 Live from the Palace Hotel in San Francisco, it's theCUBE at the HGST Press and Industry Analyst Briefing. Brought to you by headline sponsor HGST. Here are your hosts, Stu Miniman and Jeff Frick. Hi, welcome back, you're watching theCUBE. I'm Jeff Frick, we're in downtown San Francisco at the lovely Sheridan Palace for the HGST Industry Analyst and Press Day. We're here getting the inside scoop on a bunch of new product introductions, a lot of new news, joined in this segment by my co-host. Hi, I'm Stu Miniman with wikibond.org and joining us for this segment, we're pleased to have Mike Gustafson, where Gus is everyone calls you. You're the Vice President and General, sorry, Senior Vice President and General Manager of the Flash Products Group. Of course, we knew you from your virulent days, which was one of the three soft rock positions that HGST, so thanks so much for coming on theCUBE. It's awesome to be here Stu, thanks. Yeah, we were saying, we couldn't believe that you have not been on the program yet, because we always go for bringing some of the well-known industry people. We've known you a couple of jobs back, always been had some of your folks on theCUBE before, so glad to have you on, we can call you a CUBE alum, you'll be getting the LinkedIn invitation from Jeff real soon. Thanks so much, it's awesome to be here. All right, so Gus, software is eating the world. That's what we've talked about for a lot, and we've seen some really cool hardware here, but of course, you have the software component of what's lined up here. You guys announced something that you're calling device affinity. So can you explain to us for those that didn't get to watch the keynote this morning, what is device affinity, how's it not just another marketing term, what's the reality behind it? Yeah, sure, I love the opening there. I think it's the combination of the appetite of hardware and software, and so device affinity specifically allows us to take core intellectual property that we have on the hardware side, combine that with our software side, and by having the multiple layers and the depth of understanding around that, we're able to actually connect those, tightly integrate in different types of handshakes, provide benefits to customers around things like two times the endurance and two times the performance in our products, and so device affinity is a core element of our differentiation and one that we announced today with multiple examples. All right, so one of the things that really resonated with David Fleur, our CTO, and myself is your flash fabric that you announced. Of course, Wikibon wrote a market definition earlier this year about what we call server sand, which is all of the great things that the storage industry did for the last 15 years to take storage functionality and utilization in network storage, and we need to bring it closer to the compute. And of course, that's what your guy's vision is. So have you guys just been reading our playbook? I mean, I know you guys have been working on this for a while, but explain to us your vision as to where the flash fabric is. Oh, I absolutely want to give you credit where credit is due. I think as far as the industry, I think it's something that we've definitely seen and I referenced the fact that we saw this a decade or so ago with storage area networking and I think we do absolutely leverage the Wikibon views and the perspective of server side flash fabric. I think one of the things that maybe to expand on it a little bit is we view flashes of fabric not only on the server side, but the importance of our partnerships and our different connective tissue we talked about at the array side, whether it's all flash array, hybrid array in the appliances themselves with a flash accelerator or obviously with all the benefits that you know so well on the server side. And I think from an industry perspective, the vision is to do exactly what some of the things you've described is how do we actually bring the capabilities of shared flash creating a platform where you can actually provide high availability, shared pools, aggregate volumes, those types of things. And it does require a tremendous amount of software and intelligence around those devices. All right, so scalability is a big concern when we talk about this. Flash started out, I put it into a single server and that was kind of it. There's a lot of solutions out there today trying to extend that server side and asking solution, everything from just the software only components to of course VMware is doing a lot with BSAN. Can you talk a little bit about how you guys view kind of the clustering, the volume management and the scalability of your solution? Yeah, there's a lot of ways to address these and I think anything that can be done to help accelerate flash, we win. And the industry wins in terms of the benefits. We win in terms of the building blocks but we've really started with a perspective of trying to think about this as an at scale opportunity and an at scale problem to solve. And so from the foundational aspects of mirroring and HA as an example with the combination of scale out through clustering, whether that's clustering for cache or whether that's clustering also for scale out multiple devices. And you can see today with the announcement that we made around up to 16 devices and 128 servers, how we're really breaking through and taking that now to another level. All right, so that product, it's HGST Viridant space. That's correct. So it pulls your company's name into it. That's correct. And 38 terabytes, I mean that's pretty good. Talk about customers that you've been talking to as you're getting ready, it's my understanding the software's shipping now. So have people gotten their hands on this stuff? What are there real world environments that have over 100 servers running it? What can you share? So there are definitely real world environments, none today that I could report publicly that are showing over 100 servers or 38 terabytes of a single pool. But I will say that the people that we've engaged with are actively looking at multiple applications on the shared fabric and being able to scale to those levels. So one thing that we find all over the place with enterprises that nobody wants to begin to architect where they know there's either a