 Okay, we're back live here in Orlando for IBM Edge 2012. I'm John Furrier, the founder of SiliconANGLE.com. We're out scouring the hallways, the sessions. I'm joined by my co-host, John MacArthur, co-hosting with today. John, we have Keri Johns-Vano. We're from Intel Storage, who recently has a new role at Intel, but you've been on theCUBE many times. We're still in storage. Exciting to see you again. Yeah, you too. So what's new? Tell us about what's going on, your new role, and what's happened with Intel Storage. Well, storage is growing, as you know, right? And so our focus is always changing, it seems, right? We have really focused on scale-out storage. So we're really helping a lot of our partners move to scale-out storage based on the unstructured data growth. But I think one of the things that we're really embracing now is the future of storage and where it's going, based on big data, based on massive data, fast data, and how storage is going to evolve. So we're really looking at enabling the ecosystem to create ubiquitous storage or pervasive storage where storage is going to become a service on standard high-volume servers, right? Well, we had theCUBE at Intel NAB. You guys hosted us there. We were able to do theCUBE there. One of the exciting things at NAB was Thunderbolt. Moving data around really, really fast. So data is no brainer. It's growing like crazy. Machines have thrown off data. Users are throwing off data. Applications are throwing off data. Massive tsunami of data. New types of data. Dealing with unstructured data. So what is the definition of future of storage for the new guys at Intel? So that's the vision. So for us, the vision for storage in the future is that it becomes ubiquitous, right? It becomes pervasive. It becomes the data centers transform to a point where you basically have pools of resources of compute storage and networking that you can dynamically allocate and secure as you move across public and private cloud across service providers and have multi-tenancy type of tiering, right? So all of those things become important and they're very dynamic, right? So I'm interested. So where does Intel, where's the line between Intel and Intel's partners sort of lie? What do you mean, the line? What's Intel's contributions? What's Intel's primary thinking about where are you going to contribute? Because you've had a number of sort of initiatives in the storage space where you've developed solutions for the white box market. You've developed a lot of things. Yeah, I think, you know, I mean, our key role is really to provide the fundamental building block that creates that foundation of intelligence for storage. And there's an awful lot of storage that runs on Intel today. Yeah. I don't know what the percentages are. Is anybody doing numbers on that? Yeah, it's 80% market segment share. We currently have Xeon in two for, you know, shipping storage systems that are shipping today. It's about 80% market segment share. Okay. So, you know, I mean, our focus is really on being that foundation to provide the intelligence in the right. And you know, you mentioned about IO and memory and stuff. And that stuff is really important to storage. And as we saw in the keynote this morning, right? Memory's becoming more and more important as it plays a key role, especially in data analytics, right? Where you've got things real-time, you've got to do real-time compression, you've got to do real-time tweets. There's millions and millions of users accessing at the same time, right? And so, IO, memory, that stuff is really important. So, we're really looking at how do we take our multi-core architecture? How do we optimize that so that you can get quick access to the cache? You can get quick access and high bandwidth through IO and quick access to the memory. That's been one of the themes, obviously, here at IBM. It's been the whole notion that, with all the compute, with all the cores, that's cool. Now, it takes the storage, it takes the storage, enable the storage to be faster and have more diverse storage types. How do you guys look at that? And have more efficiencies, yeah. Well, we've, our E5 processor that we just released, that is the first standard server processor that has integrated storage features in there. And we continue to increase the IO and the memory. And I mean, we have 2X and 3X improvements in IO and memory today in that E5 processor. But those advanced storage features were only found in SOCs prior to the E5 generation. So, you know, our goal is to continue the integration, continue the convergence, so that servers and storage, the line between servers and storage becomes closer, right? That gap is not as far as it is today with some of your traditional purpose-built storage, right? So, you guys like the whole Flash movement? You got to be excited by what's going on with Flash. Yeah, yeah. Well, if you look at the, if you compare, you know, hard drive, you know, performance and how it's evolved over the years, right? And then you compare Moore's law and then you look at how Flash is, is really kind of bridging that gap. It's, you know, it's one of those technologies. It's not going to be, you know, the do-all end-all for everything and the storage workloads are different, right? Well, what do you say to customers that want to know what's going on in storage? Because they're bombarded with a lot of, you know, market data, all this stuff's going on. They have to re-architect their environments. We've talked extensively yesterday and today around how Flash has changed its server, you got Flash everywhere, it's kind of a theme here. When you talk to customers and they say, Kevin, just tell me the bottom line, what's going on in storage? From our customer base, I mean, take Intel IT as an example, right? Diane talked about that in her keynote this morning. But, you know, the customers that are, the IT customers out there, the end-users out there that are looking for storage efficiencies, they're really looking across their data center and how do I transform my IT to optimize my workloads? And usually that's not just one solution, usually it's a combination of various solutions, including Flash, including strategically placing SSDs, right? Including how many tiers of storage do I need so that I can optimize that storage real estate, right? And so it's, you know, that's the exciting thing to me, I think, but also the complexity about it is that it's not a one-size-fits-all, right? We don't have these standard, like in the server world, right, where you've got this big ecosystem, things are highly virtualized. But I think as storage continues to get converged with