 Live from the Oracle Conference Center in Redwood Shores, California. It's theCUBE at the Next Generation Engineered Systems Launch Event. Brought to you by headline sponsor, Oracle. Hello everyone, we are here live in Silicon Valley at Oracle's headquarters. I'm John Furrier, the co-founder of SiliconANGLE Media. My co-host Dave Vellante, we are here covering Oracle's special announcement. It's our pre-event pre-game. We're pre-gaming the announcement here. Obviously huge, big discussion around exit data, the future of engineered systems, really the future of the data center. And this all touches the cloud, mobile, social, big data world we're living in. It's exciting to be here. And we're going to hear from Larry Ellison at one o'clock for his keynote. That should be entertaining as always. But really more importantly, this is the three year journey after really the vision of this engineered system. So you start to see the fruit on the tree coming off the Oracle, Oracle product lines. And we're excited to have that. Our next guest is Mike Almett, our senior director of product engineering. Welcome to theCUBE. Thanks. Thanks, happy to be here. So engineered systems is all the rage now. You see everyone going into the higher end stuff where people want purpose built was kind of the word back in the day. But you go back a few years ago at Oracle Open World, even in the industry, it was seen as kind of a radical thing. And with the Sun Microsystems assets, you guys are doing that. So give us the latest update. What's happening? And what are you guys, the key announcements? And where are we today with that vision? It's interesting, you know. I think the vision was built around a fairly basic observation that we made, which is that when you start looking at all the work that you can do in understanding your customer data better, you start looking at all the different kinds of applications that you can build on top of a modern internet-based platform. As an enterprise, I think it raises a question, which is, you know, given all the new stuff that you can do to move the ball forward to advance your business, you know, to differentiate yourself, what are you going to stop doing? Right? And I think many of our customers have kind of come to the conclusion that there's not as much opportunity to kind of differentiate by focusing on applying expertise in the implementation of platforms at the low level, right? Putting disks and chips and memory in a box and putting a box in a rack and connecting switches and all that stuff, that's hard work. There's no question. It takes very specialized skills and it costs a lot of money to do it right. But here's the thing, if you're in healthcare or if you're in financial services, is that what you do for a living? Is that your business? Well, now it's supporting the business. Well, so then, do you need to keep doing it? Would you be better off spending your resources, developing applications, talking to your customers, you know, using the social tools, using the mobile devices to provide a better user experience, maybe that's where you want to concentrate. And I think we saw that coming a long time ago. I think Oracle saw, Larry saw, it's not just Larry, but you know, he's an important guy. Larry saw that there were some pretty big changes happening in the marketplace in general and it wasn't visible to a lot of people working on the ground day to day, but I think it was pretty clear long term that, you know, the rise of pervasive networks, new technologies that allow applications to be co-located on shared infrastructure in a very secure way, all that stuff was going to change the way customers thought about building platforms. That's why, you know, Oracle looked at database as a way of building up its application portfolio in middleware as part of that stack. You know, customers want to go up. Yeah, and the other thing we're seeing too, that narrative is interesting, you mentioned that, the customer view and that is that, you know, this idea of integration or integration's integrated systems really is about functionality. So the old thing was, whoa, I don't want to worry about that but there's so much demand with transformation going on with the technologies. There's none of resources around. So what is that key integration point? I mean, it used to be, hey, I got all these clusters everywhere, you know, I got a little sprawl on the server side, I got sprawl on VMWare now, VMs everywhere. So, you know, horizontally scalable systems are great in theory with open source and it's done a lot of great things but you have sprawl. There's a lot of hidden costs there. So, what is this new notion that you guys are talking about this today around super clusters? What is that about and how is that different from what you guys have done in the past? How is it different? Well, I don't know that it's fundamentally different. I think it's, we've made a lot of strong moves in the direction that we've been going for a long time. So I don't know that it's different. I think that it's kind of new and every year we do something new. We spend a lot of our resources on solving some pretty hard engineering challenges that frankly most enterprises aren't able or willing to take on, right? Designing microprocessors is not something that any bank's gonna do, right? But we've spent a lot of time focusing on database and middleware and ultimately building applications portfolio because that's where we think the technology meets the business. Like I was saying, if you're in financial services, you care about your data model. You care about how you're implementing your business logic. You