 Live from San Francisco, it's theCUBE. Covering Oracle Open World 2016. Brought to you by Oracle. Now, here's your host, John Furrier and Peter Burris. Welcome back, everyone. We are here live in San Francisco for theCUBE. This is SiliconANGLE Media's flagship program where we go out to the events and extract the signal from the noise. I'm John Furrier, the co-CEO of SiliconANGLE Media with Peter Burris, head of research with SiliconANGLE Media. I'm also the general manager of Wikibon Research. Check out wikibon.com for all the latest research in cloud big data infrastructure. We're at Oracle Open World 2016. I'm excited to have our next guest, Don Johnson, VP of Engineering for product development, for the infrastructure as a service, for Oracle Cloud. Welcome to theCUBE. Thank you. Thanks for spending the time to come on. Really appreciate it. Obviously, Oracle's cloud last year was obviously the announcement they're marching to the cloud. A big building block in this was infrastructure as a service. They had the SaaS, they'd taken names, taken a button there and they're transforming. Platform as a service, developing nicely. This year showed some progress. But the upgrade, if you will, or reboot or reset, however you want to call it, was fundamentally introduced the new stuff with infrastructure as a service. Kind of round everything off. Give us the update. What's the key new news for infrastructure as a service and why is it important? Well, a couple things. Let me start with the last part of your question. Why is it important? So, very broadly, I would say there's kind of two strata of cloud. There's cloud platform and there's everything that's up above, apps, SaaS, et cetera. Cloud platform, I think, is a big category. It's broad spectrum, but it's IS and PAS. And then there's lots of stuff that falls inside of there. IS is a fundamental and foundational building block and all of the characteristics that everything up above relies on or requires is basically enabled by infrastructure. If you want to run in massive scale, if you want network connectivity between place A and place B, if you want intrinsic security, that's all things that are foundational characteristics and you either have them or you don't based on where their infrastructure IS gives them to you. And so, for us, for Oracle, we're a cloud platform company. This is a foundational piece and we're investing in this very aggressively and we're driving in a very innovative direction on this. So, you've been in Amazon since 2005, just recently joined Oracle on the engineering side and infrastructure right now we're seeing is a cost and performance game. Drive the cost down as low as possible while preserving scale and performance. Real critical and almost hardening the top, if you will, creating a hardened infrastructure so that you can enable DevOps and some coolness around agility, all that good stuff up on top of the cloud. So, what's the key things this year that you guys filled in in terms of product? I mean, what was the key innovations and on the development side, what was the key we're in for you guys? Well, so what we've been announcing in IS is really our next generation infrastructure, which is two-fold. It is the infrastructure itself and what our data centers and networks and virtual network looks like and then it's a new suite of products that we put on top of this, bare metal cloud services. And this is the fruit of a big kind of back to basics foundational exercise where we have gone and redesigned everything from the ground up and we've done it with a focus on a bunch of core, core criteria, core things that we wanted to capture and that we wanted to do better, better than have been done in the industry today. And I would characterize those as two-fold. First, we are bringing along all of the best characteristics of the cloud and why the cloud is compelling and what people use it for. Self-service, pay for what you use, it's elastic, it's easy to use, there's low friction, it's high-scale, et cetera. But there's a number of things that for our core customer base actually are very challenging in moving to the cloud. When I say our core customer base, if you have a large existing, if you're an enterprise and you have a large existing infrastructure and deployment typically on-premise, you have a lot of constraints and it's difficult to actually move into this new environment and take advantage of all that it has to offer. And there are, this applies to how your applications will run there, the assumptions that they make, your security and controls. And so we've identified a number of areas that we fundamentally wanted to do better than they've been done before. Security, reliability, governance, the ability to manage if you're a large complex organization, you have a large complex footprint and deployment in the cloud, the ability to manage it. Performance, performance is a broad spectrum. Key performance, raw performance, predictable performance and a particular price performance, you're talking about performance and cost. And sort of an adjunct to performance is the ability to harness modern technologies. As if you look at where storage is going, non-pollutile RAM and technologies like Intel Crosspoint, how can you actually enable customers to get access to that and use it and harness what it offers very, very quickly. And most of all, really flexibility, sort of the choice. And what I mean by that is when you're a cloud provider, you kind of pick a certain level at which you implement and define and build your abstractions and then that has consequences in what choices you actually offer. So let me be a little bit more precise about this. A core thing that we did, sort of the keys, the special sauce in any cloud platform is the virtual network. And we made a fundamental choice that the way in which we're gonna do the virtual network is to pull the virtualization into the network itself where we think it belongs. But no hypervisor. It's not in the hypervisor. And so what that means is first, it means we're able