 Hi everybody, Jeff Frick here with theCUBE. We're having a CUBE conversation in the Palo Alto studio. It's something that we do when we get a little break in the show schedule, we can take a minute, catch our breath and still sit down with the tech leaders that you want to hear from. But now we can do it in the studio outside the context of the hustle and bustle of a show. And really excited to have a true industry veteran. He's been around for probably longer than he wants me to say on air. So I'll let him say how long. But who Yoshida, the CTO of Hitachi Data Systems, welcome. Thank you, Jeff. Great to see you. Pleasure to be here. Well, doing a little research for this interview, you've been around for a while, you've done a number of interviews and the thing that struck me was maybe the last big trend that you were so excited about, server virtualization and what a phenomenal difference that made in the marketplace as well as your business. Are we going through another one of those now? Yes, well we're going through this digital transformation and I guess IDC is the one that started that term and it's based upon the social, mobile, analytics and cloud, or SMAC as they call it. And that has brought some new technologies and be able to create some new innovations in terms of how businesses can transform themselves. Right. Hitachi Data Systems, you guys are way down in the bowels of these big systems. You guys are powering a lot of the storage and you came from the mainframe business. So how is it affecting your business? How are you seeing real concrete changes and what your customers are asking you for and how do you see their business changing? Yes, well we started as mainframes and then we transitioned to storage when the mainframe business sort of declined but we were more than storage. We have, now we have an x86 server platform, a blade server that enables us to provide a converged solution along with our networking partners like Brocade. And these converged solutions are kind of the basis for private clouds because it eliminates all the need for infrastructure connectivity and things like that. So you can roll in one of these things, plug in the power, plug in the network and actually pick an application from a table, a menu of tables, templates and be off and running. So it makes it very easy to move into this new phase of digital transformation. Yeah, cause it's funny because on the infrastructure side, you know it's kind of production line 101 as soon as you take care of one piece in the production line then you move to your next point of failure and move to your next point of failure between compute and storage and networking. Everyone seems to see that kind of networking was kind of the slowest leg of the three and kind of coming up to the modern architecture but now with this type of announcement they're really bringing their game up quite a bit. Right, yeah. Gen six is really going to open up a lot of bandwidth and IOPS for us and move a lot of the... Actually, it's the peaks that we worry about, right? We have to over configure for the peaks but they've got this 32 gigabits per second. Yeah, the old mom bill problem, right? Everybody calls mom on Mother's Day and AT&T used to have to build the whole network out from Mother's Day but Mother's Day only comes once a year. Yeah, the other huge trend that you've talked about extensively which is another driver behind this is really software defined and how software defined is spreading throughout many parts of the infrastructure and adding a whole new layer of flexibility, expandability, elasticity to what customers can do with their infrastructure. Right, yeah, software defined is key to this transformation that we're talking about and to us, software defined, you know, many times people consider software defined as a way of commoditizing the hardware and to us it's much different than that. It's really the communications between hardware and the application layer. Good example is VEVOL from VMware where we can publish our unique capabilities up through the VASA interface, API and vSphere can see our capabilities and they find a virtual volume around their capabilities and on our part, we can see into VMware and know that we're talking or configuring for a virtual machine not just presenting up LUNs and blocks but we can actually recognize that this virtual machine has higher priority than others and we can allocate to the right resources. So it's a communication process and a synergy between applications and hardware infrastructure. And then what this has enabled which you've talked about numerous times too is the ability for an individual to manage a whole lot more in terms of infrastructure storage, et cetera. So now as the kind of amount of stuff that I'm responsible for goes up, the management and the management tools and the ability to manage this bigger, more complex things becomes much more significant. Oh yeah, much simpler. The old view of infrastructure or the data center, it was sort of like a triangle with the base of it being the infrastructure costs and the operations and all that. The top of it was the smaller part was what we focused on the applications and analytics. What we have to do now is turn that triangle upside down. So we focus less on the infrastructure. Software Define helps us do that. Cloud helps us to do that and automate that so that we can spend most of our effort on the application, the end user and analytics. Right, and we hear that time and time again especially with the DevOps ethos and what Amazon has done with Swipe Your Card infrastructure that it's really the application that drives everything. And there's an expectation in the developer world that now with containers that the application or the infrastructure should just respond to what I need from the application as opposed to limiting my application development based on what I think or I got away from the spin-up of the server or whatever, that's completely flip-flopped as you said. Yeah, I mean, you make a good point. I mean, it's very disruptive. I mean, not just on the infrastructure side but it's also in the development side as you talked about. So DevOps and Agile and Scrum and those things are very important. So instead of the waterfall approach we took to development, that's too slow. We've got to go be faster. And using these, technology is the one thing but how we use that technology and innovation we put into that is what really makes a difference. And you've been in the game, like we said, you've been in the game for a while and you've mentioned in a number of your interviews that these little guys have driven kind of this last big way of innovation but there's a new one coming on. We hear about it all the time. It's IoT, Internet of Things. Now as sensors get cheap and actually a benefit of these is now all the sensors that are in them are less expensive and much more pervasive. So now we can put them on dogs. You can put them on shipping boxes from Amazon. You can put them on all kinds of things. So from your point of view as you start to see IoT build and the momentum build and it's a lot of hype probably right now but it's coming, right? And big companies like GE are behind it and a lot of players are behind it. What does that make you think? How excited are you about IoT? Are there some specific challenges you're looking forward to taking down or do you see it as just kind of the next big step function of kind of demand for the big three of compute networking and storage? Yeah, it's another integration process between the information technology we have grew up with the data centers and the operational technology that comes from those sensors. How do we bring those things together? We