server and closer to server, we could see that in the future, right? If storage becomes a service, it could be a lot more standardized. What do you think about the IBM show here this week? It's great, it's really been good. I mean, we are a diamond sponsor and we've had a lot of great collaboration with them. I think the format's been awesome. The keynotes have been great and, of course, they've... This message of portfolio is interesting, right? They're bringing in more solutions to bear. Yeah, yeah, yeah, I think the messaging was really, was really good, right? I think that they're trying to really converge or rally around a few key things, like agile cloud, et cetera, they're integrated solutions. And then there was the third pillar, I can't remember, but it's good to see that cohesiveness across all of their sessions and across the whole show. I thought that that was... Intel's customers obviously have people like IBM buying stuff from you guys. What are their number one concerns on the roadmap when they buy future products? What is the direction for you guys? The direction? Well, the direction for us from a roadmap perspective is really how do we continue to help the solution providers break the bottlenecks that they face in IT, right? For storage, a lot of it's around IO, memory. How do we optimize, when we add multiple cords, how do we optimize the cash and the capability of everybody accessing that simultaneously, right? So we're looking, and we work with a lot of our partners three, five years in advance. We have these two days architectural summits where we lock them all in a room with our partners one at a time, right? And we get their direct input to what they'd like to see in storage systems. I guess if the story is flash everywhere that IBM's promoting, your message is processors everywhere, right, cores? Because you could put cores now on everything. Absolutely, yeah, yeah. The compute power, compute is more and more important as the storage workloads become more complex. Why was that changing with some of the backup and recovery paradigms? Are you seeing any trends you can share with folks around adding more cores and more flash and more storage technology? How's it changing the backup and recovery? Absolutely, people are offloading things to, I mean, Tate's not going to go away, but they can add automated storage tiering, they can offload things to other types of media depending on the usage of the data, right? Is it really archived or is it, and the recovery time is a lot quicker. So server has been on a fairly fast refresh cycle. Storage tends to have a longer refresh cycle. Yes. Right? Yes. But the process of refresh cycle is pretty consistent, right? It's correct. Yeah, yeah. How does that impact sort of? Well, I think that's a result of the evolution of storage, right? And how storage evolved from a very purpose-built, vertically-scaled system to now scale out storage, which is, according to Atlas, it's going to be about 80% of the market segment share by 2015. So it's evolving, right? And as we look at the future, where storage becomes an application or a service, then you get closer to these refresh cycles. You think we'll get on a similar refresh cycle on storage as we've been on? I think if we can realize the future vision. Yeah. Yeah. As long as we don't have to move the data on the back end too much. We're getting closer. That's the problem. You can move a workload faster than you can move the data. Right, right. And there's improvements. You've seen at IBM Edge here today that they're implementing capabilities, storage capabilities that approve that, moving the data. Because anytime you retire or something, you have to move the data and it's... So I think one of the things that's sort of interesting with this sand volume controller technology that they're implementing in some of their solutions, that's been out there for a while, but it is the ability to maybe add new levels of services to without having to necessarily move the data right away. Yeah, yeah. And then when you add the solid state disk up at that layer, then you can improve the performance of the back end. Right, yeah. If you look at today's, the latest greatest today's storage systems out there, there's significant improvements that they've made. They're making them fast. Things are moving fast in storage. So, and many of our vendors are really striving to be more time to market. They see the benefits. It's just a matter of things moving fast enough and then moving fast enough to get there, right? And it's also going to take a little bit, storage systems have mission critical requirements, have certification requirements. So they've got to get a design cycle that works with that as well. There are a lot of Asian suppliers that are really going after the high volume market and are really crushing it on the high volume, very cheap, very stupid storage, you know, just pop it out. Yeah, we're working with, well, China's got the smart city, right? Right. Going on and we're working with many of the vendors over there to build out these data centers by due and other folks, UIT to build out these big data centers that are pretty, kind of like a, you know, cloud model, scalable like Amazon type of model. Yeah, in those cases that you've got a lot of, maybe not too bright storage, but then above that you've got a level of sort of data management capabilities and intelligence. It also probably runs on Intel, but it's not part of the storage controller. Exactly, right, right. They're really kind of a leap ahead, right? In that standardization vision that I talked about, right? And that's their vision. They want to be able to, you know, provide this capability to, with huge amount of capacity and performance, right? But they want that management layer over it so that it's simple, right? Because they want to be able to refresh on the latest, greatest hardware. And they're using mostly standard hardware today. Right. With a bunch of disks, right? Yeah, with very tight budgets. Yeah. Right. Lots and lots of capacity, but really tight budget. Right, right. So that's kind of standard architecture. And I think, you know, I mean, if we look at IT across the board, I think that, you know, some of the IT transformation analysis that people are doing are looking at that model as well, right? When you look internationally, because I mean, you're, as a global company, you get exposure to everybody. Now, you know, here in the US, sometimes we think a little too brokenly about, you know, the big systems companies that are lodged here. They may sell globally, but where do you see some of the, you've got, you've mentioned