to, the requirement that we have of something that we can plug into our cloud, your cloud, your virtual network is, it has an ethernet port. This means that we can put anything into a virtualized network. Our whole infrastructure, the presentation of customers is everything runs in a virtual overlay. It's all virtual network. But we could put any class of resource in there. We could do bare metal. We could do an engineered system. We can honestly, we could take an arbitrary middle box from any third party vendor. This lets us give our customers bare metal. Giving our customers bare metal means we can take, so we provide bare metal compute with NVMe drives. They are phenomenal. There is nothing, like we're literally giving you a server in the cloud with a provision in minutes paid by the hour. And you get, in our biggest shape, you get an excess of 4,000,000 and 4K read IOPS. Like this is phenomenal power. So really there's nothing that stands between us, between a technology and us giving it to you. So that was the key design criteria then. That was the key design criteria. And so this, in terms of sort of flexibility and preserving choice, this means, principle you can bring any OS, you can bring any hypervisor. If you have some old stuff that's difficult to move, you can't bring up our hypervisor. So you let the performance, everyone kind of speak for themselves, if you will. So just a customer can put anything on this thing. Yeah. And these are phenomenally powerful boxes. Okay, so now how does that compare with Amazon as your, because the number one question I get is, and I'd love to see if you can put some color around this, is obviously Amazon had a different thing. You guys had a clean sheet of paper and took smaller steps, compute and storage and built services and scaled up there. Azure had kind of backed into it with their existing business, from their portals and all their services, and then now moving their customers on there. So number one question I get is, well, what's different with the IaaS on Oracle, VisaV, AWS, and Microsoft Azure? How do you answer that question? Is there a distinct difference? Is there a design philosophy? Is it a? Well, the design philosophy for IaaS is what I was just articulating in a sense. It looks and acts very, very similar from the perspective of the customer, the user experience at scale, as well as it preserves choice and flexibility and is amenable, basically. It is much more friendly to the large enterprise or large business that is outside of the, oftentimes and typically outside of the sweet spot of what an infrastructure like, say, Amazon was originally designed for. So as a principle, we are trying to meet our customers where they're at from they wanna migrate over some apps and do it cautiously and maybe not change too much about them and not see that as a constraint or an obstacle to get to all of the promise and power of running modern applications in high-scale, highly available. Look, in many respects, Cloud is naturally a network-centered compute model. Yes. By putting more, by not putting network virtualization above the network, by putting it into the network, does that also at some point in time give you greater flexibility, the option to bring even more of the core work that's done down into the network so that you can actually start liberating some of the power of a real network computing model. Others can't do that right now. So if you think about it, what kinds of applications might that make possible in the future? Thinking about IoT, for example, the ability to use a network model to describe how work gets allocated within a cloud of services. Well, I think the network ultimately, what you needed to do, there's a few things you needed to do. You needed to very reliably and quickly move bits from place A to place B. You needed to have the flexibility as a typology to be able to put things in. And you needed to preserve privacy and plugability. So the fundamental thing that I see our virtual network supporting and enabling is basically building up a fabric of services and letting us say, so everyone runs in a private overlay, we want to make it easy for any provider, ourselves as well as any third party provider to inject microservices into your private network. We want to make it easy to be able to bring over traditional security controls where you want to set up bastions and set up taps and be able to introspect, do traditional IDS, IPS. So I see network virtualization really as an enabler of providing a fabric that lets you give you great flexibility and wiring things together to answer your question. So final question for you, what's next? So what's the priorities on the to-do list for you guys as you go down to 2.5 or 2.1, as they say in Microsoft, they're making an odd and even number, making a 2.1 or 2.5 or 3.0, what's next? There's a ton of things. So we're building out data centers and new geographies, we're going big, we're going to add a ton of SKUs, we're going to make bigger things, smaller things, a ton more features really all across the board. So I don't know that I see it as there's a 2.5, there's going to be a rapid pace. So more slew of announcements, very similar to Cade's we've been seeing at Oracle's and Amazon traditionally has started that trend, Larry couldn't even finish the keynote on Sunday because the announcement stream was so large. No, we have a, you'll see a constant string of releases on a weekly, monthly, quarterly basis. There's just a ton of stuff coming. We have a ton of features to add, we have a ton of interesting new services to add. The pace is fast, you're running hard. The pace is very fast. Well, congratulations and looking forward to following you guys and your success. Love the agile mindset, love to see that cadence of shipping stuff moving really, really fast. Appreciate it, spend the time. Thank you very much. See you on your insights. It's theCUBE live here at Oracle Open World. You're watching theCUBE back with more live coverage here in San Francisco after this short break.