. Yeah, well that's Cloudera. So I like that world view because it's kind of like Democrat-Republican, pick your religious view and you only know it'll play out based upon what happens on the street. The market will spin in whatever direction. Is it that though? Or is it Democrat, you can make a case for either side. You could play for my counterpart. Well, let's think what he was saying about the wrapper though, that there are use cases but there'll always be these niche deployments but what's going to really, when the world spins to whatever preferred environment, use cases developed at scale in terms of corporations. Right, so. That'll be defined by the solutions and if wrapping it and that's the package, a container if you will. But if the, what do you want to call this new way? Big data way, whatever. You call it Green Plum, whatever. This new world delivers much more corporate productivity on a growth trajectory, almost like the microprocessor revolution, then it's not like, it's almost not a win or lose. It's like this is the area that's going to now consume everybody's attention and get all the investment. It's a classic headroom, right? It's a headroom argument, right? It's like, hey, I'm going to do something today and do some headroom. So what do you think? Or there's foreclosure, you know? Yeah, I'm thinking about it. You're the expert, right? So John and I are sort of peripherally involved. I suspect that having lived through the first set of BI and data warehousing wars when we were telling main framers that those reporting systems that they had been designing were the wrong way to do it and there was this data warehouse architecture and it was a new way. You ended up with, oh, OLTP is an IT architecture. It has its whole systems architecture and philosophy around it and we renovated that in the whole Web 2.0 sort of stateless scale out stuff. Then you take the data delivery side of the equation, the business intelligence data warehouse market to date and that was all about getting data in and getting it out to eyeballs. And now you've got a third piece which is both that and machine to machine like recommendation engines or whatever or automated detection and management systems. I think you're seeing the evolution of a third piece of architecture which is processing, right? Because databases are for storing and retrieving data. They're not for processing. OLTP is for recording transactions and sort of storing it in that way but for the execution of tasks. And now we have a third leg of an IT system architecture evolving which is the processing of data at low or high latencies, at large or small scale, in real time, in batch, it doesn't really matter but it's a different beast. And so you've now got something that offers you new capabilities. So it doesn't supplant the data warehouse any more than the data warehouse supplanted the mainframe or the ERP system. Right. Now at the same time, sort of back to what I said before who sort of driving the most corporate productivity being the traditional, let's call it the traditional BI data warehouse, failed to deliver in my opinion on a lot of the promises that were made. That's 360 degree view of the business, the predictive analytics, it was just. That's called hype. It was hype but they didn't live up to it, right? Now, it'd be hard to live up to this hype. I mean, I guess the internet lived up to its hype, right? So sometimes it happens. We'll see, with big data, the practitioners that you talk to in the data warehouse, the business intelligent world, they're constantly trying to figure out how to make what they have better. They're chasing chips, they're maybe bringing in the latest and go, oh man, the tease is here, let's try that. They're always trying new things. It maybe helps a little bit for a while and then deluge of data and they're back chasing their tails. So that cycle is not going to change anytime soon, right? And they're still playing now a vital role within the organization that's tied to the processes they're locked in to the organization. But at the same time, that traditional world is being dragged into Hadoop. So those two worlds are coming together. Right now it's like the Hadoop tail wagging the dog. Do you see that flipping? I mean, it's a different version of the question that you said I don't really know on before but I'm going to ask it anyway. I'm just going to give you the same answer. Well you said you couldn't be pinned down so I'm trying to pin you. The thing is, I don't see either or I see both because Korean retrieval is one half of a problem. Think about it like Web 2.0, right? Web 1.0 was a publishing metaphor. Here are these words, you give them to this guy, he HTMLizes them, formats them, pushes a button, they go into this thing which broadcasts them out to the world but there's no rewrite. Along comes HTTP, RPC, and suddenly REST based web services develop and now we have read write web. That's a different, it changes architectures, it changes what you can accomplish. The big data stuff, not specifically Hadoop but all the processing plus the low latency which we don't have in the database world because you can't store data and then act on it in real time, you have a new way. And part of that is that, what does Hadoop? It's a storage system and a processing engine. It's not an ETL tool, it's not a pipe, it's not a database. And so I have a read write system. That's a different set of infrastructure. It's not optimized to get data out real fast for query but with that read write system, I can change so much of what I've done with information as opposed to IT being the arbiters of taste and I talk to you and then I cook up a meal and then I serve it up to you. Well, now with sort of the big data tech you're starting to give people the kitchen. Mark, what's your thoughts on sort of the structure of the industry? I've said many times that we live in this world that's almost, it's an oligopoly, right? You've got five, six, maybe seven companies controlling the chessboard. Oracle goes and buys a company, they own Java, they own MySQL, right? The list goes on, EMC could make big moves, they have made some moves with VMware. So it's almost as though the sort of innovative startup, their objective is to get, have an exit. Maybe there's an IPO in it for some of them, some small minority. But generally, I don't see that changing. And so, and I wonder if you agree or disagree with that but the question I have is, do you feel, do you agree with that? And do you feel as though those whales could continue to control that chessboard, make moves and then morph into this new world pretty, without too much disruption, you know, they'll cross the chasm. I think they've figured out how to do that. Do you agree or disagree? I agree in principle, but it could play out differently. You know, when you look at what's happened, you look at a bubble chart of sort of the size of companies in tech over the last 10 years and you just play that forward, you have four or five giant bubbles and tiny little dots all around them. And they've just essentially outsourced their R&D function and their risk-taking functions. The problem is every time they ingest them, they do a crap job. IBM, Oracle, SAP, they've all digested companies and poorly executed on the promise that those companies were expected to deliver. All these database companies, the netizas and the verticas and then this and then that, how much