 Live from Las Vegas, Nevada, it's theCUBE at IBM Edge 2014. Brought to you by IBM. Welcome back to Las Vegas, everybody. This is Dave Vellante with Jeff Frick. Andrea Nelson is here. She's the director of storage for Intel. Andrea, welcome to theCUBE. Thank you, glad to be here. So here at Edge, what's going on at Edge for you guys? Well, we have two very exciting product announcements in conjunction with IBM. The first being their real-time compression appliance with the Intel 8926 communications chipset. And we also have a new store-wise product family being announced with the Intel E5-2600B2 processors, including. Great, so good, congratulations. So we were talking off-camera about all this need for data reduction. So you're seeing it, we'll talk about flash, some more as a big disruptor, but you're certainly starting in the early days some of the deduplication appliances, obviously use some of your products, but they're very CPU-intensive operations, a lot of math-heavy stuff, whether it's deduplication or compression, particularly with the case of IBM, it's in line. So there's got to be no performance degradation. So talk about that a little bit, the progression of that technology, IBM specifically or throughout the industry. What do you guys see? Yeah, I mean, we are definitely seeing the need for more IO, more networking, more cores, more cache capability to support executing these functions in line. The latest version of our processors, the E5-2600B2 added more compute cores, more cache. And so along with that, we have the Intel Storage Acceleration Libraries, which are foundational algorithms that our customers and developers can integrate into their platforms that optimize functions such as hashing for dedu or AES and I and any of the storage security features. So a lot of people might not know about Intel's storage business. I mean, obviously you sell to the storage companies when they need compute, they go to you. Maybe talk about the business a little bit. Talk about its charter, its formation, its history. Sure, sure. So the storage business at Intel has been around for quite a while and basically this group has been working on influencing our standard server CPUs to include features and functionality that provide data security and the performance for storage workloads that our customers and partners need. We at the end of 2011 announced that we believed about 80% of new storage systems were shipping based on Intel architecture converting over from risk. So you can see that that drive for greater compute is really what was behind that foundation. Also the desire to use standard X86 platforms and benefit from that ecosystem of providers, both from a hardware and a software perspective that are available. So let's just talk about that benefit a little bit because they say there's still 20% out there, 20% hold up, so it's probably a healthy thing, right? You don't want to have 100% of that more going on. People start looking at that under the coverage, you don't want to do that. But I've seen, I've observed other organizations, whether it be IBM or guys like EMC or HP or whomever, sort of converting from risk. And it's not trivial necessarily. No, it's not. You got a big effort, we talked to Pat Gelsinger about this, he's a legend, he's been on the queue many, many times, but then the result, once you transition over, is the massive ecosystem. And software development capabilities that exist. A lot of people don't may not know about those. I wonder if we could talk about those a little bit and what role Intel plays in terms of catalyzing that. As the steward of the ecosystem. So Intel plays two roles. One, we work with our O&M customers who have proprietary software to optimize around IA. But even more importantly, we work broadly with the storage ecosystem. So BSD, major operating system, Linux. Also, we work with the customers like Microsoft to help them optimize their software around IA. Especially in storage, you're seeing a lot of new startups coming into fruition who are utilizing BSD and Linux. And so in those boxes, you'll see key storage features like in our Xeon platforms, we have asynchronous dim refresh, which protects, it's a hardware pin that connects and protects the system in case of a power failure. Those drivers are included in the Linux and BSD offerings, which makes it easier for our customers to bring their products to market. How about Flash, what are you guys seeing in Flash? You guys play in the market, you support the market? Yes, we play in the market. We have an offering of both NVMe, right? The highest performance Flash interface available and SATA SSDs and both two and a half inch foreign factors, as well as PCIe adapter cards. We have data center products that are optimized for the data center that provide consistency of in performance, they provide data protection through the different right phases. And those products have been extremely popular. And we see all of our customers announcing more and more of Flash arrays because of the value prop that they offer in terms of performance and reducing latency. What if you could talk about sort of what you're seeing, Andrea, Andrea in the data center, trends that you're seeing and how does Intel respond? You sort of, are you more tactical, customers come to you and say, hey, I need this, I want this, I want that, can you deliver? How much of a sort of horizon do you put on where you're doing big time planning with the companies, your partners and you're going to market with broader strategies, how does it work? Yeah, I think so, from our perspective, especially in the storage space, we're looking out a lot further than we used to. We look at the, I think IDC is saying 44 zettabytes now, recently released update in 2020 thinking what does the infrastructure in that timeframe need to look like to manage and store the subset of that 44 zettabytes of data that's going to be available. So we have been spending a lot of time looking at software defined infrastructure generally and then our team on software defined storage and working with OpenStack, right, on optimizations. I think the biggest thing that we see