roadblock or a cul-de-sac out there in the distance. So being able to confidently commit to 128 servers and 38 terabytes of mirrored flash is a huge advantage. Some of those real world examples are shared storage applications like Oracle Rack where we have a huge advantage there by being able to provide not only a better performance with server side with an all flash world of an environment there for Oracle Rack but also at a reduced cost point. And I think the benefit there not only to those customers but also in terms of bringing Oracle Rack to more of the masses because the price points are now much more achievable for small and medium businesses as well. Yeah, I'd like you to speak a little bit more about those use cases because Oracle Rack is one obviously, that screams to me kind of tier one. So a lot of the solutions that we've seen to expand on flash, hit high performance but don't necessarily have the services that you would put in Oracle Rack, for example. So can you speak to some of the use cases beyond that you say? Yeah, and I think you do see two extremes and I won't say that they are opposites but you do have a tier one kind of a highest level enterprise set of expectations when you're in an Oracle environment and in that environment there are no compromises whether it's on the performance or the reliability. And again, combining the capabilities of layers of software intelligence through the mirroring, et cetera is where it really allows us to provide that capability. On the other side you also see people that are looking at just general open system, clustered software capabilities where you can actually, let's take a KVM example or MySQL environment as examples where you start to see the challenges of expanding and at scale with multiple pairs of servers. In this case, we can actually take this technology and provide a shared pool where you can create a single server that's actually the slave device that provides the replication capability. And by doing that you avoid the need for these duplicate pairs and driving server consolidation in this case as much as 37%. All right, so your software when you were a standalone company with Viridant was OEM by some big storage guys. One of the big questions that people have had today is the balance between what HGST does and your customers. So your customers from HGST of course being the OEMs and the cloud service riders. Can you speak a little bit to what you see there? Yeah, absolutely. And I think this is a really important point for the industry what is driving our company and what is driving our innovation is market driven needs. And so from that standpoint, we're looking at how do we actually bring differentiation across all of the products and solutions we talked about today. And a key part of that is that the industry is looking for more from us. And so HGST is a big part of our day today is not only delivering enabling but helping shape the industry. And that's by driving the specific purpose built capabilities. It's also critically important that people understand that we're not trying to do this all by ourselves. Everything that we're doing right now is focused on the market and available and enabled also through partners if they choose to do that. And so whether it's in whole or whether that's piece parts that makes sense as you mentioned, people have worked with us in the past whether that's leveraging our open extensions and APIs or packaging up parts of our caching capabilities or being more aggressive over time. All of these options are available and things that are core tenants of our companies go to market strategy. So Gus, you've been through a couple of acquisitions as a matter of fact, I was looking at it. This is the second time you've been acquired by a company that has some relationship to Hitachi. So I was wondering if you can speak a little bit to kind of the cultural shift for your team joining the HGST, did the software guys talk to the HDD guys that are out there or do they all wear different colored shirts and play tug of war against each other? Very insightful question. And I'll say if I can I want to make another point here. One, we're very fortunate to have been one of the three companies acquired by HGST last year. So we acquired Vellabit, Estek and Verdan. And so the combination of all those today is we're all one team HGST. To your question around the software and hardware and different colored jerseys, not at all. I think you heard today and I think it really is a testament to early vision and action around a company is taking the benefits of both hardware and software. And so I think both the hardware device company whether it's been for a decade or two decades people understand that we have the opportunity and the market need to continue to build more around that. And likewise, if you come out the world from a software perspective, even though the Dilbert cartoons would all say that they're two very different worlds, we work together. And so it's been a great synergy between those companies and inside of the culture of HGST. So I want to get your insight onto just the general caching market because there are a lot of players in the software space that have been hacking at this problem. And you guys of course have a combination of hardware and software. Do you think there's room for both of those in the marketplace? Do you think you have to have the device affinity specifically to be able to make this successful? What's your view on that? No, I'd like to take a customer-centric view of this and I would say there's definitely room for multiple approaches to this. I think there are environments where one is better than the other and we'll show that with our own performance. But I think that caching has become one of those power applications where everybody knows that you can either extend the life of the sand or the performance of the sand or through new application deployments start to see the benefits. And we have with single server cache we have a software-only approach. We also have the clustered cache capabilities that we've announced in the past. And in both cases we will actually drive some device affinity and differentiation there. Important note though that while we'll generate as much as an eight times performance benefit through that device affinity in