care about your ability to analyze your business to understand what users are doing and to understand where your assets are deployed and that kind of stuff. Do you necessarily care about the details of how to build a cluster under your app tier or do you necessarily care about how to implement a relational database or build a large scale data management infrastructure? No, you did in the past because you had to. But now that you don't have to, why do you? And we're finding a lot of customers are saying, yeah, I don't want to anymore. That's where the cloud is coming from, you know? So, I want to talk about the cloud, that's super cluster, spark based system. I was at an analyst meeting about four years ago, Steve Mills, IBM, smart guy, well respected. He made a statement, he said, mark my words, spark is dead. What did he miss? Well. Is spark dead? Well, yeah, it's most certainly not dead. Okay, so what led him to underestimate spark? Well, it turns out that you can do a lot of really interesting things with software to make software systems reliable independent of the underlying hardware. And you know, you go back 10 years ago, there was a realization, you know, kind of across the industry that we didn't have to rely on really expensive what we call big iron servers anymore to get reliable applications. You could use technologies like Oracle Rack to get a highly scalable, highly reliable database. You could cluster the middle tier and do things like that. And that allowed a large ecosystem to develop around what we would have considered commodity technologies that resulted in some pretty significant gains. I mean, a lot of us have benefited from that. But I think what folks saw back then was that trend happening. What they didn't see is that eventually it was gonna run its course, that we were gonna realize all the gains that could be realized from that approach. And as is the nature of progress, you must always go forward. So we came up to this point where things are working really well and suddenly we have massive scale applications everywhere, life cycles getting shorter, the demands for agility going up. And we discovered that that strategy while it took cost out through standardization, in fact, it left a whole bunch of important value unrealized and we saw customers moving up the stack to the applications to the middle where wanting to care less about the infrastructure and what's happened to us is that the whole notion that at the end of this era, there shall only be one underlying platform architecture, that notion was undercut by the fact that folks in fact wanted to focus on the layers above the infrastructure. And so now you have companies like IBM with open power. You have Intel with its foundry. You know, this whole notion and the success of ARM and all the different platforms built on the ARM, designs, I think it's no longer given that there has to be a single platform under everything. And a lot of folks are leaning back now a little bit and saying that maybe I want to use the right tool for the job. I mean, if I'm going to run an enormous infrastructure at large scale and I want my users to concentrate on the platform, they want to talk about implementing business logic as Java and deploying the Java to a reliable platform. They want to talk about their data model. They don't want to talk about the implementation of the database. Now all of a sudden the question is, well, do I need to care which chip is underneath? Do I need to care even about the traditional operating system? Am I going to be touching it all the time? And that's changed the economics. Okay, but you care, obviously the Oracle cares. Right tool for the job. And so to carry that through, I mean Linux and X86 obviously are very popular, dominant, dominant platforms. So where are customers using things like Supercluster and Spark based systems specifically? Well, so at first I want to say that I think X86 is correct. We're a big partner with Intel. Of course Intel. Intel says the Exa stands for Xeon and that's their little joke. Well, yeah, but look, so we think X86 is great. Great architecture, Intel's a great company, done great things, Linux is a great technology, we've invested a lot in Linux. Oh my gosh, we've invested a lot. We're pushing it as far as we can push it. But like I said, the truth of the matter is that we have the option now of focusing on using the right tool for the job. So we don't have to force fit what we're doing into any particular platform architecture. We get to bring in the platform components that make sense for that particular workload. And we're doing a lot of work in engineering our own chips too. I mean Supercluster is a good example of a hybrid, right? We have a fairly complex system that has a number of subsystems that we have all optimized for their particular bit of the work of the platform as a whole. And we're using the right technologies in there, right? If you look at something like the Supercluster, we have networking, we have general purpose storage, we have exadata storage technology, we have Spark processors to run the database, we have Solaris, we have Linux. All of these technologies are brought together in this system. They make the system better, the storage functions better, the compute runs better for database workloads, the storage grid runs as well as it can run. We've built the platform using the right tools for the job and all that complexity, we've absorbed that into our own R&D spending. So as a customer, you don't see it all. What you see is essentially a turnkey system that provides a really, really reliable, really secure, really scalable platform for