have to be able to bridge that. One of the ways we can do that is with several things we have to bridge. We have to bridge the infrastructure and then that's software defined. We have to bridge the data and so we have to move more toward object stores with more enriched metadata and we've got to bridge the information. So the data that comes from IoT is different from your structured data center but you need to bring together that Oracle or SAP data together with this sensor data that comes in and integrate that together. So we acquired a company last year called Pentaho that does that, it allows me to integrate all these things and the way we have all these connectors to all these disparate types of databases is that it's open source. So open source contributes a lot of this, we just harden it and provide a subscription maintenance for that. So open source is another key driver for Enabler for this transformation. Yeah, because you even talked about the transformation that Hitachi going from proprietary ASICs, proprietary software to more open source and Intel chips and again, kind of leveraging best of breed at scale and bringing that type of capability into your core. Right, the other thing is the Intel road map, I mean that is amazing how they went to all these cores and everything and so that has enabled us to do away with a lot of the ASICs we used to have to make. We do have some ASICs and FPGAs for special purpose but primarily it's standard Intel memory and cores and what that enables us to do is to have a software hypervisor for storage. In other words, all our mid-range, do you remember how we used to have separate mid-range and enterprise storage. Now that's all running all with one hypervisor, storage hypervisor. It's interesting, I think it was at HP, maybe we were talked about this IoT, the concept of kind of IT versus OT and congratulations on the Pentaho acquisition and we're at Pentaho World's a great event, great show, a lot of traction, but the OT, the operational technology that runs shop floors that people at GE are working, that's been cranking along all the time and yeah, the IT is kind of two separate worlds and this IoT really is bringing those two worlds together and the connectivity together of the devices and the sensors in the shop floor versus the IT systems. And what's fortunate for us as Hitachi Data Systems is our parent company has been in the IoT, well, the operational world, they build nuclear reactors, they've changed locomotives and all the infrastructure types of things and so we're able to bring that expertise together with our expertise in the information systems and create this IoT solutions. It's a great spot. Right, we're in a great spot. So a little more specific about the announcement today. You're partnering with Brocade on this Gen 6. What does it mean to you for Hitachi Data Systems? What does it mean for your customers? Oh, well, it enables us, we're going to all flash. I mean, I think we've already passed the tipping point for all flash. With our 6.4 terabyte flash drives, we're actually cheaper than lower cost, total cost of ownership than hard drives. And so the cost is not a factor anymore and then all the surveys, Gartner just did a survey, he said that the users of flash reported savings, not only in power cooling maintenance and performance, normal things, but also things like licensing costs because they don't have to license as many cores or instances of databases because of performance of flash. So what this Gen 6 does is it just opens up the highway or the lanes as Jack was talking about for us to be able to drive more workload through there and possibly even reduce the footprint even further by making better utilization of what we have and not have as many cores and instances of applications. And as you were talking about a little bit online, it's beyond just flash or the all flash array but really now looking down the road at potentially the all flash data center and the impacts that that is going to have as these data centers keep getting bigger and bigger, the demands, the loads are going up and up, power continues to be an issue but this is a complete game changer in terms of an all flash data center. You know, all flash arrays were the hot thing, right? The investors or just VCs are going crazy about those things, investing a lot of money into them. But you know, those all flash arrays are really appliances. If you want an all flash data center, you still have to worry about all the enterprise things around availability, you know, replication, a disaster recovery, security features, shredding, encryption and all that. Those things come with an enterprise array. So if you're talking about all flash data center, it's more than just an all flash array. You've got to expand that requirements to include all the enterprise requirements we traditionally had. Right. And that's why Gen Brocade is so, the Gen 6 is so important to this. Right, right. Because not only does it give us a performance but it also has some additional availability features like they have forward error correction for in-stream types of error corrections. It has FCSP, they do CHAP, you know, like a challenge handshake authentication protocol that we have with Ethernet, they do that with fiber channel. And so we had those additional capabilities in the fiber channel switches now, Gen 6. Really, really just in sync with software to find everything, right? So it's not just beads, now you have management, you have software capabilities, you have all kinds of things that you can now add in. And as you said, what's the point of hooking up a really fast drive to an old legacy connection system that really wasn't built for the performance that you get out of that. And the IO insight, which is key to seeing that whole network and seeing what's there. So before I let you go, running out of time, just kind of get your perspective as to where we are today in kind of the IT industry with these massive shifts in terms of, you know, cloud and big data now being an asset and non-liability and flash, even the all flash data center and mobile and around the corner, IoT is you kind of sit back, you know, on a Friday night, maybe with a glass of wine and think, wow, this is just crazy. For all the innovation you've lived through and seen, how do you rank where we are today? And what do you think about when you look out over the dash? Yeah, I don't know, you know, I've been in this business a long time, but every year it just seems to be getting, you know, more and more, the world is just expanding. You know, we see it, you know, so much data being created that we know we can't store all that data. So part of the things that we'll have to struggle with is how what do we save and what we don't save and what can we recreate just from metadata. So metadata in an object store has become more important. But you know, today we're in this transition, we have to have sort of take a bimodal approach. We still have our core systems that we need to take care of and nurture and grow and scale, but we also need to then move into the new, the new innovations, the things that are, that are not as atomic in consistency as we have in our data center, but eventual consistency, things like that. So we have both worlds, but we need to be able to bridge the information, the data, and the infrastructure between the two. And networking is a key piece of that bridging. There's no shortage of opportunity going forward. No. All right. All right, Hugh, thanks for taking a few minutes out of your day, I appreciate it. Thank you. All right, Hugh Yashida, I'm Jeff Frick. You're watching Cube Conversations, Silicon Angles TV, The Cube Production. Thanks for watching.