have they actually done for those vendors? The answer is it varies. Really, they just become more price book entries for people to throw at the wall. And there's some point where you can't just be solving things by buying product. You have to have rethinking of architectures and those vendors are not because they have lines of business that are drawn up specifically around things. They don't want to cannibalize their own thing. We saw that like Cisco was a great example. I mean, they were on a run and on Web 1.0, they bought everyone and then the integration nightmare. I mean, I will give EMC credit. They do know how to swallow a company, but operationally, they're buying lines of business. But you're right. We ever talking all week about old way, new way and fundamentally this data platform is a new architecture and some of the things just aren't invented yet. And I think the entrepreneurial community is ripe with innovation in the sense that you got guys beavering away out there saying, hey, you know what? I don't want to go work for the big company. I got 15 years experience doing data warehouse business intelligence, data management, or whatever, I'm just going to build my own system and bring that to the market. So you're right. There is a little bit of picks and shovels Dave right now. But the database is particularly tricky. Well, database is interesting. What unstructured data gives a program sort of a bit more programmatic freedom, right? So you have, you're not schema bound, right? So you now can have some range, right? So I think, you know, our concept of data as code building off our infrastructure as code riff this morning was interesting because infrastructure as code is changing the cloud and business. Because you're talking about simultaneous deployment of code data and teams. I'd say there's a second piece to this too, right? There's the buyer side, the IT audience. Think about what's happening in IT. You mentioned cloud, right? All these business departments are running around buying apps to serve their department and it's shifting to SaaS and it's shifting to cloud and IT's role is becoming not the guys who have to do everything. It's become sort of all administer the app in a light remote way. And so what does IT's role become? They're really, the main thing is data. The data to connect to the Salesforce app, to the CRM app, to this other app, to this internal thing which may be hosted, which is my ERP solution, plus all of these other bits and pieces. Plumbing and data. And that is totally different than the role of IT in the past in many ways. It's interesting, you mentioned main frames and talking about the data warehouse in the old days and all that. The old app systems and stuff and kind of jolting on that. One thing that's very, very interesting and this is kind of popped out of my head when I saw it over and over here in that same term, data processing. Now, remember back in the textbooks 1970s, data processing was a discipline. Yeah, and data processing was punch cards, right? So the DP department, so now we're in data processing. So here we go. Here we go, we're back into data processing. So, okay, data pipelines, data processing, data management, data stewardship, we're seeing this kind of coming back. Again, coming back to your other point, the riff on is architectures. Yes. And I hate the word software main frame and I know that's been talked about with Paul Moritz all the time, but I don't like it, but I don't have another term. If data processing is going to be a discipline again, okay, in a new architecture, what do you think it's going to look like? And obviously, you know, I don't think any vendor really has the answer yet. And I will give Cloudera props. I mean, what they're proposing with Impala is academically and theoretically and being executed as a multi-resource platform that enables resource-based data processing, meaning no one vendor can do anything. You know, but Morton, we're all the same thing. So what was your take on all this? Well, I think trying to say that an architect is something that a vendor can sell you is a mistake. I don't think vendors sell architecture. They sell products and products fit into larger architectures. And I think what we're like, let's use the term data platform because that's what I like to call it. Data platform should subsume data warehouse and Hadoop and a real-time processing framework, a storm or a Kafka or whatever your favorite, you know, du jour as well. And you have to start thinking about where do you draw the boundary lines between platform and application in a new and different way? And that gets so little attention. I think that the vendors don't even understand it because part of it is you have to remove the old frame of reference. The old frame of reference says, here's a box and here's a box and here's a box and here's this thing. I'm going to shove it into a box. And preserve special interest, right? So, you know. Exactly, the gardeners and foresters of the world, they have their waves and their quadrants. That's how they make their money. We have a new report coming out called the Magic Kingdom. Yes. And the Magic Kingdom will be, I tweeted that one, it's got a lot of, why can't I see the Magic Kingdom report? Well, it's kind of bigger than the quadrant, but we haven't released it yet. No, I mean, but yeah, it was slow, but that's the magic quadrants for people who aren't paying attention, right? So, they need to see things in. I think that it actually is a sign of the difference between a stable market and an unstable market. And I look at IT markets as periods of punctuated equilibrium. We had a nice period of reasonable stability on the data side. The web stuff was going crazy. But now, all of a sudden, just like 1992 or so, this part of the market's going insane. And it's going to be chaotic for the next few years, the data side of the market. Not the, I'm taking and doing transactions, but the data infrastructure, data management, data storage, data retrieval, data delivery, all of those things because we've had several orders of magnitude of technology change and capability and the software architectures have not shifted to match. Just like in Web 2.0, software architectures went from app servers on centralized grids to stateless architectures and browsers. There's no doubt that this decade is a lot more interesting than last decade. I mean, there was some cool stuff with mobile for sure, but I mean, from an enterprise standpoint, it was compliance and I mean, it sort of gave a nice boon to the BI business. Thank you, Enron. But you're right, this is a much more disruptive and volatile. Okay, we are wrapping up day two. Mark Madsen, great way to end it and kind of reading of tea leaves, kind of stepping back, looking at the worldview, kind of riffing out here, kind of like a bunch of musicians playing our notes on the marketplace, looking at what's happening. And I think it's a really good perspective. I think, you know, the summary of this conversation is one of great innovation opportunity, great disruption, disruption kind of a lot of things coming together at the same time, multiple theaters as they say, but the platform really is the focus and we're going to be covering it all day tomorrow. Stay tuned. This is theCUBE at the Stratoconference at SiliconANGLE.com's exclusive coverage of Stratoconference and we'll be right back for a wrap right after this short break.