your business applications or your cloud. Okay, so you mentioned cloud before. So is essentially this notion of, let's talk about it. This notion of engineered systems becoming sort of the de facto infrastructure for whether it's a private cloud or public clouds or running my SaaS based applications. Can you talk about that a little bit? Well, so if you think about the original premises that most businesses, even though they spend a pretty significant portion of revenue on technology, don't consider themselves to be technology companies. So they've acknowledged this and they've looked for opportunities to divest themselves of responsibilities that they don't need to have, that don't offer an opportunity for differentiation, some advantage in the marketplace. Now, we're seeing a lot of companies looking at public cloud because there are some advantages, agility that comes from having someone else own and operate the infrastructure and sell it to you by the drink. A lot of companies are looking for ways to control their infrastructure better than they would in a public cloud or even a third party hosted private cloud. And they're thinking that there's a bit of a compromise they can make. They can get away from the peace parts. They can still have the agility and flexibility of a cloud and in many cases they can save a lot of money doing it by buying essentially ready-made building blocks, converged infrastructure, engineered systems like Supercluster. They can buy those. They're really quick to deploy. We do all the hardcore engineering up front to improve performance reliability and all of that stuff. We single patch stacks and better support at no additional cost. And so that's the relationship, I think, between engineered systems and cloud. And we use engineered systems under your cloud too. Single patch stacks with four times. You're talking about the whole stack, right? The whole stack. The whole stack. That's unique, I think. Compute virtualization storage, driver's firmware, the whole thing, the whole system. And of course with X, up to the database and the middle tier and even the applications, absolutely. I mean, that's the joy of being a company like Oracle. We have a complete, I know people say that a lot, but it's really true. We have a very complete portfolio, right? All those core business applications, like ERPs, CRM and HCM for the utilities. We have the storage technology, you know, database and exadata technology. We have the hardware, the switching that. We really have it all. I love the comments. So I'm going to get some questions. So we have some veterans on CrowdChat, ex son veterans. And I talked to a lot of son folks who live in Palo Alto and the heritage of son was about engineering and performance. Obviously, this spark is dead to comment here. Spark is dead like the mainframe is dead. Come on, right? But what the comment here is many people don't understand Oracle's strategy making Spark number one, a five year road map twice should prove R&D is there. So comment on that passion that you have in this group of people who have that heritage from son and now integrates into Oracle. What is that road map? What kind of R&D? This is not a one trick pony kind of deal. What is the R&D? Summarize kind of like what happened? What's going on? How much R&D? Share the folks out there. Yeah, so I can't give you a number spending money. But I can tell you that we have the largest engineering team working on the Solaris operating system that we've had in almost a decade. We have hundreds of engineers working full time on Spark chip development and server development. The engineering team behind the Spark and Solaris part of our platform portfolio is huge. It's a huge organization full of some of the best people that we've ever had. Some of the best brains that worked on this technology back 10 years ago when it was just starting out, those people have been retained and some of them have come back to us. A lot of new college hires bringing fresh thinking about technologies like OpenStack. They've joined the team now. We're doing really good business with this technology. We're seeing super cluster growth take off. I mean, it's growing incredibly rapidly. Triple digit growth. Customers love this product. Because of performance? What is the aha moment for the customers? They just need to make the stuff go faster. Well, see, a lot of it is about performance. And the technology, say in the excavated, the reliability and the performance optimizations that come from that part of the portfolio are inherited. There are a number of really interesting capabilities of the Solaris operating system. And I think more importantly, there's an opportunity for us because we have control over the intellectual property behind the microprocessor. There's an opportunity for us to do things in the chip that will optimize it for certain workloads that are really highly, highly high value and it doesn't matter a lot to us. So I'm not going to. Talk about the OS, because one of the things, first of all, I love the old comparison of the iPhone for the data centers, Oracle, Larry's, Caprivatto there. You know, putting that aside, it is interesting, right? What the OS did for Mac, if you could look at the history of Apple, you know, they were struggling and they took the awesome OS and they bolted it in while Microsoft had the old virus craze and dominant market share. What the power of the OS, talk about the power of the OS in this new normal of, hey, integrated systems, I got large scale performance, scale up, scale out, lowering costs, what is the OS role? Well, you know, so what an operating system used to