is a need for OpenStack and Seth for block storage. They don't have the enterprise capabilities to be picked up and adopted broadly today. So there's a lot of effort and work that's going on. So what's your role in OpenStack? Are you sort of watching it? Are you guys directly contributing? Are you writing code? We're contributing. We are contributing, yeah. So we're going to be contributing an erasure code technology and Swift, yeah. So for the distributed object? Distributed object store, yes. All right, so that's, what role? I mean, you guys contributing the code, you're helping write the code. We're writing it, we're writing it and contributing it to the open source community. And then we'll also work with our customers to help them with their implementations and optimization. Okay, can you share with us kind of, so how's that work? Are those folks in the storage division that are doing that or is that broader Intel? This is, well, Intel has our software solutions group, which maintains most of our developers. But yes, they work with the open source community just like Intel's done. This isn't a new game for Intel, right? I mean, Intel's been one of the top contributors to Linux for many, many years. And so we're setting up and we'll be driving a similar effort to OpenStack. We're also contributing on OpenFlow and some of the other networking protocols as well. Now, what do you make of all this software-defined meme that's going on? Everybody's going with software-defined crazy and software is eating the world and software-defined is eating storage and everything else, so. Well, I view it as absolutely necessary, right? You have to break the lock between hardware and software to truly unleash and enable the flexibility in a data center that's required. I mean, we look at, we have an example internally at Intel where we're working with our own IT group to set up a sync and share file application for employees similar to like a Dropbox. And scoping that effort, we thought we'd need about two petabytes of storage. Well, Intel has about 20 petabytes in the ecosystem today, but we can't reclaim any of it, right? For this new application. So when you look at a software-defined future, providing that break and ability for an application to come in, take control of a pool of resources, utilize it, turn it back over so another application can come in and grab it, it's absolutely vital to supporting what we need in the future. Well, so 10% to create your own internal Dropboxes of the existing storage that you've built up over all these years. So I always said. Sorry? You said it would take two. We're thinking about two petabytes over the lifetime of the application, right? You said you've got about 20 now. Not on that particular application, but within storage, yeah, overall. Okay, but not all of Intel. Okay, I was going to say, that would seem like a awfully large percent. Andrew, you're saying you see that software-defined as a positive, right? There's a flip side in there. So there's the elephant in the room is a lot of people make a lot of money by not having things software-defined, by having all this function sort of buried inside the box. That's true. Are you nervous about that? Are your partners nervous about that? You think it's a good thing because if prices drop, people will buy more? What are your thoughts? I think that it's the latter, more of what you said, but absolutely, I think some of our partners are nervous. But if you go out and you talk to many of our major customers, they all have their software-defined stories. So from an Intel perspective, our focus is really looking at what are the features and functionality that's needed in the underlying billing blocks and then how can we work with our partners to optimize their solution, right? When you think about things like hyper-scale and massive scale-out, mega data centers, you guys, you got your fingers in every pie, right? So you have good visibility on what's going on in the industry and you were seeing that a lot of the trends in hyper-scale bleed into the enterprise. How real is that? How fast is that? What can we learn from what's going on in the big mega data centers and how can IT practitioners in the data center borrow from that, learn from that, benefit from that? Yeah, I think what's happening, hyper-scale from a public cloud service provider perspective is very real, right? I mean, they have unique problems that they were trying to solve and they found that a lot of the existing storage infrastructure was optimized for something else, right? It's not wrong, it's just, you know, was built to fix a different problem. And so what they were able to do very successfully, I think, was take reliability to a different level and utilize standard high-volume servers and some of the open source software that's available to create very innovative storage solutions. Now, they were building something brand new. They didn't have the enterprise IT legacy applications that they had to deal with, right? They didn't have the baggage. Yeah, you know, they didn't have, so I think that some of those best-owned practices being developed in the hyper-scale cloud space will migrate their way into the enterprise data center, but it's going to be a journey, not an immediate, immediate flip. What should we be watching for your business, for Intel storage business, near-term, mid-term, even long-term? What are some of the objectives that you're setting for the business? What should observers be looking for as indicators of success? So I think from a storage perspective, we're still going after the remaining, hold out of TPU, market segmentary, but we're really focusing in on software-defined infrastructure, software-defined storage, looking at the orchestration layer and the software-defined storage controller and making sure that those are optimized and ready for enterprise consumption. Great. All right, Andrea, well, thanks very much for coming on theCUBE. Appreciate your time and keep it right there, everybody. Jeff Frick and I will be back with our next guest. We're live at IBM Edge. This is day one. This is theCUBE and we'll be right back.