examples like Microsoft SQL, we also provide those caching capabilities on non-HGST devices because it's important that we're open in that respect. When you're out in the field talking to customers, how many of them are really starting to kind of get the opportunity that's provided by Flash in writing new applications to take advantage of what Flash provides versus kind of the traditional low latency, high performance, high value app? Are they getting it? Can you share some interesting stories about what you're seeing out there? Yeah. Kind of native Flash from the ground up. No, it's really an interesting one because right now I think you, we've all talked about when is the tipping point of Flash and I mentioned that today over 50% of enterprises have deployed in one way or another. What's tended to happen and I think we're about to go even more faster on this is the people have deployed in kind of point problems or solutions or when they needed that little boost of performance up to this point. And I think some of the things that have prevented more widespread step back, whiteboard and architect, the next generation of architecture are the things that we announced today. All of those capabilities around high availability and clustering, et cetera. And those are required for people to actually take that next step. I did mention LinkedIn as an example of a customer though that has been not only aggressively driving that but delivering some significant business advantage through the Flash strategy and the vision that they have. And I do see, Jeff, I see more customers doing this now, I think the aha moments really across the world when you actually sit down with the core architects of the storage architects and the application experts inside these companies now, starting to break down what we're barriers before and understanding what's possible with Flash has really kind of opened up a little bit of that aha capability. And then the first place they go is yeah, but and we're now able to answer those yeah, but questions. Yeah, that's a good point. So Gus, seeing that there's changes that need to be happening in the stack, are you guys working with the folks that are creating the next generation architectures for cloud mobile deployments? Some of the traditional ISVs are also going there. So what is the relationship between HGST and kind of the application creators? Yeah, so the short answer is yes. So I think this is another benefit of being in a larger company, one that has a platform to have a discussion around much more than just HDD or SSD or software. It's really the whole part of how do we influence and shape the ecosystem. So specifically on the ISV side, many of those folks are customers of ours today, the HDD or device side. And so we're continuing to have the same type of discussions Jeff just asked about. Being able to sit down and understand what we're doing today and then carve out a part of that meaning to discuss what's possible on the application side. So on some of those ISVs, I think people are beginning to realize what sort of benefits we can get by being more direct in terms of access to or more efficient in terms of direct access to or maybe leveraging some of the capabilities around multiple copies or garbage collection or snapshot capabilities of things that are kind of in there already and how to leverage those even further. And I think that's true in the ISV side. And we're also seeing the same with some of our largest OEMs and the same with some of our largest cloud customers. We're now having a much more forward discussion around where the architectures are going and how we might contribute more aggressively to that. Yeah, since the acquisition, I'm curious, do you have more access to, especially I think the cloud guys because that's one of the things that really came out to me in our interviews today is HDST has relationships with those big guys whereas as you know, Mike, the traditional storage companies aren't there and can't necessarily get there. No, it's absolutely yes again. And I think part of that is the disruptive nature of what the cloud companies have done on not only the supply chain, but the economics of the models that they generate and the fact that they bring so much of their own capability to the party. Their own number of resources and insight around what exactly they're trying to do. So our company, HDST, has had long and deep relationships in providing hard drive technology and solutions to these folks. And definitely with the addition now of our combined HDD and SSD portfolio with that software we're much more actively engaged in those dialogues. All right, so we're getting towards the close of our coverage of this show. A lot of different announcements. There's so many different pieces. As we drive away from this, what's the bumper sticker you would want people to take away and to know kind of who HDST and how they fit in the market. You know, I thought Mike had a great wrap up this morning so I'm going to borrow it. I think this is not your father's disk drive company. I think what we're aiming to do and what we're actually demonstrating today is the delivery of a more forward market shaping insight. We want to be viewed as that thought leader but more so we want to be viewed as those that deliver and help shape the market. So I think you can see that across a combined platform, combined capability of products as well. So partner of choice, not your father's HDD company only, we're going to build upon that. That's a big bumper sticker. It's a big bumper sticker but Oldsmobile makes big cars so that's okay. But guys, thanks for, that's a great way to close and I think you're really in an interesting position where you get the benefit of the hyperscale companies and where they're really pushing the envelope. Now, as you said, you've introduced some technology with the enterprises. They say, yeah, that's all fine and dandy but and now you guys have some of the interest of the buts. So that's terrific. So guys, thanks for stopping by. I'm sure we'll see you again. So I'm Jeff Frick. Again, we're at the HGST Press and Analyst Event in downtown San Francisco, the Sheridan Palace Hotel joined by Stu Miniman. We'll be back with our next guest after this short break. Long live data. Long live data. Long live data.