be isn't what an operating system is today. So when we think about operating systems, you know, we don't just think about, you know, whether it's Windows versus Solaris versus Linux or anything like that. We think more in terms of the infrastructure software that ties the system together and makes the database and makes them aware and makes the applications run. And so when you talk about something like a super cluster, you know, what's the operating system? Well, the truth is it's got, depending on how you look at it, four or five operating systems together. And they do work together. You could think of them as kind of a new hybrid operating system. And in the future, we're going to be getting away from the kind of view of the operating system as something that runs on a server and thinking about the operating system as something that runs the data center. And so again, technologies like OpenStack are really important to us because it offers an opportunity to kind of unify and expose the infrastructure in a way that matters to the application users and application developers but shields people from the complexity of the underlying implementation of the infrastructure. Focus on what matters to you, hide what doesn't. Okay, so what's the enablement opportunity? So in data center, I get that, it's certainly changing. But now the stuff happening outside the data certainly in cloud and big data, internet of things. Right now it's probably maximizing. It's on its way to the apex of the hype scale. I mean, but it's very relevant. I mean, it's just connected to the edge of the network. You know, which is not a new concept in computing edge of the network. So it's just now a lot of edges. How does that impact this system? Does it help it? Does it highlight the benefits? So there's one really important way that it matters. So I mean, you know, the efficiency of operation, you know, run and maintain simplicity and all that stuff performance, that's all really good. But what we've seen really rise to the surface are concerns about performance. You know, with the data center kind of getting turned inside out. That's what cloud sort of does, right? For many companies now suddenly the data center is a thing out there and all your branch offices talk to is not inside your walls anymore. There's a real focus on security. Like a lot of efficiency in implementing systems that way and architecting applications so that they can be used in a kind of cloudy way. But that exposes a lot of surfaces for exploit. And so when we look at technologies like Spark and Solaris that have the, you know, a long tradition of being really, really secure, really reliable, really hard to compromise enterprise class systems. We think that plays a big role in the success of Supercluster. We've seen a lot of customers implementing e-commerce, implementing their core ERP, their core HCM database as a service, Java platform as a service, right? These are the things that are driving Supercluster forward because it gives them all the operational benefits, the performance and so on and that security. You know, we have tools in the product for end-to-end compliance validation. We have Solaris which has something called Zones and the notion of an immutable zone. We have some great technology in the database itself to prevent, you know, SQL injection attacks. So we're really concentrating on security and we think that frankly everybody, who doesn't care about security? Yeah, I think, I mean, it's tried and true, you know, it's a standard computing paradigm. I want more speeds and fees, but the real trend that you highlighted there is this inside-out data center. This is absolutely the number one issue most people are dealing with right now is because of the transformation, which is the combination of the shift and the inflection point with mobile, social, native things, the highlights, that inside-out. It absolutely does, you know, and we talked at Open World, I don't know if you've caught it at Open World, we talked about something, the next generation spark chip, which we call M7. We've been working hard to extend the instruction set of that chip and I'm going to say this just to give people a sense of kind of where we're going with our own microprocessor technology. We've been taking core functions into the chip, things that are extraordinarily hard to do in software, pulling them into the chip and one area is security. We talked about something called application data integrity and the idea is that there's a lot of software out there, C++ C software written in languages that require the programmer to manage memory themselves. That's a notoriously common source of defects in software. Really, really hard to find, happens all the time. But it's also, I believe it's the third most common vector of attack on enterprise systems is what we call buffer overflow. Now, we're engineering into the chip a solution for buffer overflow that will effectively make it impossible for that vector of attack. It's simply, you won't be able to do it. Things like Heartbleed can't be done, it can't happen. Irrespective of how well written the software is, it simply can't happen. We think that those kinds of advances are what are really going to differentiate our whole portfolio in the long term. The perimeter is not secure, it's a whole new world inside up. Mike, thanks for coming on the queue. We are here live in Silicon Valley at the Oracle Headquarters Conference Center for the big announcement waiting to hear from Larry Ellison. But we're doing the pre-game. I'm John Furrier with Dave Vellante. This is theCUBE. We'll be right back after this short break.