China. Are there other areas where you're seeing sort of advanced innovation that is going to try to change? Yeah, yeah. Actually, another emerging geo is Brazil. Yeah. You know, Brazil is growing like crazy. And they're looking- As both a consumer and a developer of technology. Yeah, absolutely. Can you talk a little bit more about that? I'm interested in. Well, they're, you know, they're looking at, I wouldn't say funding, they're looking at bringing companies into Brazil to do research on technologies in conjunction with some of their local partners. Right. Okay, so there's some initiative steps that are going on with respect to that. And that's really meant to, you know, for innovation from their local companies. Right, and some of the local companies invest in creating their own schools and to educate the workforce. Right. So we'll see how it all plays out, right? But, you know, that's definitely a growing geo. Brazil, I was just down there a month or so ago and it's amazing how fast they're growing. And they're- Yeah, prepping for the Olympics will be a big deal already. For the Olympics and the World Cup and that was some of their, you know- Now talk about your new role, that's why your new role. My new role? Yeah. In storage? Yeah. That's the same role. You had to travel to more, you said you were- Well, it's, we've been growing so much in the storage group at Intel that we've basically taken the marketing role and we've split it up into marketing alliances and marketing products. So enabling the vendors on one side versus creating marketing alliances with not only our partners but with emerging markets and so that's kind of an expansion. Okay, so it's not changing, it's more expansion. Yeah, it's more of an expansion because we've grown so fast. So we've been hearing a lot of conversations around tiering, obviously. Even tiering is a great concept. You guys talk a lot about tiering. What do people need to know about the tierings, the different tiers of performance improvements? Well, tiering really gives you, you know, tiering really gives you that capability to optimize where your data resides and how it gets processed. So if you have data at rest or data at motion, there's two tiers there, right? If you have data that you're working on, like for instance, when we do tape outs, well, there's a lot of parallel processing, right? And so, you know, you could treat that as a different tier. So the tiering that, like we implemented it Intel IT, we had one or two tiers and now we have five. And it's based on business process needs that we have and how we can optimize all of our resources across the board depending on what the data's doing. And I think when you talk about big data analytics, that's where it gets really exciting because the way we do things today versus having big data being in fast data and massive data being prevalent in our data centers is totally different in how we're gonna treat that data, right? The value of the data and the processing of the data and where it resides is gonna become even more important because you're not gonna need it all. You might need bursts of like DNA sequencing, right? Burst of huge data, but at the end of the day, you only need a little piece of that data to identify you from me, from him, right? So it's, you know- Keep pairs all the way until they talk about big data. Yeah, like Mayo Clinic and ASU, we were talking to them a bit ago about their DNA sequencing and genomics and what they're doing with that. And you know, at the end of the day, they're like, you know what would be really cool is if we could have like a pay-as-you-go model in storage so that when we need those bursts and we need that big analytics, we can utilize it, know that it's safe. You know, they're highly regulated environment, right? But not have to hold that overhead all the time because we only need to store this little piece of data here. So, you know, people are thinking about things totally different. What are the top three trends that you're seeing in storage? Because obviously, IBM's recognized with this event that stores at the center of the conversation, they're spanning around it. But you guys are Intel, you guys are super geeks building out the cores and all the performance moors, locks, et cetera. What are the trends, the top three trends from Intel's seed? From a technology perspective, I think the top three are really, you know, continued focus on compute. Obviously, I mean, the core is and the core optimizations are important. But because of the way things are changing, I think I owe and memory are the other two, you know, key technology things that we're really focusing on and how do we, you know, innovate and, you know, get a leap ahead in those spaces because there's a lot of new. Yeah, because you can enable a lot of stuff from the data, app data, for example, the big data tsunami. Right, right. So we've got some things in research there that can't really disclose, but. Come on, tell us. But faster, smaller, less expensive, typical Moore's law. Yeah, we can say that. You know, it's Moore's law, you got to try to stay. So Intel, too, sales is going faster in storage or servers. You can run a Kenya, so that either. You let it go. Okay, we've got. You can read in that in the animal. We're getting the break from Mark and the team here. So, Cary, John's van, I know you got to catch a plane heading out again to be discovered. We'll see you there. Yeah. We've got the cube there as well. Great. So on our summer tour. Yeah. We're going to be at the desk, the orange forum after. Also the cloud forum with Intel. We're going to do some cube action there, kind of a mini cube, kind of hallway cube. And then Hadoop Summit. Oh yeah, great, great. So next week, another busy week. Yeah, we're working on optimizing IA on Hadoop. So that's, so that it works best on IA. Can you share any details or just do work on it? That's all you can share. That's what we're optimizing it for. Hadoop is great. I mean, Hadoop is getting better and faster and stronger. The availability is coming out, which seems to be a big analysis of cloud era. Hortonworks has got some big analysis coming out as well. So that ecosystem, that community is growing very, very fast. It's good to see what the industry's doing with that. They're going to go from batch to real time very fast. So very disruptive price points. The economics will be pretty interesting to watch. All of course, work from filers all the way to dinner. Absolutely. Hadoop warehousing. Okay, Kerry Johns-Vano from Intel here inside the cube. SiliconANGLE.tv, John MacArthur and John Furrier here. We'll be right back after this break.